Fpr2Deficiency Alleviates Diet-Induced Insulin ResistanceThrough Reducing Body Weight Gain and InhibitingInflammation Mediated by Macrophage Chemotaxis andM1 PolarizationXiaofang Chen,1 Shu Zhuo,1 Tengfei Zhu,1 Pengle Yao,1 Mengmei Yang,1 Hong Mei,1 Na Li,1 Fengguang Ma,1

Ji Ming Wang,2 Shiting Chen,1 Richard D. Ye,3 Yu Li,1 and Yingying Le1,4

Diabetes 2019;68:1130–1142 | https://doi.org/10.2337/db18-0469

Obesity and related inflammation are critical for thepathogenesis of insulin resistance, but the underlyingmechanisms are not fully understood. Formyl peptidereceptor 2 (FPR2) plays important roles in host immuneresponses and inflammation-related diseases. We foundthat Fpr2 expression was elevated in the white adiposetissue of high-fat diet (HFD)–induced obese mice anddb/db mice. The systemic deletion of Fpr2 alleviatedHFD-induced obesity, insulin resistance, hyperglycemia,hyperlipidemia, and hepatic steatosis. Furthermore, Fpr2deletion in HFD-fed mice elevated body temperature,reduced fat mass, and inhibited inflammation by reduc-ing macrophage infiltration and M1 polarization in met-abolic tissues. Bone marrow transplantations betweenwild-type and Fpr22/2 mice and myeloid-specific Fpr2deletion demonstrated that Fpr2-expressing myeloidcells exacerbated HFD-induced obesity, insulin resis-tance, glucose/lipid metabolic disturbances, and inflam-mation. Mechanistic studies revealed that Fpr2 deletionin HFD-fed mice enhanced energy expenditure probablythrough increasing thermogenesis in skeletal muscle;serum amyloid A3 and other factors secreted by adipo-cytes induced macrophage chemotaxis via Fpr2; andFpr2 deletion suppressed macrophage chemotaxisand lipopolysaccharide-, palmitate-, and interferon-g–induced macrophage M1 polarization through block-ing their signals. Altogether, our studies demonstrate thatmyeloid Fpr2 plays critical roles in obesity and related

metabolic disorders via regulating muscle energy expen-diture, macrophage chemotaxis, and M1 polarization.

Obesity is a major risk factor for the development ofinsulin resistance and type 2 diabetes (1,2). Chroniclow-grade inflammation in obesity is one of the mostimportant causes of obesity-related complications (3).Macrophages play crucial roles in obesity-related inflam-mation and insulin resistance (4). In obese individuals,macrophages accumulate in adipose tissue and formcrown-like structures (CLSs) by surrounding the deadadipocytes (5). Adipose tissue macrophages undergo a phe-notypic switch from the alternatively activated M2 phe-notype to the classically activated M1 phenotype duringobesity (4). M1 macrophages secrete proinflammatorycytokines that aggravate tissue inflammation and causelocal and systemic insulin resistance through interferingwith insulin-signaling pathways (6). In contrast, M2 mac-rophages are constitutively present in the lean adiposetissue and express high levels of anti-inflammatory cyto-kines that are positively associated with insulin sensitivity(7). Studies have shown that the reduced macrophageinfiltration into tissues and inhibited macrophageM1 polarization could attenuate obesity-related insulinresistance (8,9). Therefore, the identification of key mol-ecules contributing to macrophage infiltration and M1

1CAS Key Laboratory of Nutrition, Metabolism and Food Safety, Shanghai Instituteof Nutrition and Health, Shanghai Institutes for Biological Sciences, University ofChinese Academy of Sciences, Chinese Academy of Sciences, Shanghai, China2Cancer and Inflammation Program, Center for Cancer Research, National CancerInstitute at Frederick, Frederick, MD3Institute of Chinese Medical Sciences, University of Macau, Macau SpecialAdministrative Region, China4Key Laboratory of Food Safety Risk Assessment, Ministry of Health, Beijing, China

Corresponding author: Yingying Le, yyle@sibs.ac.cn

Received 24 April 2018 and accepted 17 February 2019

This article contains Supplementary Data online at http://diabetes.diabetesjournals.org/lookup/suppl/doi:10.2337/db18-0469/-/DC1.

© 2019 by the American Diabetes Association. Readers may use this article aslong as the work is properly cited, the use is educational and not for profit, and thework is not altered. More information is available at http://www.diabetesjournals.org/content/license.

1130 Diabetes Volume 68, June 2019

METABOLISM

polarization may provide therapeutic targets against in-sulin resistance and type 2 diabetes.

Formyl peptide receptor 2 (FPR2) is a chemoattractantreceptor that belongs to the FPR family. Human FPR2 and itsmouse homolog Fpr2 are highly expressed in macrophages/monocytes and neutrophils, interact with peptide and lipidligands, and transduce pro- or anti-inflammatory actions(10). The different effects of FPR2 on inflammation aredependent on the context and ligands that activate differentsignaling pathways (11,12). FPR2 is involved in multiplediseases, including bacterial infection, inflammation, asthma,Alzheimer disease, and cancers (13,14). Studies with Fpr2knockout or Fpr2/Fpr3 double-knockout mice showed thatFpr2 played protective or detrimental roles in differentdisease models (15–18). Annexin A1 (ANXA1) and lipoxinA4 (LXA4) are Fpr2 ligands with anti-inflammatory prop-erties. ANXA1 reduces body weight gain in obese miceand inhibits hepatic inflammation during nonalcoholicsteatohepatitis progression (19,20). LXA4 attenuates obesity-induced inflammation in adipose tissue and related diseases(21). Because ANXA1 and LXA4 can activate other recep-tors in addition to Fpr2 (11), it is of interest to determinethe contribution of Fpr2 in obesity-related chronic in-flammation and metabolic disorders by using Fpr2 knock-out mice.

In the current study, our in vivo data from Fpr2-deficient mouse models and in vitro results from bonemarrow-derived macrophages (BMDMs) demonstratedthat Fpr2 plays vital roles in diet-induced obesity(DIO), inflammation, and metabolic disorders.

RESEARCH DESIGN AND METHODS

Animal ExperimentsFpr22/2 mice and wild-type (WT) littermates were generatedby intercrossing Fpr2+/2 mice (16). Myeloid-specific Fpr2knockout (Fpr2MKO) mice were developed by crossingFpr2flox/flox mice with LysM-Cre mice. C57BL/6 and db/db mice were obtained from Shanghai Laboratory AnimalCompany (Shanghai, China). Mice of different genotypeswere housed in different cages. Eight-week-old male micewere fed a chow diet (10% fat calories) or high-fat diet (HFD;60% fat calories) (Research Diets, New Brunswick, NJ) for 9–12 weeks.

