Visceral Fat Adipokine Secretion Is Associated With Systemic Inflammation in Obese Humans

Twenty-five subjects with class III upper-body obesity (6 male and 19 female, BMI 54.7 ± 12.6 kg/m², waist circumference 150 ± 10 cm, age 42 ± 9 years), who were scheduled to undergo open gastric bypass surgery, participated in this study. Subjects completed a comprehensive medical evaluation, which included history, physical examination, electrocardiogram, and standard blood and urine tests. All obese subjects had evidence of insulin resistance, based on either a history of type 2 diabetes or high homeostasis model assessment score (13). All women who participated in this study were premenopausal. Six obese subjects had type 2 diabetes and were treated with insulin and metformin; none were being treated with thiazolidinediones. Each subject provided written informed consent before enrolling in the study, which was approved by the Human Studies Committee of the Washington University School of Medicine in St. Louis, Missouri.

Open gastric bypass surgery and upper gastrointestinal tract surgery were performed in the morning at Barnes-Jewish Hospital after subjects had fasted overnight. During the operation, blood samples were obtained simultaneously from the radial artery and portal vein before gastric stapling or intestinal resection was initiated. Blood samples were immediately transferred to sterile glass EDTA tubes (BD Vacutainer; BD Biosciences, Oxford, U.K.), placed on ice, and centrifuged at 2,200g for 10 min at 4°C. Aliquots of plasma were then placed in sterile cryovials, snapfrozen in liquid nitrogen, and stored at −80°C until subsequent analyses were performed.

Commercial radioimmunoassay kits were used to measure plasma total adiponectin (Linco Research, St. Louis, MO), and commercial ELISA kits were used to measure IL-6, TNF-α, resistin, MCP-1 (Quantakine High Sensitive; R&D Systems, Minneapolis, MN), and plasma C-reactive protein (CRP) (ALPCO Diagnostics, Windham, NH) concentrations. Plasma high–molecular weight (HMW) and low–molecular weight (LMW) adiponectin were determined by velocity sedimentation and quantitative Western blot analysis (14). Plasma insulin and leptin concentrations were measured by radioimmunoassay (Linco Research, St. Louis, MO).

The statistical significance of differences between blood sampling sites was evaluated by using Student's paired t test for data that was normally distributed with approximately equal SD values. The statistical significance of differences between groups that were not normally distributed or that had unequal SDs was evaluated by using the Wilcoxon two-samples test for variables. Pearson correlation was used to assess associations between continuous variables. Correlation data were log transformed when the data were not normally distributed. A P value <0.05 was considered statistically significant. All data were analyzed by using SPSS for Windows software, version 12.0 (SPSS, Chicago, IL). All values are expressed as means ± SD.

Portal vein and peripheral artery blood samples were obtained simultaneously from 25 subjects with extreme obesity (BMI 54.7 ± 12.6 kg/m²) and large visceral fat mass (waist circumference 150 ± 10 cm) during open gastric bypass surgery. All obese subjects had evidence of insulin resistance based on the homeostasis model assessment of insulin resistance (13).

In obese subjects, mean plasma insulin concentration was more than twofold greater in the portal vein than in the peripheral artery (34.4 ± 21 and 15.2 ± 8 μU/ml, respectively; P = 0.0004). Plasma IL-6 concentration was ∼50% higher in the portal vein than in the peripheral artery (Table 1). Plasma leptin concentration was ∼20% lower in the portal vein than in the peripheral artery. In contrast, portal vein plasma TNF-α, MCP-1, resistin, and total adiponectin concentrations were not significantly different from their peripheral artery concentrations. The percentage of adiponectin present as the HMW form in the portal vein and radial artery were evaluated in a subset of 10 obese subjects; no significant difference in percentage of HMW adiponectin was found between the portal vein (33.1 ± 11%) and artery (28.6 ± 17%). Portal vein IL-6 concentrations correlated directly with arterial CRP concentrations (Fig. 1).

The importance of adipokines released from visceral fat in the pathogenesis of insulin resistance and systemic inflammation has not been carefully studied in human subjects because of the difficulty in gaining access to portal vein blood. The portal vein, which drains visceral fat, is the major source of blood supply to the liver and is responsible for ∼80% of total liver blood flow (15). In the present study, we obtained blood samples from the portal vein and radial artery during open gastric bypass surgery in morbidly obese subjects who had a large amount of intra-abdominal fat. Plasma IL-6 concentrations were much higher in portal vein than in peripheral artery blood in obese subjects, demonstrating that visceral fat is an important source of IL-6 production in obese people. These data are consistent with previous ex vivo findings that IL-6 secretion is greater in omental than in subcutaneous adipose tissue samples (16,17). Moreover, portal vein IL-6 concentrations correlated directly with arterial CRP concentrations in obese subjects. These results provide the first evidence of a potential mechanistic link between visceral fat mass and systemic inflammation in human subjects.

