140 Diabetes Volume 65, January 2016

Stephan Wueest,1,2 Flurin Item,1,2 Fabrizio C. Lucchini,1,2,3 Tenagne D. Challa,1,2 Werner Müller,4 Matthias Blüher,5 and Daniel Konrad1,2,3

Mesenteric Fat Lipolysis Mediates Obesity-Associated Hepatic Steatosis and Insulin Resistance

Diabetes 2016;65:140–148 | DOI: 10.2337/db15-0941

Hepatic steatosis and insulin resistance are among the suggests that -6 (IL-6) contributes to the devel- most prevalent metabolic disorders and are tightly associ- opment of fatty disease and hepatic insulin resistance, ated with obesity and type 2 diabetes. However, the theexactroleofIL-6inthepathogenesisofthesedisordersis underlying mechanisms linking obesity to hepatic lipid thesubjectofintensedebate(2–5). IL-6 is secreted by var- accumulation and insulin resistance are incompletely un- ious cells and tissues such as adipocytes, immune cells, and derstood. Glycoprotein 130 (gp130) is the common signal skeletal muscle, thereby modulating metabolism under both transducer of all (IL-6) . We provide physiological and pathophysiological conditions (2,6). To ac- evidence that gp130-mediated adipose tissue lipolysis tivate its intracellular signaling pathways, IL-6 either binds to promotes hepatic steatosis and insulin resistance. In obese its membrane-bound receptor (mIL-6R; classic signaling) or mice, adipocyte-specific gp130 deletion reduced basal to its soluble receptor (sIL-6; trans-signaling pathway). Of lipolysis and enhanced insulin’s ability to suppress lipolysis note, mIL-6R is expressed in the liver and immune cells from mesenteric but not epididymal adipocytes. Consis- butisabsentinadipocytes(7).Inturn,thisligand/receptor tently, free fatty acid levels were reduced in portal but not in systemic circulation of obese knockout mice. Of note, complex associates with a homodimer of glycoprotein 130 adipocyte-specific gp130 knockout mice were protected (gp130), which is a common signal transducer of all PATHOPHYSIOLOGY from high-fat diet–induced hepatic steatosis as well as IL-6 cytokines (8,9). Subsequently, the /signal – from insulin resistance. In humans, omental but not sub- transducer and activator transcription and extracellular signal cutaneous IL-6 mRNA expression correlated positively with related kinase (ERK) 1/2 pathways are activated (9,10). liver lipid accumulation (r = 0.31, P < 0.05) and negatively with In the liver, chronically elevated circulating IL-6 levels hyperinsulinemic-euglycemic clamp glucose infusion rate have been proposed to induce insulin resistance and (r = 20.28, P < 0.05). The results show that IL-6 cytokine- inflammation in mice (11,12). In contrast, another study induced lipolysis may be restricted to mesenteric white proposed the opposite after finding IL-6 to reduce inflam- adipose tissue and that it contributes to hepatic insulin re- mation and improve insulin sensitivity in the liver (13). sistance and steatosis. Therefore, blocking IL-6 Therefore, the hepatic derangements associated with ele- signaling in (mesenteric) adipocytes may be a novel ap- vated IL-6 levels as observed in the former studies (11,12) proach to blunting detrimental fat-liver crosstalk in obesity. may be mediated through an indirect effect of IL-6 through, for example, induction of adipose tissue (AT) lipolysis. In- The prevalence of obesity and associated diseases such as deed, IL-6 was shown to induce lipolysis in vivo and in vitro insulin resistance and hepatic steatosis are increasing to (14,15), and a recent study suggested that IL-6 promotes epidemic proportions (1), but the underlying pathological hepatic insulin resistance indirectly through increased free mechanisms are not fully understood. Although evidence fatty acid (FFA) release from AT in rodents (16). Of note,

1Division of Pediatric Endocrinology and Diabetology, University Children’s Received 8 July 2015 and accepted 8 September 2015. Hospital, Zurich, Switzerland This article contains Supplementary Data online at http://diabetes 2 ’ ’ Children s Research Centre, University Children s Hospital, Zurich, Switzerland .diabetesjournals.org/lookup/suppl/doi:10.2337/db15-0941/-/DC1. 3Zurich Center for Integrative Human Physiology, University of Zurich, Zurich, © 2016 by the American Diabetes Association. Readers may use this article as Switzerland long as the work is properly cited, the use is educational and not for profit, and 4Faculty of Life Sciences, University of Manchester, Manchester, U.K. the work is not altered. 5Department of Medicine, University of Leipzig, Leipzig, Germany Corresponding author: Daniel Konrad, [email protected], or Stephan Wueest, [email protected]. diabetes.diabetesjournals.org Wueest and Associates 141 obesity-induced IL-6 increase in other studies was higher in 58% calories from fat, 28% from carbohydrates, and 16% mesentericfatthaninotheradipose depots (17,18), suggest- from protein. All protocols conformed to Swiss animal pro- ing that IL-6–induced AT lipolysis plays a distinct role in tection laws and were approved by the cantonal veterinary visceral fat. In the current study, we sought to determine office in Zurich, Switzerland. 