Body compositionwas analyzed using EchoMRI (EchoMRI,Houston, TX). Glucose tolerance tests (GTTs) and insulintolerance tests (ITTs) were conducted as described previously(22). Rectal temperatures were measured using a rectal probeattached to a digital thermometer (Physitemp Instruments,Clifton, NJ). To examine the tissue response to insulin, micefasted for 4 h were injected with insulin (2 units/kg bodywt) or PBS in the inferior vena cava, followed by collectionof liver tissues at 3 min, epididymal white adipose tissue(WAT) at 5 min, and gastrocnemius muscle at 7 min afterthe injection. Akt phosphorylation was examined by West-ern blotting. All animal experiments were performed inaccordance with the guidelines of the Institutional AnimalCare and Use Committee at Shanghai Institute of Nutrition

and Health, Shanghai Institutes for Biological Sciences,Chinese Academy of Sciences.

Tissue Uptake of 18F-FluorodeoxyglucoseMice fasted for 16 h were intravenously injected with 18F-fluorodeoxyglucose (5.8 6 0.7 MBq/mouse). After 1 h, thetissue samples were collected and measured for radioactivity.The differential uptake ratio (DUR) was used as an index ofradiotracer uptake in tissues. DUR = (tissue counts [cpm] perg of tissue)/(injected dose counts per g of body weight) (23).

Energy Expenditure MeasurementVO2, VCO2, and locomotive activities of mice were de-termined in the Comprehensive Lab Animal MonitoringSystem (Columbus Instruments, Columbus, OH). Datawere collected for 48 h after the mice were acclimatedto the system for 24 h with free access to food and water.The rate of energy expenditure (calories/min) was calcu-lated as VO2 3 (3.815 + [1.232 3 {VCO2/VO2}]).

Biochemical Parameter AnalysisSerum and liver triglycerides (TGs), serum aspartate ami-notransferase (AST), and alanine aminotransferase (ALT)were determined using the respective kits (Shensuoyoufu,Shanghai, China). Serum levels of nonesterified fatty acids(NEFA), insulin, and proinflammatory cytokines weremeasured with a NEFA assay kit (Wako Pure Chemicals,Osaka, Japan) and ELISA kits (Millipore, Billerica, MA;R&D Systems, Minneapolis, MN), respectively.

Histology and ImmunohistochemistryThe tissues sections were stained with hematoxylin andeosin (H&E) or Oil Red O or immunostained with anti-bodies against CD68 and F4/80. Adipocyte diameter andCLSs in the WAT sections were analyzed using Image-ProPlus software (Media Cybernetics, Rockville, MD). Macro-phage infiltration in the liver sections was quantified byaverage optical density with Image-Pro Plus software.

mRNA and Protein AnalysisQuantitative real-time PCR was performed as describedpreviously (24). The primer sequences are presented in theSupplementary Data. Western blot analysis was performedwith primary antibodies against phosphorylated and totalforms of Akt, nuclear factor-kB (NF-kB) p65, p38, extra-cellular signal–regulated kinase (ERK), c-Jun N-terminalkinase (JNK), transforming growth factor b-activatedkinase 1 (TAK1) and STAT1, HSP90 (Cell Signaling, Dan-vers, MA), and b-actin (Sigma-Aldrich, St. Louis, MO).Signaling was visualized with ECL Plus Western BlottingDetection System (GE Healthcare, Salem, CT).

Cell Isolation, Culture, and TreatmentBMDMs prepared from bone marrow cells (25) were stim-ulated with lipopolysaccharide (LPS; Sigma-Aldrich), palmi-tate, IFN-g (Peprotech, Rocky Hill, NY), or serum amyloid A3(Saa3; CUSABIO, Wuhan, China), and examined for thephosphorylation ofmitogen-activated protein (MAP) kinases,

diabetes.diabetesjournals.org Chen and Associates 1131

NF-kB p65, STAT1, and TAK1 by Western blotting, forexpression of M1 macrophage-related proinflammatory cyto-kines by quantitative real-time PCR and ELISAs, respectively.Sodium palmitate (Sigma-Aldrich) was dissolved in fatty acid–free BSA (Sigma-Aldrich) solution with a 2.5:1 mol/L ratio.

Fpr2 was introduced into BMDMs by infecting the cellswith Fpr2-lentiviruses. 3T3-L1 fibroblasts were differen-tiated into mature adipocytes as described previously (26)and transfected with Saa3 siRNA or negative control usingLipofectamine 3000 (Invitrogen, Thermo Fisher Scientific,Waltham, MA). The sequence of Saa3 siRNA is 59-GCUG-GUCAAGGGUCUAGAG-39.

FACS AnalysisStromal vascular cells (SVCs) were isolated from epididy-mal WAT by type I collagenase digestion (27,28), stainedwith an antibody cocktail containing anti–CD45-FITC,F4/80-phosphatidylethanolamine (PE), CD11b-eFluor 450,CD206-allophycocyanin (APC), and CD11c-PE-Cyanine7(eBioscience, Thermo Fisher Scientific). M1 andM2macro-phages were identified with FACS analysis using the gatestrategy as previously published (28).

Chemotaxis AssayChemotaxis of BMDMs in response to adipose tissuelysate, culture medium from adipocytes, or Saa3 was ana-lyzed using a polycarbonate membrane with an 8-mm poresize in 48-well chemotaxis chambers (NeuroProbe, Gai-thersburg, MD) (29). The migrated cells were analyzedwith Image-Pro Plus software.

Bone Marrow TransplantationsBone marrow transplantations (BMTs) between mice wereperformed as previously described (30). Fpr2 expression inneutrophils and monocytes isolated from bone marrowswas detected to evaluate the efficiency of BMTs.

StatisticsAll experiments were repeated at least three times. Theresults are presented as the mean 6 SD or SEM. Thestatistical analysis was performed by using unpaired two-tailed Student t tests for two-group comparison and usingtwo-way ANOVA or two-way repeated-measures ANOVA formultiple group comparison. The residual method (31) wasused to control body weight in analyzing data of GTTs andITTs and semiquantitative data of immunostainings.Body weight–adjusted blood glucose in GTTs/ITTs andsemiquantitative data of immunostainings were computedas the residuals from the regression model with bloodglucose or semiquantitative data of immunostainings asthe independent variable and body weight as the dependentvariable. Significance was accepted at P , 0.05.

RESULTS

Fpr2 Is Upregulated in Adipose Tissues of DiabeticMiceWe examined Fpr2 expression in major metabolic tissuesof diabetic mice and found that Fpr2 expression was

upregulated in WAT and gastrocnemius muscle of DIOmice compared with that of chow-fed mice (Fig. 1A). Fpr2expression was also upregulated inWAT of db/dbmice (Fig.1A). We next examined the cellular source of Fpr2 in WATand found that DIO mice had a higher level of Fpr2 mRNAin SVCs (Fig. 1A). These results show that Fpr2 expressionis significantly increased in WAT of obese mice, especiallyin SVCs.