The direct secretion of IL-6 into the portal vein has important metabolic consequences because IL-6 stimulates hepatic acute-phase reactant production (18), impairs insulin-mediated glycogenesis (19), and stimulates hepatic gluconeogenesis (20). Increased serum IL-6 concentration is also associated with increased risk of developing type 2 diabetes and cardiovascular disease (21–23). In our obese subjects, portal vein IL-6 concentrations correlated directly with systemic CRP concentrations, suggesting that IL-6 delivery to the liver contributes to the regulation of CRP production. This observation provides a potential mechanism for the relationship between visceral fat and systemic insulin resistance and supports the possibility that visceral fat is involved in regulating the hepatic production of acute-phase reactants that activate inflammatory pathways (24). These findings are consistent with data obtained from LIKK mice, showing that a localized increase in nuclear factor-κB and inflammation in the liver can cause peripheral insulin resistance in skeletal muscle (25). In addition, we have previously found that there is a considerable net release of IL-6 from subcutaneous abdominal fat in vivo, and this release is greater in obese than in lean subjects (26). Therefore, both visceral and subcutaneous fat depots in obese persons produce a large portion of circulating IL-6 during basal conditions.

Plasma leptin concentrations were lower in portal vein than in peripheral artery blood in our subjects. The lower portal vein leptin concentration is consistent with data obtained in vitro from isolated adipose tissue, showing that the expression of the ob gene, which produces leptin, and leptin secretion are lower in omental than subcutaneous fat (27,28).

In contrast with IL-6 and leptin, the plasma concentrations of other potential inflammatory adipokines, such as TNF-α and resistin, were similar in the portal vein and peripheral artery. These results are not surprising, based on what is known about the production and breakdown of these adipokines. Although TNF-α production is upregulated in visceral fat (29), it is likely that this cytokine acts locally and is not primarily released into the bloodstream. In fact, we have previously found that there is no net release of TNF-α from subcutaneous abdominal fat in vivo (26). Resistin is exclusively produced by adipocytes in mice (30), but resistin expression in humans is more prominent in monocytes and macrophages (31). Adiponectin circulates in plasma primarily as a LMW hexamer and as a larger, multimeric HMW complex (32). The HMW protein, but not the LMW form, has direct beneficial metabolic effects in the liver by increasing hepatic insulin sensitivity and decreasing hepatic glucose production (33–34). In addition, treatment with recombinant adiponectin has been shown to decrease hepatomegaly, steatosis, hepatic inflammation, and serum transaminase concentrations in a mouse model of nonalcoholic steatohepatitis (35). Therefore, preferential production of HMW adiponectin by visceral fat could have direct beneficial metabolic effects on the liver that might not be identified by only measuring total adiponectin concentrations. However, we found no differences between portal and systemic total, HMW, or LMW adiponectin concentrations. These findings suggest that visceral fat is not a major site for adiponectin production.

Fasting plasma insulin concentration was more than twofold greater in the portal vein than in the peripheral artery blood in the obese subjects, consistent with data reported previously (36). This observation underscores the important strategic anatomical location of both the pancreas and visceral fat as endocrine organs that regulate hepatic glucose and lipid metabolism. Both tissues are drained by the portal circulation, which provides an efficient system for delivering protein hormones (e.g., insulin) and inflammatory cytokines (e.g., IL-6) directly to the liver, where they can modulate endogenous glucose production and the production of acute-phase inflammatory reactants.

Our study has several important limitations. First, it was conducted in subjects during postabsorptive conditions, which might not reflect adipose tissue adipokine secretion into the systemic or portal circulations during postprandial conditions. Second, blood samples were obtained during surgery, which could have affected plasma adipokine concentrations. General anesthesia can decrease portal blood flow (37), which could increase portal vein adipokine concentrations. However, a decrease in blood flow should have simultaneously affected all adipokines secreted by visceral fat, whereas we found that portal vein IL-6 concentrations were higher and leptin concentrations lower than in radial artery blood. Third, our study subjects were limited to extremely obese subjects (BMI >40 kg/m²) who had massive amounts of intra-abdominal fat. Therefore, these results might not necessarily apply to obese people who have lower BMI values and lesser amounts of intra-abdominal fat. However, we did not find any significant correlation among BMI or waist circumference and portal vein or arterial adipokine concentrations. Fourth, the association we observed between portal vein IL-6 and systemic CRP concentrations does not prove a causal relationship. Additional and very complex studies, involving portal vein infusion of IL-6 inhibitors and recombinant portal vein IL-6, will be needed to prove that portal vein IL-6 is a major regulator of CRP production.

The results of the present studies support the notion that visceral fat is an important endocrine organ that is involved in the complex interrelationship between obesity and systemic inflammation. Our findings suggest that increased IL-6 secretion from visceral fat into the portal circulation is involved in the pathogenesis of systemic metabolic abnormalities associated with abdominal obesity.

This study was supported by National Institutes of Health Grant DK-37948, General Clinical Research Center Grant RR-00036, Diabetes Research and Training Center Grant DK-20579, and Clinical Nutrition Research Unit Grant DK56341. The funding agency had no role in the analysis or interpretation of the data or in the decision to submit the report for publication.

The study design was developed by S.K. and L.F.; data collection was performed and supervised by L.F., J.C.E., M.E.T., P.E.S., and S.K.; data analyses and interpretation were performed by L.F., M.E.T., P.E.S., and S.K.; writing was performed by L.F. and S.K. L.F. had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

The authors thank Jennifer McCrea and Jennifer Shew for technical assistance.