- Email: [email protected]
Elevated soluble tumor necrosis factor receptor levels in non-obese adults with the atherogenic dyslipoproteinemia
Atherosclerosis 177 (2004) 77–81
Elevated soluble tumor necrosis factor receptor levels in non-obese adults with the atherogenic dyslipoproteinemia Robert S. Rosenson a,∗ , Christine C. Tangney b , Daniel M. Levine c , Thomas S. Parker c , Bruce R. Gordon c a
Departments of Medicine and Preventive Medicine, Division of Cardiology, Preventive Cardiology Center, The Feinberg School of Medicine, Northwestern University, 201 East Huron Street, Galter Pavilion, Suite 11-120, Chicago, IL 60611, USA b Department of Clinical Nutrition, Rush University, Chicago, IL, USA c The Rogosin Institute, New York, NY, USA Received 21 August 2003; received in revised form 4 May 2004; accepted 17 May 2004 Available online 28 July 2004
Abstract Adipose tissue expression of tumor necrosis factor-alpha (TNF-␣) has been implicated in the pathogenesis of obesity-linked insulin resistance and the dyslipoproteinemia of insulin resistance. This study has two aims: (1) to compare select inflammatory mediators in non-smoking, normoglycemic male subjects with and without the atherogenic dyslipoproteinemia (ADL), and (2) to determine the effects of statin therapy on select inflammatory mediators. ADL subjects had higher levels of insulin (16.7 ± 7.5 versus 11.6 ± 5.9 IU/mL, P = 0.008), soluble TNF receptor superfamily 1B (sTNFRSF1B) (3.3 ± 0.7 versus 2.7 ± 0.5 ng/mL, P = 0.005), and interleukin-6 (IL-6) (2.6 ± 2.2 versus 1.3 ± 1.8 pg/mL, P = 0.006) as compared to those of the non-ADL subjects. After adjustment for age, sTNFRSF1B (P = 0.003) was more predictive of ADL than high-sensitivity C-reactive protein (hs-CRP) (P = 0.047). Statin therapy did not change sTNFRSF1B, TNF-␣, IL-6, hs-CRP, whereas soluble TNF receptor superfamily 1A (sTNFRSF1A) increased slightly (P = 0.048). A high level of sTNFRSF1B is a strong marker of the pro-inflammatory state in this sample of male ADL subjects. © 2004 Elsevier Ireland Ltd. All rights reserved. Keywords: Atherogenic dyslipoproteinemia; Tumor necrosis factor receptors; Tumor necrosis factor-alpha; Metabolic syndrome
1. Introduction Tumor necrosis factor-alpha (TNF-␣) is a multifunctional cytokine that effectuates pleiotropic biological responses in different cells [1]. TNF-␣ expression and protein levels are increased in human adipose and skeletal muscle cells of obese and insulin-resistant subjects [2–4]. The expression of TNF-␣ expression in human adipose [5,6] and skeletal muscle cells [7] is positively correlated with body mass index (BMI) and the level of hyperinsulinemia, and inversely correlated with activity of adipose tissue lipoprotein lipase. Adipose tissue expression of TNF-␣ has been implicated as a causal factor in the pathogenesis of obesity-linked insulin resistance [8]. Cellular actions of TNF-␣ are mediated by the ∗
Corresponding author. Tel.: +1 312 695 0013; fax: +1 312 695 0047. E-mail address: [email protected] (R.S. Rosenson).
tumor necrosis factor receptor (TNFR) superfamily [9–12]. Insulin signaling is mediated by two distinct membrane receptors, TNFR superfamily 1A (previously p55 in rodents, p60 in humans) and TNFR superfamily 1B (previously p75 in rodents, p80 in humans). Several studies have shown that TNF-␣ reduces the activity of lipoprotein lipase [13]. Deficient lipoprotein lipase activity results in high plasma triglycerides, low high-density lipoprotein (HDL) cholesterol, and small low-density lipoprotein (LDL) particle size, or a pattern known as the atherogenic dyslipoproteinemia (ADL). In this study, we examine the relationships between TNF-␣, soluble TNFR superfamily 1A (sTNFRSF1A), and soluble TNFR superfamily 1B (sTNFRSF1B) in apparently healthy, overweight men with normal glucose levels. In addition, we investigate whether statin therapy lowers levels of sTNFRSF1A and/or sTNFRSF1B in men and women with ADL.
