NATAP: IL-18 & HIV-Associated Lipodystrophy

[Guest guest]

NATAP - www.natap.org WELCOME TO 2005

IL-18 & HIV-Associated Lipodystrophy

This story contains two published articles in JAIDS & AIDS about IL-18 & lipodystrophy. The two articles report finding IL-18 expression in adipose tissue (fat tissue) for the first time the authors say and in blood (plasma) in people with lipodystrophy; and authors feel IL-18 is associated with lipodystrophy.

Author Lindegaard suggests:

IL-18 induces apoptosis in other cell types, and it is possible that IL-18 in HIV patients can induce adipocyte apoptosis and thereby be mechanistically involved in the lipodystrophy syndrome. Alternatively, increased IL-18 mRNA levels may be a secondary phenomenon to a general inflammation in AT in lipodystrophic patients induced by HAART.

No correlation existed between IL-18 mRNA and circulating cytokines, whereas IL-18 mRNA correlated with the expression of IL-6, TNF-α and IL-8 mRNA in AT. As TNF-α, IL-6 and IL-8 mRNA are downstream targets of IL-18 mRNA, this may indicate that IL-18 is produced in AT and induces TNF-α, IL-8 and IL-6 expression in AT.

In conclusion, given that IL-18 is involved in apoptosis, the present finding that IL-18 is expressed in AT and is associated with lipoatrophy suggests that IL-18 is involved in the pathogenesis of the lipodystrophy syndrome in HAART-treated HIV patients.

Adipose tissue expression of IL-18 and HIV-associated lipodystrophy

AIDS: Volume 18(14) 24 September 2004

Lindegaard, Birgittea,b; Hansen, Ann-Brit Ega; Pilegaard, Henrietteb,c; Keller, Pernillea,b; Gerstoft, Jana; Pedersen, Bente Klarlunda,b

aDepartment of Infectious Diseases, bThe Copenhagen Muscle Research Centre, Rigshosptalet and cAugust Krogh Institute, University of Copenhagen, Copenhagen, Denmark.

Summary. IL-18 is an inducer of apoptosis/tissue injury. IL-18 messenger RNA expression was examined in adipose tissueobtained from HIV patients with lipodystrophy, without lipodystrophy and healthy controls. IL-18 mRNA was expressed in AT at increased levels in lipodystrophy-positive compared with lipodystrophy-negative patients and healthy controls. Higher levels of IL-18 mRNA were found in femoral-gluteal AT compared with abdominal AT, and correlated with limb fat loss. These findings suggest that IL-18 is linked to HIV-associated lipodystrophy.

More than 50% of HIV patients receiving highly active antiretroviral therapy (HAART) develop lipodystrophy, characterized by the subcutaneous loss of adipose tissuefrom the limbs, buttocks and face. The pathogenesis of lipodystrophy remains elusive. Subcutaneous AT from HIV patients with lipodystrophy demonstrates adipocyte apoptosis and signs of inflammation with high expressions of TNF-α, IL-6 and IL-8. We recently found increased plasma levels of IL-18 in HIV patients with lipodystrophy. IL-18, a pro-inflammatory cytokine, is believed to be involved in apoptosis and ultimately tissue injury in inflammatory diseases. IL-18 induces TNF-a, and TNF-α also stimulates IL-18 production. IL-18 also induces IL-8, a chemotactic factor for neutrophil infiltration, and the production of IL-6. The circulating level of IL-18 is elevated in obesity, declines with weight loss [12], and is elevated in HIV patients with lipodystrophy. Given that IL-18 is associated with altered AT mass, inflammation and apoptosis, we hypothesized that IL-18 belongs to the new class of AT-derived cytokines, the adipocytokines, and that lipodystrophy and especially limb fat loss is associated with the high AT expression of IL-18.

