DIABETES, VOL. 47, NOVEMBER 1998 1757 Plasma Levels of the Soluble Fraction of Tu m or Necrosis Factor Receptor 2 and Insulin Resistance José-Manuel Fernández-Real, Montserrat Broch, Wifredo Ricart, Roser Casamitjana, Cristina Gutierrez, Joan Vendrell, and Cristobal Richart Recent studies have shown that the tumor necrosis fac- tor ( 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 pre- menopausal 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 sim- ilar 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 .0 0 7 ) , 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 sam- pled intravenous glucose tolerance test with minimal model analysis ( r = –0.38, P = 0.02). Plasma sTNFR1 lev- els 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 · mU · l – 1 , P = 0.001; and 1.15 ± 1.1 vs. 1.8 ± 0.8 · min – 1 · mU · 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 con- founder in the evaluation of insulin sensitivity) , basal and postload glucose levels, and insulin resistance. These findings support TNF- as a system regulating insulin action in human obesity. Diabetes 47:1757–1762, 1998 I n the last years, it has been demonstrated that tumor necrosis factor (TNF)- has important effects on whole-body lipid and glucose metabolism (1,2). Fat tis- sue is a significant source of endogenous TNF- pro- duction, and the expression of this cytokine is elevated in human obesity in both adipose (3,4) and muscle (5) tissues. TNF- actions in obesity appear to occur via an autocrine- paracrine mechanism in these latter tissues (4,5). Neither the mechanisms that control this restricted activity nor the receptor systems that are involved in TNF- signaling are clearly understood (6,7). TNF signals through at least two known cell-surface receptors (6,7) (TNFRs), TNFR1 (p60) and TNFR2 (p80), that are present in virtually all cells of higher mammals, including adipocytes. It appears that TNFR1 can signal for virtually all known activities o f TNF, inc luding apo - ptosis, differentiation, and proliferation. TNFR2 seems to signal metabolic actions (8,9). Cytokines circulate bound to a number of binding pro- teins that appear to alter their clearance rates and/or enhance their bio logical activity. Both TNFRs exist also in so l- uble forms (10–13), apparently derived by proteolytic cleav- age from the cell-surface forms (14). These soluble TNFRs (sTNFRs) can compete with the cell-surface receptors and thus block TNF activity, safeguarding against its potentially harmful effects. However, sTNFRs affect TNF function also by stabilizing its activity (15), most likely by preventing dis- sociation of the homotrimeric TNF molecules (16) to inactive monomers. It has been suggested that sTNFR1 and sTNFR2 represent a buffer system that prolongs the biologic effects of TNF- by forming a “slow release reservoir” and impeding spontaneous denaturation of the cytokine (15). Limited info rmatio n is c urrently available c o nc erning the effects of the two TNFRs in energy metabolism and glucose homeostasis. In a recent study, obese women expressed approximately twofold more TNFR2 mRNA in fat tissue and appro x imately sixfold more sTNFR2 in circulation relative From the Department of Endocrinology (J.M.F.-R., W.R.), University Ho s- pital o f Girona “Dr. Josep Trueta,” Girona; the Department of Endocrinol- o gy (M.B., C.G., J.V., C.R.), University Ho spital o f Tarragona Jo an XXIII, Tar- ragona; and the Hormonal Laboratory (R.C.), University Hospital Clínic, Barcelona, Spain. Address correspondence and reprint requests to J.M. Fernández-Real, MD, Department of Endocrinology, Hospital de Girona, Ctra. França s/n, 17007 Girona, Spain. E-mail: hosptrueta@ comgir.c o m . Received for publication 27 March 1998 and accepted in revised form 8 July 1998. ASF, abdominal skinfold thickness; AUC, area under the curve; BSF, biceps skinfold thickness; CV, coefficient of variation; FFM, fat-free mass; FSIG TT, frequently sampled intravenous glucose tolerance test; IGT, impaired glucose tolerance; MAMA, mid-arm muscle area; MAMC, mid- arm muscle circumference; NGT, normal glucose tolerance; OGTT, oral glucose tolerance test; S G , glucose effectiveness; S I , insulin sensitivity index; SSF, subscapular skinfold thickness; sTNFR, soluble tumor necro- sis factor receptor; TNF, tumor necrosis factor; TNFR, tumor necrosis fac- tor receptor; TSF, triceps skinfold thickness; WHR, waist-to-hip ratio.