Deletion of Fpr2 Alleviates DIO, Insulin Resistance, andImpairment of Glucose and Lipid MetabolismTo study the contribution of Fpr2 to metabolic regulation,we fed WT and Fpr22/2mice a chow diet or HFD. We foundthat systemic Fpr2 deletion had no significant effect onbody weight and composition, food intake, energy expen-diture, glucose, and lipid metabolism in chow-fed mice (Fig.1B and Supplementary Fig. 1). When fed the HFD, Fpr22/2

mice presented similar food intake to WT mice but hadlower body weight than WT mice after 3 weeks (Fig. 1B andC). Compared with WT mice, Fpr22/2 mice had lower fatmass percentage, higher leanmass percentage (Fig. 1D), andsmaller adipocytes in WAT (Fig. 1E). Fpr2 deletion increasedO2 consumption (O2), CO2 production (CO2), energy ex-penditure rate, and rectal temperature, but had no signif-icant effect on physical activities (Fig. 1F–H). We furtherevaluated the expression of thermogenic genes in sub-cutaneous WAT (sWAT), brown adipose tissue (BAT), andthe gastrocnemius muscle. Compared with WT mice, HFD-fed Fpr22/2 mice expressed higher levels of Pgc1a, Ppara,Ucp2, and Cd36 in the muscle, but expressed comparablethermogenic genes in sWAT and BAT (Fig. 1I). Theseresults indicate that Fpr2 deficiency reduces body weightgain in HFD-fed mice through enhancing energy expendi-ture, especially in skeletal muscle.

Further studies showed that HFD-fed Fpr22/2 micedisplayed lower blood glucose and serum insulin levels (Fig.2A and B) and improved glucose tolerance and insulinsensitivity (Fig. 2C and E) compared with WT mice.Consistently, higher Akt phosphorylation in response toinsulin was observed in WAT and gastrocnemius muscle ofFpr22/2 mice (Fig. 2G). After adjusting the data of GTTsand ITTs with body weight, the differences between WTand Fpr22/2 mice were decreased but still statisticallysignificant (Fig. 2D and F), indicating that Fpr2 deficiencyattenuates insulin resistance through reducing bodyweight gain and other mechanisms. In addition, Fpr22/2

deficiency reduced serum TG and NEFA levels (Fig. 2H),hepatic lipid accumulation (Fig. 2J and K), and serum ASTand ALT levels (Fig. 2I). Collectively, these data demon-strate that Fpr2 deficiency improves insulin sensitivity andalleviates lipid and glucose dysregulation in DIO mice.

Deletion of Fpr2 Reduces HFD-Induced SystemicInflammation, Tissue Macrophage Infiltration, and M1PolarizationWe next evaluated the contribution of Fpr2 to obesity-related inflammation. When fed the HFD, Fpr22/2 mice

1132 Fpr2 in Inflammation and Insulin Resistance Diabetes Volume 68, June 2019

showed lower serum levels of interleukin 6 (IL-6), tumornecrosis factor-a (TNF-a), and chemokine (C-C motif)ligand 2 (CCL2) than WT mice (Fig. 3A). CLSs andCD68-positive macrophages in WAT and F4/80-positivemacrophages in liver (Fig. 3B and C) were markedly re-duced in Fpr22/2 mice. After adjustment with bodyweight, the differences of CLSs in WAT and infiltratedmacrophages in liver were still statistically significantbetween WT and Fpr22/2 mice (Fig. 3D). Consistently,obese Fpr22/2mice showed reduced expression of macro-phage and M1 macrophage markers in WAT (Fig. 3E).However, the expression of M2 macrophage markers (Il-10, Ym1 except Retnla) in WAT was comparable betweenFpr22/2 and WT mice (Fig. 3E). The expression of mac-rophage and M1 macrophage markers was slightly down-regulated in the liver and muscle of Fpr22/2mice (Fig. 3E).Peritoneal macrophages from HFD-fed Fpr22/2mice alsoexpressed lower mRNA levels of M1 macrophage-relatedgenes thanWTmice (Fig. 3F). These data demonstrate thatFpr2 deletion reduces tissue and systemic inflammation in

DIO mice by inhibiting macrophage infiltration and M1polarization.

Restoring Fpr2 Expression in Immune Cells in Fpr2-Deficient Mice Exacerbates DIO, Insulin Resistance,and InflammationBecause deletion of Fpr2 improved inflammation in HFD-fed mice, we further investigated the contribution of Fpr2-expressing immune cells to metabolic dysregulation andinflammation in DIO mice through BMT. Bone marrowsfrom Fpr22/2 mice were transplanted into irradiated WTrecipients (Fpr22/2-WT), or vice versa (WT-Fpr22/2), andBMTs between WT mice (WT-WT) were performed asa control. The expression of Fpr2 mRNA in monocytesand neutrophils isolated from bone marrow was signifi-cantly decreased in Fpr22/2-WT mice and was restored inWT-Fpr22/2mice (Supplementary Fig. 2A). Food intake ofWT-WT, Fpr22/2-WT, and WT-Fpr22/2 mice was similar(Supplementary Fig. 2B). Compared with WT-WT mice,Fpr22/2-WT mice were resistant to DIO (Fig. 4A and B),

Figure 1—Deletion of Fpr2 attenuates DIO. A: Fpr2 mRNA expression in liver, epididymal WAT (eWAT), and gastrocnemius muscle of HFDmice (n = 6–10 per group), db/dbmice, and their corresponding control mice (n = 6–7 per group), and in adipocytes (Ads) and SVCs isolatedfrom adipose tissue of mice fed the HFD or chow (n = 3–4 per group). B–I: Fpr22/2 and WT mice were fed the HFD or control chow. B: Bodyweight wasmeasuredweekly. Food intake over 24 h (C), body composition (D), and rectal temperature (RT) (F ) were determined at 8 weeks onthe HFD (n = 8–13 per group). Other parameters were examined at 10 weeks on HFD as follows: adipocyte diameter in H&E-stained eWATsections (n = 5–9 per group) (E); average VO2, VCO2, the rate of energy expenditure, and locomotor activities of mice over 24 h (n = 6–8 pergroup) (G andH); and expression of thermogenic genes in gastrocnemiusmuscle, BAT, and sWAT (n = 7–12 per group) (F ). Data are shown asthe mean6 SEM. *P, 0.05, **P, 0.01 for comparison betweenWT and db/dbmice (A) or between Fpr22/2 andWTmice fed the HFD (B–I).Two-way repeated-measures ANOVA was used in B, and the Student t test was used in A and C–I.

diabetes.diabetesjournals.org Chen and Associates 1133

hyperglycemia, insulin resistance, dyslipidemia (Fig. 4C–Fand Supplementary Fig. 2C and D), and systemic/tissueinflammation (Fig. 4G–L). The metabolic and inflammatoryphenotypes of HFD-fed Fpr22/2-WT mice were similar tothose of HFD-fed Fpr22/2mice. In contrast, WT-Fpr22/2

mice were prone to DIO (Fig. 4A and B), metabolic dis-turbances (Fig. 4C–F and Supplementary Fig. 2C andD), andtissue/systemic inflammation (Fig. 4G–L). Collectively, BMTstudies demonstrate that Fpr2 in immune cells is deeplyinvolved in DIO, metabolic disturbances, and inflammation.