0021-9150/$ – see front matter © 2004 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.atherosclerosis.2004.05.027
78
R.S. Rosenson et al. / Atherosclerosis 177 (2004) 77–81
2. Research design and methods 2.1. Selection of patients and study design Statin therapy and immunological markers (STIM) was as a two-center, randomized, double blind, placebo-controlled study with four parallel groups. Subjects were recruited at two centers from April 1999 to May 2000. The trial was specifically powered to examine the influence of different statin regimens on lipopolysaccharide-stimulated TNF-␣ production. We enrolled 70 healthy non-smoking non-obese adults aged 35–65 years with LDL cholesterol levels between 3.4–4.9 mmol/L (130–189 mg/dL), triglycerides <4.5 mmol/L (400 mg/dL), and glucose <6.9 mmol/L (126 mg/dL). ADL was defined by the baseline levels of triglycerides ≥1.69 mmol/L (150 mg/dL) and HDL cholesterol <1.04 mmol/L (40 mg/dL) in men and <1.30 mmol/L (50 mg/dL) in women, and/or the presence of predominantly small LDL size (≤20.5 nm) [14]. Subjects were excluded if they had acute infection or trauma, inflammatory or connective tissue disorders, were women of childbearing potential, or had known intolerance to statins. Subjects reporting the use of aspirin, non-steroidal inflammatory agents, or corticosteroids within the previous 4 weeks or use of lipid lowering medications, vitamins, and nutritional supplements within the preceding 6 weeks, were excluded unless they consented to discontinue these medications. A total of five women treated with stable doses of hormonal replacement therapy and they were continued on this therapy throughout the study. There were two screening visits scheduled for all subjects in order to provide adequate time of 6 weeks for supplement and medication washouts. All subjects received instruction on how to follow at least a Step I American Heart Association cholesterol lowering diet at one of these visits. During the screening visits, we obtained a plasma lipid profile, and safety laboratory studies. Subjects were assigned to one of four treatments: placebo (n = 18), pravastatin 40 mg/daily (n = 17), simvastatin 20 mg/daily (n = 17), or simvastatin 80 mg/daily (n = 18). Of the 70 subjects, 10 subjects were excluded from the final analyses. Two subjects withdrew consent before randomization. One subject randomized to simvastatin 80 mg/daily developed myositis, and seven subjects reported acute infections that were accompanied by persistently elevated CRP levels and there were insufficient sample volumes for two persons and three specimens that were unavailable for subjects who did not complete the study. 2.2. Laboratory studies Fasting lipid profiles were obtained at the screening and randomization visit and twice at study completion (weeks 7–8). Plasma lipids were analyzed by standard enzymatic procedures at each medical center. In addition, NMR lipoprofiles were acquired at the randomization visit
and week 8 visit. High-sensitivity CRP (hs-CRP) was measured by the Roche COBAS Integra (Roche Diagnostics), TNF-␣ by ELISA (Amersham Pharmacia Biotech), and interleukin-6 (IL-6) and soluble cytokine receptors (sTNFRSF1A, sTNFRSF1B) were measured by enzyme immunoassay (BioSource International) on the Roche COBAS Core II (Roche Diagnostics). Coefficients of variation for these assays in our laboratory are as follows: hs-CRP, 2.9%; TNF-␣, <10%; IL-6, ≤6.0%; sTNFRSF1A and sTNFRSF1B, ≤7.5%. 2.3. Statistical analyses Descriptive statistics were performed using SAS statistical package, version 8.1 (Cary, NC). Distributions of all variables were examined for normality and transformed if necessary. Where appropriate, t-tests or Mann–Whitney U-tests were performed to examine differences in baseline outcomes between subjects with and without ADL. For non-continuous data, Chi-square tests were used. Logistic regression was performed with selected variables to determine the variables that best predicted ADL status. Because there were so few women in our sample, comparison between subjects with and without ADL, as well as logistic models, included male subjects only. Wilcoxon tests were used to examine the significance of changes in inflammatory markers on active statin therapy for all subjects regardless of ADL.
3. Results The baseline characteristics of the study group are described in Table 1. ADL male subjects were predominately overweight, middle-aged, and Caucasian. There was one African–American and one Hispanic male. There were no differences in BMI, blood pressure, pulse, or fasting glucose levels. Only 5 of 60 subjects fulfilled criteria for the metabolic syndrome based on criteria established by the ATP III guidelines [15]. Subjects with ADL had significantly higher levels of sTNFRSF1B (P = 0.005) and IL-6 (P = 0.006), but non-significantly higher levels of TNF-␣ (P = 0.54), sTNFRSF1A (P = 0.07), and hs-CRP (P = 0.09) (Table 1). All baseline values of inflammatory markers have been adjusted for age (Table 2). Logistic models were performed with selected cytokines (BMI and these variables available for selection). After adjustment for age, sTNFRSF1B, sTNFRSF1A, and hs-CRP were significant predictors of ADL (Table 2). The best age adjusted model for classifying subjects with ADL included sTNFRSF1B (75.6%, P = 0.003). The inclusion of sTNFRSF and hs-CRP did not improve the model. As there were no differences between active statin agents on sTNFRSF1A and sTNFRSF1B outcomes, data from all active therapies were combined. In Table 3, when data
R.S. Rosenson et al. / Atherosclerosis 177 (2004) 77–81
79
Table 1 Baseline characteristics of male study participants stratified according to atherogenic dyslipoproteinemia (ADL) statusa
N Age (years) BMI (kg/m2 ) Blood pressure (mm Hg) Systolic Diastolic Pulse Glucose (mmol/L) (mg/dL) Insulin (IU/mL) Cholesterol (mmol/L) (mg/dL) Total LDL HDL Triglycerides (mmol/L) (mg/dL) [median] LDL size (nm) Basal TNF-␣ (ng/mL) [median] sTNFRSF1A (ng/mL) sTNFRSF1B (ng/mL) hs-CRP (mg/L) [median] IL-6 (pg/mL)
With ADL
Without ADL
P-value
15 49.1 ± 9.3 28.6 ± 4.3
26 48.7 ± 7.6 27.4 ± 4.1
– 0.97 0.37
122 78 75 4.88 16.7
± ± ± ± ±
15 8 8 0.67 (87.9 ± 12) 7.5
113 75 69 4.80 11.6
± ± ± ± ±
12 7 11 0.33 (86.5 ± 6) 5.9
0.07 0.31 0.11 0.74 0.008
6.5 4.5 0.91 2.90 19.9 6.9 1.1 3.3 1.7 2.6
± ± ± ± ± ± ± ± ± ±
1.10 (252.2 ± 42.5) 0.85 (173.0 ± 32.8) 0.24 (35.3 ± 9.1) 1.08 (256.6 ± 95.4) [2.50 (221.0)] 0.5 21.0 [0.005] 0.3 0.7 1.2 [1.8] 2.2
5.6 3.9 1.06 1.23 21.1 0.5 0.9 2.7 1.0 1.3
± ± ± ± ± ± ± ± ± ±
0.73 (216.6 ± 28.0) 0.57 (151.7 ± 22.0) 0.30 (40.9 ± 11.2) 0.31 (109.0 ± 27.7) [1.22 (107.6)] 0.2 1.2 [0.01] 0.2 0.5 0.7 [0.8] 1.8
0.006 0.03 0.07 <0.001 <0.001 0.54 0.07 0.005 0.09 0.006
BMI = body mass index, LDL = low-density lipoprotein, HDL = high-density lipoprotein, TNF-␣ = tumor necrosis factor-alpha, sTNFRSF = soluble tumor necrosis factor super family, hs-CRP = high-sensitivity C-reactive protein, IL-6 = interleukin-6. a Data are means ± S.D. except for those variables that were skewed in which medians are given in parentheses. Data for 14 women are not provided.
