| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Endocrinology and Metabolism |
1 Department of Internal Medicine, Division of Endocrinology, Metabolism, Pathobiochemistry, and Clinical Chemistry, 2 Department of Diagnostic Radiology, and 3 Section on Experimental Radiology, Eberhard-Karls-University of Tübingen, Tübingen, Germany.
aAddress correspondence to this author at: Department of Internal Medicine, Division of Endocrinology, Metabolism, Pathobiochemistry, and Clinical Chemistry, University of Tübingen, Otfried-Müller-Strasse 10, D-72076 Tübingen, Germany. Fax 49-7071-295974; e-mail norbert.stefan{at}med.uni-tuebingen.de.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Methods: We measured plasma adiponectin concentrations, serum lipids, and serum markers of inflammation, atherosclerosis, and endothelial function in 242 Caucasians without type 2 diabetes. We also measured visceral adipose tissue with magnetic resonance tomography and liver and intramyocellular fat with 1H magnetic resonance spectroscopy.
Results: We divided the study participants into 2 groups: lean [mean (SE) total body fat, 26% (0.6%); n = 119] and obese [36% (0.6%); n = 123]. In the obese group, plasma adiponectin concentrations showed a strong positive association with concentrations of HDL cholesterol (P <0.0001) and negative associations with LDL cholesterol, triglycerides, high-sensitivity C-reactive protein, interleukin 6, apolipoprotein B100, soluble E-selectin, soluble vascular cellular adhesion molecule 1, plasminogen activator inhibitor 1, leukocyte count, and liver and intramyocellular fat (all P <0.03). In the lean group, adiponectin showed a less strong association with HDL cholesterol (P = 0.005) and liver fat (P = 0.03) and no significant associations with the other variables (all P >0.10). High visceral adipose tissue was a strong predictor of low adiponectin concentrations, particularly in the obese group, and attenuated many of the significant relationships.
Conclusions: High adiponectin plasma concentrations are associated with favorable lipid profiles, decreased subclinical inflammation, decreased markers of atherosclerosis and endothelial function, and low ectopic fat accumulation, particularly in obese persons. Adiponectin may also have a concentration-related effect on the relationship between visceral adipose tissue and these metabolic characteristics, especially in obese persons.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Emerging evidence from animal and human studies suggests that the beneficial effect of adiponectin on metabolism becomes stronger with increasing adiposity. Adiponectin knockout mice exhibited severe insulin resistance only when fed with a high-fat/high-carbohydrate diet (21). In humans, an adiponectin-encoding gene haplotype leading to low plasma adiponectin concentrations was associated with type 2 diabetes in obese and morbidly obese but not lean persons (22). Furthermore, the relationship of plasma adiponectin with insulin resistance (14)(23), HDL-c, and triglycerides (14) was strengthened with increasing adiposity.
Modulation of adiponectin effects by adiposity might explain inconsistent results of previous studies and further elucidate the role of adiponectin in human metabolism. Therefore, we investigated the effect of adiposity on the relationships of adiponectin with plasma lipids and lipoprotein concentrations, inflammatory markers, and accumulation of fat in liver and muscle, commonly referred to as ectopic fat.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
2 = 0.54, 0.59, and 0.42, respectively) in the 4 subgroups of participants. None of the participants regularly consumed alcohol. Participants were considered healthy according to results of a physical examination and routine laboratory tests. Informed written consent was obtained from all participants, and the local medical ethics committee approved the protocol.
body composition and body fat distribution
Total body fat was measured by bioelectrical impedance (RJL Systems, Inc.). In this method, electrodes are attached to various parts of the patient’s body. A small electric signal is circulated, and the impedance or resistance to the signal as it travels through the water in muscle and fat is measured. The more fat a person has, the more resistance to the current exists.
We measured visceral fat with an axial T1-weighted fast-spin echo technique as previously described (24) with a 1.5 T whole-body imager (Magnetom Sonata, Siemens Medical Solutions).
1h magnetic resonance spectroscopy
Liver fat and intramyocellular fat of the tibialis anterior muscle were measured with 1H magnetic resonance spectroscopy as previously described (25).
oral glucose tolerance test
All study participants underwent a 75-g oral glucose tolerance test. We obtained venous plasma samples at 0, 30, 60, 90, and 120 min for determination of plasma glucose and insulin. We determined glucose tolerance according to the 1997 WHO diagnostic criteria. We used the method proposed by Matsuda and DeFronzo (26) to calculate insulin sensitivity from glucose and insulin values during the oral glucose tolerance test: (10 000 /
(mean insulin x mean glucose) x (fasting insulin x fasting glucose).
analytical procedures
We measured blood glucose with a bedside glucose analyzer (glucose-oxidase method; Yellow Springs Instruments). For measurements of insulin, lipoproteins, apoB100, and apo A-I, blood was placed on ice after drawing, immediately transferred to the laboratory, and subsequently analyzed. We measured plasma insulin with a microparticle enzyme immunoassay (Abbott Laboratories); serum total cholesterol, HDL-c, and LDL-c concentrations with a standard colorimetric method on a Bayer analyzer (Bayer HealthCare); and serum concentrations of apo B100 and apo A-I with an immunonephelometric method (Behringwerke).