Myeloid-Specific Deletion of Fpr2 Alleviates DIO,Insulin Resistance, and InflammationWe generated Fpr2MKO mice to investigate which type ofFpr2-expressing immune cells participated in DIO, insulin

resistance, and inflammation (Supplementary Fig. 3A).Consistent with the results from Fpr22/2 mice, myeloid-specific deletion of Fpr2 had no significant effect onmetabolic phenotypes in mice fed the chow diet (Fig.5A and Supplementary Fig. 3B–G) but alleviated obesity,insulin resistance, and glucose/lipid dysmetabolism andelevated body temperature, energy expenditure, and mus-cle thermogenesis in mice fed the HFD (Fig. 5 and Sup-plementary Fig. 3H–K). After adjustment with bodyweight, the improvement of glucose tolerance and insulinsensitivity in HFD-fed Fpr2MKO mice was slightly reducedbut still statistically significant (Fig. 5G and I), indicatingthat myeloid Fpr2 deletion alleviates insulin resistancethrough mechanisms dependent and independent of bodyweight.

Figure 2—Fpr22/2mice are protected fromHFD-induced insulin resistance and hepatic steatosis. Fpr22/2 andWTmicewere fed theHFD for10weeks.Metabolic phenotypeswere examined as follows: blood glucose levels under fed and fasted conditions (A), serum insulin levels (n =7–12 per group) (B); GTT and average area under the curve (AUC) adjusted with or without body weight (C and D); ITT and AUC adjusted withor without body weight (n = 10–14 per group) (E and F ); Akt phosphorylation (p-) in the liver, epididymal WAT, and gastrocnemiusmuscle afterinsulin administration in the inferior vena cava (n = 4–6 per group) (G); serum levels of TG and NEFA (H); serum levels of AST and ALT (I); liverTG content (J); and H&E (HE) and Oil Red O staining of liver sections (n = 6–11 per group) (K). The results represent the mean 6 SEM. *P ,0.05, **P, 0.01, ***P, 0.001 for the comparison between Fpr22/2 andWTmice fed the HFD. Images in K are representatives of H&E and OilRed O staining of liver sections. Scale bar = 100 mm. The Student t test was used in A and B and H–J. Two-way repeated-measures ANOVAwas used inC–F, and two-way ANOVAwas used inG. Body weight–adjusted blood glucose levels (D) and blood glucose change (%) (F) werecomputed as the residuals from the regressionmodel, with blood glucose level or change as the independent variable and bodyweight as thedependent variable.

1134 Fpr2 in Inflammation and Insulin Resistance Diabetes Volume 68, June 2019

Myeloid-specific deletion of Fpr2 in HFD-fed mice alsorelieved systemic and tissue inflammation (Fig. 6A–C)and reduced the expression of M1 macrophage markersin WAT, liver, and muscle (Fig. 6E). The FACS analysis ofSVCs from WAT of these mice consistently revealeda lower percentage of M1 macrophages and a comparablepercentage of M2 macrophages compared with those ofHFD-fed WT mice (Fig. 6F and G). The reduction of CLSsin adipose tissue and macrophage infiltration in hepatictissues by myeloid Fpr2 deletion remained after adjust-ment with body weight (Fig. 6D). Collectively, these dataindicate that myeloid Fpr2 plays an important role inDIO, insulin resistance, glucose/lipid dysmetabolism,and inflammation.

Adipocyte-Secreted Saa3 Induces MacrophageChemotaxis Through Fpr2

We examined whether obese adipose tissue produced Fpr2ligands to induce macrophage infiltration. We found thatthe chemotactic activity of the obese adipose tissue lysate(OATL) to BMDMs is higher than that of lean adiposetissue lysate. The chemotactic response of Fpr22/2

BMDMs to OATL is lower than that of WT BMDMs(Fig. 7A). These results indicate that obese adipose tissuecontains a higher level of Fpr2 ligands than lean adiposetissue.

Saa3 upregulation has been reported in adipose tissueof obese mice (32). We found that Saa3 significantly in-duced WT BMDMs chemotaxis but had less effect on

Figure 3—Decreased tissue macrophage infiltration and M1 polarization in HFD-fed Fpr22/2 mice. Fpr22/2 and WT mice fed the HFDfor 10 weeks were examined for serum levels of proinflammatory cytokines (n = 7–12 per group) (A); CD68-positive cells and CLSsin WAT adjusted with or without body weight (B and D); F4/80-positive cells in hepatic tissues adjusted with or without body weight (n =6 per group) (C and D); mRNA expression of M1 and M2 macrophage markers in the epididymal WAT, liver, and gastrocnemiusmuscle (n = 9–12 per group) (E ); and expression of M1 macrophage markers in peritoneal macrophages (pMac) (n = 5–10 per group)(F ). Data are the mean6 SEM. *P, 0.05, **P, 0.01, ***P, 0.001. Images in B and C are representative immunostainings of WAT andliver sections with antibodies against CD68 and F4/80, respectively; the scale bar = 100 mm in B and 50 mm in C. CLSs and F4/80-positive cells per field were quantified using Image-Pro Plus software. The Student t test was used. Body weight–adjusted CLSs andF4/80-positive cells in D were computed as the residuals from the regression model, with CLSs or F4/80-positive cells asthe independent variable and body weight as the dependent variable.

diabetes.diabetesjournals.org Chen and Associates 1135

Figure 4—Effect of deletion or restoration of Fpr2 in hematopoietic cells by BMT on HFD-induced insulin resistance and inflammation. WT micereceived bone marrows from WT mice or Fpr22/2 mice (WT-WT, Fpr22/2-WT), and Fpr22/2 mice received bone marrows from WT mice (WT-Fpr22/2) fed the HFD for 9 weeks. Metabolic phenotypes were examined as follows: body weight (A), body composition (B), fed and fasted bloodglucose levels (C), serum insulin levels (D), serum levels of TG andNEFA (E), and liver TG content (n= 7–12 per group) (F). Inflammatory parameterswere measured as follows: serum levels of IL-6 and TNFa (n = 5–10 per group) (G), CLSs in epididymal WAT (H and J), F4/80-positive cells inhepatic tissues (n = 7–11 per group) (I andK), andmRNA expression of macrophagemarkers, andM1 andM2macrophagemarkers inWAT, liver,andgastrocnemiusmuscle (n=7–9per group) (L). Dataare themean6SEM. *P,0.05, **P,0.01, ***P,0.001 for the comparison betweenHFD-fed Fpr22/2-WT andWT-WTmice; †P, 0.05 for the comparison between HFD-fed Fpr22/2-WT andWT-Fpr22/2 mice. Scale bar = 100 mm in Jand 50mm inK. CLSs and F4/80-positive cells per fieldwere quantified using Image-Pro Plus software. Two-way repeated-measures ANOVAwasused in A and two-way ANOVA was used in B–I and L.