from all subjects assigned to statin therapy were compared across time, there was a small increase in sTNFRSF1A, but no significant changes in other inflammatory markers. When we compared the changes in lipids and inflammatory markers associated with statin therapy between ADL subjects and those without ADL, the reduction in triglycerides was significantly (P = 0.001) greater among those with ADL (2.90 ± 1.11 mmol/L to 1.50 ± 0.10 mmol/L [256.7 ± 98.2 mg/dL to 132.8 mg/dL ± 8.7 mg/dL]) than for those
Table 2 Predictors of atherogenic dyslipoproteinemia in 41 male study participants: multivariate logistic regression analysis
Age Additional adjustment for BMI TNF-␣ sTNFRSF1B sTNFRSF1A IL-6 hs-CRP sTNFRSF1A + sTNFRSF1B sTNFRSF1B + sTNFRSF1A + hs-CRP
Overall classification accuracy (%)
Overall P-value for model
63.4
NSa
63.4 69.7 75.6 70.7 58.5 68.3 75.6
NS NS 0.003 0.013 NS 0.047 0.012
75.6
0.009
BMI = body mass index, TNF-␣ = tumor necrosis factor-alpha, sTNFRSF = soluble tumor necrosis factor super family, IL-6 = interleukin-6, hs-CRP = high-sensitivity C-reactive protein. a NS = non-significant.
without ADL (1.20 ± 0.33 mmol/L to 0.92 ± 0.32 mmol/L [106.2 ± 29.2 mg/dL to 81.4 ± 28.3]).
4. Discussion ADL is a common metabolic disorder that predicts increased risk of coronary heart disease and an increased propensity for the development of type 2 diabetes [16]. The major lipoprotein abnormalities that characterize ADL include high triglycerides, low levels of HDL cholesterol, and increased concentrations of small LDL particles [14]. Often, ADL is accompanied by other characteristics of the metabolic syndrome such as hyperglycemia, central adiposity, elevated blood pressure, high levels of plasminogen activator inhibitor-1 levels, and systemic inflammatory markers [15]. Elevated TNF-␣ production by human adipocytes and skeletal muscle cells has been linked to obesity induced type 2 diabetes. The major cellular responses of TNF-␣ are mediated by sTNFRSF1B. Circulating levels of TNF-␣ are often undetectable or highly variable in healthy subjects. In our report, the median TNF-␣ levels were null regardless of ADL status (Table 1). For this reason, sTNFR levels may be a better marker for the biological effects of TNF-␣. In this study of healthy subjects with normal glucose levels, sTNFRSF1B levels strongly correlated with ADL. In a case-control series of 37 (18 controls, 19 obese) premenopausal women, circulating levels of sTNFRSF1B were increased 6.3-fold in obese females compared to lean controls [17]. In contrast, TNFR superfamily 1A expression and
80
R.S. Rosenson et al. / Atherosclerosis 177 (2004) 77–81
Table 3 Inflammatory markers in all subjects assigned to active statin therapya,b Before treatment Basal TNF-␣ (ng/mL) [median] sTNFRSF1A (ng/mL) sTNFRSF1B (ng/mL) IL-6 (pg/mL) [median] hs-CRP (mg/L) [median] Cholesterol (mmol/L) (mg/dL) Total LDL HDL Triglycerides (mmol/L) (mg/dL) [median]
4.2 1.0 2.9 2.9 1.14
± ± ± ± ±
15.0 [0.0] 0.2 0.6 6.5 [1.4] 0.95 [0.89]
6.0 4.2 1.1 1.76
± ± ± ±
0.90 0.75 0.28 1.02
After 7–8 weeks of statin treatment
(230.9 ± 38.1) (160.5 ± 28.8) (40.7 ± 10.7) (155.6 ± 90.3) [1.45 (128.2)]
3.4 1.1 3.0 2.6 1.06
± ± ± ± ±
10.9 [0.0] 0.3 0.7 4.1 [1.6] 1.00 [0.69]
4.3 2.6 1.2 1.13
± ± ± ±
0.82 0.66 0.32 0.65
(165.4 ± 31.6) (100.7 ± 25.4) (45.8 ± 12.4) (99.8 ± 57.6) [0.95 (83.7)]
P-value 0.818 0.048 0.110 0.700 0.437 <0.001 <0.001 <0.001 <0.001
TNF-␣ = tumor necrosis factor-alpha, sTNFRSF = soluble tumor necrosis factor super family, IL-6 = interleukin-6, hs-CRP = high-sensitivity C-reactive protein, LDL = low-density lipoprotein, HDL = high-density lipoprotein. a The number of subjects assigned to 8 weeks active statin therapy with both before and after values is 38. b Data are presented as means ± S.D.
