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

The prevalence of obesity and associated diseases such as insulin resistance and hepatic steatosis are increasing to epidemic proportions (1), but the underlying pathological mechanisms are not fully understood. Although evidence suggests that interleukin-6 (IL-6) contributes to the development of fatty liver disease and hepatic insulin resistance, the exact role of IL-6 in the pathogenesis of these disorders is the subject of intense debate (2–5). IL-6 is secreted by various cells and tissues such as adipocytes, immune cells, and skeletal muscle, thereby modulating metabolism under both physiological and pathophysiological conditions (2,6). To activate its intracellular signaling pathways, IL-6 either binds to its membrane-bound receptor (mIL-6R; classic signaling) or to its soluble receptor (sIL-6; trans-signaling pathway). Of note, mIL-6R is expressed in the liver and immune cells but is absent in adipocytes (7). In turn, this ligand/receptor complex associates with a homodimer of glycoprotein 130 (gp130), which is a common signal transducer protein of all IL-6 cytokines (8,9). Subsequently, the Janus kinase/signal transducer and activator transcription and extracellular signal–related kinase (ERK) 1/2 pathways are activated (9,10).

In the liver, chronically elevated circulating IL-6 levels have been proposed to induce insulin resistance and inflammation in mice (11,12). In contrast, another study proposed the opposite after finding IL-6 to reduce inflammation and improve insulin sensitivity in the liver (13). Therefore, the hepatic derangements associated with elevated IL-6 levels as observed in the former studies (11,12) may be mediated through an indirect effect of IL-6 through, for example, induction of adipose tissue (AT) lipolysis. Indeed, IL-6 was shown to induce lipolysis in vivo and in vitro (14,15), and a recent study suggested that IL-6 promotes hepatic insulin resistance indirectly through increased free fatty acid (FFA) release from AT in rodents (16). Of note, obesity-induced IL-6 increase in other studies was higher in mesenteric fat than in other adipose depots (17,18), suggesting that IL-6–induced AT lipolysis plays a distinct role in visceral fat. In the current study, we sought to determine whether 1) mesenteric AT is more sensitive to IL-6–induced lipolysis than epididymal fat and whether 2) IL-6–induced FFA release from (mesenteric) fat contributes to obesity-associated hepatic steatosis and insulin resistance, making use of adipocyte-specific gp130-depleted mice.

IL-6 mRNA expression was measured in abdominal omental and subcutaneous AT samples obtained in parallel from 63 men (n = 34) and women (n = 29) who underwent open abdominal surgery for Roux-en-Y bypass, sleeve gastrectomy, explorative laparotomy, or elective cholecystectomy. Liver and AT biopsy specimens were taken during surgery, immediately snap-frozen in liquid nitrogen, and stored at −80°C until further preparations. Measurement of total body and liver fat, tissue sample handling, and analysis of blood samples, including measurement of serum adiponectin and leptin concentrations, was performed as previously described (19,20). IL-6 plasma concentrations were measured by a high-sensitivity Human IL-6 Quantikine ELISA Kit (R&D Systems, Oxford, U.K.). Insulin sensitivity was assessed with the hyperinsulinemic-euglycemic clamp method using a previously described protocol (21). Glucose infusion rate (GIR) was calculated from the last 45 min of the clamp during which it could be kept constant to achieve the target plasma glucose concentration of 5.5 (± 5%) mmol/L. Therefore, the duration of the clamp varied among individuals (range 120–200 min). GIR was normalized to lean body mass. In premenopausal women, clamp studies were performed during the luteal phase of the menstrual cycle. All participants gave written informed consent before taking part in the study. All investigations were approved by the ethics committee of the University of Leipzig (363-10-13122010 and 017-12-230112) and were carried out in accordance with the Declaration of Helsinki.

Human IL-6 mRNA expression was measured by quantitative RT-PCR using an Assay-on-Demand gene expression kit (Hs00985639_m1; Applied Biosystems, Darmstadt, Germany), and fluorescence was detected on an ABI PRISM 7000 Sequence Detection System (Applied Biosystems). IL-6 mRNA expression was calculated relative to the mRNA expression of HPRT1 mRNA (Hs01003267_m1; Applied Biosystems).

