This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Fernández-Real, J. M. Articles by Ricart, W. Search for Related Content PubMed PubMed Citation Articles by Fernández-Real, J. M. Articles by Ricart, W.

Insulin Resistance and Chronic Cardiovascular Inflammatory Syndrome

Section of Diabetes, Endocrinology and Nutrition, University Hospital of Girona "Dr. Josep Trueta," 17007 Girona, Spain

Correspondence: Address all correspondence and requests for reprints to: J. M. Fernández-Real, M.D., Ph.D., Unitat de Diabetes, Endocrinologia i Nutrició., Hospital de Girona "Dr. Josep Trueta", Carretera de Francia s/n, 17007 Girona, Spain. E-mail: uden.jmfernandezreal{at}htrueta.scs.es

	Abstract

Top Abstract I. Introduction II. Hypertension and Chronic... III. Dyslipidemia as a... IV. Insulin Resistance and... V. Evidence Linking... VI. Inflammatory Markers Predict... VII. Antiinflammatory Agents... VIII. Possible Interpretations,... IX. Summary References

 
Insulin resistance is increasingly recognized as a chronic, low-level, inflammatory state. Hyperinsulinemia and insulin action were initially proposed as the common preceding factors of hypertension, low high-density lipoprotein cholesterol, hypertriglyceridemia, abdominal obesity, and altered glucose tolerance, linking all these abnormalities to the development of coronary heart disease. The similarities of insulin resistance with another inflammatory state, atherosclerosis, have been described only in the last few decades. Atherosclerosis and insulin resistance share similar pathophysiological mechanisms, mainly due to the actions of the two major proinflammatory cytokines, TNF- and IL-6. Genetic predisposition to increased transcription rates of these cytokines is associated with metabolic derangement and simultaneously with coronary heart disease. Dysregulation of the inflammatory axis predicts the development of insulin resistance and type 2 diabetes mellitus. The knowledge of how interactions between metabolic and inflammatory pathways occur will be useful in future therapeutic strategies. The effective administration of antiinflammatory agents in the treatment of insulin resistance and atherosclerosis is only the beginning of a promising approach in the management of these syndromes.

I. Introduction

II. Hypertension and Chronic Inflammation

A. Hypertension and the inflammatory cascade

B. Hypertension and TNF-

C. Hypertension, IL-6, and other cytokines

III. Dyslipidemia as a Chronic Inflammatory Disease

A. Dyslipidemia and TNF-

B. Lipid metabolism, IL-6 and other cytokines

IV. Insulin Resistance and Inflammation

A. Insulin action on adipose tissue and inflammation

B. Insulin action on muscle and inflammation

C. Insulin action on liver and inflammation

D. General insulin action and inflammation

V. Evidence Linking Proinflammatory Cytokines to Cardiovascular Disease through Metabolic Pathways

A. TNF- and cardiovascular disease

B. IL-6 and cardiovascular disease

VI. Inflammatory Markers Predict Type 2 Diabetes Mellitus and Its Complications

VII. Antiinflammatory Agents Improve Insulin Resistance and Prevent Development of Type 2 Diabetes Mellitus

VIII. Possible Interpretations, Unifying Hypothesis

IX. Summary

	I. Introduction

Top Abstract I. Introduction II. Hypertension and Chronic... III. Dyslipidemia as a... IV. Insulin Resistance and... V. Evidence Linking... VI. Inflammatory Markers Predict... VII. Antiinflammatory Agents... VIII. Possible Interpretations,... IX. Summary References

 
IN THE LAST few decades, a cluster of metabolic factors that can lead to coronary heart disease (CHD) have been identified. Hyperinsulinemia was proposed as the common preceding factor of hypertension, low high-density lipoprotein (HDL) cholesterol, hypertriglyceridemia, abdominal obesity, and altered glucose tolerance, linking all of these abnormalities to the development of CHD (1). Although different potential genes may lead to the insulin resistance syndrome (2), prospective cohort studies confirmed hyperinsulinemia, at first in univariate analysis and more recently in multivariate analysis, as a major underlying abnormality driving cardiovascular disease (3). In 1996, a link between a direct measure of insulin resistance itself and atherosclerosis was definitively demonstrated (4).

Impressive evidence has accumulated over the past decade that the atherosclerotic process is regulated by inflammatory mechanisms (reviewed in Refs. 5 and 6). Insulin resistance has also been increasingly recognized as having an important role in inflammatory pathways (7, 8, 9, 10). Initially, in the late 1980s, active chronic inflammatory disease was found to lead to peripheral insulin resistance (11). After achieving remission of the inflammatory process, normalization of the glucose handling and insulin sensitivity was observed. The data were interpreted as follows: "The linkage between inflammatory indices and glucose metabolism might reflect a special consequence of inflammation ... " (12). In 1993, circulating insulin was found to be associated with white blood cell count and C-reactive protein (CRP) in patients with angina pectoris (13). More recently, chronic subclinical inflammation has been proposed as a part of the insulin resistance syndrome (10). Epidemiological evidence that inflammatory markers predict the development of diabetes and glucose disorders is also beginning to accumulate (14, 15, 16). In both insulin resistance and atherosclerosis, the acute-phase response is enhanced. The study of the factors that regulate the acute-phase response in apparently healthy subjects has yielded consistent results implicating cytokines and growth factors in the pathophysiology of insulin resistance and atherosclerosis and in their complications. Two viewpoints exist on the initiation of this exaggerated acute-phase response. The first holds that the acute-phase response is activated by ongoing intraarterial inflammation, in which arterial wall-resident macrophages secrete proinflammatory cytokines in response to multiple stimuli. According to the second view, extravascular stimuli induce a chronic, low-level activation of the acute-phase response. The low-level activators would include smoking, mucosal infections (bronchitis, gastritis, or periodontitis), aging, and obesity (17). The final result of both views would be the triggering of the inflammatory cascade leading to insulin resistance and atherosclerosis (Fig. 1).

View larger version (26K): [in this window] [in a new window]   Figure 1. Possible pathways leading to chronic inflammation, resulting in atherosclerosis.

It should be stated that cytokines act primarily as autocrine/paracrine factors, and circulating levels may have nothing to do with their pathophysiological roles. IL-6 is possibly the exception to this rule. It is thought that IL-6 is secreted by a specific tissue (adipose tissue, among others), circulates in the bloodstream, and acts on distant tissues (the definition of hormone). It should also been kept in mind that whether or not TNF- or IL-6 is elevated in one condition vs. another may not be relevant to pathophysiology of insulin resistance. The majority of published studies do little to indicate which factors are primary and which are acquired in the evolution of the syndrome. From current available evidence, it is not still clear whether inflammatory parameters are markers or mediators of insulin resistance and/or cardiovascular disease, i.e., whether these parameters are elevated secondary to ongoing atherosclerosis or represent the direct cause of accelerated atherogenesis. Thus, any inference on causality should still be demonstrated with adequately designed studies.