protein levels did not differ among obese and lean women. TNFR superfamily 1B mRNA levels in adipose tissue of obese women were strongly correlated with BMI (r = 0.65, P < 0.001) and fasting plasma insulin levels (r = 0.71, P < 0.001) [5]. A weaker correlation was observed between TNFR superfamily 1B expression and waist-to-hip ratio (r = 0.47, P = 0.03). Although adipocytes express TNF-␣, sTNFRSF1A, and sTNFRSF1B, BMI did not improve our classification of ADL. Our data suggest that TNFR superfamily 1B receptors contribute to the dyslipidemia of insulin resistance, and this association is not related to overweightedness. Statin therapy reduces cellular inflammation and circulating levels of the inflammatory marker CRP. In the West of Scotland Coronary Prevention Study (WOSCOPS), hypercholesterolemic subjects randomized to pravastatin had a reduced incidence of type 2 diabetes mellitus [18]. The enhanced cardioprotective effects of statins in type 2 diabetes include a reduction in pro-inflammatory cytokine production [18] and an improvement in insulin sensitivity [19]; however, the clinical studies are inconsistent. In a 12-month placebo-controlled trial, low dose pravastatin (15 mg/daily) reduced fasting insulin levels from 89.0 to 61.5 pmol/L (P < 0.05) in 96 elderly hypertensive hypercholesterolemic subjects. In the present short-term study, statin therapy did not significantly reduce circulating insulin levels. To our knowledge, we are unaware of published data in which the effects of statins on soluble TNF receptors have been evaluated. In this study of overweight hypercholesterolemia subjects, there were no changes in concentrations of either sTNFRSF1B or sTNFRSF1A with statin therapy. Previously, we reported that pravastatin (40 mg daily) reduced lipopolysaccharide stimulated TNF-␣ production by 31% in older subjects. Elevated sTNFRSF1B level is a marker of the proinflammatory state in subjects with the ADL. Since sTNFRSF1B blocks insulin action at the insulin receptor, this inflammatory marker may be involved in the pathogenesis of the insulin resistance syndrome.
Acknowledgements This clinical trial was supported by a grant from the Bristol-Myers Squibb Company, Princeton, NJ. Some of the findings in this report were presented at the American College of Cardiology 52nd Annual Scientific Sessions, held in Chicago, IL on April 1, 2003. Financial Support: Financial support for this study was provided by Bristol-Myers Squibb Company through an unrestricted research grant.
References [1] Vilcek J, Lee TH. Tumor necrosis factor. New insights into the molecular mechanisms of its multiple actions. J Biol Chem 1991;266:7313– 6. [2] Old LJ. Tumor necrosis factor (TNF). Science 1985;230:630–2. [3] Beutler B, Cerami A. The biology of cachectin/TNF – a primary mediator of the host response. Annu Rev Immunol 1989;7:625–55. [4] Grunfeld C, Feingold KR. The metabolic effects of tumor necrosis factor and other cytokines. Biotherapy 1991;3:143–58. [5] Hotamisligil GS, Arner P, Atkinson RL, Spiegelman BM. Differential regulation of the p80 tumor necrosis factor receptor in human obesity and insulin resistance. Diabetes 1997;46:451–5. [6] Fernandez-Real JM, Broch M, Ricart W, et al. Plasma levels of the soluble fraction of tumor necrosis factor receptor 2 and insulin resistance. Diabetes 1998;47:1757–62. [7] Saghizadeh M, Ong JM, Garvey WT, Henry RR, Kern PA. The expression of TNF alpha by human muscle. Relationship to insulin resistance. J Clin Invest 1996;97:1111–6. [8] Hotamisligil GS, Shargill NS, Spiegelman BM. Adipose expression of tumor necrosis factor-alpha: direct role in obesity-linked insulin resistance. Science 1993;259:87–91. [9] Locksley RM, Killeen N, Lenardo MJ. The TNF and TNF receptor superfamilies: integrating mammalian biology. Cell 2001;104:487– 501. [10] Liu LS, Spelleken M, Rohrig K, Hauner H, Eckel J. Tumor necrosis factor-alpha acutely inhibits insulin signaling in human adipocytes: implication of the p80 tumor necrosis factor receptor. Diabetes 1998;47:515–22. [11] Guo D, Donner DB. Tumor necrosis factor promotes phosphorylation and binding of insulin receptor substrate 1 to phosphatidylinositol 3-kinase in 3T3-L1 adipocytes. J Biol Chem 1996;271:615–8.
R.S. Rosenson et al. / Atherosclerosis 177 (2004) 77–81 [12] Kanety H, Hemi R, Papa MZ, Karasik A. Sphingomyelinase and ceramide suppress insulin-induced tyrosine phosphorylation of the insulin receptor substrate-1. J Biol Chem 1996;271:9895–7. [13] Kern PA, Saghizadeh M, Ong JM, et al. The expression of tumor necrosis factor in human adipose tissue. Regulation by obesity, weight loss, and relationship to lipoprotein lipase. J Clin Invest 1995;95:2111–9. [14] Sniderman AD, Scantlebury T, Cianflone K. Hypertriglyceridemic hyperapob: the unappreciated atherogenic dyslipoproteinemia in type 2 diabetes mellitus. Ann Intern Med 2001;135:447–59. [15] Executive summary of the Third Report of The National Cholesterol Education Program (NCEP) Expert Panel on detection, evaluation, and treatment of high blood cholesterol in adults (Adult Treatment Panel III). JAMA 285;2001:2486–97.
81
[16] Austin MA, King MC, Vranizan KM, Krauss RM. Atherogenic lipoprotein phenotype. A proposed genetic marker for coronary heart disease risk. Circulation 1990;82:495–506. [17] Hotamisligil GS, Arner P, Caro JF, Atkinson RL, Spiegelman BM. Increased adipose tissue expression of tumor necrosis factor-alpha in human obesity and insulin resistance. J Clin Invest 1995;95:2409– 15. [18] Freeman DJ, Norrie J, Sattar N, et al. Pravastatin and the development of diabetes mellitus: evidence for a protective treatment effect in the West of Scotland coronary prevention study. Circulation 2001;103:357–62. [19] Chan P, Tomlinson B, Lee CB, Pan WH, Lee YS. Beneficial effects of pravastatin on fasting hyperinsulinemia in elderly hypertensive hypercholesterolemic subjects. Hypertension 1996;28:647–51.