Adipocyte-specific gp130 knockout mice (gp130^Δadipo) on a C57BL/6J background were generated by crossing gp130 floxed (gp130^F/F) mice (22) to animals expressing the Cre recombinase controlled by the Adipoq promoter (AdipoqCre mice; The Jackson Laboratory). Six-week-old male mice were fed ad libitum with standard rodent diet (chow) or high-fat diet (HFD) (D12331; Research Diets, New Brunswick, NJ) for 12 weeks. The HFD comprised 58% calories from fat, 28% from carbohydrates, and 16% from protein. All protocols conformed to Swiss animal protection laws and were approved by the cantonal veterinary office in Zurich, Switzerland.

Glucose clamp studies were performed as previously described (23). Clamps were performed in freely moving mice. GIR was calculated once glucose infusion reached a more or less constant rate, with blood glucose levels at 5 mmol/L (80–90 min after the start of insulin infusion). Thereafter, blood glucose level was kept constant at 5 mmol/L for 15–20 min, and GIR was calculated. The glucose disposal rate was calculated by dividing the rate of [3-³H]glucose infusion by the plasma [3-³H]glucose-specific activity (24,25). Endogenous glucose production (EGP) during the clamp was calculated by subtracting the GIR from the glucose disposal rate (24,25). To assess tissue-specific glucose uptake, a bolus (10 μCi) of 2-[1-¹⁴C]deoxyglucose was administered through a catheter at the end of the steady-state period. Blood was sampled 2, 15, 25, and 35 min after bolus delivery. Area under the curve of disappearing plasma 2-[1-¹⁴C]deoxyglucose was used together with tissue concentration of phosphorylated 2-[1-¹⁴C]deoxyglucose to calculate glucose uptake.

Adipocytes were isolated and lipolysis assessed as previously described (26). Isolated adipocytes were incubated in the absence or presence of 100 nmol/L insulin, 100 ng/mL recombinant murine IL-6 (R&D Systems), or 1 μmol/L isoproterenol (Sigma, Buchs, Switzerland) for 1 h. FFA levels were measured using the ACS-ACOD-MEHA method (Wako Chemicals GmbH, Neuss, Germany).

Cell size of isolated adipocytes was analyzed by a Multisizer 3 Coulter Counter as previously described (27).

Mouse abdominal cavity was opened immediately after kill, and the portal vein was exposed. The portal vein was punctured by a syringe (0.30 mm [30G] × 8 mm; BD, Franklin Lakes, NJ), and the blood was collected. Systemic blood was sampled after cardiac puncture using similar syringes. Blood was added to an Eppendorf tube with a final concentration of 5 mmol/L EDTA. After centrifugation, plasma was stored at −80°C until further processing.

Plasma triglyceride (TG) and FFA levels were determined as described elsewhere (28). Plasma insulin levels were measured using an ELISA kit as previously described (29).

Liver or AT samples were homogenized and Western blotting performed as previously described (30). The following primary antibodies were used: anti-gp130 and anti–suppressor of cytokine signaling 3 (SOCS3) (Santa Cruz Biotechnology, Dallas, TX); anti-phosphop38, anti-phosphoERK, and anti-ERK (Cell Signaling, Danvers, MA); and anti-actin (Millipore, Billerica, MA). Membranes were analyzed with a Luminescent Image Analyzer running Image Reader software (FujiFilm, Dielsdorf, Switzerland).