In a cross-sectional study, we included 17 HIV patients with lipodystrophy, 17 without lipodystrophy and 22 healthy controls without the metabolic syndrome, all men and matched for age. All patients were on stable and effective nucleoside analogue-based HAART with no changes during the preceding 8 weeks. Of the patients with lipodystrophy, 12 were receiving protease inhibitors and nine were receiving non-nucleoside analogues. In those without lipodystrophy, nine patients were receiving protease inhibitors and three were receiving non-nucleoside analogues. Fat and fat-free tissue masses for the whole body, trunk and extremities were measured using dual-energy X-ray absorptiometry.

IL-18 mRNA was measured in subcutaneous abdominal and femoral-gluteal AT biopsies. AT RNA was purified and IL-18 mRNA was measured by real-time polymerase chain reaction. Human IL-18 Taqman probe and primers were amplified using pre-developed assay reagents obtained from Applied Biosystems ( City, CA, USA). The gene expression levels were normalized to the housekeeping gene 18S. Cytokines were measured in plasma using enzyme-linked immunosorbent assay.

HIV patients did not differ with regard to the duration of HIV infection, CD4 cell counts and viral loads. The mean duration of therapy was 96 months versus 67 months (P < 0.01, Student's t-test), and the mean body mass index was 21 versus 24 kg/m2 (P < 0.01, analysis of variance; ANOVA) in patients with lipodystrophy and those without lipodystrophy, respectively, but did not differ between lipodystrophy-negative patients and controls.

The dual-energy X-ray absorptiometry scans supported the clinical diagnosis of lipodystrophy. HIV patients were characterized by a reduced total fat mass [mean ± 95% confidence interval (CI) lipodystrophy-positive 9.85 kg; 8.04-11.7; lipodystrophy-negative 14.8 kg; 12.5-17.1; controls 19.9 kg; 17.8-21.9; P < 0.0001, ANOVA], truncal fat mass (lipodystrophy-positive 5.86 kg; 4.46-7.27; lipodystrophy-negative 8.02 kg; 6.43-9.62; controls 10.57 kg; 9.38-11.8, P < 0.0001) and limb fat mass (lipodystrophy-positive 3.15 kg; 2.59-3.72; lipodystrophy-negative 5.88 kg; 5.13-6.64; controls 8.41 kg; 7.55-9.28; P < 0.0001). The limb fat loss in patients with lipodystrophy was more pronounced compared with those without lipodystrophy (P < 0.0001), and the trunk : limb ratio was higher in patients with lipodystrophy compared with those without lipodystrophy and controls (lipodystrophy-positive 1.93; 1.53-2.34; lipodystrophy-negative 1.35; 1.18-1.52; controls 1.26; 1.17-1.36; P < 0.01). All groups had the same lean body mass.

IL-18 mRNA was expressed in subcutaneous AT in all three groups. IL-18 mRNA expression in abdominal subcutaneous AT was elevated in patients with lipodystrophy compared with those without lipodystrophy [geometric mean 1.75 (95% CI 1.10-2.77) versus 0.71 (95% CI 0.49-1.01), P = 0.008, ANOVA]. There was no difference between controls (0.95; 95% CI 0.69-1.32) and lipodystrophy-negative patients. In the study group, there was an overall greater expression of IL-18 mRNA in subcutaneous femoral-gluteal AT (1.39; 95% CI 1.11-1.73) compared with abdominal AT (0.96; 95% CI 0.73-1.26; P = 0.003, paired t-test), which was primarily caused by a difference within lipodystrophy-negative patients (P = 0.013). The mean plasma IL-18 was 416 pg/ml (95% CI 341.9-491.4); 347.0 pg/ml (95% CI 272.4-421.6) and 279.6 pg/ml (95% CI 233.9-325.3) in patients with lipodystrophy, those without lipodystrophy and controls, respectively, being significantly different between lipodystrophy-positive patients and controls (P = 0.005, ANOVA).