Elevated soluble tumor necrosis factor receptor levels in non-obese adults with the atherogenic dyslipoproteinemia Robert S. Rosenson a,∗ , Christine C. Tangney b , Daniel M. Levine c , Thomas S. Parker c , Bruce R. Gordon c a
Departments of Medicine and Preventive Medicine, Division of Cardiology, Preventive Cardiology Center, The Feinberg School of Medicine, Northwestern University, 201 East Huron Street, Galter Pavilion, Suite 11-120, Chicago, IL 60611, USA b Department of Clinical Nutrition, Rush University, Chicago, IL, USA c The Rogosin Institute, New York, NY, USA Received 21 August 2003; received in revised form 4 May 2004; accepted 17 May 2004 Available online 28 July 2004
Abstract Adipose tissue expression of tumor necrosis factor-alpha (TNF-␣) has been implicated in the pathogenesis of obesity-linked insulin resistance and the dyslipoproteinemia of insulin resistance. This study has two aims: (1) to compare select inflammatory mediators in non-smoking, normoglycemic male subjects with and without the atherogenic dyslipoproteinemia (ADL), and (2) to determine the effects of statin therapy on select inflammatory mediators. ADL subjects had higher levels of insulin (16.7 ± 7.5 versus 11.6 ± 5.9 IU/mL, P = 0.008), soluble TNF receptor superfamily 1B (sTNFRSF1B) (3.3 ± 0.7 versus 2.7 ± 0.5 ng/mL, P = 0.005), and interleukin-6 (IL-6) (2.6 ± 2.2 versus 1.3 ± 1.8 pg/mL, P = 0.006) as compared to those of the non-ADL subjects. After adjustment for age, sTNFRSF1B (P = 0.003) was more predictive of ADL than high-sensitivity C-reactive protein (hs-CRP) (P = 0.047). Statin therapy did not change sTNFRSF1B, TNF-␣, IL-6, hs-CRP, whereas soluble TNF receptor superfamily 1A (sTNFRSF1A) increased slightly (P = 0.048). A high level of sTNFRSF1B is a strong marker of the pro-inflammatory state in this sample of male ADL subjects. © 2004 Elsevier Ireland Ltd. All rights reserved. Keywords: Atherogenic dyslipoproteinemia; Tumor necrosis factor receptors; Tumor necrosis factor-alpha; Metabolic syndrome
1. Introduction Tumor necrosis factor-alpha (TNF-␣) is a multifunctional cytokine that effectuates pleiotropic biological responses in different cells [1]. TNF-␣ expression and protein levels are increased in human adipose and skeletal muscle cells of obese and insulin-resistant subjects [2–4]. The expression of TNF-␣ expression in human adipose [5,6] and skeletal muscle cells [7] is positively correlated with body mass index (BMI) and the level of hyperinsulinemia, and inversely correlated with activity of adipose tissue lipoprotein lipase. Adipose tissue expression of TNF-␣ has been implicated as a causal factor in the pathogenesis of obesity-linked insulin resistance [8]. Cellular actions of TNF-␣ are mediated by the ∗
Corresponding author. Tel.: +1 312 695 0013; fax: +1 312 695 0047. E-mail address: [email protected] (R.S. Rosenson).
tumor necrosis factor receptor (TNFR) superfamily [9–12]. Insulin signaling is mediated by two distinct membrane receptors, TNFR superfamily 1A (previously p55 in rodents, p60 in humans) and TNFR superfamily 1B (previously p75 in rodents, p80 in humans). Several studies have shown that TNF-␣ reduces the activity of lipoprotein lipase [13]. Deficient lipoprotein lipase activity results in high plasma triglycerides, low high-density lipoprotein (HDL) cholesterol, and small low-density lipoprotein (LDL) particle size, or a pattern known as the atherogenic dyslipoproteinemia (ADL). In this study, we examine the relationships between TNF-␣, soluble TNFR superfamily 1A (sTNFRSF1A), and soluble TNFR superfamily 1B (sTNFRSF1B) in apparently healthy, overweight men with normal glucose levels. In addition, we investigate whether statin therapy lowers levels of sTNFRSF1A and/or sTNFRSF1B in men and women with ADL.