	II. Hypertension and Chronic Inflammation

Top Abstract I. Introduction II. Hypertension and Chronic... III. Dyslipidemia as a... IV. Insulin Resistance and... V. Evidence Linking... VI. Inflammatory Markers Predict... VII. Antiinflammatory Agents... VIII. Possible Interpretations,... IX. Summary References

 
A. Hypertension and the inflammatory cascade
The mechanism by which primary, essential, or idiopathic forms of hypertension occur in humans is multifactorial and is based on a chronic metabolic disturbance. A large percentage of human hypertensive patients are salt sensitive, referring to the dependence of hypertension on sodium intake, but the cause of the salt sensitivity is not known. Although several mechanisms may contribute to salt-sensitive hypertension, the nitric oxide (NO) system appears to play a major role (18). NO is an important messenger molecule that plays a critical role in a wide variety of physiological functions, including immune modulation, vascular relaxation, neuronal transmission, and cytotoxicity. There are at least three isotypes of NO synthase (NOS): endothelial cell NOS, the neuronal type NOS, and the so-called inducible NOS (iNOS). iNOS is implicated in host defense and is synthesized de novo in response to a variety of inflammatory stimuli. Studies in humans indicate that NO production is decreased during hypertension (19). Interestingly, a polymorphism within the promoter of the iNOS candidate gene, NOS2A, revealed both increased allele sharing among sibpairs and positive association of NOS2A to essential hypertension (20). These facts are linked to insulin resistance because insulin stimulation of glucose uptake in skeletal muscles and adipose tissues in vivo seems to be NO dependent (21). Moreover, iNOS has recently been shown to be crucial for the development of insulin resistance (22).

NO antagonizes the effects of angiotensin II on vascular tone and growth and also down-regulates the synthesis of angiotensin converting enzyme (ACE) and angiotensin II type 1 receptors (23). The role that inflammation plays in atherosclerosis is amplified by the renin angiotensin system via the effects on adhesion molecules, growth factors, and chemoattractant molecules, which modulate the migration of inflammatory cells into the subendothelial space. Clinical and basic research has increased our knowledge of the actions of the vasoactive hormone angiotensin II, showing that it has multifunctional properties beyond its hemodynamic effects. A new aspect of this peptide is coming into focus: its potential role as a proinflammatory modulator. Angiotensin II is important in stimulating the production of reactive oxygen species and the activation of ancient inflammatory mechanisms through its angiotensin II type 1 receptor (24). The relationships between these events and insulin resistance seem to play a role in humans, given the protective role of ACE inhibitors on the development of type 2 diabetes (see Section VII).

Immunopathogenic mechanisms are increasingly recognized to be involved in the pathogenesis of hypertensive disease (25, 26). Abnormalities in immune system function, in both humoral and cellular immunity, and inflammatory mediators have been claimed to be responsible for the onset of hypertension (25, 26). A common finding in patients with hypertension is an elevated level of serum Ig, which is found in 20–40% of hypertensive subjects (27, 28, 29). Patients with borderline or established essential hypertension also display a delayed-type hypersensitivity to vascular antigens (30).

The complement system, which plays an important role in the overall responsiveness of the innate system and in the initiation and regulation of inflammation, also seems to be involved in blood pressure regulation. A substantial part of patients with essential hypertension express the C3F protein (31), which binds more avidly to mononuclear cells than the other allele, C3S. Expression of the C3F protein more than doubles the risk of developing hypertension and increases the risk of ischemic heart disease (IHD) by more than 10-fold in hypertensive patients (32). Plasma concentrations of the third complement component (C3) have been found to be associated with blood pressure (33, 34) in parallel to insulin resistance (35).

Evidence for immune dysregulation in animal hypertensive models is extensive (reviewed in Ref. 25), with the more prominent observation being that IL-2 significantly attenuated the elevated arterial pressure in the spontaneously hypertensive rat (36).

It remains to be established whether immune alterations are primary, concomitant, or secondary to the hypertensive process.

B. Hypertension and TNF-
The available information on the effects of TNF- in experimental models suggests that it is involved in the pathophysiology of hypertension. TNF- stimulates the production of endothelin-1 (37) and angiotensinogen (38) in vitro. In the spontaneously hypertensive rat model, TNF- synthesis and secretion are increased in response to lipopolysaccharide (LPS) stimulation compared with those in nonhypertensive control rats, and fat angiotensinogen mRNA increased after LPS in the former but not the latter (39).

In humans, the TNF- gene locus seems to be involved in insulin resistance-associated hypertension (40). However, in this study, a multivariant analysis of this gene locus effect on obesity and hypertension was not performed, and an obesity-independent effect of this gene on blood pressure is still obscure. A positive correlation has been found between serum TNF- concentration and both systolic blood pressure and insulin resistance in subjects with a wide range of adiposity (41). Up-regulation of TNF- secretion has also been observed in peripheral blood monocytes from hypertensive patients (42). TNF- also determines endothelial dysfunction linked to insulin resistance (43).

TNF- signals through at least two known cell surface receptors (TNFRs), TNFR1 (p60) and TNFR2 (p80; Refs. 44 and 45). The soluble fractions of these receptors, sTNFR1 and sTNFR2, result from a proteolytic cleavage of the cell surface forms (46, 47) when TNF- binds to its receptors. Measurements of the sTNFR concentrations in healthy individuals at different time lapses showed that the levels in the same subject were quite stable over time (48) and have been validated as sensitive indicators of TNF- system activation (49). The ratio of soluble TNF- receptors (sTNFR2/sTNFR1) correlated positively with systolic and diastolic blood pressure (P < 0.01) in a recent study (50). This ratio was significantly greater in type 2 diabetic patients than in type 1 diabetic patients and was greater in both than in control nondiabetic subjects. Interestingly, shedding of TNFR1 and TNFR2 was found to be associated with insulin resistance and vascular dysfunction in type 2 diabetic patients. After exercise-induced lowering of blood pressure, a concomitant decrease in the sTNFR2/sTNFR1 ratio was observed. It was concluded that insulin resistance and blood pressure are linked to altered shedding of TNFRs in type 2 diabetes mellitus (50).

C. Hypertension, IL-6, and other cytokines
IL-6 is a multifunctional cytokine produced by many different cell types, including immune cells, endothelial cells, fibroblasts, myocytes, and adipose tissue, mediating inflammatory as well as stress-induced responses. In recent studies, blood pressure was a significant and independent predictor of circulating IL-6 concentrations in women but not in men (51, 52), but not all studies are concordant (53). A polymorphism in the promoter of the IL-6 gene has also been found to show divergent associations with blood pressure (54, 55). IL-6 stimulates the central nervous system and the sympathetic nervous system, which may result in hypertension (56, 57). The administration of IL-6 led to increased heart rate in healthy women and increased norepinephrine levels and heart rate in women with fibromyalgia (58). However, other mechanisms cannot be excluded. IL-6 might increase in concert with the modification of the redox state of the vascular wall in chronic hypertension, as occurs in some hypertensive animal models (59), and in this vessel wall IL-6 also can lead to increased collagen (60). IL-6 is a well-characterized acute inducer of fibrinogen, and fibrinogen is a major determinant of blood viscosity (61). Finally, IL-6 might result in hypertension via effects on angiotensinogen expression (62), leading to higher concentration of angiotensin II, which is a potent vasoconstrictor.