1 – whether ) mesenteric AT is more sensitive to IL-6 induced Glucose Clamp Studies lipolysis than epididymal fat and whether 2)IL-6–induced Glucose clamp studies were performed as previously de- FFA release from (mesenteric) fat contributes to obesity- scribed (23). Clamps were performed in freely moving mice. associated hepatic steatosis and insulin resistance, making fi GIR was calculated once glucose infusion reached a more or use of adipocyte-speci c gp130-depleted mice. less constant rate, with blood glucose levels at 5 mmol/L (80–90minafterthestartofinsulininfusion).Thereafter, RESEARCH DESIGN AND METHODS blood glucose level was kept constant at 5 mmol/L for 15– Determination of Total Liver Fat and AT IL-6 mRNA 20 min, and GIR was calculated. The glucose disposal rate 3 Expression in Humans was calculated by dividing the rate of [3- H]glucose infusion 3 fi IL-6 mRNA expression was measured in abdominal omental by the plasma [3- H]glucose-speci c activity (24,25). Endog- and subcutaneous AT samples obtained in parallel from 63 enous glucose production (EGP) during the clamp was calcu- men (n =34)andwomen(n = 29) who underwent open lated by subtracting the GIR from the glucose disposal rate fi abdominal surgery for Roux-en-Y bypass, sleeve gastrectomy, (24,25). To assess tissue-speci c glucose uptake, a bolus 14 explorative laparotomy, or elective cholecystectomy. Liver and (10 mCi) of 2-[1- C]deoxyglucose was administered through AT biopsy specimens were taken during surgery, immediately a catheter at the end of the steady-state period. Blood was snap-frozen in liquid nitrogen, and stored at 280°C until sampled 2, 15, 25, and 35 min after bolus delivery. Area 14 further preparations. Measurement of total body and liver under the curve of disappearing plasma 2-[1- C]deoxyglucose fat, tissue sample handling, and analysis of blood samples, was used together with tissue concentration of phosphory- 14 including measurement of serum adiponectin and leptin con- lated 2-[1- C]deoxyglucose to calculate glucose uptake. centrations, was performed as previously described (19,20). Lipolysis Assays IL-6 plasma concentrations were measured by a high-sensi- Adipocytes were isolated and lipolysis assessed as pre- tivity Human IL-6 Quantikine ELISA Kit (R&D Systems, Ox- viously described (26). Isolated adipocytes were incubated ford, U.K.). Insulin sensitivity was assessed with the in the absence or presence of 100 nmol/L insulin, 100 ng/mL hyperinsulinemic-euglycemic clamp method using a previously recombinant murine IL-6 (R&D Systems), or 1 mmol/L described protocol (21). Glucose infusion rate (GIR) was cal- isoproterenol (Sigma, Buchs, Switzerland) for 1 h. FFA culated from the last 45 min of the clamp during which it levels were measured using the ACS-ACOD-MEHA method could be kept constant to achieve the target plasma glucose (Wako Chemicals GmbH, Neuss, Germany). 6 concentration of 5.5 ( 5%) mmol/L. Therefore, the duration Cell Size Determination of the clamp varied among individuals (range 120–200 min). Cell size of isolated adipocytes was analyzed by a Multi- GIR was normalized to lean body mass. In premenopausal sizer 3 Coulter Counter as previously described (27). women, clamp studies were performed during the luteal phase of the menstrual cycle. All participants gave written Blood Sampling informed consent before taking part in the study. All inves- Mouse abdominal cavity was opened immediately after kill, tigations were approved by the ethics committee of the Uni- and the portal vein was exposed. The portal vein was 3 versity of Leipzig (363-10-13122010 and 017-12-230112) and punctured by a syringe (0.30 mm [30G] 8 mm; BD, were carried out in accordance with the Declaration of Helsinki. Franklin Lakes, NJ), and the blood was collected. Systemic Human IL-6 mRNA expression was measured by quantita- blood was sampled after cardiac puncture using similar fi tive RT-PCR using an Assay-on-Demand gene expression syringes. Blood was added to an Eppendorf tube with a nal (Hs00985639_m1; Applied Biosystems, Darmstadt, Germany), concentration of 5 mmol/L EDTA. After centrifugation, 2 and fluorescence was detected on an ABI PRISM 7000 Se- plasma was stored at 80°C until further processing. quence Detection System (Applied Biosystems). IL-6 mRNA Determination of Plasma Insulin, Triglyceride, and FFA expression was calculated relative to the mRNA expression Levels of HPRT1 mRNA (Hs01003267_m1; Applied Biosystems). Plasma triglyceride (TG) and FFA levels were determined Animals as described elsewhere (28). Plasma insulin levels were Adipocyte-specific gp130 knockout mice (gp130Dadipo) measured using an ELISA kit as previously described (29). on a C57BL/6J background were generated by crossing Western Blotting gp130 floxed (gp130F/F) mice (22) to animals expressing LiverorATsampleswerehomogenizedandWesternblotting the Cre recombinase controlled by the Adipoq promoter performed as previously described (30). The following pri- (AdipoqCre mice; The Jackson Laboratory). Six-week-old mary were used: anti-gp130 and anti–suppressor male mice were fed ad libitum with standard rodent diet of cytokine signaling 3 (SOCS3) (Santa Cruz Biotechnology, (chow) or high-fat diet (HFD) (D12331; Research Diets, Dallas, TX); anti-phosphop38, anti-phosphoERK, and anti- New Brunswick, NJ) for 12 weeks. The HFD comprised ERK (, Danvers, MA); and anti-actin (Millipore, 142 Visceral Lipolysis Mediates Fat-Liver Crosstalk Diabetes Volume 65, January 2016