High levels of IL-18 mRNA in subcutaneous abdominal AT were associated with a high waist : hip ratio (r s = 0.40; P < 0.01, Spearman's correlation test), a high trunk : limb ratio (r s = 0.35; P < 0.05) and a low percentage limb fat mass (r s = -0.35; P < 0.02). The same trend was seen for plasma IL-18. IL-18 mRNA levels correlated positively with TNF-α mRNA (r s = 0.37; P < 0.02), IL-6 mRNA (r s = 0.26; P < 0.05) and IL-8 mRNA (r s = 0.63, P < 0.0001). Plasma IL-18 correlated positively with TNF-α mRNA (r s = 0.53, P < 0.001) and with plasma IL-6 (r s = 0.61; P < 0.0001), but not with IL-18 mRNA (r s = 0.12; NS).

The present study demonstrated for the first time the expression of IL-18 mRNA in human AT. Furthermore, patients with lipodystrophy had increased IL-18 mRNA levels in AT compared with those without lipodystrophy. The finding that high levels of IL-18 mRNA were associated with limb fat loss supports a previous finding of an association between plasma IL-18 and limb fat loss in another cohort of HIV patients. In patients with lipodystrophy, fat atrophy is most pronounced in the peripheral fat regions. The finding of increased IL-18 mRNA in subcutaneous femoral-gluteal AT compared with abdominal AT, is further in accordance with the hypothesis that IL-18 is involved in AT injury.

HIV patients with lipodystrophy demonstrate apoptosis of subcutaneous adipocytes in lipoatrophic areas. IL-18 induces apoptosis in other cell types, and it is possible that IL-18 in HIV patients can induce adipocyte apoptosis and thereby be mechanistically involved in the lipodystrophy syndrome. Alternatively, increased IL-18 mRNA levels may be a secondary phenomenon to a general inflammation in AT in lipodystrophic patients induced by HAART.

No correlation existed between IL-18 mRNA and circulating cytokines, whereas IL-18 mRNA correlated with the expression of IL-6, TNF-α and IL-8 mRNA in AT. As TNF-α, IL-6 and IL-8 mRNA are downstream targets of IL-18 mRNA, this may indicate that IL-18 is produced in AT and induces TNF-α, IL-8 and IL-6 expression in AT.

In conclusion, given that IL-18 is involved in apoptosis, the present finding that IL-18 is expressed in AT and is associated with lipoatrophy suggests that IL-18 is involved in the pathogenesis of the lipodystrophy syndrome in HAART-treated HIV patients.

High Plasma Level of Interleukin-18 in HIV-Infected Subjects With Lipodystrophy

JAIDS Journal of Acquired Immune Deficiency Syndromes: Volume 36(1) 1 May 2004

Lindegaard, Birgitte MD*†; Hansen, Ann-Brit Eg MD*; Gerstoft, Jan MD, DrSci*; Pedersen, Bente Klarlund MD, DrSci*†

From *Department of Infectious Diseases, Rigshospitalet, University of Copenhagen, Denmark, and †the Copenhagen Muscle Research Centre, Rigshospitalet, University of Copenhagen, Denmark.

SUMMARY:

The study examined men with fat accumulation (mixed group) (n = 12) and one without fat accumulation (lipoatrophic group) (n = 15). Controls were HIV-positive men without LD (n = 15) and HIV-negative, age-matched men (n = 12). The study found that the lipoatrophic group had the highest IL-18.

The main finding of this study was that the level of circulating IL-18 is elevated in HIV-infected patients with lipoatrophy (both with and without central fat accumulation) compared with HIV-infected patients without LD.

High levels of IL-18 were found especially in patients who had received antiretroviral therapy for a long period. This could reflect lipoatrophy or that IL-18 was related to a direct effect of >=1 antiretroviral treatment.

A positive correlation was found between the total duration of NRTI use and IL-18 in HIV subjects without the mixed group as well as in the lipoatrophic group. The present study was not designed to study the specific effect of stavudine or other treatments. Patients currently receiving stavudine (n = 11) had significantly higher levels of plasma IL-18 compared with patients currently on regimens without stavudine. Patients who had stavudine switched to other antiretroviral drugs did not have a decrease in the level of IL-18 We found that IL-18 was correlated with TNF-a, thus supporting its role as a true proinflammatory cytokine. Given that IL-18 induces tissue destruction, it may be involved in adipocyte apoptosis occurring in patients with LD

IL-18 is a cytokine with proinflammatory features and is involved in tissue destruction. Thus, it may play a role in the host defense against infectious diseases and tumor growth, but it seems also to play a pathogenetic role in inflammatory conditions.