1136 Fpr2 in Inflammation and Insulin Resistance Diabetes Volume 68, June 2019

Fpr22/2 BMDMs (Fig. 7B). The conditioned medium (CM)from Saa3-knockdown adipocytes had lower chemotacticactivity to WT BMDMs than CM from control adipocytes(Fig. 7C and D). The chemotactic response of Fpr22/2

BMDMs to the CM from Saa3-knockdown adipocytes islower than that of WT BMDMs (Fig. 7D). These resultsindicate that adipocyte-secreted Saa3 and other factorsinduce BMDM migration through Fpr2. We found thatSaa3 significantly induced Tnfa, Il-1b, Il-6, and Ccl2 ex-pression in an Fpr2-independent manner (Fig. 7E). Theexpression of Saa3 in WAT was significantly decreasedin HFD-fed Fpr22/2, Fpr22/2-WT, and Fpr2MKOmice but

was recovered in Fpr22/2 mice who received BMTs fromWT mice (Fig. 7F).

Fpr2 Promotes Macrophage M1 Polarization andProinflammatory Cytokine Expression In VitroWe further investigated the role of Fpr2 in macrophage M1polarization and proinflammatory cytokine expression bystimulating BMDMs with LPS. Data showed that deletionof Fpr2 in BMDMs significantly reduced the expression ofM1 macrophage-related genes (Il-1b, Il-6, Ccl2, and Tnfa)and proinflammatory cytokines (IL-6, CCL2, and TNF-a)(Fig. 8A and B). NF-kB and MAP kinases are important

Figure 5—Myeloid-specific deletion of Fpr2 improves HFD-induced obesity and insulin resistance. Fpr2flox/flox and Fpr2MKO mice were fedcontrol chow or the HFD for 10 weeks and examined for body weight (A), body composition (B), fed and fasted blood glucose levels (C), andserum insulin levels (n = 6–13 per group) (D); the DUR of 18F-fluorodeoxyglucose in the liver, WAT, and gastrocnemius muscle (E ) (n = 5–7 pergroup); GTT and average area under the curve (AUC) adjusted with or without body weight (F and G); ITT and AUC adjusted with or withoutbody weight (H and I); insulin stimulated Akt phosphorylation (p-) in liver, WAT, and gastrocnemius muscle (J); serum levels of TG and NEFA(K); serum levels of AST and ALT (L); and liver TG content (n = 6–13 per group) (M). The results represent the mean6 SEM, *P, 0.05, **P,0.01, ***P , 0.001 for the comparison between Fpr2MKO and Fpr2flox/flox mice fed the HFD with or without insulin stimulation. Two-wayrepeated-measures ANOVA was used in A and F–I, and two-way ANOVA was used in J. The Student t test was used in B–E and K–M.

diabetes.diabetesjournals.org Chen and Associates 1137

signaling molecules mediating proinflammatory cytokineexpression and macrophage polarization. We found thatFpr2 deletion markedly inhibited NF-kB p65 and p38phosphorylation and slightly inhibited ERK and JNKphosphorylation induced by LPS (Fig. 8C). The Fpr2 an-tagonist WRW4 consistently suppressed LPS-inducedMAP

kinases and NF-kB activation as well as M1 macrophage-related gene expression in BMDMs (Supplementary Fig. 4Band C). In contrast, Fpr2 overexpression enhanced LPS-stimulated signal transduction and proinflammatory cy-tokine expression (Fig. 8D–F). Fpr2 deficiency in BMDMsalso reduced LPS-induced phosphorylation of TAK1 (Fig. 8C),

Figure 6—Myeloid Fpr2 deficiency reduces tissuemacrophage infiltration andM1 polarization. Fpr2flox/flox and Fpr2MKOmice fed the HFD for10 weeks were examined for serum levels of IL-6 and TNF-a (A); CD68-positive cells and CLSs in epididymal WAT adjusted with or withoutbody weight (B and D); F4/80-positive cells in hepatic tissues adjusted with or without body weight (C and D); mRNA expression ofmacrophage markers, M1 and M2 macrophage markers in WAT, liver, and gastrocnemius muscle (E) (n = 7–12 per group); and flowcytometric analysis of M1 (CD45+F4/80+CD11b+CD11c+CD2062) and M2 (CD45+F4/80+CD11b+CD11c2CD206+) macrophages amongadipose tissue macrophages (CD45+F4/80+CD11b+) (n = 5–7 per group) (F and G). Data are the mean6 SEM. *P , 0.05, **P, 0.01, ***P ,0.001. Images in B and C are representative immunostainings of WAT and liver sections with antibodies against CD68 and F4/80,respectively. The scale bar = 100 mm in B and 50 mm in C. CLSs and F4/80-positive cells per field were quantified using Image-Pro Plussoftware. The Student t test was used. Body weight–adjusted CLSs and F4/80-positive cells in D were computed as the residuals from theregression model, with CLSs or F4/80-positive cells as the independent variable and body weight as the dependent variable.

1138 Fpr2 in Inflammation and Insulin Resistance Diabetes Volume 68, June 2019

an upstream molecule of NF-kB and MAP kinases inthe LPS-stimulated signaling pathway (33). We also exam-ined the involvement of Fpr2 in macrophage M1 polari-zation induced by palmitate and IFN-g. Fpr2 deletion inBMDMs reduced M1 macrophage–related gene expressionand phosphorylation of NF-kB p65 and MAP kinases inresponse to palmitate (Fig. 8G–I) and reduced STAT1phosphorylation by IFN-g (Supplementary Fig. 4D). Theseresults indicate that Fpr2 deletion suppresses macrophagepolarization toward an M1 phenotype via inhibitingTAK1-MAP kinase/NF-kB and STAT1-related signalingpathways.

DISCUSSION

In the current study, we found that Fpr2 was highlyexpressed in WAT of obese mice models, especially inthe SVCs of WAT. Further studies of systemic Fpr2 de-letion, BMTs, and myeloid-specific Fpr2 deletion in micedemonstrated that myeloid Fpr2 plays critical roles in DIO,metabolic disturbances, and inflammation.