0021-9150/$ – see front matter © 2004 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.atherosclerosis.2004.05.027
78
R.S. Rosenson et al. / Atherosclerosis 177 (2004) 77–81
2. Research design and methods 2.1. Selection of patients and study design Statin therapy and immunological markers (STIM) was as a two-center, randomized, double blind, placebo-controlled study with four parallel groups. Subjects were recruited at two centers from April 1999 to May 2000. The trial was specifically powered to examine the influence of different statin regimens on lipopolysaccharide-stimulated TNF-␣ production. We enrolled 70 healthy non-smoking non-obese adults aged 35–65 years with LDL cholesterol levels between 3.4–4.9 mmol/L (130–189 mg/dL), triglycerides <4.5 mmol/L (400 mg/dL), and glucose <6.9 mmol/L (126 mg/dL). ADL was defined by the baseline levels of triglycerides ≥1.69 mmol/L (150 mg/dL) and HDL cholesterol <1.04 mmol/L (40 mg/dL) in men and <1.30 mmol/L (50 mg/dL) in women, and/or the presence of predominantly small LDL size (≤20.5 nm) [14]. Subjects were excluded if they had acute infection or trauma, inflammatory or connective tissue disorders, were women of childbearing potential, or had known intolerance to statins. Subjects reporting the use of aspirin, non-steroidal inflammatory agents, or corticosteroids within the previous 4 weeks or use of lipid lowering medications, vitamins, and nutritional supplements within the preceding 6 weeks, were excluded unless they consented to discontinue these medications. A total of five women treated with stable doses of hormonal replacement therapy and they were continued on this therapy throughout the study. There were two screening visits scheduled for all subjects in order to provide adequate time of 6 weeks for supplement and medication washouts. All subjects received instruction on how to follow at least a Step I American Heart Association cholesterol lowering diet at one of these visits. During the screening visits, we obtained a plasma lipid profile, and safety laboratory studies. Subjects were assigned to one of four treatments: placebo (n = 18), pravastatin 40 mg/daily (n = 17), simvastatin 20 mg/daily (n = 17), or simvastatin 80 mg/daily (n = 18). Of the 70 subjects, 10 subjects were excluded from the final analyses. Two subjects withdrew consent before randomization. One subject randomized to simvastatin 80 mg/daily developed myositis, and seven subjects reported acute infections that were accompanied by persistently elevated CRP levels and there were insufficient sample volumes for two persons and three specimens that were unavailable for subjects who did not complete the study. 2.2. Laboratory studies Fasting lipid profiles were obtained at the screening and randomization visit and twice at study completion (weeks 7–8). Plasma lipids were analyzed by standard enzymatic procedures at each medical center. In addition, NMR lipoprofiles were acquired at the randomization visit
and week 8 visit. High-sensitivity CRP (hs-CRP) was measured by the Roche COBAS Integra (Roche Diagnostics), TNF-␣ by ELISA (Amersham Pharmacia Biotech), and interleukin-6 (IL-6) and soluble cytokine receptors (sTNFRSF1A, sTNFRSF1B) were measured by enzyme immunoassay (BioSource International) on the Roche COBAS Core II (Roche Diagnostics). Coefficients of variation for these assays in our laboratory are as follows: hs-CRP, 2.9%; TNF-␣, <10%; IL-6, ≤6.0%; sTNFRSF1A and sTNFRSF1B, ≤7.5%. 2.3. Statistical analyses Descriptive statistics were performed using SAS statistical package, version 8.1 (Cary, NC). Distributions of all variables were examined for normality and transformed if necessary. Where appropriate, t-tests or Mann–Whitney U-tests were performed to examine differences in baseline outcomes between subjects with and without ADL. For non-continuous data, Chi-square tests were used. Logistic regression was performed with selected variables to determine the variables that best predicted ADL status. Because there were so few women in our sample, comparison between subjects with and without ADL, as well as logistic models, included male subjects only. Wilcoxon tests were used to examine the significance of changes in inflammatory markers on active statin therapy for all subjects regardless of ADL.
3. Results The baseline characteristics of the study group are described in Table 1. ADL male subjects were predominately overweight, middle-aged, and Caucasian. There was one African–American and one Hispanic male. There were no differences in BMI, blood pressure, pulse, or fasting glucose levels. Only 5 of 60 subjects fulfilled criteria for the metabolic syndrome based on criteria established by the ATP III guidelines [15]. Subjects with ADL had significantly higher levels of sTNFRSF1B (P = 0.005) and IL-6 (P = 0.006), but non-significantly higher levels of TNF-␣ (P = 0.54), sTNFRSF1A (P = 0.07), and hs-CRP (P = 0.09) (Table 1). All baseline values of inflammatory markers have been adjusted for age (Table 2). Logistic models were performed with selected cytokines (BMI and these variables available for selection). After adjustment for age, sTNFRSF1B, sTNFRSF1A, and hs-CRP were significant predictors of ADL (Table 2). The best age adjusted model for classifying subjects with ADL included sTNFRSF1B (75.6%, P = 0.003). The inclusion of sTNFRSF and hs-CRP did not improve the model. As there were no differences between active statin agents on sTNFRSF1A and sTNFRSF1B outcomes, data from all active therapies were combined. In Table 3, when data
R.S. Rosenson et al. / Atherosclerosis 177 (2004) 77–81
79
Table 1 Baseline characteristics of male study participants stratified according to atherogenic dyslipoproteinemia (ADL) statusa
N Age (years) BMI (kg/m2 ) Blood pressure (mm Hg) Systolic Diastolic Pulse Glucose (mmol/L) (mg/dL) Insulin (IU/mL) Cholesterol (mmol/L) (mg/dL) Total LDL HDL Triglycerides (mmol/L) (mg/dL) [median] LDL size (nm) Basal TNF-␣ (ng/mL) [median] sTNFRSF1A (ng/mL) sTNFRSF1B (ng/mL) hs-CRP (mg/L) [median] IL-6 (pg/mL)
With ADL
Without ADL
P-value
15 49.1 ± 9.3 28.6 ± 4.3
26 48.7 ± 7.6 27.4 ± 4.1
– 0.97 0.37
122 78 75 4.88 16.7
± ± ± ± ±
15 8 8 0.67 (87.9 ± 12) 7.5
113 75 69 4.80 11.6
± ± ± ± ±
12 7 11 0.33 (86.5 ± 6) 5.9
0.07 0.31 0.11 0.74 0.008
6.5 4.5 0.91 2.90 19.9 6.9 1.1 3.3 1.7 2.6
± ± ± ± ± ± ± ± ± ±
1.10 (252.2 ± 42.5) 0.85 (173.0 ± 32.8) 0.24 (35.3 ± 9.1) 1.08 (256.6 ± 95.4) [2.50 (221.0)] 0.5 21.0 [0.005] 0.3 0.7 1.2 [1.8] 2.2
5.6 3.9 1.06 1.23 21.1 0.5 0.9 2.7 1.0 1.3
± ± ± ± ± ± ± ± ± ±
0.73 (216.6 ± 28.0) 0.57 (151.7 ± 22.0) 0.30 (40.9 ± 11.2) 0.31 (109.0 ± 27.7) [1.22 (107.6)] 0.2 1.2 [0.01] 0.2 0.5 0.7 [0.8] 1.8
0.006 0.03 0.07 <0.001 <0.001 0.54 0.07 0.005 0.09 0.006
BMI = body mass index, LDL = low-density lipoprotein, HDL = high-density lipoprotein, TNF-␣ = tumor necrosis factor-alpha, sTNFRSF = soluble tumor necrosis factor super family, hs-CRP = high-sensitivity C-reactive protein, IL-6 = interleukin-6. a Data are means ± S.D. except for those variables that were skewed in which medians are given in parentheses. Data for 14 women are not provided.