Interestingly, a cytokine-like molecule increasingly recognized to regulate several inflammatory pathways acting on a receptor of the IL-6 family (leptin; reviewed in Ref. 63) seems also to be associated with hypertension. The leptin signal, via central leptin receptors, is believed to interact with the central sympathetic nervous system (64). Infusion of leptin leads to increases in blood pressure (65). Transgenic mice overexpressing leptin had elevated blood pressure, normalized by -adrenergic blockade (66). Recent findings implicate the leptin receptor gene locus with blood pressure regulation in men (67).

Other cytokines might also play a role. The secretion of IL-1-ß was significantly increased in peripheral blood monocytes from hypertensive patients vs. healthy individuals after stimulation with LPS (42). Similar findings were observed after stimulation with angiotensin II (42). Up-regulation of this cytokine was also seen at the RNA level in hypertensive patients. TGF-ß1 protein might also play a role in blood pressure regulation in humans (reviewed in Ref. 68). The potential link between hypertension and inflammatory mechanisms is summarized in Fig. 2.

View larger version (26K): [in this window] [in a new window]   Figure 2. Proposed pathophysiology of hypertension through inflammatory mechanisms (see Section II). CNS, Central nervous system; SNS, sympathetic nervous system.

	III. Dyslipidemia as a Chronic Inflammatory Disease

Top Abstract I. Introduction II. Hypertension and Chronic... III. Dyslipidemia as a... IV. Insulin Resistance and... V. Evidence Linking... VI. Inflammatory Markers Predict... VII. Antiinflammatory Agents... VIII. Possible Interpretations,... IX. Summary References

A. Dyslipidemia and TNF-
1. TNF- and triglycerides.
TNF- has important effects on whole-body lipid metabolism. The mechanisms and dynamics of cytokine-induced hypertriglyceridemia have been reviewed elsewhere (69, 70, 71). TNF- acutely raises serum triglyceride levels in vivo by stimulating very low-density lipoprotein (VLDL) production (72). Hypertriglyceridemia is well described in patients with frequent infections and chronic secretion of cytokines such as those with AIDS (73) or in patients with cystic fibrosis (74). In AIDS, other cytokines like interferon- also contribute to hypertriglyceridemia (75).

In apparently healthy people, a positive association between plasma concentration of the soluble fraction of TNFR-2 (sTNFR2, a surrogate of previous TNF- effects) and total triglycerides has been described, in parallel to a negative one with HDL cholesterol (76). Because plasma sTNFR2 is thought to reflect insulin resistance (77, 78), the possible contribution of the latter to increased plasma triglyceride levels should not be ignored.

Plasma TNF- correlated positively with VLDL triglycerides in healthy men and postinfarction patients (79, 80) and negatively with HDL cholesterol in the latter (80). However, difficulties in measuring TNF- in plasma should be recognized, where it is normally in a very low concentration (in the range of picograms per milliliter) and usually below the linear range of the assay.

2. TNF- and cholesterol.
The influence of cytokines on total and LDL cholesterol metabolism has been less well studied and seems species specific. TNF- administration to cynomolgus monkeys resulted in hypocholesterolemia (81), and, in humans with chronic infections such as AIDS or with cystic fibrosis, decreased total, LDL, and HDL cholesterol levels are usually found (73, 74). In contrast to this evidence, TNF- has been found to produce a 25% increase in serum cholesterol levels and a 2.3-fold increase in hepatic hydro-3-methyl-glutaryl coenzyme A reductase activity in C57BL/6 mice (82). TNF- is also capable of inducing sterol regulatory element binding protein-1 maturation, a key transcription factor in cholesterol biosynthesis, in a time- and dose-dependent manner in human hepatocytes (83).

The intensity, duration, and timing of TNF- hypersecretion might contribute to explaining these opposite actions of TNF- on cholesterol metabolism. In pathological conditions such as chronic infection, moderate to heavily increased TNF- concentrations may activate pathways of cholesterol metabolism such as increased LDL-receptor expression leading to increased lipoprotein clearance, increased conversion of newly synthesized cholesterol into bile acids, or enhanced esterification and storage of cholesterol (reviewed in Refs. 84 and 85). In other chronic low-level inflammatory diseases, such as those caused by intracellular pathogens, the reverse might also be true. Chlamydia species have been found to induce production of TNF-, inhibiting the action of lipoprotein lipase (LPL), leading to accumulation of serum triglycerides and a decrease in serum HDL cholesterol (reviewed in 86). LPS, a bacterial component, binds to both HDL cholesterol and LDL cholesterol, which buffers its toxic capacity. However, in the long term, when this mechanism is overwhelmed, LPS makes LDL immunogenic or toxic to endothelial cells. In a vicious cycle, the accumulation of cholesteryl esters in macrophages exposed to LDL-immune complexes is again associated with increased synthesis and release of TNF- (85), and activated macrophages are better able to form foam cells. In this context, it is interesting to note that increased plasma cholesterol per se [as observed in hypercholesterolemic rabbits (87), LDL-receptor knockout mice (88), or diet-induced (89)] is associated with increased plasma TNF- concentration or enhanced endotoxin-stimulated TNF- and IL-1 gene expression in aortae (89).

In apparently healthy subjects, sTNFRs circulate in proportion to total and LDL cholesterol (76, 90). High cholesterol could lead to an increased activity of TNF- axis, of which increased sTNFRs would be its reflection. This rise in soluble fraction of cytokine receptors parallels increases in other soluble factors, such as soluble endothelial leukocyte adhesion molecules and soluble intercellular adhesion molecule-1 (sICAM-1) observed in other situations of hypercholesterolemia (91). Damage to the endothelium could provide a link explaining simultaneously increased sTNFRs, soluble endothelial leukocyte adhesion molecules, and sICAM-1, because elevated cholesterol levels are associated with endothelial dysfunction (92). An alternative explanation is that high levels of sTNFRs block the hypocholesterolemic action of TNF- by inhibiting its interaction with cell surface receptors, resulting in high cholesterol levels.

Another possible mechanism is one related to the stimulatory effect of insulin on LDL-receptor activity through the enhancement of LDL-receptor mRNA expression (93). In this sense, increased LDL cholesterol might be the result of hampered insulin action (possibly induced by TNF-) on LDL-receptor activity.

B. Lipid metabolism, IL-6, and other cytokines
Recent investigations have shown that locally produced cytokines possess important autocrine/paracrine properties that influence diverse functions of the adipose tissue in addition to possible effects on other tissues. In this sense, IL-6 has been hypothesized to be responsible for the lipid abnormalities occurring in subjects with the insulin resistance syndrome (7, 8). This hypothesis is based on the findings of increased blood concentrations of IL-6 and markers of the acute-phase response, including CRP and cortisol in parallel with dyslipidemia (decreased plasma HDL cholesterol and increased plasma triglyceride concentration) in patients with this syndrome (7, 8).