For other measurements, serum and plasma samples were frozen immediately and stored at –80 °C. We used RIA (LINCO Research) to measure plasma adiponectin and ELISAs to measure serum concentrations of tumor necrosis factor (TNF) 
, IL-6 (R&D Systems Inc.), soluble E-selectin, soluble intercellular adhesion molecule (sICAM) 2, soluble vascular cellular adhesion molecule (sVCAM) 1, and plasminogen activator inhibitor (PAI) 1 (Bender MedSystems). Published data indicate that storage at –70 °C for several years does not affect plasma adiponectin measurements (27), and serum for measurements of IL-6, TNF-, and markers of endothelial function can also be stored at –70 °C (19)(28)(29).
statistical analyses
Data are given as mean (SE). IL-6 and TNF- measurements were available for 178 and soluble E-selectin (sE-selectin), sICAM-1, sVCAM-1, and PAI-1 for 134 study participants. Data that did not show gaussian distribution (Shapiro-Wilk W test) were logarithmically transformed. First, males (n = 98) and females (n = 144) were separately divided into 2 groups on the basis of the median percentage of body fat (26% for males and 36% for females) measured by bioelectrical impedance. Lean males (n = 48) and lean females (n = 71) and obese males (n = 50) and obese females (n = 73) were then combined into the lean group (n = 119) and the obese group (n = 123). For each of the 2 groups, we performed separate correlation analyses of the association of plasma adiponectin concentrations with the variables of interest. In multivariate regression models, the dependent variables were adjusted for age, sex, and percentage of total body fat, and for some analyses, for visceral fat. To more clearly depict the different relationships of adiponectin with the most important variables, we used median plasma adiponectin concentrations to subdivide the lean and the obese groups into 2 additional groups. We then used multivariate regression models to test differences in the dependent variables after inclusion of adiponectin as a categorical variable in the lean and in the obese group. A P value <0.05 was considered statistically significant. We used the statistical software package JMP 4.0 (SAS Institute Inc).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
. Correlation analyses were first performed in the lean (n = 119) and obese (n = 123) groups. The obese group, after adjustment for age, sex, and total body fat, showed a strong positive association of plasma adiponectin concentrations with HDL-c and a negative association with LDL-c and triglycerides. In contrast, in the lean group, only a weaker, positive association of adiponectin concentrations with HDL-c was observed (Table 2
).
|
|
Similarly, when the relationships of adiponectin plasma concentrations with markers of inflammation, atherosclerosis, and endothelial function were investigated, we found negative correlations of adjusted adiponectin concentrations with leukocyte count, hsCRP, IL-6, PAI-1, sE-selectin, sVCAM-1, apo B100, and apo A-I/ apo B100 ratio in the obese group, but none of these relationships were significant in the lean group.
We further investigated whether adiposity modulates the well-documented negative relationship between plasma adiponectin concentrations and liver and intramyocellular fat, both measured by 1H magnetic resonance spectroscopy. The obese group showed a negative association of adjusted adiponectin concentrations with liver fat and a less strong association with intramyocellular fat of the tibialis anterior muscle. In contrast, in the lean group, the association of adiponectin concentrations with liver fat was weaker, and the association of adiponectin concentrations with intramyocellular fat was not significant (Table 2
).
When we analyzed data from males and females separately, we found similarly strong relationships for females. In the male groups, the relationships were in the same direction, but weaker, probably because of the smaller sample size (data not shown).
Given that most reported data suggest a prominent role of visceral adipose tissue in adiponectin production, we investigated whether high visceral fat in obese persons accounted for the stronger associations between adiponectin and the variables mentioned above. Visceral fat, adjusted for age and sex, was positively associated with BMI (r = 0.58, P <0.0001), waist circumference (r = 0.61, P <0.0001), total body fat measured by bioelectrical impedance (r = 0.43, P <0.0001), liver fat (r = 0.54, P <0.0001), and intramyocellular fat (r = 0.20, P = 0.05) in lean and in obese study participants (r = 0.47, P <0.0001; r = 0.51, P <0.0001; r = 0.22, P = 0.02; r = 0.56, P <0.0001; r = 0.18, P = 0.07).