from all subjects assigned to statin therapy were compared across time, there was a small increase in sTNFRSF1A, but no significant changes in other inflammatory markers. When we compared the changes in lipids and inflammatory markers associated with statin therapy between ADL subjects and those without ADL, the reduction in triglycerides was significantly (P = 0.001) greater among those with ADL (2.90 ± 1.11 mmol/L to 1.50 ± 0.10 mmol/L [256.7 ± 98.2 mg/dL to 132.8 mg/dL ± 8.7 mg/dL]) than for those
Table 2 Predictors of atherogenic dyslipoproteinemia in 41 male study participants: multivariate logistic regression analysis
Age Additional adjustment for BMI TNF-␣ sTNFRSF1B sTNFRSF1A IL-6 hs-CRP sTNFRSF1A + sTNFRSF1B sTNFRSF1B + sTNFRSF1A + hs-CRP
Overall classification accuracy (%)
Overall P-value for model
63.4
NSa
63.4 69.7 75.6 70.7 58.5 68.3 75.6
NS NS 0.003 0.013 NS 0.047 0.012
75.6
0.009
BMI = body mass index, TNF-␣ = tumor necrosis factor-alpha, sTNFRSF = soluble tumor necrosis factor super family, IL-6 = interleukin-6, hs-CRP = high-sensitivity C-reactive protein. a NS = non-significant.
without ADL (1.20 ± 0.33 mmol/L to 0.92 ± 0.32 mmol/L [106.2 ± 29.2 mg/dL to 81.4 ± 28.3]).
4. Discussion ADL is a common metabolic disorder that predicts increased risk of coronary heart disease and an increased propensity for the development of type 2 diabetes [16]. The major lipoprotein abnormalities that characterize ADL include high triglycerides, low levels of HDL cholesterol, and increased concentrations of small LDL particles [14]. Often, ADL is accompanied by other characteristics of the metabolic syndrome such as hyperglycemia, central adiposity, elevated blood pressure, high levels of plasminogen activator inhibitor-1 levels, and systemic inflammatory markers [15]. Elevated TNF-␣ production by human adipocytes and skeletal muscle cells has been linked to obesity induced type 2 diabetes. The major cellular responses of TNF-␣ are mediated by sTNFRSF1B. Circulating levels of TNF-␣ are often undetectable or highly variable in healthy subjects. In our report, the median TNF-␣ levels were null regardless of ADL status (Table 1). For this reason, sTNFR levels may be a better marker for the biological effects of TNF-␣. In this study of healthy subjects with normal glucose levels, sTNFRSF1B levels strongly correlated with ADL. In a case-control series of 37 (18 controls, 19 obese) premenopausal women, circulating levels of sTNFRSF1B were increased 6.3-fold in obese females compared to lean controls [17]. In contrast, TNFR superfamily 1A expression and
80
R.S. Rosenson et al. / Atherosclerosis 177 (2004) 77–81
Table 3 Inflammatory markers in all subjects assigned to active statin therapya,b Before treatment Basal TNF-␣ (ng/mL) [median] sTNFRSF1A (ng/mL) sTNFRSF1B (ng/mL) IL-6 (pg/mL) [median] hs-CRP (mg/L) [median] Cholesterol (mmol/L) (mg/dL) Total LDL HDL Triglycerides (mmol/L) (mg/dL) [median]
4.2 1.0 2.9 2.9 1.14
± ± ± ± ±
15.0 [0.0] 0.2 0.6 6.5 [1.4] 0.95 [0.89]
6.0 4.2 1.1 1.76
± ± ± ±
0.90 0.75 0.28 1.02
After 7–8 weeks of statin treatment
(230.9 ± 38.1) (160.5 ± 28.8) (40.7 ± 10.7) (155.6 ± 90.3) [1.45 (128.2)]
3.4 1.1 3.0 2.6 1.06
± ± ± ± ±
10.9 [0.0] 0.3 0.7 4.1 [1.6] 1.00 [0.69]
4.3 2.6 1.2 1.13
± ± ± ±
0.82 0.66 0.32 0.65
(165.4 ± 31.6) (100.7 ± 25.4) (45.8 ± 12.4) (99.8 ± 57.6) [0.95 (83.7)]
P-value 0.818 0.048 0.110 0.700 0.437 <0.001 <0.001 <0.001 <0.001
TNF-␣ = tumor necrosis factor-alpha, sTNFRSF = soluble tumor necrosis factor super family, IL-6 = interleukin-6, hs-CRP = high-sensitivity C-reactive protein, LDL = low-density lipoprotein, HDL = high-density lipoprotein. a The number of subjects assigned to 8 weeks active statin therapy with both before and after values is 38. b Data are presented as means ± S.D.