Billerica, MA). Membranes were analyzed with a Luminescent Histology Image Analyzer running Image Reader software (FujiFilm, Diel- Liver tissues were fixed in 4% buffered formalin and sdorf, Switzerland). embedded in paraffin. Sections were cut and stained with hematoxylin-eosin. RNA Extraction and Quantitative RT-PCR Total RNA was extracted using an RNeasy Lipid Tissue Data Analysis Mini Kit (QIAGEN, Basel, Switzerland), and concentration Statistical analyses were performed using Student t test or was determined spectrophotometrically (NanoDrop 1000; ANOVA with Newman-Keuls post hoc test. In human stud- NanoDrop Instruments, Boston, MA). One microgram of ies, linear relationships were assessed by Spearman RNA was reverse transcribed with SuperScript III Reverse correlation analyses. P , 0.05 was considered significant. Transcriptase (Invitrogen, Basel, Switzerland) using ran- RESULTS dom hexamer primer (Invitrogen). TaqMan (Applied Bio- systems, Rotkreuz, Switzerland) was used for real-time Similar Weight Gain and Adipogenesis in Adipocyte- fi PCR amplification. The following PCR primers (Applied Speci c gp130 Knockout Mice and Control Littermates – Biosystems) were used: -a (TNF-a), To investigate the role of IL-6 induced lipolysis in meta- bolic derangements in vivo, we generated adipocyte-specific Mm00443258_m1; IL-6, Mm00446190_m1; F4/80, Dadipo Mm00802529_m1; CD11b, Mm00434455_m1; fatty acid gp130 knockout mice (gp130 ) using the Cre-Lox sys- tem. As controls, littermate mice with floxed gp130 but synthase (FAS), Mm00662319_m1; peroxisome prolifera- F/F tor–activated receptor-a (PPARa), Mm00627559_m1; absent Cre-recombinase expression were used (gp130 ). SREBP1, Mm00550338_m1; SCD-1, Mm01197142_m1; Although we are aware that knockout of gp130 in adipo- CPT-1, Mm00550438_m1; and AOX, Mm00443579_m1. cytes affects the signaling pathway of all members of the Relative gene expression was obtained after normalization IL-6 cytokines, the mouse model still seems adequate to 2 to 18S RNA (Applied Biosystems) using the equation 2 DDcp investigate our hypothesis because the mIL-6R is not fi (31). expressed in adipocytes (7), precluding cell-speci cdeple- tion of IL-6 signaling in adipocytes similar to liver-specific Liver TG and Total Lipid Determination IL-6R knockout mice (13). As expected, gp130 protein Liver tissue (20–30 mg) was homogenized in PBS, and lipids levels were highly reduced in isolated adipocytes of were extracted in a chloroform-methanol (2:1) mixture. Total gp130Dadipo mice; decreased in white AT (WAT), which liver lipids were determined by a sulfo-phospho-vanillin re- also contains nonadipocyte cell types; and unchanged in action as previously described (32). Liver TG levels were de- skeletal muscle and liver compared with control mice termined in 50 mg of liver tissue according to the method of (Fig. 1A). To induce obesity, 6-week-old mice were either Bligh and Dyer (33) and quantified with an enzymatic assay fed a control chow diet or an HFD for 12 weeks. As (Roche Diagnostics, Rotkreuz, Switzerland). expected, HFD feeding increased body and fat pad weights,

Figure 1—Similar weight gain and adipogenesis in gp130F/F and gp130Dadipo mice. A: Western blot analysis of gp130 protein levels in respective cells and tissues of gp130F/F and gp130Dadipo mice. B: Body weight of chow-fed (●,●) and HFD-fed (■,■) gp130F/F and gp130Dadipo mice (n =5–24). Fat pad weights of epididymal (C) and mesenteric (D) depots of chow- and HFD-fed gp130F/F and gp130Dadipo mice (n =6–18). Diameter of epididymal (E) and mesenteric (F) adipocytes of chow-fed (●,●) and HFD-fed (■,■) gp130F/F and gp130Dadipo mice (n =4–6). Data are mean 6 SEM. *P < 0.05, **P < 0.01, ***P < 0.001 (ANOVA). Δad, gp130Dadipo; Adipo, adipocytes; F/F, gp130F/F; SkM, skeletal muscle. diabetes.diabetesjournals.org Wueest and Associates 143 whereas lack of IL-6 cytokine signaling in adipocytes did not FFA release from mesenteric adipocytes during obesity. affect body weight or adipogenesis as assessed by fat pad Functionality of isolated adipocytes was confirmed by in- weights and adipocyte size (Fig. 1B–F). Likewise, protein creased lipolysis after stimulation with the b-adrenergic levels of PPARg in epididymal and mesenteric WAT were receptor agonist isoproterenol in all depots examined similar in both genotypes (data not shown). Thus, adipocyte- (Supplementary Fig. 1A and B). Furthermore, phosphoryla- specific gp130 deficiency affected neither HFD-induced in- tion of ERK was significantly reduced in mesenteric but not crease in body and fat pad weight nor adipogenesis. in epididymal WAT of obese knockout mice, supporting the notion that IL-6 cytokine-induced lipolysis in vivo is medi- Reduced Basal Lipolysis and Portal FFA Levels ated by ERK (Fig. 2C and D)andconfirming previous find- in Obese gp130Dadipo Mice ings in adipocytes in vitro (15). Ex vivo findings of reduced To assess the impact of IL-6 cytokine signaling on lipolytic lipolysis in portally drained mesenteric adipocytes were con- activity, FFA release was analyzed in