Plasma levels of the soluble fraction of tumor necrosis factor receptor 2 and insulin resistance.

Recent studies have shown that the tumor necrosis factor (TNF) system is implicated in the insulin resistance of human obesity. Plasma concentrations of the soluble fraction of the TNF receptors 1 and 2 (sTNFR1 and sTNFR2) are thought to reflect the degree of activation of the TNF system. The purpose of this study was to explore whether this activation, as measured by the levels of circulating sTNFR1 and sTNFR2, is associated with insulin resistance. A total of 19 men (mean age 36.2 +/- 1.9; BMI 28.8 +/- 1.2, range 22.2-35.7) and 17 premenopausal women (age 34.9 +/- 1.4; BMI 28.1 +/- 0.8, range 19-37.9) were studied. Men showed higher levels of plasma sTNFR1 and sTNFR2 than women. However, obese men showed increased levels of sTNFR2 but similar levels of sTNFR1 in comparison with obese women. In fact, sTNFR2 levels correlated with BMI (r = 0.50, P = 0.002), fat-free mass (FFM) (r = 0.61, P < 0.0001), and waist-to-hip ratio (WHR) (r = 0.39, P = 0.02), but not with fat mass or percent fat mass. sTNFR2 levels correlated with basal glucose levels (r = 0.45, P = 0.007), area under the curve (AUC) for glucose during an oral glucose tolerance test (r = 0.42, P = 0.013), and with the quotient AUC glucose/log AUC insulin (r = 0.41, P = 0.015). sTNFR2 also correlated negatively with insulin sensitivity (S(I)), evaluated using the frequently sampled intravenous glucose tolerance test with minimal model analysis (r = -0.38, P = 0.02). Plasma sTNFR1 levels were not associated with any of these variables. Because WHR influenced both S(I) and sTNFR2 levels, we constructed a multiple linear regression to predict S(I), with WHR and sTNFR2 as independent variables. In this model, both WHR (P = 0.0078) and sTNFR2 levels (P = 0.025) contributed to 47% of the variance in S(I). In parallel with higher FFM, lean and obese men showed a lower S(I) (2.9 +/- 0.9 vs. 5.2 +/- 1.3 min(-1) x mU x l(-1), P = 0.001; and 1.15 +/- 1.1 vs. 1.8 +/- 0.8 min(-1) x mU x l(-1), P = 0.035, respectively) and higher sTNFR2 levels in comparison with lean and obese women, respectively. After controlling for FFM, the correlation between S(I) and sTNFR2 levels disappeared, indicating that FFM was significantly influencing these associations. In summary, plasma sTNFR2 levels, but not sTNFR1, were proportional to BMI, WHR, FFM (a well-known confounder in the evaluation of insulin sensitivity), basal and postload glucose levels, and insulin resistance. These findings support TNF-alpha as a system regulating insulin action in human obesity.