Total RNA was extracted using an RNeasy Lipid Tissue Mini Kit (QIAGEN, Basel, Switzerland), and concentration was determined spectrophotometrically (NanoDrop 1000; NanoDrop Instruments, Boston, MA). One microgram of RNA was reverse transcribed with SuperScript III Reverse Transcriptase (Invitrogen, Basel, Switzerland) using random hexamer primer (Invitrogen). TaqMan (Applied Biosystems, Rotkreuz, Switzerland) was used for real-time PCR amplification. The following PCR primers (Applied Biosystems) were used: tumor necrosis factor-α (TNF-α), Mm00443258_m1; IL-6, Mm00446190_m1; F4/80, Mm00802529_m1; CD11b, Mm00434455_m1; fatty acid synthase (FAS), Mm00662319_m1; peroxisome proliferator–activated receptor-α (PPARα), Mm00627559_m1; SREBP1, Mm00550338_m1; SCD-1, Mm01197142_m1; CPT-1, Mm00550438_m1; and AOX, Mm00443579_m1. Relative gene expression was obtained after normalization to 18S RNA (Applied Biosystems) using the equation 2^−ΔΔcp (31).

Liver tissue (20–30 mg) was homogenized in PBS, and lipids were extracted in a chloroform-methanol (2:1) mixture. Total liver lipids were determined by a sulfo-phospho-vanillin reaction as previously described (32). Liver TG levels were determined in 50 mg of liver tissue according to the method of Bligh and Dyer (33) and quantified with an enzymatic assay (Roche Diagnostics, Rotkreuz, Switzerland).

Liver tissues were fixed in 4% buffered formalin and embedded in paraffin. Sections were cut and stained with hematoxylin-eosin.

Statistical analyses were performed using Student t test or ANOVA with Newman-Keuls post hoc test. In human studies, linear relationships were assessed by Spearman rank correlation analyses. P < 0.05 was considered significant.

To investigate the role of IL-6–induced lipolysis in metabolic derangements in vivo, we generated adipocyte-specific gp130 knockout mice (gp130^Δadipo) using the Cre-Lox system. As controls, littermate mice with floxed gp130 but absent Cre-recombinase expression were used (gp130^F/F). Although we are aware that knockout of gp130 in adipocytes affects the signaling pathway of all members of the IL-6 cytokines, the mouse model still seems adequate to investigate our hypothesis because the mIL-6R is not expressed in adipocytes (7), precluding cell-specific depletion of IL-6 signaling in adipocytes similar to liver-specific IL-6R knockout mice (13). As expected, gp130 protein levels were highly reduced in isolated adipocytes of gp130^Δadipo mice; decreased in white AT (WAT), which also contains nonadipocyte cell types; and unchanged in skeletal muscle and liver compared with control mice (Fig. 1A). To induce obesity, 6-week-old mice were either fed a control chow diet or an HFD for 12 weeks. As expected, HFD feeding increased body and fat pad weights, whereas lack of IL-6 cytokine signaling in adipocytes did not affect body weight or adipogenesis as assessed by fat pad weights and adipocyte size (Fig. 1B–F). Likewise, protein levels of PPARγ in epididymal and mesenteric WAT were similar in both genotypes (data not shown). Thus, adipocyte-specific gp130 deficiency affected neither HFD-induced increase in body and fat pad weight nor adipogenesis.

To assess the impact of IL-6 cytokine signaling on lipolytic activity, FFA release was analyzed in adipocytes isolated from gp130 knockout and control littermates. Compared with chow-fed mice, basal lipolysis was significantly increased in epididymal and mesenteric adipocytes of obese control mice (Fig. 2A and B). Of note, basal FFA release was blunted in mesenteric but not in epididymal adipocytes isolated from obese gp130^Δadipo mice (Fig. 2A and B), suggesting that IL-6 cytokine signaling contributes to enhanced FFA release from mesenteric adipocytes during obesity. Functionality of isolated adipocytes was confirmed by increased lipolysis after stimulation with the β-adrenergic receptor agonist isoproterenol in all depots examined (Supplementary Fig. 1A and B). Furthermore, phosphorylation of ERK was significantly reduced in mesenteric but not in epididymal WAT of obese knockout mice, supporting the notion that IL-6 cytokine-induced lipolysis in vivo is mediated by ERK (Fig. 2C and D) and confirming previous findings in adipocytes in vitro (15). Ex vivo findings of reduced lipolysis in portally drained mesenteric adipocytes were confirmed in vivo, as the obesity-induced increase in circulating FFA levels was blunted in portal but not in systemic circulation of knockout mice (Fig. 2E and F). In contrast to lipolysis, gp130 depletion did not affect IL-6 mRNA expression in mesenteric WAT (Fig. 2G), suggesting similar portal delivery of IL-6 to the liver in both genotypes. In addition, no major change was found in transcript levels of the macrophage markers CD11b and F4/80 (34), suggesting similar macrophage content between both genotypes (Fig. 2H). Taken together, IL-6 cytokine signaling stimulates lipolysis of portally drained mesenteric but not of systemically drained epididymal fat.