The body weight change is associated with an imbalancebetween food intake and energy expenditure. Deletion ofFpr2 systemically or in myeloid cells in mice fed the HFDdid not affect food intake but increased energy expendi-ture and body temperature. That BAT and skeletal muscleare important players in regulating thermogenesis is wellknown. Our results showed that Fpr2 deletion in obese

mice increased fatty acid oxidation–related and thermo-genic gene expression in muscle. Skeletal muscle is thelargest organ in the body and is a major determinant ofbasal metabolic rate. An increase of energy expenditure inmuscle through nonshivering thermogenesis can sub-stantially affect whole-body metabolism and weightgain (34,35). Therefore, the increase of lean mass per-centage and thermogenic gene expression in skeletalmuscle may contribute to higher energy expenditureand lower body weight gain in HFD-fed Fpr22/2 andFpr2MKO mice.

The mechanisms underlying the regulation of thermo-genesis by Fpr2 in the muscle of obese mice is not clear.Studies by systemic knockout of TNF-a receptor or hypo-thalamic immunoneutralization of TNF-a in rodents indi-cate that obesity-related inflammation impairs whole-bodyenergy expenditure, mitochondrial biogenesis, ATP produc-tion, and thermogenesis in BAT and muscle (36,37).Whether Fpr2 deficiency in DIO mice promotes musclethermogenesis through alleviating inflammation remainsto be further investigated.

In addition to energy expenditure, lipid accumulation inthe liver is also associated with the changes of body fatmass (38). Therefore, the alleviation of hepatic steatosis inHFD-fed Fpr22/2 and Fpr2MKO mice may also contributeto lower fat mass. Body weight changes and inflammationhave been reported to influence insulin sensitivity (39,40).

Figure 7—Saa3 induces macrophage migration via Fpr2. WT and Fpr22/2 BMDMmigration in response to adipose tissue lysate from chow-fed mice (lean adipose tissue lysate [LATL]) and HFD-fed mice (OATL) (n = 4–6 per group) (A), or 10 mg/mL Saa3 (n = 8–10 per group) (B).Adipocytes differentiated from 3T3-L1 cells were transfected with siSaa3 or control siRNA (Scr) and then examined for gene expression (C)and the chemotactic activity of the culture medium to WT and Fpr22/2 BMDMs (n = 4–6 per group) (D). E: Proinflammatory cytokineexpression in BMDMs induced by 100 mg/mL Saa3 for 6 h (n = 3 per group). F: Saa3 mRNA expression in the epididymal WAT of HFD-fedFpr22/2, Fpr2MKO and their corresponding control mice, as well as mice with BMT (n = 7–12 per group). Data are the mean6 SD (A–E) or themean6SEM (F ). *P, 0.05, **P, 0.01, ***P, 0.001, †P, 0.05, †††P, 0.001. Two-way ANOVAwas used inA, B, D, and the right panel of F.The Student t test was used in C and the left and middle panels of F.

diabetes.diabetesjournals.org Chen and Associates 1139

Our studies showed that adjustment with body weight onlyslightly reduced the protective effect of Fpr2 deletion onHFD-induced insulin resistance and tissue inflammation,indicating that Fpr2 deficiency improves insulin resistancepartly through reducing body weight gain and mainlythrough inhibiting inflammation.

Macrophage infiltration and polarization toward an M1phenotype are the main drivers of insulin resistance in thecontext of obesity (3). We found that myeloid Fpr2 playedcritical roles in macrophage accumulation in the metabolictissues of DIO mice. This conclusion is supported by thefollowing evidence: First, Fpr2 was highly expressed in

Figure 8—Fpr2 mediates macrophage M1 polarization. WT and Fpr22/2 BMDMs were treated with 100 ng/mL LPS and examined forexpression of M1 macrophage markers (treated for 6 h) (A), production of proinflammatory cytokines in the culture supernatant (treated for9 h) (B), and phosphorylation of NF-kB p65, MAP kinases (ET, exposure time), and TAK1 (treated for 30 min) (C) (n = 3 per group). D–F:BMDMs infected with Fpr2 expressing lentiviruses (Lenti-Fpr2) or control viruses (Lenti-NC) and examined for the expression of M1macrophagemarkers (iNos, inducible nitric oxide synthase) (D) and phosphorylation (p-) of NF-kB p65 andMAP kinases after stimulationwith100 ng/mL LPS for 30 min (E and F ).G–I: BMDMswere stimulated with 400 mmol/L palmitate (Pal) (dissolved in BSA buffer) and examined forexpression of M1-related proinflammatory cytokines (stimulated for 3 h) (G), and phosphorylation of NF-kB p65 andMAP kinases (treated for1 h) (H and I) (n = 3 per group). Values shown represent themean6SD. *P, 0.05, **P, 0.01, ***P, 0.001. Two-wayANOVAwas used inA,B,D, F, G, and I, and two-way repeated-measures ANOVA was used in C. Veh, vehicle.

1140 Fpr2 in Inflammation and Insulin Resistance Diabetes Volume 68, June 2019

adipose tissue SVCs of DIO mice (Fig. 1A). Second, studies ofFpr2 deletion in three independent experiments demon-strated that myeloid Fpr2 contributed to macrophage in-filtration in the metabolic tissues (Figs. 3B–D, 4H–K, and6B–D). Third, obese adipose tissues contained higher levels ofchemotactic ligands, which induced macrophage migration.

Our study for the first time demonstrated that adipo-cyte-released Saa3 induced macrophage chemotaxis in anFpr2-dependent manner. Because Saa3 could stimulatemacrophage migration in a short time (2 h) (Fig. 7B),we propose that Saa3may be a chemotactic ligand for Fpr2.The direct interaction between Saa3 and Fpr2 awaitsfurther verification. In addition, we found that systemic,bone marrow-, and myeloid-specific Fpr2 deletion in HFD-fed mice significantly reduced the expression of Saa3 inWAT and that the reduction of Saa3 was reversed afterrecovering Fpr2 expression in bone marrow cells (Fig. 7F).In obese mice, Saa3 has been reported to be highlyexpressed in the adipose tissue (41) and contributes tomacrophage infiltration (32). Saa3 is upregulated by pal-mitate acid, glucose, LPS, TNF-a, and IL-1b in adipocytes(42–45). Our studies showed that Saa3 induced proin-flammatory cytokines expression in macrophages indepen-dent of Fpr2 (Fig. 7E). We thus hypothesize that Fpr2deficiency alleviates DIO and inflammation, which resultsin the decrease of Saa3 in WAT and in turn blocks itscontribution to WAT inflammation. Mitochondrial pepti-des released from ruptured cells have been reported toinduce phagocyte migration and activation through Fpr2(46). Therefore, Fpr2 ligands released by dead adipocytesand other cells of obese mice may also contribute tomacrophage infiltration in metabolic tissues.