IL-18 induces interferon-g production and IL-6. In addition, IL-18 induces tumor necrosis factor-a (TNF-a) and TNF-a also stimulates IL-18 production. Thus, the two cytokines IL-18 and TNF-a seem to work in synergy.

The level of interleukin-18 (IL-18) is elevated in patients with HIV infection as well as in people with insulin resistance (IR). The level of IL-18 is elevated in patients with LD and closely linked to limb atrophy, whereas it is not associated with cholesterol or IR.

Elevated levels of IL-18 have recently been demonstrated in HIV-infected patients and especially in those with immunodeficiency and advanced clinical disease and to decline with HAART for 24 months. Recently, the levels of IL-18 were found to be elevated in persons with obesity, type 2 diabetes, and arteriosclerosis and to decline with weight loss. Given that these conditions share common features with HIV-associated LD, we hypothesized that IL-18 would be elevated in these patients. Furthermore, we specifically evaluated whether IL-18 was associated with fat mass, fat distribution, and NRTI treatment and whether IL-18 was associated with other cytokines such as TNF-a, IL-6, and adiponectin.

Abstract:

The level of interleukin-18 (IL-18) is elevated in patients with HIV infection as well as in people with insulin resistance (IR). As HIV-associated lipodystrophy (LD) shares metabolic characteristics with the metabolic syndrome, it was hypothesized that IL-18 would be elevated in patients with LD.

Two groups of HIV-infected men with LD, one with fat accumulation (mixed group) (n = 12) and one without fat accumulation (lipoatrophic group) (n = 15) were included. Controls were HIV-positive men without LD (n = 15) and HIV-negative, age-matched men (n = 12).

The levels of plasma IL-18 were elevated in all 3 HIV groups compared with HIV-negative controls (P <0.01). In the HIV groups the lipoatrophic group had the highest IL-18, followed by the mixed group and the HIV-positive controls. Only the differences between the lipoatrophic group and the HIV-positive controls were significant (P <0.01).

Plasma IL-18 correlated with tumor necrosis factor-a (P <0.05), but not IL-6, adiponectin, or HOMA-IR (homeostasis model of insulin resistance). In contrast to the HIV-negative controls, IL-18 did not correlate with total or low-density cholesterol in either of the HIV groups. An inverse correlation was observed between IL-18 and limb fat (P <0.05).

In conclusion, the level of IL-18 is elevated in patients with LD and closely linked to limb atrophy, whereas it is not associated with cholesterol or IR.

INTRODUCTION

A syndrome of lipodystrophy (LD), including peripheral subcutaneous fat loss, lipoatrophy, and central fat accumulation has been described in HIV-infected patients receiving highly active antiretroviral therapy (HAART). It is still not established whether loss of subcutaneous adipose tissue on the one hand and central fat accumulation on the other constitutes one pathogenetic entity. However, in many studies most patients presented a mixture of peripheral fat loss and central fat accumulation, named mixed LD. As other types of lipodystrophies (congenital as well as acquired) are characterized by selective loss of subcutaneous fat tissue only, some authors recognize HIV-associated LD as being primarily due to loss of peripheral fat, with fat accumulation as a secondary phenomenon. In support of this, a prospective study has recently demonstrated that HAART may induce a selective loss of limb fat. The development of LD was first observed in patients treated with a combination of protease inhibitors (PIs) and nucleoside reverse transcriptase inhibitors (NRTIs). Recently, LD has also been recognized in patients only receiving NRTIs. While the NRTI stavudine has especially been associated with a high risk for the development of lipoatrophy, the pathogenesis of HIV-associated LD is not known.

The HIV-associated LD syndrome has characteristics in common with the metabolic syndrome: hypertriglyceridemia, hypercholesterolemia, increased lipolysis, and insulin resistance, but the metabolic abnormalities are not only a result of the fat distribution in HIV. The metabolic alterations may also result from a direct effect from the PIs.