Mesenteric (and omental) AT is mainly drained through the portal vein to the liver (35), where FFAs have been previously reported to activate the stress kinase p38 mitogen-activated protein kinase (MAPK) (36,37). Supporting this notion, phosphorylation of the p38 MAPK was reduced in livers of obese gp130^Δadipo mice (Fig. 3A). In contrast, mRNA expression of SOCS3, which can be induced in the liver by IL-6, thereby promoting hepatic insulin resistance (38), was similar in HFD-fed gp130^Δadipo and control littermates (1.0 ± 0.3 in gp130^F/F mice vs. 1.2 ± 0.2 in gp130^Δadipo mice, P = 0.7). In addition, SOCS3 protein levels were not different between groups (Fig. 3B). This finding supports the notion of similar portal delivery of IL-6 to the liver in both gp130^Δadipo and control mice. WAT lipolysis is believed to be an important source of FFA leading to hepatic steatosis (39). Accordingly, in humans, up to 59% of FFA bound in liver TGs may arise from plasma FFA (40). In the current study, the observed reduction of mesenteric WAT lipolysis in obese gp130^Δadipo mice was associated with a significant decrease in hepatic steatosis as determined by total liver lipid and TG content (Fig. 3C and D) as well as by histological examination (Fig. 3E). As expected, HFD feeding significantly increased hepatic lipid accumulation (Fig. 3C and D), and circulating TG content was notably similar between the genotypes (80.5 ± 3.7 mg/dL in HFD-fed gp130^F/F mice vs. 84.1 ± 3.8 mg/dL in HFD-fed gp130^Δadipo mice, P = 0.5). Besides FFA uptake, hepatic lipid content is influenced by de novo lipogenesis and β-oxidation (39–41). Therefore, hepatic mRNA expression of enzymes involved in lipogenesis and β-oxidation were determined in HFD-fed mice and found to be similar between genotypes (Fig. 3F). These data further support the assumption that the reduced FFA flux is responsible for the observed decrease in hepatic lipid accumulation of knockout mice. No difference in hepatic mRNA levels of IL-6, TNF-α, and F4/80 were noted, suggesting a similar degree of hepatic inflammation in knockout and control mice (Fig. 3G).

Steatosis is closely associated with liver insulin resistance that may also be affected by WAT lipolysis (39). In this regard, increased portal delivery of FFA may induce liver insulin resistance through the accumulation of lipid metabolites and/or hepatic acetyl CoA or directly through activation of p38 MAPK (16,36,37,42). As mentioned previously, phosphorylation of p38 MAPK was decreased in livers of obese gp130^Δadipo mice (Fig. 3A). To assess hepatic insulin sensitivity, hyperinsulinemic-euglycemic clamp studies were performed, revealing significantly improved insulin sensitivity in HFD-fed gp130^Δadipo mice as determined by a higher GIR (Fig. 4A and Supplementary Fig. 2A–C). Insulin-induced suppression of EGP (mainly reflecting hepatic glucose production) was significantly higher in HFD-fed gp130^Δadipo than in gp130^F/F mice (Fig. 4B), indicating improved hepatic insulin sensitivity in knockout mice. In contrast, glucose uptake capacity of skeletal muscle was not affected (Fig. 4C). In addition, gp130 depletion had no impact on insulin’s ability to suppress FFA release from isolated epididymal adipocytes (Fig. 4D). In contrast and in support of depot-specific differences in adipocyte metabolism regarding IL-6 cytokine action, insulin’s ability to blunt lipolysis was significantly improved in mesenteric adipocytes isolated from obese knockout compared with control mice (Fig. 4E).