Adiponectin concentrations, adjusted for age and sex, were negatively associated with BMI (r = –0.34, P = 0.0002), waist circumference (r = –0.32, P = 0.0005), and visceral fat (r = –0.35, P = 0.0001) in lean and in obese study participants (r = –0.15, P = 0.10; r = –0.21, P = 0.02; r = –0.42, P <0.0001).
Adiponectin concentrations were more closely associated with visceral fat in obese (r = –0.41, P <0.0001) than in lean (r = –0.28, P = 0.0029) individuals. In these models, total body fat did not significantly contribute to the variability in plasma adiponectin concentrations in either group. Thus, although correlation analyses cannot fully address whether adiponectin is predominantly produced in abdominal fat, our data suggest that visceral fat may be a stronger determinant of adiponectin concentrations in obese than in lean individuals.
Visceral fat was significantly associated with most of the variables tested in this study (in the obese group, with adiponectin, HDL-c, LDL-c, apo B100, triglycerides, hsCRP, liver fat, and intramyocellular fat; all P 
0.02). When visceral fat was included in the multivariate regression models (Table 2
), in the obese group, the relationships of adiponectin with leukocyte count, HDL-c, triglycerides, and apo B100 were still significant. All other relationships between adiponectin and the variables tested were rendered nonsignificant.
To graphically better depict the different relationships between adiponectin plasma concentrations and the variables of interest, we further analyzed the data and present our results with a different approach, for which individuals in the lean and in the obese groups were further divided into 2 groups according to median plasma adiponectin concentrations (Table 3
). Investigation of differences between serum lipid variables, adjusted for age, sex, and percentage of body fat, revealed that in the obese group HDL-c concentrations were higher (Fig. 1A
) and LDL-c (Fig. 1B
) and triglyceride concentrations (Fig. 1D
) were lower in the group with high vs low adiponectin plasma concentrations. No statistical difference was seen for total cholesterol concentrations (Fig. 1C
). Except for a small difference in HDL-c (Fig. 1A
), no significant differences were observed between the lean groups with high vs low adiponectin (all P >0.36, Fig. 1
, B–D). In the obese but not the lean group with high vs low adiponectin plasma concentrations, hsCRP (Fig. 1E
), IL-6 (Fig. 1F
), and apo B100 (Fig. 1G
) concentrations and the apo B100/apo A-I ratio (Fig. 1H
) were significantly lower, as were liver fat (Fig. 2A
) and intramyocellular fat (Fig. 2B
).
|
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
A positive relationship of adiponectin with HDL-c and a negative relationship with triglycerides have been reported in many studies (6)(7)(8)(9)(10)(11)(12)(13)(14)(15), but only a few tested whether adiposity modulates these relationships. A recent study (14) found adiponectin to be associated with HDL-c and triglycerides in both lean and obese adolescents, with the relationships becoming stronger with increasing adiposity. Our results confirm these findings and extend them to adults. Two other studies (9)(13) also found associations of adiponectin concentrations with serum HDL-c and triglycerides. The relationship with HDL-c was stronger in the nonobese group (9), however, and the relationship with triglycerides was statistically significant only in individuals with BMI <30 kg/m2 (13). This study included only women with type 2 diabetes, which may account for the difference between their results and ours. The lack of a correlation of lipid-related variables with adiponectin concentrations in obese diabetic women may be attributable to a bottom effect, in which plasma adiponectin concentrations are very low and there is little variability in adiponectin concentrations in individuals with type 2 diabetes (2). In men with type 2 diabetes, there also was no statistically significant effect of obesity on the association of adiponectin with HDL-c (12). In addition to the confounding effect of type 2 diabetes, there was less variation in BMI in the latter study population than in our cohort.
In the presence of dyslipidemia, adiponectin concentrations may be more strongly correlated with HDL-c and triglycerides in lean individuals, but we could not test this hypothesis because of the small number of dyslipidemic participants in our study (total n = 53 in the lean and in the obese group according to National Cholesterol Education Program Adult Treatment Panel III criteria of the metabolic syndrome for triglycerides and HDL-c). This issue must be addressed in future studies.
Ambiguous data have been reported regarding the association of adiponectin with apo B (8)(10)(11)(12)(13). The studies that did not find significant correlations included mainly nonobese individuals (8)(11). Possibly for the same reason, as with HDL-c and triglycerides, no interaction of obesity with adiponectin on apo B was found in the study by Schulze et al. (12).