IL-6 inhibits adipocyte LPL activity (94) and induces increases in hepatic triglyceride secretion in rats (95). In man, IL-6 infusion leads to increased free fatty acid concentration (96), and fasting triglycerides, VLDL triglycerides, and post-glucose load free fatty acids are linked to serum IL-6 concentration (97).

The link between inflammation, insulin resistance, and CHD might be mediated through different pathways, including fatty acid (FA) metabolism. Dietary FA appear to modulate the release of different cytokines (98). The production of IL-1 ß, TNF-, IL-6, and granulocyte and macrophage colony-stimulating factor by peripheral mononuclear cells decreases after dietary polyunsaturated FA supplementation in women (99, 100). Docosahexaenoic acid (DHA) and eicosapentaenoic acid inhibited in vitro human endothelial cell production of IL-6 (101). DHA also reduced endothelial expression of IL-6 in response to different stimuli (102). In contrast, the consumption of a diet high in hydrogenated fat increases production of IL-6 and TNF- (103). Hence, the profile of dietary FA strongly influences cytokine production. In fact, the proportion of saturated and polyunsaturated -3 FA in serum of healthy volunteers was associated with circulating IL-6 concentration (104). Other circulating cytokines such as plasma granulocyte and macrophage colony-stimulating factor concentration appear to be linked to serum DHA and eicosapentaenoic acid levels in healthy volunteers (105).

	IV. Insulin Resistance and Inflammation

Top Abstract I. Introduction II. Hypertension and Chronic... III. Dyslipidemia as a... IV. Insulin Resistance and... V. Evidence Linking... VI. Inflammatory Markers Predict... VII. Antiinflammatory Agents... VIII. Possible Interpretations,... IX. Summary References

 
Defects in insulin action on the main insulin-sensitive tissues (adipose tissue, muscle, and liver) are proposed to lead to a worsening of the chronic, low-grade inflammatory state. Independent of the triggering agent and of the initial events, the relationship is bidirectional; any process linked to chronic inflammation will decrease insulin action, and insulin resistance will lead to worsening of inflammation in a vicious cycle.

A. Insulin action on adipose tissue and inflammation
There exists increased evidence that generalized and abdominal obesity constitute low-grade inflammatory states. Adipose tissue, long being misconstrued as a mere tissue of fat storage, is progressively acknowledged to be an active participant in energy homeostasis. The term "adipocytokines" was recently coined to describe the adipose-derived bioactive factors that modulate the physiological function of the other tissues in our body (106).

1. Abdominal obesity and TNF-.
TNF- seems to play a key role in regulating adipose tissue metabolism (107, 108, 109, 110, 111, 112). In obese humans (111, 112, 113) and numerous rodent models (107, 108, 109) of obesity-diabetes syndromes, TNF- is overexpressed in the adipose tissue, as compared with tissues from lean individuals. A decreased processing rate of transmembrane TNF- in mature adipocytes combined with an increase in TNF- production may be a potential mechanism resulting in elevated membrane-associated TNF- in adipose tissue in obesity (114).

The TNF- gene locus seems to influence the distribution of body fat according to sex. Although this locus exerted the most significant effects on waist circumference and suprailiac skinfold in men, the most significant impact in women was on upper thigh circumference and thigh skinfold (40). In fact, the only element in the TNF- cascade that is known to have gender-specific regional effects is LPL (40). The mRNA level and the enzyme activity of LPL are higher in abdominal than in thigh adipose cells in men and vice versa in women (115).

Obese women express approximately 2-fold more TNFR2 mRNA in fat tissue and approximately 6-fold more sTNFR2 in circulation relative to lean control subjects (77). Interestingly, adipose tissue expression of TNFR2 strongly correlates with body mass index (BMI) and with measures of abdominal obesity [waist-to-hip ratio (WHR); Refs. 77 and 78 ]. However, there exists some discrepancy in the relationships among sTNFR1, sTNFR2, and adiposity measures depending on study design and inclusion or not of morbidly obese subjects (116, 117, 118, 119).

In one study, plasma sTNFR2 concentration was described to cosegregate with measures of obesity but not with insulin resistance in twins (120). However, mechanistically speaking, circulating sTNFR2 concentration changes with systemic insulin action (see Section IV.B). Subcutaneous adipose tissue also produced sTNFR1 as shown by arteriovenous differences (121), and diet-induced weight loss led to significantly decreased sTNFR1 levels (122).

Administration of LPS or the recombinant cytokines TNF- and IL-1 has been reported to induce leptin expression and secretion by adipose tissue (123, 124). Recent works have suggested the existence of a TNF--leptin axis, in which leptin and TNF- would be in a mutual interrelationship. TNF- stimulates leptin secretion in cultured adipocytes and obese mice, and, as a feedback loop, leptin administration to rats decreased TNF- expression by 40% (123, 124, 125, 126). In fact, TNF- administration increases serum leptin levels in humans (127), and plasma sTNFR1 concentration circulates in proportion to leptin (128). However, the possible role of leptin resistance in these interactions is still confusing in humans. These facts are important in understanding pathophysiology of abdominal obesity because leptin production is dependent on the distribution of body fat (129). Moreover, although leptin was first described for its role in modulating food intake and energy expenditure, there is now substantial evidence that leptin is also involved in immune function (63), as evidenced by its effect to enhance cytokine production and phagocytosis by macrophages (130). In fact, increased leptin concentrations correlate with increased concentrations of inflammatory markers in morbidly obese individuals (131). However, the role of leptin in human insulin action or obesity-associated inflammation is probably very small (132).

2. Abdominal obesity and IL-6.
IL-6 is secreted from adipose tissue during noninflammatory conditions in humans. Omental adipose tissue produces 3-fold more IL-6 than sc adipose tissue (133). Dynamic studies of IL-6 concentration in humans showed that IL-6 increases postprandially, in parallel to glucose and insulin levels in the interstitial fluid of sc adipose tissue (134). This finding suggests that IL-6 might modulate adipose glucose metabolism in the fed state.

It has been calculated that one third of total circulating concentrations of IL-6 originate from adipose tissue (135). A positive association between different measures of obesity and plasma IL-6 levels has been described in men and postmenopausal women (51, 52, 133). Because venous drainage from omental tissue flows directly into the liver, the increased physiological WHR of men is expected to have more metabolic impact. Abdominal arterial IL-6 was also associated with BMI (133). In contrast, plasma IL-6 levels were higher in obese patients with sleep apnea but not in obese controls in comparison with normal weight controls (136). In another study, the relationship between BMI and serum IL-6 was only observed in postmenopausal women, and this relationship was lost among those women receiving hormone replacement (52). In fact, estrogens are well-known inhibitors of IL-6 secretion (137). Adipose tissue-derived estrogens in postmenopausal women would not be sufficient to reduce IL-6 in a similar way as endogenous estrogens in premenopausal women (52). It should be stated that an important variable is smoking, which was a significant confounding factor in the relationship between measures of body fat and circulating IL-6 concentration (51). Recently, the IL-6-174G/C polymorphism has been found to be associated with indices of obesity in men (138).