adipocytes isolated firmed in vivo, as the obesity-induced increase in circulating from gp130 knockout and control littermates. Compared FFA levels was blunted in portal but not in systemic circu- with chow-fed mice, basal lipolysis was significantly in- lation of knockout mice (Fig. 2E and F). In contrast to creased in epididymal and mesenteric adipocytes of obese lipolysis, gp130 depletion did not affect IL-6 mRNA expres- control mice (Fig. 2A and B). Of note, basal FFA release was sion in mesenteric WAT (Fig. 2G), suggesting similar portal blunted in mesenteric but not in epididymal adipocytes delivery of IL-6 to the liver in both genotypes. In addition, isolated from obese gp130Dadipo mice (Fig. 2A and B), sug- no major change was found in transcript levels of the mac- gesting that IL-6 cytokine signaling contributes to enhanced rophage markers CD11b and F4/80 (34), suggesting similar

Figure 2—Reduced basal lipolysis and portal FFA levels in obese gp130Dadipo mice. Basal FFA release from epididymal (A) and mesenteric (B) adipocytes of chow- and HFD-fed gp130F/F and gp130Dadipo mice (n =5–10). Protein levels of phospho-ERK, ERK, and actin in epididymal (C) and mesenteric (D) WAT of HFD-fed gp130F/F and gp130Dadipo mice (n = 4). FFA concentration was determined in systemic (E) and portal (F) plasma samples of chow- and HFD-fed gp130F/F and gp130Dadipo mice (n =3–8). IL-6 (G) as well as CD11b and F4/80 (H) mRNA expression in mesenteric fat of HFD-fed gp130F/F and gp130Dadipo mice (n = 6). Data are mean 6 SEM. *P < 0.05, **P < 0.01, ***P < 0.001 (Student t test in D or ANOVA in A, B, E, and F). Δad, gp130Dadipo; F/F, gp130F/F; pERK, phospho-ERK. 144 Visceral Lipolysis Mediates Fat-Liver Crosstalk Diabetes Volume 65, January 2016 macrophage content between both genotypes (Fig. 2H). SOCS3 protein levels were not different between groups Taken together, IL-6 cytokine signaling stimulates lipolysis (Fig. 3B). This finding supports the notion of similar portal of portally drained mesenteric but not of systemically delivery of IL-6 to the liver in both gp130Dadipo and control drained epididymal fat. mice. WAT lipolysis is believed to be an important source of FFA leading to hepatic steatosis (39). Accordingly, in Reduced Hepatic Steatosis in HFD-Fed gp130Dadipo humans, up to 59% of FFA bound in liver TGs may arise Mice from plasma FFA (40). In the current study, the observed Mesenteric (and omental) AT is mainly drained through reduction of mesenteric WAT lipolysis in obese gp130Dadipo the portal vein to the liver (35), where FFAs have been mice was associated with a significant decrease in hepatic previously reported to activate the stress kinase p38 mito- steatosis as determined by total liver lipid and TG content gen-activated protein kinase (MAPK) (36,37). Supporting (Fig. 3C and D) as well as by histological examination (Fig. this notion, of the p38 MAPK was re- 3E). As expected, HFD feeding significantly increased he- duced in of obese gp130Dadipo mice (Fig. 3A). In patic lipid accumulation (Fig. 3C and D), and circulating contrast, mRNA expression of SOCS3, which can be in- TG content was notably similar between the genotypes duced in the liver by IL-6, thereby promoting hepatic in- (80.5 6 3.7 mg/dL in HFD-fed gp130F/F mice vs. 84.1 6 sulin resistance (38), was similar in HFD-fed gp130Dadipo 3.8 mg/dL in HFD-fed gp130Dadipo mice, P = 0.5). Besides and control littermates (1.0 6 0.3 in gp130F/F mice vs. FFA uptake, hepatic lipid content is influenced by de novo 1.2 6 0.2 in gp130Dadipo mice, P = 0.7). In addition, lipogenesis and b-oxidation (39–41). Therefore, hepatic

Figure 3—Reduced hepatic steatosis in HFD-fed gp130Dadipo mice. A: Protein levels of phospho-p38 and actin in livers of HFD-fed gp130F/F and gp130Dadipo mice (n = 10). B: Protein levels of SOCS3 and actin in livers of HFD-fed gp130F/F and gp130Dadipo mice (n = 6). C and D: Total liver lipid and liver TG levels of chow- and HFD-fed gp130F/F and gp130Dadipo mice (n =4–6). E: Representative histological hematoxylin-eosin–stained liver sections of HFD-fed gp130F/F and gp130Dadipo mice. F and G: mRNA expression of respective genes in livers of HFD-fed gp130F/F and gp130Dadipo mice (n = 10). Data are mean 6 SEM. *P < 0.05, **P < 0.01, ***P < 0.001 (Student t test in A or ANOVA in C and D). Δad, gp130Dadipo; F/F, gp130F/F. diabetes.diabetesjournals.org Wueest and Associates 145 mRNA expression of enzymes involved in lipogenesis and adipocyte metabolism regarding IL-6 cytokine action, insu- b-oxidation were determined in HFD-fed mice and found lin’s ability to blunt lipolysis was significantly improved in to be similar between genotypes (Fig. 3F). These data fur- mesenteric adipocytes isolated from obese knockout com- ther support the assumption that the reduced FFA flux is pared with control mice (Fig. 4E). responsible for the observed decrease in hepatic lipid accu- mulation of knockout mice. No difference in hepatic mRNA Positive Correlation of Omental IL-6 Expression With levels of IL-6, TNF-a, and F4/80 were noted, suggesting Hepatic Steatosis and Insulin Resistance in Humans a similar degree of hepatic inflammation in knockout and Similar to observations in rodents, IL-6 expression was control