protein levels did not differ among obese and lean women. TNFR superfamily 1B mRNA levels in adipose tissue of obese women were strongly correlated with BMI (r = 0.65, P < 0.001) and fasting plasma insulin levels (r = 0.71, P < 0.001) [5]. A weaker correlation was observed between TNFR superfamily 1B expression and waist-to-hip ratio (r = 0.47, P = 0.03). Although adipocytes express TNF-␣, sTNFRSF1A, and sTNFRSF1B, BMI did not improve our classification of ADL. Our data suggest that TNFR superfamily 1B receptors contribute to the dyslipidemia of insulin resistance, and this association is not related to overweightedness. Statin therapy reduces cellular inflammation and circulating levels of the inflammatory marker CRP. In the West of Scotland Coronary Prevention Study (WOSCOPS), hypercholesterolemic subjects randomized to pravastatin had a reduced incidence of type 2 diabetes mellitus [18]. The enhanced cardioprotective effects of statins in type 2 diabetes include a reduction in pro-inflammatory cytokine production [18] and an improvement in insulin sensitivity [19]; however, the clinical studies are inconsistent. In a 12-month placebo-controlled trial, low dose pravastatin (15 mg/daily) reduced fasting insulin levels from 89.0 to 61.5 pmol/L (P < 0.05) in 96 elderly hypertensive hypercholesterolemic subjects. In the present short-term study, statin therapy did not significantly reduce circulating insulin levels. To our knowledge, we are unaware of published data in which the effects of statins on soluble TNF receptors have been evaluated. In this study of overweight hypercholesterolemia subjects, there were no changes in concentrations of either sTNFRSF1B or sTNFRSF1A with statin therapy. Previously, we reported that pravastatin (40 mg daily) reduced lipopolysaccharide stimulated TNF-␣ production by 31% in older subjects. Elevated sTNFRSF1B level is a marker of the proinflammatory state in subjects with the ADL. Since sTNFRSF1B blocks insulin action at the insulin receptor, this inflammatory marker may be involved in the pathogenesis of the insulin resistance syndrome.
Acknowledgements This clinical trial was supported by a grant from the Bristol-Myers Squibb Company, Princeton, NJ. Some of the findings in this report were presented at the American College of Cardiology 52nd Annual Scientific Sessions, held in Chicago, IL on April 1, 2003. Financial Support: Financial support for this study was provided by Bristol-Myers Squibb Company through an unrestricted research grant.
References [1] Vilcek J, Lee TH. Tumor necrosis factor. New insights into the molecular mechanisms of its multiple actions. J Biol Chem 1991;266:7313– 6. [2] Old LJ. Tumor necrosis factor (TNF). Science 1985;230:630–2. [3] Beutler B, Cerami A. The biology of cachectin/TNF – a primary mediator of the host response. Annu Rev Immunol 1989;7:625–55. [4] Grunfeld C, Feingold KR. The metabolic effects of tumor necrosis factor and other cytokines. Biotherapy 1991;3:143–58. [5] Hotamisligil GS, Arner P, Atkinson RL, Spiegelman BM. Differential regulation of the p80 tumor necrosis factor receptor in human obesity and insulin resistance. Diabetes 1997;46:451–5. [6] Fernandez-Real JM, Broch M, Ricart W, et al. Plasma levels of the soluble fraction of tumor necrosis factor receptor 2 and insulin resistance. Diabetes 1998;47:1757–62. [7] Saghizadeh M, Ong JM, Garvey WT, Henry RR, Kern PA. The expression of TNF alpha by human muscle. Relationship to insulin resistance. J Clin Invest 1996;97:1111–6. [8] Hotamisligil GS, Shargill NS, Spiegelman BM. Adipose expression of tumor necrosis factor-alpha: direct role in obesity-linked insulin resistance. Science 1993;259:87–91. [9] Locksley RM, Killeen N, Lenardo MJ. The TNF and TNF receptor superfamilies: integrating mammalian biology. Cell 2001;104:487– 501. [10] Liu LS, Spelleken M, Rohrig K, Hauner H, Eckel J. Tumor necrosis factor-alpha acutely inhibits insulin signaling in human adipocytes: implication of the p80 tumor necrosis factor receptor. Diabetes 1998;47:515–22. [11] Guo D, Donner DB. Tumor necrosis factor promotes phosphorylation and binding of insulin receptor substrate 1 to phosphatidylinositol 3-kinase in 3T3-L1 adipocytes. J Biol Chem 1996;271:615–8.
R.S. Rosenson et al. / Atherosclerosis 177 (2004) 77–81 [12] Kanety H, Hemi R, Papa MZ, Karasik A. Sphingomyelinase and ceramide suppress insulin-induced tyrosine phosphorylation of the insulin receptor substrate-1. J Biol Chem 1996;271:9895–7. [13] Kern PA, Saghizadeh M, Ong JM, et al. The expression of tumor necrosis factor in human adipose tissue. Regulation by obesity, weight loss, and relationship to lipoprotein lipase. J Clin Invest 1995;95:2111–9. [14] Sniderman AD, Scantlebury T, Cianflone K. Hypertriglyceridemic hyperapob: the unappreciated atherogenic dyslipoproteinemia in type 2 diabetes mellitus. Ann Intern Med 2001;135:447–59. [15] Executive summary of the Third Report of The National Cholesterol Education Program (NCEP) Expert Panel on detection, evaluation, and treatment of high blood cholesterol in adults (Adult Treatment Panel III). JAMA 285;2001:2486–97.
81
[16] Austin MA, King MC, Vranizan KM, Krauss RM. Atherogenic lipoprotein phenotype. A proposed genetic marker for coronary heart disease risk. Circulation 1990;82:495–506. [17] Hotamisligil GS, Arner P, Caro JF, Atkinson RL, Spiegelman BM. Increased adipose tissue expression of tumor necrosis factor-alpha in human obesity and insulin resistance. J Clin Invest 1995;95:2409– 15. [18] Freeman DJ, Norrie J, Sattar N, et al. Pravastatin and the development of diabetes mellitus: evidence for a protective treatment effect in the West of Scotland coronary prevention study. Circulation 2001;103:357–62. [19] Chan P, Tomlinson B, Lee CB, Pan WH, Lee YS. Beneficial effects of pravastatin on fasting hyperinsulinemia in elderly hypertensive hypercholesterolemic subjects. Hypertension 1996;28:647–51.