Of the aforementioned, obesity is expected to result in increased secretion of IL-6, with its detrimental metabolic effects (see Section IV.D.2.b). In contrast, mice lacking the gene encoding IL-6 (IL-6^-/-) developed mature-onset obesity and disturbed carbohydrate and lipid metabolism, which were reversible by IL-6 replacement (139). The antiobesity effect of IL-6 was mainly exerted at the level of the central nervous system, being inactive when administered peripherally (139). At first glance, this is puzzling. However, IL-6 acts as a terminator as well as a prompter of inflammation. It should again be remembered that cytokines act in cascade, and total depletion of IL-6 might be detrimental because other proinflammatory cytokines (TNF-) are not adequately down-regulated, as is the case in other knockout models (140). Unfortunately, other cytokines were not studied in that report (139).

B. Insulin action on muscle and inflammation
1. Muscle and TNF-.
Obesity is not the only condition in which circulating sTNFR2 concentration is proportional to insulin resistance. Plasma sTNFR2 concentration has been described as being linked to markers of the muscle compartment as fat-free mass and midarm muscle circumference (78, 141, 142). Circulating sTNFR2 concentration is also associated with insulin resistance in other diseases characterized by muscle disease (143).

TNF- is a strong inducer of TNFR2 expression in adipocytes and other cell types (144), and, in this context, the association between the muscle compartment and insulin resistance might be attributed to increased production by the muscle of sTNFR2, leading to stabilization of TNF- homotrimers, resulting in insulin resistance at the level of the adipocyte. These facts are related to the thrifty genotype hypothesis. Neel (145) postulated that a thrifty genotype existed that had a selection advantage as hunter-gatherers fluctuated between feast and famine. The thrifty genotype in type 2 diabetes contributes to the insulin resistance seen in muscle (146). A selective insulin resistance in muscle would have the effect of blunting the hypoglycemia that occurs during fasting but would allow energy storage in fat and liver during feeding. Both of these features could allow hunter-gatherers to have survival advantages during periods of food shortage (147, 148, 149). Although mild exercise seems to produce health benefits (150), strenuous exercise, taken to the extreme or during prolonged fasting conditions, initiates an immune and vascular proinflammatory response. In fact, acute strenuous exercise is considered to be a model of the acute-phase response that occurs in parallel to insulin resistance (150, 151, 152), and the cytokine response to strenuous exercise [both plasma TNF- and sTNFR2 concentrations increase during exercise (151)] is similar to that found in sepsis and trauma (150, 152). As recently hypothesized (9), the induction of muscle insulin resistance would allow a food-seeking behavior and would prevent the wasting of glucose to nonvital organs, protecting the brain and the immunological system.

In the absence of food shortage, regular physical exercise led to a consistent decrease in circulating sTNFR2 in obese women (141, 153) and in type 2 diabetic patients (50), in parallel to improved insulin sensitivity.

2. Muscle, IL-6, and other cytokines.
Several cytokines can be detected in plasma during and after strenuous exercise. The increase in TNF-, sTNFR2, IL-1ß, IL-1 receptor antagonist, IL-8, and IL-10 is accompanied by a dramatic increase in IL-6 (reviewed in Ref. 154). After exercise, IL-6 is produced in larger amounts by the contracting muscle than any other cytokine examined. Both IL-6 mRNA [which increases more than 10-fold after exercise (155)] and IL-6 receptor mRNA (156) have been detected in muscle. It has been suggested that depleted glycogen content or an energy crisis in the contracting muscle, rather than muscle damage, may be one stimulus for the IL-6 release (154). Thus, the possibility exists that the elevated serum IL-6 is a consequence of an increased production and release of IL-6 from muscle in response to the impaired insulin sensitivity, allowing hunter-gatherers to have survival advantages during periods of food shortage, in a similar way to the events described above.

C. Insulin action on liver and inflammation
1. Acute-phase proteins (APP) and insulin resistance.
After trauma or infection, the human body mounts a highly complex acute-phase response as part of the homeostatic response to injury. In the acute phase, the acute-phase response is protective because it counteracts the effects of injury and improves survival. Long-term exposure to stressful stimuli (mucositis, aging, increased fat intake, periodontitis, etc.) may result in disease rather than repair. The liver is the target of systemic inflammatory mediators and is also the organ responsible for determining the level of essential metabolites provided to the organism during the critical stages of stress. Among the most important aspects of this response is the reprioritization of hepatic protein synthesis with the increased production of a number of plasma proteins (positive APP) and reduced production of a number of normal export proteins (negative APP). It should be stated that expression of acute-phase reactants at high levels has also been recently identified in adipose tissue of mice, and this was especially remarkable in the diabetic state (157). The induction of APP production is thought to be regulated at the transcriptional level, and at least two signaling pathways have been identified within the hepatocyte. One pathway activates transcription of class I acute-phase genes such as CRP, serum amyloid A, and complement C3, whereas the other pathway activates transcription of class II acute-phase genes such as ₁-antitrypsin and fibrinogen (158).

Although the concentrations of multiple components of the acute-phase response increase together, not all of them increase uniformly in all patients. These variations indicate that the components of the acute-phase response are individually regulated, and this may be caused in part by differences in the pattern of production of specific cytokines (159, 160). These facts would explain increased susceptibility to increased inflammatory activity among healthy volunteers with genetically increased rates of some cytokines (161).

The hormonal and cytokine networks acting on the hepatocyte may lend a degree of fine-tuning to the spectrum of APP in response to different stimuli (158). Insulin seems to be one of the main regulators of the cytokine-associated acute-phase reaction (162, 163). Liver cells respond to many of the factors via their cell surface receptors. According to one vision (158), the inflammatory mediators fall into four major categories: 1) IL-6 type cytokines, of which IL-6 is the major representative; 2) IL-1 type cytokines (including IL-1, IL-1ß, TNF-, and TNF-ß); 3) glucocorticoids; and 4) growth factors (including insulin). The cytokines would act as primary stimulators of APP gene expression, whereas the glucocorticoids and growth factors function more as modulators of cytokine action. An adequate balance between these opposite pathways will result in resolution of the acute-phase process (Fig. 3).

View larger version (35K): [in this window] [in a new window]   Figure 3. Balance of proinflammatory and antiinflammatory agents regulating the acute-phase response. An adequate balance will lead to resolution of the process.