mice (Fig. 3G). elevated in visceral/omental compared with subcutaneous fat depots in obese humans (43). Moreover, increased omen- Improved Hepatic Insulin Sensitivity in HFD-Fed tal but not subcutaneous lipolysis was linked to obesity- gp130Dadipo Mice induced fatty liver disease in humans with morbid obesity Steatosis is closely associated with liver insulin resistance (44). To unravel whether IL-6–induced omental fat lipolysis that may also be affected by WAT lipolysis (39). In this may contribute to obesity-associated hepatic steatosis and regard, increased portal delivery of FFA may induce liver insulin resistance, IL-6 mRNA expression was analyzed in insulin resistance through the accumulation of lipid metab- AT of lean (n = 12, BMI 24.5 6 0.2 kg/m2)andoverweight olites and/or hepatic acetyl CoA or directly through activa- or obese (n = 51, BMI 34.4 6 0.8 kg/m2)individualsand tion of p38 MAPK (16,36,37,42). As mentioned previously, correlated to liver fat content. Basic clinical characteristics of phosphorylation of p38 MAPK was decreased in livers of these subjects are shown in Table 1. Supporting previous obese gp130Dadipo mice (Fig. 3A). To assess hepatic insulin findings (43), IL-6 mRNA was increased in omental but not sensitivity, hyperinsulinemic-euglycemic clamp studies in subcutaneous AT in obese compared with lean individuals were performed, revealing significantly improved insulin (Fig. 5A). Of note, we found a significant positive correlation sensitivity in HFD-fed gp130Dadipo mice as determined between omental (r = 0.31, P , 0.05) but not subcutaneous by a higher GIR (Fig. 4A and Supplementary Fig. 2A–C). (r =0.09,P =0.49)IL-6 expression and hepatic steatosis Insulin-induced suppression of EGP (mainly reflecting (Fig. 5B and Supplementary Fig. 3A). In addition, GIR dur- hepatic glucose production) was significantly higher in ing hyperinsulinemic-euglycemic clamp studies revealed HFD-fed gp130Dadipo than in gp130F/F mice (Fig. 4B), asignificant negative correlation with omental but not indicating improved hepatic insulin sensitivity in knock- with subcutaneous IL-6 mRNA expression (Fig. 5C and out mice. In contrast, glucose uptake capacity of skeletal Supplementary Fig. 3B). Of note, a significant correlation muscle was not affected (Fig. 4C). In addition, gp130 between liver fat content and clamp GIR (r = 20.53, P , depletion had no impact on insulin’sabilitytosuppress 0.001) was found. These data support a role for omental FFA release from isolated epididymal adipocytes (Fig. 4D). IL-6 in the development of obesity-associated hepatic stea- In contrast and in support of depot-specific differences in tosis as well as of insulin resistance in humans.

Figure 4—Improved hepatic insulin sensitivity in HFD-fed gp130Dadipo mice. GIR (A), EGP (B), and glucose uptake into quadriceps muscle (C) during hyperinsulinemic-euglycemic clamps in HFD-fed gp130F/F and gp130Dadipo mice (n =4–6). Insulin-inhibited FFA release from epididymal (D) and mesenteric (E) adipocytes of HFD-fed gp130F/F and gp130Dadipo mice (n =8–10). Data are mean 6 SEM. *P < 0.05, **P < 0.01 (Student t test). Δad, gp130Dadipo; F/F, gp130F/F; inh., inhibition; Sk., skeletal. 146 Visceral Lipolysis Mediates Fat-Liver Crosstalk Diabetes Volume 65, January 2016

Table 1—Basal clinical characteristics of human subjects shown to induce FFA release in obesity, thereby contribut- ing to increased hepatic glucose production and, hence, Overweight/ fi Lean (n = 12) obese (n = 51) hepatic insulin resistance (16). The current ndings suggest – Age (years) 62 6 12 70 6 12 that the IL-6 signaling mediated increase in lipolysis is re- stricted to the portally drained mesenteric AT depot be- No. men/women 8/4 26/25 cause depletion of gp130, the common signal transducer No. with type 2 diabetes 0 22 of IL-6, reduced FFA release only from isolated mesenteric 2 BMI (kg/m ) 24.5 6 0.2 34.4 6 0.8*** but not from epididymal adipocytes. Because IL-6 is the Body fat (%) 22.4 6 0.5 36.7 6 1.3*** gp130 cytokine with the best-known lipolytic function (9), Liver fat content (%) 2.0 6 0.1 22.5 6 1.8*** this finding may be explained by an increased sensitivity of Fasting plasma glucose mesenteric adipocytes to IL-6–mediated lipolysis. However, (mmol/L) 5.3 6 0.1 5.7 6 0.1 IL-6–induced FFA release was only slightly elevated in mes- Fasting plasma insulin enteric compared with epididymal adipocytes isolated from (pmol/L) 16.8 6 3.5 204.1 6 19.1*** wild-type mice (1.25 6 0.14-fold in mesenteric vs. 1.12 6 GIR, clamp (mmol/kg/min) 101.2 6 3.3 49.7 6 3.5*** 0.02-fold in epididymal, P = 0.4). Alternatively, a selective