IL-18 is a cytokine with proinflammatory features and is involved in tissue destruction. Thus, it may play a role in the host defense against infectious diseases and tumor growth, but it seems also to play a pathogenetic role in inflammatory conditions. IL-18 is synthesized as an inactive precursor and converted to its biologic form by the cysteine protease caspase-1. IL-18 induces interferon-g production and IL-6. In addition, IL-18 induces tumor necrosis factor-a (TNF-a) and TNF-a also stimulates IL-18 production. Thus, the two cytokines IL-18 and TNF-a seem to work in synergy. The cytokine adiponectin is associated with both diabetes and LD, but its possible relationship to IL-18 has not been evaluated.

Elevated levels of IL-18 have recently been demonstrated in HIV-infected patients and especially in those with immunodeficiency and advanced clinical disease and to decline with HAART for 24 months. IL-18 has been reported both to inhibit and to induce HIV-1 expression in monocytes. Recently, the levels of IL-18 were found to be elevated in persons with obesity, type 2 diabetes, and arteriosclerosis and to decline with weight loss. Given that these conditions share common features with HIV-associated LD, we hypothesized that IL-18 would be elevated in these patients. Furthermore, we specifically evaluated whether IL-18 was associated with fat mass, fat distribution, and NRTI treatment and whether IL-18 was associated with other cytokines such as TNF-a, IL-6, and adiponectin.

RESULTS

The HIV groups did not differ with regard to duration of HIV infection, CD4 cell count, or HIV RNA. The lipoatrophic group had received therapy for a longer period than the HIV controls and had a lower body mass index compared with the other groups. The DXA scans supported the clinical diagnosis of LD and revealed 3 distinct different HIV groups with regard to phenotype.

The lipoatrophic group was characterized by reduced limb fat mass compared with the mixed group, the HIV controls, and the healthy controls. All groups had the same lean body mass. All HIV-infected patients had reduced limb fat mass compared with the healthy controls. The mixed group had increased truncal fat mass compared with the lipoatrophic group and the HIV controls but not compared with the healthy controls.

The levels of plasma IL-18 were elevated in all 3 HIV groups compared with HIV-negative controls (P <0.01). In the HIV groups, the lipoatrophic group had the highest IL-18, followed by the mixed group, and the HIV controls. Only the differences between the lipoatrophic group and the HIV controls were significant: 345.9 (264.3-427.3 pg/mL) vs. 212.6 (151.0-259.5 pg/mL), P <0.01.

A positive correlation between IL-18 and TNF-a was found in all HIV-infected patients. IL-18 did not correlate with IL-6, adiponectin, or HOMA-IR in any of the groups. In the healthy group, IL-18 correlated with the total cholesterol, LDL, and triglycerides. However, such correlations did not exist in either of the HIV groups, although triglycerides correlated with IL-18 within HIV subjects without the mixed group. Within all subjects a negative correlation between IL-18 and limb fat mass was found, but no correlation within the HIV subgroups except for HIV subjects with exclusion of the mixed group. In contrast to the HIV-infected patients, IL-18 correlated positively to the total fat mass and to the truncal fat mass in the healthy controls.

Moreover, a positive correlation was found between the total duration of NRTI use and IL-18 in HIV subjects without the mixed group as well as in the lipoatrophic group.

A negative correlation was also found between total duration of NRTI use and limb fat mass within the HIV groups (HIV subjects without the mixed group: rs = -0.497, n = 26, P <0.02; all HIV subjects: rs = -0.484, n = 35, P <0.001). There was no correlation between the duration of NRTI use and trunk fat mass in either of the HIV groups (HIV subjects without the mixed group: rs = -0.270, n = 26, P = NS; All HIV subjects: rs =-0.188, n = 35, P = NS) (data not shown).

Patients currently receiving stavudine (n = 11) had significantly higher levels of plasma IL-18 compared with patients currently on regimens without stavudine (n = 28, 369.4 [323.3-409.5] vs. 221.5 [170.2-351.49] pg/mL, P <0.01).