In addition to revealing the contribution of Fpr2 toobesity and macrophage infiltration in metabolic tissuesof DIO mice, another important finding of our study is thatFpr2 could promote macrophage M1 polarization in DIOmice, which was supported by in vivo studies with Fpr22/2

and Fpr2MKO mice as well as mice with BMTs. This con-clusion was further supported by in vitro studies withBMDMs stimulated by LPS, palmitate acid, and IFN-g. First,Fpr2 expression in macrophages was upregulated by LPS(Supplementary Fig. 4A) and IFN-g (47). Second, inhibitionof Fpr2 by antagonist reduced LPS-stimulated proinflam-matory cytokine expression and NF-kB/MAP kinase phos-phorylation (Supplementary Fig. 4B and C). Third, gain- andoff-Fpr2 studies showed that Fpr2 modulated the expres-sion of M1 macrophage marker genes and proinflammatorycytokines and the activation of proinflammatory signalingpathways by LPS, palmitate, and IFN-g (Fig. 8 and Supple-mentary Fig. 4D).

Jablonski et al. (48) recently reported that Fpr2 ex-pression was upregulated during macrophage M1 differ-entiation but downregulated during M2 differentiation.They proposed Fpr2 as a new marker for M1 macrophages(48). These data support our finding that Fpr2 is animportant regulator in macrophage M1 polarization. Inaddition, our studies showed that Fpr2 deletion had

neither effect on hemogram profile and M2 macrophagesin metabolic tissues of obese mice (Supplementary Fig. 5Aand B and Figs. 3E, 4L, and 6E) nor effect on macrophagedifferentiation in vitro (Supplementary Fig. 5C), indicatingthat the effect of Fpr2 deficiency on macrophage M1polarization is not mediated by altering macrophage dif-ferentiation and M2 polarization. Additional studies areneeded to investigate the cross talk between Fpr2 andsignaling pathways involved in macrophage M1 polariza-tion induced by LPS, palmitate, and IFN-g.

Finally, we checked whether other mechanisms are in-volved in the improvement of insulin resistance by Fpr2deletion in obese mice. In vitro studies with 3T3L1 adipo-cytes showed that Fpr2 agonist and antagonist had nosignificant effect on insulin-induced phosphorylation ofInsR, Akt, and GSK3b (Supplementary Fig. 6A). Insulinstimulation of primary adipocytes isolated from WT andFpr2 KO mice consistently induced comparable phosphor-ylation of these proteins (Supplementary Fig. 6B). Theseresults indicate that there is no cross talk between signalingsof Fpr2 and insulin in adipose tissue. In addition, we foundthat depletion of gut microbiota with antibiotics had noeffect on the improvement of obesity and insulin resistanceby Fpr2 deficiency in HFD-fed mice (Supplementary Fig. 7),indicating that Fpr2 deletion regulates glucose and lipidmetabolism in obese mice independent of gut microbiota.

Taken together, our study demonstrates that myeloidFpr2 plays critical roles in DIO and its related complica-tions by modulating energy expenditure as well as in-flammation mediated by macrophage accumulation andM1 polarization in metabolic tissues. Our findings indicatethat myeloid Fpr2 is a potential therapeutic target againstobesity and related metabolic disorders. LXA4 and ANXA1,two anti-inflammatory ligands of Fpr2, have been reportedto attenuate obesity-related inflammation and metabolicdiseases (19–21). Clarifying whether these two moleculesinhibit obesity-related inflammation via activating the biassignaling of Fpr2 will be helpful for developing strategiesagainst obesity-related metabolic disorders.

Acknowledgments. The authors thank Shengzhong Duan (Shanghai JiaoTong University, Shanghai, China) for providing LysM-Cre mice and the animalfacility staff at the Shanghai Institute of Nutrition and Health, Shanghai Institute forBiological Sciences, Chinese Academy of Science, for their support.Funding. This study was supported by grants from the National Key Researchand Development Program of China (2017YFC1601702) and the National NaturalScience Foundation of China (31671232).

The funders had no role in the study design, data collection and analysis,decision to publish, or preparation of the manuscript.Duality of Interest. No potential conflicts of interest relevant to this articlewere reported.Author Contributions. X.C. researched the data and wrote and edited themanuscript. S.Z. contributed to the experimental design. T.Z., P.Y., M.Y., H.M.,N.L., F.M., and S.C. researched the data. J.M.W. provided the Fpr2flox/flox andFpr22/2 mice and contributed to discussion. R.D.Y. and Y.Li contributed todiscussion. Y.Le directed the project, contributed to discussions, and wrote,reviewed, and edited the manuscript. Y.Le is the guarantor of this work and, as

diabetes.diabetesjournals.org Chen and Associates 1141

such, had full access to all the data in the study and takes responsibility for theintegrity of the data and the accuracy of the data analysis.

References1. Després JP, Lemieux I. Abdominal obesity and metabolic syndrome. Nature2006;444:881–8872. Wilmot E, Idris I. Early onset type 2 diabetes: risk factors, clinical impact andmanagement. Ther Adv Chronic Dis 2014;5:234–2443. Saltiel AR, Olefsky JM. Inflammatory mechanisms linking obesity andmetabolic disease. J Clin Invest 2017;127:1–44. Lackey DE, Olefsky JM. Regulation of metabolism by the innate immunesystem. Nat Rev Endocrinol 2016;12:15–285. Murano I, Barbatelli G, Parisani V, et al. Dead adipocytes, detected as crown-like structures, are prevalent in visceral fat depots of genetically obese mice. JLipid Res 2008;49:1562–15686. Olefsky JM, Glass CK. Macrophages, inflammation, and insulin resistance.Annu Rev Physiol 2010;72:219–2467. Gordon S, Martinez FO. Alternative activation of macrophages: mechanismand functions. Immunity 2010;32:593–6048. Lee Y, Ka SO, Cha HN, et al. Myeloid sirtuin 6 deficiency causes insulinresistance in high-fat diet-fed mice by eliciting macrophage polarization toward anM1 phenotype. Diabetes 2017;66:2659–26689. Shin KC, Hwang I, Choe SS, et al. Macrophage VLDLR mediates obesity-inducedinsulin resistance with adipose tissue inflammation. Nat Commun 2017;8:108710. Ye RD, Boulay F, Wang JM, et al. International Union of Basic and ClinicalPharmacology. LXXIII. Nomenclature for the formyl peptide receptor (FPR) family.Pharmacol Rev 2009;61:119–16111. He HQ, Ye RD. The formyl peptide receptors: diversity of ligands andmechanism for recognition. Molecules 2017;22:45512. Cooray SN, Gobbetti T, Montero-Melendez T, et al. Ligand-specific confor-mational change of the G-protein-coupled receptor ALX/FPR2 determines pro-resolving functional responses. Proc Natl Acad Sci U S A 2013;110:18232–1823713. Chen K, Bao Z, Gong W, Tang P, Yoshimura T, Wang JM. Regulation ofinflammation by members of the formyl-peptide receptor family. J Autoimmun2017;85:64–7714. Li Y, Ye D. Molecular biology for formyl peptide receptors in human diseases.J Mol Med (Berl) 2013;91:781–78915. Dufton N, Hannon R, Brancaleone V, et al. Anti-inflammatory role of themurine formyl-peptide receptor 2: ligand-specific effects on leukocyte responsesand experimental inflammation. J Immunol 2010;184:2611–261916. Chen K, Le Y, Liu Y, et al. A critical role for the G protein-coupled receptormFPR2 in airway inflammation and immune responses. J Immunol 2010;184:3331–333517. Chen K, Liu M, Liu Y, et al. Formylpeptide receptor-2 contributes to colonicepithelial homeostasis, inflammation, and tumorigenesis. J Clin Invest 2013;123:1694–170418. Gobbetti T, Coldewey SM, Chen J, et al. Nonredundant protective propertiesof FPR2/ALX in polymicrobial murine sepsis. Proc Natl Acad Sci U S A 2014;111:18685–1869019. Akasheh RT, Pini M, Pang J, Fantuzzi G. Increased adiposity in annexin A1-deficient mice. PLoS One 2013;8:e8260820. Locatelli I, Sutti S, Jindal A, et al. Endogenous annexin A1 is a novel protectivedeterminant in nonalcoholic steatohepatitis in mice. Hepatology 2014;60:531–54421. Börgeson E, Johnson AM, Lee YS, et al. Lipoxin A4 attenuates obesity-induced adipose inflammation and associated liver and kidney disease. CellMetab 2015;22:125–13722. Zhuo S, Yang M, Zhao Y, et al. MicroRNA-451 negatively regulates hepaticglucose production and glucose homeostasis by targeting glycerol kinase-mediated gluconeogenesis. Diabetes 2016;65:3276–328823. Cheng C, Nakamura A, Minamimoto R, et al. Evaluation of organ-specificglucose metabolism by 18F-FDG in insulin receptor substrate-1 (IRS-1) knockoutmice as a model of insulin resistance. Ann Nucl Med 2011;25:755–761