Similar to observations in rodents, IL-6 expression was elevated in visceral/omental compared with subcutaneous fat depots in obese humans (43). Moreover, increased omental but not subcutaneous lipolysis was linked to obesity-induced fatty liver disease in humans with morbid obesity (44). To unravel whether IL-6–induced omental fat lipolysis may contribute to obesity-associated hepatic steatosis and insulin resistance, IL-6 mRNA expression was analyzed in AT of lean (n = 12, BMI 24.5 ± 0.2 kg/m²) and overweight or obese (n = 51, BMI 34.4 ± 0.8 kg/m²) individuals and correlated to liver fat content. Basic clinical characteristics of these subjects are shown in Table 1. Supporting previous findings (43), IL-6 mRNA was increased in omental but not in subcutaneous AT in obese compared with lean individuals (Fig. 5A). Of note, we found a significant positive correlation between omental (r = 0.31, P < 0.05) but not subcutaneous (r = 0.09, P = 0.49) IL-6 expression and hepatic steatosis (Fig. 5B and Supplementary Fig. 3A). In addition, GIR during hyperinsulinemic-euglycemic clamp studies revealed a significant negative correlation with omental but not with subcutaneous IL-6 mRNA expression (Fig. 5C and Supplementary Fig. 3B). Of note, a significant correlation between liver fat content and clamp GIR (r = −0.53, P < 0.001) was found. These data support a role for omental IL-6 in the development of obesity-associated hepatic steatosis as well as of insulin resistance in humans.

The current study suggests that IL-6 signaling–mediated lipolysis may be restricted to omental/mesenteric AT and contributes to hepatic insulin resistance and steatosis. This notion is based on the following findings: 1) Lipolysis from mesenteric but not epididymal AT is reduced in HFD-fed gp130^Δadipo mice compared with control littermates, and 2) gp130 depletion specifically in adipocytes reduces HFD-induced hepatic insulin resistance and steatosis.

IL-6 plays a complex role in inflammation and metabolism, having both beneficial and adverse properties. In support of its detrimental effects, IL-6 has recently been shown to induce FFA release in obesity, thereby contributing to increased hepatic glucose production and, hence, hepatic insulin resistance (16). The current findings suggest that the IL-6 signaling–mediated increase in lipolysis is restricted to the portally drained mesenteric AT depot because depletion of gp130, the common signal transducer of IL-6, reduced FFA release only from isolated mesenteric but not from epididymal adipocytes. Because IL-6 is the gp130 cytokine with the best-known lipolytic function (9), this finding may be explained by an increased sensitivity of mesenteric adipocytes to IL-6–mediated lipolysis. However, IL-6–induced FFA release was only slightly elevated in mesenteric compared with epididymal adipocytes isolated from wild-type mice (1.25 ± 0.14-fold in mesenteric vs. 1.12 ± 0.02-fold in epididymal, P = 0.4). Alternatively, a selective obesity-induced increase in IL-6 content of mesenteric WAT may be responsible for the observed difference (17,18). The latter may be the result of unique endogenous properties of mesenteric WAT and/or depend on the proximity to the gut, leading to more pronounced inflammation in mesenteric/omental compared with other fat depots (35). Along this line, staphylococcal enterotoxin A was reported to bind to gp130 in adipocytes, thereby interfering with insulin sensitivity and lipolysis (45). Of note, Staphylococcus species is detectable in gut microbiota, and translocation of intestinal bacteria to mesenteric WAT was observed upon high-fat feeding in mice (46,47). Clearly, further studies are needed to unravel a potential involvement of staphylococcal enterotoxin A in obesity-induced AT lipolysis.