HbA1c (%) 5.28 6 0.05 5.81 6 0.07*** obesity-induced increase in IL-6 content of mesenteric WAT Total cholesterol (mmol/L) 4.42 6 0.19 5.42 6 0.12*** may be responsible for the observed difference (17,18). The latter may be the result of unique endogenous properties of LDL cholesterol (mmol/L) 2.74 6 0.26 3.13 6 0.13 mesenteric WAT and/or depend on the proximity to the gut, 6 6 HDL cholesterol (mmol/L) 1.45 0.1 1.32 0.03 leading to more pronounced inflammation in mesenteric/ TG (mmol/L) 0.9 6 0.12 2.08 6 0.1*** omental compared with other fat depots (35). Along this FFA (mmol/L) 0.38 6 0.06 0.63 6 0.04* line, staphylococcal enterotoxin A was reported to bind to IL-6 (pmol/L) 1.11 6 0.24 3.87 6 0.39** gp130 in adipocytes, thereby interfering with insulin sensi- Adiponectin (ng/mL) 9.27 6 1.25 4.49 6 0.35** tivity and lipolysis (45). Of note, Staphylococcus species is Leptin (ng/mL) 4.75 6 1.15 26.9 6 1.91*** detectable in gut microbiota, and translocation of intestinal bacteria to mesenteric WAT was observed upon high-fat Data are mean 6 SEM unless otherwise indicated. *P , 0.05 between lean and obese subjects. **P , 0.01 between lean and feeding in mice (46,47). Clearly, further studies are needed obese subjects. ***P , 0.001 between lean and obese subjects. to unravel a potential involvement of staphylococcal entero- toxin A in obesity-induced AT lipolysis. Besides IL-6, other cytokines signaling through gp130 DISCUSSION may have contributed to the observed phenotype. Several The current study suggests that IL-6 signaling–mediated IL-6 cytokines were shown to individually affect differentiation lipolysis may be restricted to omental/mesenteric AT and and adipogenesis, rendering gp130 a potential therapeutic contributes to hepatic insulin resistance and steatosis. target to combat obesity and its associated diseases (9). This notion is based on the following findings: 1) Lipolysis Accordingly, it was recently shown that blocking IL-6 trans- from mesenteric but not epididymal AT is reduced in signaling using a soluble gp130 Fc protein reduces obesity- HFD-fed gp130Dadipo mice compared with control littermates, induced infiltration of macrophages into AT (48). We found and 2) gp130 depletion specifically in adipocytes reduces in the current study that lack of IL-6 cytokine signaling in HFD-induced hepatic insulin resistance and steatosis. adipocytes had no impact on fat mass and adipocyte size. IL-6 plays a complex role in inflammation and metab- This observation suggests that IL-6 cytokine signaling does olism, having both beneficial and adverse properties. In not affect adipogenesis. However, Cre expression in our support of its detrimental effects, IL-6 has recently been mousemodelwasunderthecontroloftheadiponectin

Figure 5—Positive correlation of omental IL-6 expression with hepatic steatosis and insulin resistance in humans. A: IL-6 mRNA expression in omental and subcutaneous WAT of lean (n = 12) and obese (n = 51) human subjects. Omental IL-6 mRNA expression correlates with liver fat content (B) and GIR during the steady state of a hyperinsulinemic-euglycemic clamp (C). Data are mean 6 SEM. #P = 0.11 (Student t test). sc, subcutaneous. diabetes.diabetesjournals.org Wueest and Associates 147 promoter and, hence, only induced during late stages of review, and editing of the manuscript. S.W. and D.K. are the guarantors of this adipocyte differentiation because it is located downstream work and, as such, had full access to all the data in the study and take re- of C/EBP (49). Thus, potential effects of individual IL-6 sponsibility for the integrity of the data and the accuracy of the data analysis. cytokines on early adipocyte differentiation (9) may not Prior Presentation. Parts of this study were presented in poster form at the 75th Scientific Sessions of the American Diabetes Association, Boston, MA, be reflected in gp130Dadipo mice. 5–9 June 2015. Chronically elevated circulating IL-6 levels were proposed fl to contribute to obesity-associated hepatic in ammation and References insulinresistanceinmice(11,12).Incontrast,IL-6wasre- fl 1. Lazo M, Hernaez R, Eberhardt MS, et al. Prevalence of nonalcoholic fatty cently reported to reduce in ammation and to improve in- liver disease in the United States: the Third National Health and Nutrition Ex- fi sulin sensitivity in liver (13). The current ndings propose an amination Survey, 1988-1994. Am J Epidemiol 2013;178:38–45 indirect negative impact of IL-6 on hepatic insulin resistance 2. Carey AL, Febbraio MA. Interleukin-6 and insulin sensitivity: friend or foe? and steatosis through induction of AT lipolysis mainly in Diabetologia 2004;47:1135–1142 mesenteric/omental fat depots. Accordingly, depletion of 3. Carter-Kent C, Zein NN, Feldstein AE. Cytokines in the pathogenesis of fatty IL-6 cytokine signaling in adipocytes reduced portal FFA con- liver and disease progression to steatohepatitis: implications for treatment. Am J centration and improved hepatic lipid accumulation as well Gastroenterol 2008;103:1036–1042 as insulin resistance in mice. Moreover, omental (but not 4. Leclercq IA, Da Silva Morais A, Schroyen B, Van Hul N, Geerts A. Insulin subcutaneous) IL-6 expression correlated with hepatic stea- resistance in hepatocytes and sinusoidal liver cells: mechanisms and con- – tosis and insulin resistance in humans. Because increased sequences. J Hepatol 2007;47:142 156 5. Wieckowska A, Papouchado BG, Li Z, Lopez R, Zein NN, Feldstein AE. In- omental (but not subcutaneous) lipolysis has been shown creased hepatic and circulating interleukin-6 levels in human nonalcoholic to be linked to obesity-induced fatty liver disease in humans steatohepatitis. Am J Gastroenterol 2008;103:1372–1379 with morbid obesity (44) and visceral lipolysis may contrib- 6. Ellingsgaard H, Ehses JA, Hammar EB, et al. Interleukin-6 