Systemic inflammation, measured by increased serum acute-phase reactants, has been recognized to occur in type 2 diabetes since the early work by McMillan (164). Significantly higher serum concentration of CRP, fibrinogen, ₁-acid glycoprotein, amyloid A, sialic acid, and orosomucoid have been described in patients with type 2 diabetes mellitus (7, 164). Serum CRP concentration and the acute-phase reaction have been significantly associated with clinical and biochemical indexes of insulin resistance (51, 165, 166, 167, 168, 169, 170, 171, 172). The relationship between increased CRP and decreased insulin action might be intrinsically due to insulin resistance itself (51, 162, 163). Elevated serum CRP concentrations have been demonstrated consistently in overweight and obese adults, even among young adults aged 17–39 yr (165). From 34–39% of U.S. people with diabetes had elevated CRP levels, and this association was not completely explained by increases in BMI (166). Of note was that CRP and IL-6 decreased significantly after improvement of metabolic control in type 2 diabetic patients, indicating that the inflammatory pathways are modulated by insulin (173). CRP may not measure all of the relevant effectors of inflammation because the concentration of CRP may be subject to posttranscriptional regulation (159). In one study, CRP levels were associated with several metabolic parameters in men and women, but in a multiple linear regression analysis, CRP was associated independently with IL-6 only in men (51). Interestingly, both constitutive and IL-6-dependent acute-phase expression of the human CRP transgene require testosterone in transgenic mice, implying a potential mechanism for this gender dimorphism (174).

Other acute-phase reactants, such as fibrinogen, plasminogen activator inhibitor-1, and amyloid A are associated with insulin resistance and predict development of type 2 diabetes (Ref. 167 ; also see Section IV).

2. Corticosteroid binding globulin (CBG).
CBG is the major blood transport protein for cortisol, the major antiinflammatory hormone in humans. Scarce data in the literature have suggested that CBG is a negative acute-phase reactant (175, 176). CBG level was negatively associated with insulin secretion in obese and glucose-tolerant subjects who also had lower CBG levels than obese and glucose-intolerant subjects (177). It has been shown that both insulin secretion and insulin sensitivity independently contributed to CBG changes after glucose-induced insulin stimulation (178). The insulin response after a glucose challenge was linked to acute CBG changes in lean subjects.

CBG has been recently found to be negatively associated with several indexes of insulin resistance such as BMI, WHR, and homeostasis model assessment, and with inflammatory parameters such as sTNFR1, sTNFR2, and IL-6 concentrations (179). A polymorphism of the IL-6 gene promoter, which is linked to increased IL-6 levels and to insulin resistance, was also associated with low CBG levels (161). In women, both decreased CBG and free cortisol independently contributed to homeostasis model assessment variance in a multiple linear regression analysis, suggesting that both mechanisms would be metabolically additive (179).

Because CBG secretion has been shown to be negatively regulated by both insulin (180) and IL-6 (181, 182), it is tempting to propose that CBG concentration is an index of insulin resistance and inflammation. On the other hand, constitutive low CBG levels might also contribute to insulin resistance by increasing cortisol biodisposal to target cells, including muscular cells. Interestingly, physiological increments in plasma insulin concentrations have been described to affect synthesis of other hepatic proteins in normal humans (183). In addition, insulin is also able to promote albumin distribution to peripheral tissues by increasing the protein transcapillary escape rate (184). As shown for low serum albumin in the Atherosclerosis Risk in Communities Study (14), it has been suggested that low CBG, another liver protein, could be an interesting index for the development of type 2 diabetes as well as for incident cardiovascular disease as reported early in postmenopausal women (185). It remains to be established whether low circulating or tissue CBG concentration impacts on systemic insulin action and metabolism.

D. General insulin action and inflammation
Aging is usually associated with increased insulin resistance (186). In parallel to age-related insulin resistance, the production of circulating APP (187), the secretion of cytokines from monocytes and macrophages (188), and the production of TNF- (186) and sTNFR2 (189) are all increased with age.

We will review the evidence according to which cytokines are involved in general insulin action.

1. TNF- and insulin resistance
a. TNF- gene and insulin resistance.
The relative importance of TNF- in insulin action has been tested by inducing a targeted null mutation in the TNF- gene. Mice with this mutation were spared from obesity-induced deficiencies in insulin-receptor signaling in fat and muscle tissues (190, 191). However, it should be recognized that insulin resistance is alleviated but not eliminated in these models. In humans, similar information can be obtained by comparing individuals with different transcription rates of the TNF- gene. Some substitutions in the TNF- promoter gene lead to different constitutive and inducible levels of transcription of the TNF- gene than the wild-type allele (192, 193). Those subjects that are homozygotes for the absence of the restriction site (resulting from a guanine to adenine substitution) at position -308 of the TNF- promoter showed an increased percentage of fat mass and leptin levels and a decreased insulin sensitivity index (194). The 308G/A TNF- polymorphism has subsequently been linked to obesity in an epidemiological basis (195, 196) to body fat content (197), to insulin resistance (198, 199), and to fasting glucose (200). Secretion of TNF- from adipose tissue also differed among nonobese subjects according to TNF- -863C/A polymorphism (193). Adipose tissue from subjects with the rare allele -863/A secreted significantly less TNF- than adipose tissue from nonobese subjects carrying the -863C allele. This indicated that C to A substitution at position -863 represented a functional polymorphism, which leads to decreased TNF- gene expression and thereby less production and secretion of the cytokine (193). In parallel to decreased TNF- secretion, fasting serum triglycerides were significantly lower, and insulin sensitivity was significantly higher in subjects with the rare allele -863/A in a subsequent study (201). Thus, different adipocyte secretion rates of TNF- according to -863C/A TNF- gene polymorphism are mirrored at the level of insulin action, at least in moderately obese subjects. Another substitution, the homozygous -857T allele, tended to be higher in obese patients with diabetes than in lean subjects with normal glucose tolerance (202).

Mutations within regulatory elements of the TNF- gene were not associated with an increase in the prevalence of noninsulin-dependent diabetes mellitus (202, 203). However, in the latter reports, insulin resistance was not evaluated. In other studies, the inclusion of subjects with morbid obesity probably led to false-negative associations when evaluating TNF- -308G/A polymorphism and obesity phenotypes (204), in contrast to the studies that have not included them (40, 194) or that have evaluated a small proportion of these subjects (195, 197, 200). Some associations among obesity, obesity-associated phenotypes, and cytokines have been found to be most significant in nonmorbidly obese individuals (40). Enhanced activity of cytokines due to the development of obesity is, on one hand, predicted to contribute to the development of obesity-associated phenotypes but is, on the other hand, expected to limit the progression of obesity. These mechanistic relationships are probably lost in subjects with morbid obesity. Differences in study design have also contributed to disparity of results: insulin sensitivity or insulin secretion was not significantly different in healthy young relatives of type 2 diabetic patients with the A allele (205, 206, 207). Finally, interethnic differences, gene-gene and gene-environment interactions are important confounders of any association between a single polymorphism and disease (208, 209, 210).

Other components of the TNF- axis seem to be genetically linked to insulin resistance. Mice lacking the TNFR-2 gene, TNFR2 (p75^-/-), fed a high-fat diet, consistently gained less weight and displayed reduced insulin levels, as expression of improved insulin sensitivity, in comparison with wild-type mice that followed a similar diet (211). In humans, a mutation in the TNFR2 gene has been associated with increased BMI and leptin levels in parallel to insulin resistance in nondiabetic subjects and increased BMI and leptin concentration in diet-treated type 2 diabetic patients (212). Interestingly, other mutations in this locus have been associated with other components of the insulin resistance syndrome such as hypertension (213), dyslipidemia (214), and cardiovascular disease (90), and with other degenerative diseases such as osteoporosis (215). This locus has also been associated with polycystic ovary syndrome, an entity with known defects in insulin action (216).