Besides IL-6, other cytokines signaling through gp130 may have contributed to the observed phenotype. Several IL-6 cytokines were shown to individually affect differentiation and adipogenesis, rendering gp130 a potential therapeutic target to combat obesity and its associated diseases (9). Accordingly, it was recently shown that blocking IL-6 trans-signaling using a soluble gp130 Fc protein reduces obesity-induced infiltration of macrophages into AT (48). We found in the current study that lack of IL-6 cytokine signaling in adipocytes had no impact on fat mass and adipocyte size. This observation suggests that IL-6 cytokine signaling does not affect adipogenesis. However, Cre expression in our mouse model was under the control of the adiponectin promoter and, hence, only induced during late stages of adipocyte differentiation because it is located downstream of C/EBP (49). Thus, potential effects of individual IL-6 cytokines on early adipocyte differentiation (9) may not be reflected in gp130^Δadipo mice.

Chronically elevated circulating IL-6 levels were proposed to contribute to obesity-associated hepatic inflammation and insulin resistance in mice (11,12). In contrast, IL-6 was recently reported to reduce inflammation and to improve insulin sensitivity in liver (13). The current findings propose an indirect negative impact of IL-6 on hepatic insulin resistance and steatosis through induction of AT lipolysis mainly in mesenteric/omental fat depots. Accordingly, depletion of IL-6 cytokine signaling in adipocytes reduced portal FFA concentration and improved hepatic lipid accumulation as well as insulin resistance in mice. Moreover, omental (but not subcutaneous) IL-6 expression correlated with hepatic steatosis and insulin resistance in humans. Because increased omental (but not subcutaneous) lipolysis has been shown to be linked to obesity-induced fatty liver disease in humans with morbid obesity (44) and visceral lipolysis may contribute significantly to the portal delivery of FFA to the liver (50), the mechanism we propose may be conserved between species. Besides reduced lipolysis, lower intestinal absorption of FFA (35) or increased fatty acid uptake by adipocytes may have contributed to decreased portal FFA levels in HFD-fed gp130^Δadipo mice. Although the current data suggest that increased mesenteric FFA release into the portal vein impairs liver metabolism, such a direct link may be questioned (51). Changed release of other non-FFA factors into portal and/or systemic circulation may have also contributed to the observed hepatic phenotype in obese gp130^Δadipo mice.

In conclusion, the obesity-associated rise in mesenteric/omental AT IL-6 cytokine signaling promotes FFA release into the portal circulation, inducing hepatic steatosis and/or insulin resistance. Hence, blocking IL-6 cytokine signaling in adipocytes may be a novel approach to blunting detrimental fat-liver crosstalk in obesity. To this end, the development of cell-specific adeno-associated virus vectors may be a promising tool in the future (52,53).

Acknowledgments. The authors thank Eugen Schoenle (University Children's Hospital, Zurich, Switzerland) and Giatgen Spinas (University Hospital, Zurich, Switzerland) for continuous support and Alexandra Grob (University Children's Hospital, Zurich, Switzerland) for help with mouse breeding and genotyping.

Funding. This work was supported by a grant from Deutsche Forschungsgemeinschaft (SFB 1052/1, “Obesity Mechanisms,” to M.B.) and research grants from the Swiss National Science Foundation (# 310030_160129 to D.K.) and the Olga Mayenfisch Stiftung, Zurich (to S.W.).

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. S.W. contributed to the study concept, experimental work, discussion, and writing, review, and editing of the manuscript. F.I., F.C.L., T.D.C., and M.B. contributed to the experimental work, discussion, and review and editing of the manuscript. W.M. provided the gp130^Δadipo mice, gave conceptual advice, and contributed to the discussion and review and editing of the manuscript. D.K. contributed to the study concept, discussion, and writing, review, and editing of the manuscript. S.W. and D.K. are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. Parts of this study were presented in poster form at the 75th Scientific Sessions of the American Diabetes Association, Boston, MA, 5–9 June 2015.