regulates pan- ute significantly to the portal delivery of FFA to the liver creatic alpha-cell mass expansion. Proc Natl Acad Sci U S A 2008;105:13163– (50), the mechanism we propose may be conserved between 13168 species. Besides reduced lipolysis, lower intestinal absorption 7. Wolf J, Rose-John S, Garbers C. Interleukin-6 and its receptors: a highly of FFA (35) or increased fatty acid uptake by adipocytes may regulated and dynamic system. Cytokine 2014;70:11–20 have contributed to decreased portal FFA levels in HFD-fed 8. Febbraio MA. gp130 receptor ligands as potential therapeutic targets for gp130Dadipo mice. Although the current data suggest that obesity. J Clin Invest 2007;117:841–849 increased mesenteric FFA release into the portal vein impairs 9. White UA, Stephens JM. The gp130 receptor cytokine family: regulators of – liver metabolism, such a direct link may be questioned (51). adipocyte development and function. Curr Pharm Des 2011;17:340 346 10. Yeoh GC, Ernst M, Rose-John S, et al. Opposing roles of gp130-mediated Changed release of other non-FFA factors into portal and/or STAT-3 and ERK-1/ 2 signaling in liver progenitor cell migration and proliferation. systemic circulation may have also contributed to the ob- Hepatology 2007;45:486–494 Dadipo served hepatic phenotype in obese gp130 mice. 11. Franckhauser S, Elias I, Rotter Sopasakis V, et al. Overexpression of Il6 In conclusion, the obesity-associated rise in mesenteric/ leads to hyperinsulinaemia, liver inflammation and reduced body weight in mice. omental AT IL-6 cytokine signaling promotes FFA release Diabetologia 2008;51:1306–1316 into the portal circulation, inducing hepatic steatosis and/or 12. Klover PJ, Zimmers TA, Koniaris LG, Mooney RA. Chronic exposure to interleukin- insulin resistance. Hence, blocking IL-6 cytokine signaling in 6 causes hepatic insulin resistance in mice. Diabetes 2003;52:2784–2789 adipocytes may be a novel approach to blunting detrimental 13. Wunderlich FT, Ströhle P, Könner AC, et al. Interleukin-6 signaling in liver- fat-liver crosstalk in obesity. To this end, the development parenchymal cells suppresses hepatic inflammation and improves systemic in- – of cell-specific adeno-associated virus vectors may be a sulin action. Cell Metab 2010;12:237 249 promising tool in the future (52,53). 14. van Hall G, Steensberg A, Sacchetti M, et al. Interleukin-6 stimulates lipolysis and fat oxidation in humans. J Clin Endocrinol Metab 2003;88:3005–3010 15. Wueest S, Item F, Boyle CN, et al. Interleukin-6 contributes to early fasting- induced free fatty acid mobilization in mice. Am J Physiol Regul Integr Comp Acknowledgments. The authors thank Eugen Schoenle (University Children’s Physiol 2014;306:R861–R867 Hospital, Zurich, Switzerland) and Giatgen Spinas (University Hospital, Zurich, 16. Perry RJ, Camporez JP, Kursawe R, et al. Hepatic acetyl CoA links adipose Switzerland) for continuous support and Alexandra Grob (University Children’s tissue inflammation to hepatic insulin resistance and type 2 diabetes. Cell 2015; Hospital, Zurich, Switzerland) for help with mouse breeding and genotyping. 160:745–758 Funding. This work was supported by a grant from Deutsche Forschungsge- 17. Lam YY, Ha CW, Campbell CR, et al. Increased gut permeability and mi- meinschaft (SFB 1052/1, “Obesity Mechanisms,” to M.B.) and research grants crobiota change associate with mesenteric fat inflammation and metabolic from the Swiss National Science Foundation (# 310030_160129 to D.K.) and the dysfunction in diet-induced obese mice. PLoS One 2012;7:e34233 Olga Mayenfisch Stiftung, Zurich (to S.W.). 18. Li H, Lelliott C, Håkansson P, et al. Intestinal, adipose, and liver in- Duality of Interest. No potential conflicts of interest relevant to this article flammation in diet-induced obese mice. Metabolism 2008;57:1704–1710 were reported. 19. de Guia RM, Rose AJ, Sommerfeld A, et al. microRNA-379 couples glu- Author Contributions. S.W. contributed to the study concept, experi- cocorticoid hormones to dysfunctional lipid homeostasis. EMBO J 2015;34:344– mental work, discussion, and writing, review, and editing of the manuscript. F.I., 360 F.C.L., T.D.C., and M.B. contributed to the experimental work, discussion, and 20. Klöting N, Fasshauer M, Dietrich A, et al. Insulin-sensitive obesity. Am J review and editing of the manuscript. W.M. provided the gp130Dadipo mice, gave Physiol Endocrinol Metab 2010;299:E506–E515 conceptual advice, and contributed to the discussion and review and editing of 21. Blüher M, Unger R, Rassoul F, Richter V, Paschke R. Relation between the manuscript. D.K. contributed to the study concept, discussion, and writing, glycaemic control, hyperinsulinaemia and plasma concentrations of soluble 148 Visceral Lipolysis Mediates Fat-Liver Crosstalk Diabetes Volume 65, January 2016