The mechanisms underlying the observed relationships between low plasma adiponectin concentrations and dyslipidemia are unknown. A hypothesis proposed by Cnop et al. (8) implicates central obesity and insulin resistance, both of which are associated with hypoadiponectinemia. The relationships of adiponectin with HDL-c and triglycerides, however, were found to be independent of whole-body insulin resistance (6)(7). Furthermore, adiponectin was shown to be an independent predictor of VLDL-apo B catabolism (30). A recent study also suggested that central obesity plays a major role in the relationships of adiponectin with triglycerides, C-reactive protein (CRP), and tissue plasminogen activator (31). Our results support the view that the associations of adiponectin with lipids are partially independent of visceral adipose tissue, and our data are in agreement with the results from other studies (30)(32), supporting the notion that adiponectin has direct effects on lipoprotein metabolism. The apparent inconsistency in our finding that the association of adiponectin with apoB100 remained statistically significant after additional adjustment for visceral fat, in contrast to the association of adiponectin with LDL-c, may be attributed to the effects of adiponectin on increasing hepatic insulin sensitivity, which lead to decreased apo B concentrations (8). No such clear relationships were reported for LDL-c. Nevertheless, we provide novel data suggesting that low adiponectin concentrations are strongly predicted by visceral adipose tissue, particularly in obese persons, supporting the hypothesis that in obese individuals, who commonly have increased visceral fat, adiponectin concentrations steeply decrease with increasing abdominal fat mass, thus permitting the accruement of an unfavorable metabolic profile. In light of this finding, it is apparent that adiponectin is an important mediator between visceral fat and metabolism, particularly in obese persons.
Hepatic lipase may be a modulator between adiponectin and lipoproteins. This enzyme was found to regulate lipoprotein metabolism through several pathways (33), including interaction with cell-surface heparan sulfate proteoglycans and the LDL receptor in promoting uptake of apo B-containing remnant lipoproteins. This role is supported by findings that overproduction of catalytically active human hepatic lipase in the liver in mice led to decreased HDL-c (34).
An inverse relationship, independent of insulin resistance, between plasma adiponectin concentrations and plasma hepatic lipase activity was found in humans (35). Hypoadiponectinemia was also shown to be associated with decreased plasma lipoprotein lipase activity and thus with increased triglycerides, again independently of insulin resistance (36).
We also found that adiposity modulated the association of adiponectin with markers of subclinical inflammation. An inverse correlation of adiponectin with IL-6 was recently reported (18)(19). All studies (12)(13)(16)(17) except 1 (20) demonstrated an inverse correlation of adiponectin with hsCRP concentrations. CRP is mainly produced in the liver, and IL-6 is important in controlling hepatic CRP production (37), suggesting that the association of adiponectin with CRP is similar to that of adiponectin with IL-6, i.e., becoming stronger with increasing adiposity. This theory may also explain the results of the single study (20) that did not find an association of adiponectin with hsCRP. In that study, lean females with a mean BMI of 20 were included (20).
In obese individuals, plasma adiponectin was associated with the endothelial markers sICAM-1 and sE-selectin and with the proatherothrombotic PAI-1. After adjustment for visceral fat, however, only the relationship with sE-selectin was still statistically significant. sE-selectin was shown to be inversely correlated with adiponectin, (13)(19). In 1 study (19), which included Pima Indians with a mean BMI of 36, adiponectin was associated only with sE-selectin but not with sICAM or sVCAM independent of BMI and waist circumference, in accordance with our data. We do not have a conclusive explanation for this finding, given that adiponectin was found to down-regulate production of all 3 adhesion molecules in aortic endothelial cells in vitro (38). A putative explanation is that sE-selectin may more closely reflect E-selectin production on endothelial cells than the soluble form of the other 2 markers. SICAM-1 was found to correlate with adiponectin in 1 study (13) but not in others (12)(19)(31). One of these studies showed a strong association of adiponectin with PAI-1 (31). None of these studies tested obesity interactions.
We found that the degree of adiposity modulates the relationship of plasma adiponectin with liver and intramyocellular fat. Liver fat in particular is closely associated with visceral obesity (39). In our analyses, the relationships of adiponectin with liver fat and intramyocellular fat were not independent of visceral fat, possibly because of strong intercorrelations of these variables. Nevertheless, the relationship of adiponectin with ectopic fat is probably causal, because adiponectin, by activating AMP-activated protein kinase, regulates many enzymes involved in fatty acid oxidation (40).