The differences in insulin resistance seem to be restricted to the TNF- gene because polymorphisms of the TNF-ß gene, a closely associated gene, were not associated with this phenotype (217). It should be remembered that much of TNF- is secreted, whereas most of TNF-ß is on the membrane of the lymphocytes. Again, interethnic differences have also been described. Insulin resistance was significantly lower in variant TNF-ß homozygotes vs. wild-type allele in Japanese subjects (209), whereas hyperinsulinemia, attributed to TNF-ß gene polymorphism, was described in Caucasians with CHD (218).

b. TNF- action and insulin resistance.
TNF- blocks the action of insulin in cultured cells and whole animals (107, 108, 109). The induction of insulin resistance is mediated through its ability to produce serine phosphorylation of insulin receptor substrate 1, decreasing the tyrosine kinase activity of the insulin receptor (109). Neutralization of TNF- in obese fa/fa rats by iv administration of a sTNFR-IgG chimeric protein substantially improved insulin sensitivity and restored the tyrosine kinase activity in fat and muscle (107). It also reverted the insulin-induced phosphorylation of insulin receptor substrate 1 to levels observed in lean animals (109). In contrast, treatment of type 2 diabetic or obese human subjects with an antibody specific for TNF had no effect on insulin sensitivity (219, 220). However, one of these studies was performed using one single administration of antibody in adults with established diabetes (219). Moreover, this approach does not affect the autocrine and paracrine effect of TNF- and is not directed against the primary endogenous stimulus for increased TNF- secretion.

Again, it should be remembered here that increased sTNFR2 levels circulate in association with insulin resistance in healthy volunteers (77, 78), in lean nondiabetic offspring of type 2 diabetic subjects (221), and in young obese subjects with normal and impaired glucose tolerance (142), but not in older subjects (222).

2. IL-6 and insulin resistance
a. IL-6 gene and insulin resistance.
Mice with a targeted null mutation in the IL-6 gene, made obese by a high-fat diet, became more insulin resistant compared with wild-type controls (139). However, as stated above, total depletion of IL-6 might be detrimental because other proinflammatory cytokines (TNF-) are not adequately down-regulated.

This information is apparently at odds with what is observed in humans with different transcription rates of the IL-6 gene. Subjects with increased constitutive transcription rate of this cytokine, associated with -174 C/G IL-6 promoter substitution, showed decreased insulin sensitivity (161), and this polymorphism was more frequently present among Pima Indians and Caucasians with type 2 diabetes (223). Again, divergent results have been described in other populations (224) in keeping with the observation that IL-6 promoter haplotypes (rather than simply single variant sites) influence IL-6 gene expression in vitro (225). Importantly, at least two allelic polymorphisms in the IL-6 promoter region were found to cooperate in the regulation of IL-6 activity in vivo (225).

b. IL-6 action and insulin resistance.
Available data suggest that although TNF- functions locally at the level of the adipocyte in a paracrine fashion, IL-6 circulates in plasma at high concentrations. In this sense, IL-6 may be more important systemically and perhaps represents a hormonal factor that induces muscle insulin resistance. In fact, IL-6 is named the endocrine cytokine (226). Although adipose cells contribute to one third of circulating IL-6 concentration (135), other sources are potentially important. In this sense, glucose stimulated IL-6 production by human peripheral blood monocytes has been demonstrated, but 33 mmol/liter glucose induced only a 1.56-fold increase in IL-6 compared with treatment with 11 mmol/liter glucose (227). Although this mechanism is not operative with normal fasting glucose levels, it could be of significance in patients with type 2 diabetes.

IL-6 is a pleiotropic cytokine, and some of its metabolic actions have been evaluated after IL-6 administration. In in vitro studies, IL-6 induced a dose-dependent inhibition of the glucose-stimulated insulin release of rat pancreatic islets (228, 229). In vivo, administration of recombinant human IL-6 to normal subjects induced metabolic changes usually found in catabolic states, increasing plasma glucose levels in a dose-dependent fashion without altering significantly plasma insulin or C-peptide concentrations (230). In another study in cancer patients, however, recombinant human IL-6 administration led to an increase in the metabolic clearance of glucose (96). To integrate these opposite actions, one has to consider that these metabolic effects of IL-6 have been studied mainly after exogenous treatment at relatively high doses. It is also important to keep in mind the metabolic milieu in which IL-6 is exerting its effects; cytokines act in cascade, and any single change in a given step could change the final result.

Another way to evaluate IL action is to infer it from IL-6 concentration, but this approach cannot conclude which is the cause and which is the consequence. For instance, plasma IL-6 levels are elevated in type 2 diabetic patients, particularly in those with features of the insulin resistance syndrome (7, 8). One interpretation could be that type 2 diabetes mellitus and the insulin resistance syndrome lead to an ongoing acute-phase response through increased IL-6 derived from unsuppressed adipose or immune secretion acting on hepatocytes that are oversensitive (Fig. 4). In fact, this hypothesis is derived from the findings of increased blood concentrations of markers of the acute-phase response, including CRP, serum amyloid-A, -1 acid glycoprotein, sialic acid, and cortisol in these conditions (7, 8). But the reverse could also be true. According to the opposite vision, increased IL-6 and markers of the acute-phase response are perhaps counteracting hyperglycemia and insulin resistance (Fig. 4, step 1). When proinflammation is enduring, chronic, or uncontrolled and when a challenge becomes overwhelming, the failure to reach the desirable effect results in worsening of hyperglycemia and insulin resistance (Fig. 4, step 2). Lack of exposure would result in proinflammation only when insulin resistance is severe (step 3). These possibilities are summarized in Fig. 4.

View larger version (25K): [in this window] [in a new window]   Figure 4. Proposed mechanisms leading to insulin resistance-related inflammation. Type 2 diabetes mellitus and the insulin resistance syndrome lead to an ongoing acute-phase response through increased IL-6 derived from unsuppressed adipose or immune secretion acting on hepatocytes that are oversensitive (step 3). In fact, this hypothesis is derived from the findings of increased blood concentrations of markers of the acute-phase response, including CRP, serum amyloid-A, -1 acid glycoprotein, sialic acid, and cortisol in these conditions. According to the opposite vision, increased IL-6 and markers of the acute-phase response are perhaps counteracting hyperglycemia and insulin resistance (step 1). When proinflammation is enduring, chronic, or uncontrolled, and when a challenge becomes overwhelming, the failure to reach the desirable effect results in worsening of hyperglycemia and insulin resistance (step 2). Lack of exposure would result in proinflammation only when insulin resistance is severe (step 3).

Plasma IL-6 and insulin sensitivity relationships seem to occur in parallel to increases in plasma nonesterified FA (233). A strong linear relationship between the secretions of IL-6 and TNF- from the adipose tissue (r = 0.81; P < 0.0001) was also reported (233). Circulating IL-6 concentration has been described to predict the development of type 2 diabetes mellitus in women. The relative risk of future type 2 diabetes mellitus for women in the highest vs. lowest quintile of these inflammatory markers was 7.5 (95% confidence interval, 3.7–15.4; Ref. 237).