24. Bustin SA, Benes V, Garson JA, et al. The MIQE guidelines: minimum in-formation for publication of quantitative real-time PCR experiments. Clin Chem2009;55:611–62225. Marim FM, Silveira TN, Lima DS Jr., Zamboni DS. A method for generation ofbone marrow-derived macrophages from cryopreserved mouse bone marrowcells. PLoS One 2010;5:e1526326. Kohn AD, Summers SA, Birnbaum MJ, Roth RA. Expression of a constitutivelyactive Akt Ser/Thr kinase in 3T3-L1 adipocytes stimulates glucose uptake andglucose transporter 4 translocation. J Biol Chem 1996;271:31372–3137827. Lumeng CN, Bodzin JL, Saltiel AR. Obesity induces a phenotypic switch inadipose tissue macrophage polarization. J Clin Invest 2007;117:175–18428. Cho KW, Morris DL, Lumeng CN. Flow cytometry analyses of adipose tissuemacrophages. Methods Enzymol 2014;537:297–31429. Mei H, Yao P, Wang S, et al. Chronic low-dose cadmium exposure impairscutaneous wound healing with defective early inflammatory responses after skininjury. Toxicol Sci 2017;159:327–33830. Lesniewski LA, Hosch SE, Neels JG, et al. Bone marrow-specific Cap genedeletion protects against high-fat diet-induced insulin resistance. Nat Med 2007;13:455–46231. Willett WC, Howe GR, Kushi LH. Adjustment for total energy intake in ep-idemiologic studies. Am J Clin Nutr 1997;65(Suppl.):1220S–1228S; discussion1229S–1231S32. Han CY, Subramanian S, Chan CK, et al. Adipocyte-derived serum amyloid A3and hyaluronan play a role in monocyte recruitment and adhesion. Diabetes 2007;56:2260–227333. Takeda K, Akira S. TLR signaling pathways. Semin Immunol 2004;16:3–934. Palmer BF, Clegg DJ. Non-shivering thermogenesis as a mechanism tofacilitate sustainable weight loss. Obes Rev 2017;18:819–83135. Periasamy M, Herrera JL, Reis FCG. Skeletal muscle thermogenesis and itsrole in whole body energy metabolism. Diabetes Metab J 2017;41:327–33636. Valerio A, Cardile A, Cozzi V, et al. TNF-alpha downregulates eNOS ex-pression and mitochondrial biogenesis in fat and muscle of obese rodents. J ClinInvest 2006;116:2791–279837. Arruda AP, Milanski M, Coope A, et al. Low-grade hypothalamic inflammationleads to defective thermogenesis, insulin resistance, and impaired insulin se-cretion. Endocrinology 2011;152:1314–132638. Machado MV, Michelotti GA, Jewell ML, et al. Caspase-2 promotes obesity, themetabolic syndrome and nonalcoholic fatty liver disease. Cell Death Dis 2016;7:e209639. Lee YS, Li P, Huh JY, et al. Inflammation is necessary for long-term but notshort-term high-fat diet-induced insulin resistance. Diabetes 2011;60:2474–248340. Samuel VT, Shulman GI. Mechanisms for insulin resistance: common threadsand missing links. Cell 2012;148:852–87141. Scheja L, Heese B, Zitzer H, et al. Acute-phase serum amyloid A as a markerof insulin resistance in mice. Exp Diabetes Res 2008;2008:23083742. Sanada Y, Yamamoto T, Satake R, et al. Serum amyloid A3 gene expression inadipocytes is an indicator of the interaction with macrophages. Sci Rep 2016;6:3869743. Chiba T, Han CY, Vaisar T, et al. Serum amyloid A3 does not contribute tocirculating SAA levels. J Lipid Res 2009;50:1353–136244. Yeop Han C, Kargi AY, Omer M, et al. Differential effect of saturated andunsaturated free fatty acids on the generation of monocyte adhesion and che-motactic factors by adipocytes: dissociation of adipocyte hypertrophy from in-flammation. Diabetes 2010;59:386–39645. de Oliveira EM, Ascar TP, Silva JC, et al. Serum amyloid A links endotoxaemiato weight gain and insulin resistance in mice. Diabetologia 2016;59:1760–176846. Cattaneo F, Parisi M, Ammendola R. Distinct signaling cascades elicited bydifferent formyl peptide receptor 2 (FPR2) agonists. Int J Mol Sci 2013;14:7193–723047. Chen K, Iribarren P, Huang J, et al. Induction of the formyl peptide receptor2 in microglia by IFN-gamma and synergy with CD40 ligand. J Immunol 2007;178:1759–176648. Jablonski KA, Amici SA, Webb LM, et al. Novel markers to delineate murineM1 and M2 macrophages. PLoS One 2015;10:e0145342

1142 Fpr2 in Inflammation and Insulin Resistance Diabetes Volume 68, June 2019