It is not clear why the relationship between adiponectin and liver fat is particularly strong in obese individuals. TNF- and IL-6 may play roles in this relationship; both are known to exhibit reciprocally antagonistic actions with adiponectin at the tissue level (41). TNF- in particular suppresses expression of genes involved in fatty acid oxidation and has been implicated in the pathogenesis of alcoholic and nonalcoholic steatohepatitis (41). In our study and others (18)(19)(31), plasma TNF- concentrations were not associated with adiponectin concentrations. In other studies, TNF- receptor concentrations also were not correlated (13) or were only weakly correlated(15) with adiponectin concentrations. Nevertheless, hepatic TNF- production was found to be increased in obese individuals with fatty liver (41). High TNF- concentrations leading to decreased adiponectin production were also found in adipose tissue of obese individuals (42). Thus, both decreased adiponectin plasma concentrations and inhibition of lipid oxidation in the liver are net results of locally increased TNF- concentrations, making the association between adiponectin plasma concentrations and liver fat particularly strong in obese individuals.
A limitation of our study was the cross-sectional character of the examination, which restricted our ability to determine causal relationships between plasma adiponectin concentrations and the variables investigated. Another limitation is that our lean group contained not only lean individuals, but also few individuals with body fat percentage similar to individuals in the obese group. In a previous study in which we investigated the relationship of adiponectin concentrations with insulin sensitivity estimated from the oral glucose tolerance test, we found no statistically significant correlation in individuals with body fat in the lowest quartile [mean (SE), 16.63% (0.22%)] (23). In the present study, mean body fat was higher in the lean group [26.51% (0.58%)], and thus, in this group, a significant relationship existed between adiponectin concentrations and insulin sensitivity, albeit weaker than in the obese group (data not shown). Nevertheless, this finding supports the hypothesis that the effects of plasma adiponectin on metabolic markers are weak or absent in lean individuals.
Implications of this study are that lifestyle and pharmacological interventions (e.g., thiazolidinediones) aimed at decreasing the risk of atherosclerotic vascular disease by increasing plasma adiponectin concentrations may be more effective in obese than in lean persons. Moreover, such interventions would be more effective if they decreased visceral adipose tissue, as has been discussed for endocannabinoid receptor antagonists (43).
In conclusion, we provide novel information that the associations of adiponectin with plasma lipids and lipoproteins, markers of inflammation and endothelial function, and ectopic fat become stronger with increasing adiposity. This finding may explain some of the inconsistent results of previous studies. Furthermore, adiponectin concentrations may be an important mediator between visceral adipose tissue and metabolic characteristics, especially in obese persons.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
1 Nonstandard abbreviations: HDL-c, HDL cholesterol; apo, apolipoprotein; LDL-c, LDL cholesterol; hsCRP, high sensitivity C-reactive protein; IL-6, interleukin-6; BMI, body mass index; TNF-, tumor necrosis factor ; sICAM-2, soluble intercellular adhesion molecule-2; sVCAM-1, soluble vascular cellular adhesion molecules; PAI-1, plasminogen activator inhibitor 1; CRP, C-reactive protein; sE-selectin, soluble E-selectin.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
. Cytokine 1990;2:353-356.[CrossRef][Medline]
[Order article via Infotrieve]The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  A. Cabre, I. Lazaro, J. Girona, J. M. Manzanares, F. Marimon, N. Plana, M. Heras, and L. Masana Plasma fatty acid binding protein 4 is associated with atherogenic dyslipidemia in diabetes J. Lipid Res., August 1, 2008; 49(8): 1746 - 1751. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Kotronen and H. Yki-Jarvinen Fatty Liver: A Novel Component of the Metabolic Syndrome Arterioscler. Thromb. Vasc. Biol., January 1, 2008; 28(1): 27 - 38. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Sir-Petermann, B. Echiburu, M M. Maliqueo, N. Crisosto, F. Sanchez, C. Hitschfeld, M. Carcamo, P. Amigo, and F. Perez-Bravo Serum adiponectin and lipid concentrations in pregnant women with polycystic ovary syndrome Hum. Reprod., July 1, 2007; 22(7): 1830 - 1836. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. D. Knudson, G. M. Dick, and J. D. Tune Adipokines and Coronary Vasomotor Dysfunction Experimental Biology and Medicine, June 1, 2007; 232(6): 727 - 736. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Stefan, A. M. Hennige, H. Staiger, J. Machann, F. Schick, E. Schleicher, A. Fritsche, and H.-U. Haring High Circulating Retinol-Binding Protein 4 Is Associated With Elevated Liver Fat but Not With Total, Subcutaneous, Visceral, or Intramyocellular Fat in Humans Diabetes Care, May 1, 2007; 30(5): 1173 - 1178. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