Acute infections determine insulin resistance, and even after clinical recovery, some impairment in carbohydrate metabolism persists (238). Both IL-6 action (229) and acute infections (238) are characterized by a defect in insulin-stimulated glucose use, despite normal carbohydrate oxidation. It cannot be excluded that chronic or subclinical infections have contributed simultaneously to increased IL-6 levels and insulin resistance. Of note is that a higher peripheral white blood cell count has been associated with insulin resistance since the preliminary observations in 1992 (239). In particular, it was observed that peripheral white blood cell count correlated significantly with insulin-mediated glucose disposal during a euglycemic clamp (239). In subsequent studies, it was demonstrated that neutrophil and lymphocyte count correlated positively with several components of the insulin resistance syndrome and that plasma insulin concentration was specifically associated with the number of lymphocytes and monocytes (240). These associations were confirmed in healthy subjects. Given that IL-6 is involved in hematopoiesis (241), it has been suggested that these associations might in part be due to different IL-6 concentration. Those subjects with increased constitutive transcription rates of IL-6 would be prone, simultaneously, to insulin resistance and increased peripheral white blood cell count (161). Interestingly, increased platelet count, another component of hematopoiesis, occurs in parallel to white blood cell count in healthy subjects (242).

3. Other cytokines and insulin resistance.
Recently, a low production capacity of IL-10, a centrally operating cytokine with strong antiinflammatory properties by antagonizing IL-6 and TNF-, was found to be associated with the metabolic syndrome and type 2 diabetes in old age (243). The production capacity of the antiinflammatory cytokine IL-10 was assessed in a whole-blood assay in which LPS was used as a stimulus. Serum concentrations of total cholesterol, LDL cholesterol, triglycerides, glucose, and hemoglobin A_1c gradually decreased over strata representing higher IL-10 production capacity, whereas the concentration of HDL cholesterol gradually increased. The odds ratio for type 2 diabetes was 2.7 (95% confidence interval, 1.5–4.9) when subjects with the lowest IL-10 production capacity were compared with those with the highest IL-10 production capacity (243).

The overall picture relating insulin resistance to inflammatory pathways could be summarized as shown in Fig. 5.

View larger version (20K): [in this window] [in a new window]   Figure 5. Integrated regulative pathways leading to decreased insulin action through modulation of the inflammatory cascade. Dotted line, Negative effect. NEFA, Nonesterified FA; IR, insulin resistance.

	V. Evidence Linking Proinflammatory Cytokines to Cardiovascular Disease through Metabolic Pathways

Top Abstract I. Introduction II. Hypertension and Chronic... III. Dyslipidemia as a... IV. Insulin Resistance and... V. Evidence Linking... VI. Inflammatory Markers Predict... VII. Antiinflammatory Agents... VIII. Possible Interpretations,... IX. Summary References

A. TNF- and cardiovascular disease
TNF- decreases collagen synthesis and increases matrix metalloproteinase activity in vitro, perhaps leading to plaque rupture (249). Human atherosclerotic lesions have been found to contain TNF- mRNA (250, 251). Early accumulation of monocytes and lymphocytes in the aortic intima with cytokine release may be essential in the development of the disease. The accumulation of cholesteryl esters in macrophages exposed to LDL is associated with increased synthesis and release of TNF- (85).

The associations of TNF- gene polymorphism and myocardial infarction have been investigated in Northern Ireland and France (195). The TNFA2 allele, linked to increased transcription rate of the cytokine, was associated with parental history of myocardial infarction. The odds ratio for premature myocardial infarction in the presence of TNF-/308A was 1.76 (1.14–2.70). In Northern Ireland, the prevalence of TNFA2 homozygosity tended to be higher in IHD patients than in controls (P = 0.06; Ref. 195). In another study, TNF- gene polymorphism was not associated with clinical and angiographically assessed coronary stenosis (252). However, in the latter report, nearly 80% of IHD patients were male, in contrast to 43% of control subjects (252). The odds ratio for myocardial infarction tended to be higher (albeit not significantly) in TNF2 homozygotes in Brazil (253). These authors also described a tendency toward increased risk of myocardial infarction conferred by obesity. However, the specific effect of the homozygous state for the mutation could not be fully addressed in their investigation because of the small number of TNF2 homozygotes among Brazilian subjects (253). In another study, the TNF2 allele was associated with increased plasma homocysteine levels, a known potentiator of lipid-related oxidation, although it was not related to the prevalence of IHD (254). TNF2 homozygotes (n = 10) tended to have more fibrous lesions and calcification in their coronary arteries in an autopsy series (255). No association attributed to this polymorphism was found in angiographically examined patients and controls (256). Recently, a robust association between -308 TNF- gene polymorphism and IHD has been found, mainly attributed to women with type 2 diabetes (257). The latter would be even more susceptible by their accompanying charge of concomitant insulin resistance, hypertension, and dyslipidemia, amplifying the constitutively increased TNF- transcription rate present in TNF2 carriers. Interestingly, this gender dimorphism also occurs in the association between several genetic polymorphisms and the development of obesity. For example, ß3-adrenoceptor, ß2-adrenoceptor, LDL-receptor, and -308 TNF- gene polymorphisms have all been described to be linked to obesity in women but not in men (reviewed in Ref. 258). It could be hypothesized that women are prone to TNF--induced metabolic derangements by their physiologically increased fat mass, ultimately leading to atherosclerosis. The gender-specific effect of the TNF- gene could be the result of a gender difference in the regional expression of the gene itself or any other element involved in the cascade of events that leads from activation of the gene to its action in the target tissue.

B. IL-6 and cardiovascular disease
IL-6 has been speculated to play a key role in the development of coronary disease through a number of metabolic, endothelial, and procoagulant mechanisms (reviewed in Refs. 259 and 260). Damage to the vessel wall results in endothelial cell disruption, resulting in exposure of the underlying vascular smooth muscle cells. Endothelial and smooth muscle cells produce IL-6 and IL-6 gene transcripts that are expressed in human atherosclerotic lesions (261, 262). Prospective studies of apparently healthy and high-risk individuals indicate that increased IL-6 levels (54, 263, 264) and elevated CRP concentration (265, 266), a surrogate of IL-6 activity, predicted cardiovascular mortality (263) and future myocardial infarction (264). Substantive interrelationships among circulating IL-6, CRP, and traditional risk factors have been described in women (267). IL-6 has been demonstrated that is a strong independent marker of increased mortality in unstable coronary artery disease and identifies patients who benefit most from a strategy of early invasive management (268). Promoter polymorphisms regulating IL-6 gene expression have been found to be simultaneously associated with circulating levels of CRP (269), carotid artery atherosclerosis (270), and peripheral artery occlusive disease (271), but not with coronary artery disease (272), probably reflecting the heterogeneity of this disease and study designs. Recently, genetic variants of Toll-l