adhesion molecules in patients with impaired glucose tolerance or type II di- 37. Liu HY, Collins QF, Xiong Y, et al. Prolonged treatment of primary hepato- abetes. Diabetologia 2002;45:210–216 cytes with oleate induces insulin resistance through p38 mitogen-activated 22. Betz UA, Bloch W, van den Broek M, et al. Postnatally induced inactivation of protein kinase. J Biol Chem 2007;282:14205–14212 gp130 in mice results in neurological, cardiac, hematopoietic, immunological, 38. Senn JJ, Klover PJ, Nowak IA, et al. Suppressor of cytokine signaling-3 hepatic, and pulmonary defects. J Exp Med 1998;188:1955–1965 (SOCS-3), a potential mediator of interleukin-6-dependent insulin resistance in 23. Chin SH, Item F, Wueest S, et al. Opposing effects of reduced kidney mass hepatocytes. J Biol Chem 2003;278:13740–13746 on liver and skeletal muscle insulin sensitivity in obese mice. Diabetes 2015;64: 39. Perry RJ, Samuel VT, Petersen KF, Shulman GI. The role of hepatic lipids in 1131–1141 hepatic insulin resistance and type 2 diabetes. Nature 2014;510:84–91 24. Fisher SJ, Kahn CR. Insulin signaling is required for insulin’sdirect 40. Donnelly KL, Smith CI, Schwarzenberg SJ, Jessurun J, Boldt MD, Parks EJ. and indirect action on hepatic glucose production. J Clin Invest 2003;111: Sources of fatty acids stored in liver and secreted via lipoproteins in patients with 463–468 nonalcoholic fatty liver disease. J Clin Invest 2005;115:1343–1351 25. Kim JK, Michael MD, Previs SF, et al. Redistribution of substrates to adipose 41. Kawano Y, Cohen DE. Mechanisms of hepatic triglyceride accumulation in tissue promotes obesity in mice with selective insulin resistance in muscle. J Clin non-alcoholic fatty liver disease. J Gastroenterol 2013;48:434–441 Invest 2000;105:1791–1797 42. Summers SA. Ceramides in insulin resistance and lipotoxicity. Prog Lipid 26. Wueest S, Yang X, Liu J, Schoenle EJ, Konrad D. Inverse regulation of basal Res 2006;45:42–72 lipolysis in perigonadal and mesenteric fat depots in mice. Am J Physiol Endo- 43. Fried SK, Bunkin DA, Greenberg AS. Omental and subcutaneous adipose crinol Metab 2012;302:E153–E160 tissues of obese subjects release interleukin-6: depot difference and regulation 27. Wiedemann MS, Wueest S, Grob A, Item F, Schoenle EJ, Konrad D. Short- by glucocorticoid. J Clin Endocrinol Metab 1998;83:847–850 term HFD does not alter lipolytic function of adipocytes. Adipocyte 2014;3:115– 44. Thörne A, Löfgren P, Hoffstedt J. Increased visceral adipocyte lipolysis– 120 a pathogenic role in nonalcoholic fatty liver disease? J Clin Endocrinol Metab 28. Wueest S, Rapold RA, Schumann DM, et al. Deletion of Fas in adipocytes 2010;95:E209–E213 relieves adipose tissue inflammation and hepatic manifestations of obesity in 45. Banke E, Rödström K, Ekelund M, et al. Superantigen activates the gp130 mice. J Clin Invest 2010;120:191–202 receptor on adipocytes resulting in altered adipocyte metabolism. Metabolism 29. Konrad D, Rudich A, Schoenle EJ. Improved glucose tolerance in mice 2014;63:831–840 receiving intraperitoneal transplantation of normal fat tissue. Diabetologia 2007; 46. Amar J, Chabo C, Waget A, et al. Intestinal mucosal adherence and 50:833–839 translocation of commensal bacteria at the early onset of type 2 diabetes: mo- 30. Wueest S, Rapold RA, Rytka JM, Schoenle EJ, Konrad D. Basal lipolysis, not lecular mechanisms and probiotic treatment. EMBO Mol Med 2011;3:559–572 the degree of insulin resistance, differentiates large from small isolated adipo- 47. Bervoets L, Van Hoorenbeeck K, Kortleven I, et al. Differences in gut mi- cytes in high-fat fed mice. Diabetologia 2009;52:541–546 crobiota composition between obese and lean children: a cross-sectional study. 31. Pfaffl MW. A new mathematical model for relative quantification in real-time Gut Pathog 2013;5:10 RT-PCR. Nucleic Acids Res 2001;29:e45 48. Kraakman MJ, Kammoun HL, Allen TL, et al. Blocking IL-6 trans-signaling 32. Knight JA, Anderson S, Rawle JM. Chemical basis of the sulfo-phospho- prevents high-fat diet-induced adipose tissue macrophage recruitment but does vanillin reaction for estimating total serum lipids. Clin Chem 1972;18:199–202 not improve insulin resistance. Cell Metab 2015;21:403–416 33. Bligh EG, Dyer WJ. A rapid method of total lipid extraction and purification. 49. Park SK, Oh SY, Lee MY, Yoon S, Kim KS, Kim JW. CCAAT/enhancer binding Can J Biochem Physiol 1959;37:911–917 protein and nuclear factor-Y regulate adiponectin gene expression in adipose 34. Nguyen MT, Favelyukis S, Nguyen AK, et al. A subpopulation of macro- tissue. Diabetes 2004;53:2757–2766 phages infiltrates hypertrophic adipose tissue and is activated by free fatty acids 50. Nielsen S, Guo Z, Johnson CM, Hensrud DD, Jensen MD. Splanchnic li- via Toll-like receptors 2 and 4 and JNK-dependent pathways. J Biol Chem 2007; polysis in human obesity. J Clin Invest 2004;113:1582–1588 282:35279–35292 51. Frayn KN. Visceral fat and insulin resistance–causative or correlative? Br J 35. Konrad D, Wueest S. The gut-adipose-liver axis in the metabolic syndrome. Nutr 2000;83(Suppl. 1):S71–S77 Physiology (Bethesda) 2014;29:304–313 52. O’Neill SM, Hinkle C, Chen SJ, et al. Targeting adipose tissue via systemic 36. Gao D, Nong S, Huang X, et al. The effects of palmitate on hepatic insulin gene therapy. Gene Ther 2014;21:653–661 resistance are mediated by NADPH oxidase 3-derived reactive oxygen species 53. Uhrig-Schmidt S, Geiger M, Luippold G, et al. Gene delivery to adipose through JNK and p38MAPK pathways. J Biol Chem 2010;285:29965–29973 tissue using transcriptionally targeted rAAV8 vectors. PLoS One 2014;9:e116288