Plasma IL-18 did not differ between HIV-infected patients with HIV-1 RNA >200 copies/mL (n = 8) and patients with HIV-1 RNA <200 copies/mL (n = 31). Furthermore, IL-18 did not correlate to CD4 cell count in any of the HIV groups.

DISCUSSION by authors

The main finding of this study was that the level of circulating IL-18 is elevated in HIV-infected patients with lipoatrophy (both with and without central fat accumulation) compared with HIV-infected patients without LD.

Elevated levels of IL-18 have recently been demonstrated in HIV-infected patients and especially in those with immunodeficiency and advanced clinical disease. This could indicate that IL-18 was related to high viremia. However, we did not find that viral load had any influence on the level of IL-18, presumably because all patients were on stable and successful HAART. In addition, the level of IL-18 was not related to the CD4 cell count. Chronic inflammation is well documented in HIV disease, even in patients who are successfully treated with HAART. The elevated level of IL-18 correlated well with circulating levels of TNF-a and thus is likely to reflect such inflammation. In addition, inflammation may be a cause of HAART-induced metabolic changes. Although longitudinal prospective studies are required to fully answer this question, we found indications for a relationship between inflammation and altered metabolism.

The finding that IL-18 was elevated in both the mixed and the lipoatrophic groups indicates that IL-18 is linked with the peripheral fat atrophy. In support, IL-18 correlated inversely with limb fat mass. The fact that the levels of IL-18 did not differ between patients in the mixed group and the HIV control group would be in accordance with the possibility that IL-18 was mechanistically linked to the low total limb fat mass, as the DXA scans revealed that the HIV controls also had low limb fat mass, despite their phenotype. This DXA scan finding is in agreement with recent observations in Fat Redistribution and Metabolic Change (FRAM) Study. IL-18 did not correlate to LDL and trunk fat in HIV patients as opposed to the healthy controls, indicating that IL-18 is differently regulated in patients with and patients without HIV, a suggestion further supported by the fact that IL-18 is positively correlated to the total fat mass in the healthy; but when all the HIV-infected subjects are included in the analysis the correlation is negative. A strong correlation exists between IL-18 and triglycerides in healthy subjects. Such a correlation is also present in HIV subjects if patients with fat accumulation are excluded.

We found an association between IL-18 and the duration of NRTI therapy. In a previous study the level of IL-18 declined following HAART for 24 months. However, in the present study, all patients received HAART, and high levels of IL-18 were found especially in patients who had received antiretroviral therapy for a long period. This could reflect the lipoatrophy phenotype or that IL-18 was related to a direct effect of >=1 antiretroviral treatment.

Stavudine has been associated with the development of lipoatrophy. The present study was not designed to study the specific effect of stavudine or other treatments. However, patients on current treatment with stavudine had higher levels of IL-18 than other patients. Such a possible relationship is not likely to be due to an acute effect of stavudine as patients who had stavudine switched to other antiretroviral drugs did not have a decrease in the level of IL-18 (median and 25 and 75% quartiles, 247 (186-366) vs. 273 (166-399) pg/mL, before and after, respectively stavudine switching, unpublished data). We found that IL-18 was correlated with TNF-a, thus supporting its role as a true proinflammatory cytokine. Given that IL-18 induces tissue destruction, it may be involved in adipocyte apoptosis occurring in patients with LD. Thus, IL-18 could represent a link between NRTI therapy and fat atrophy.

In conclusion, the level of IL-18 is elevated in patients with LD (both with and without fat accumulation) and closely linked to limb atrophy, whereas it is not associated with cholesterol or IR.

Patients and Methods

Forty-two HIV-infected men were recruited from the outpatient clinic of the Department of Infectious Disease, Rigshospitalet in Copenhagen. Informed consent was obtained from all patients according to declaration of the local ethical committee. LD was defined clinically by physical examination of peripheral lipoatrophy (fat loss from face, arms, buttocks, or legs) with or without central accumulation (abdomen, dorsocervical fat pad). Age-matched patients were enrolled into the following 3 groups according to clinical examination: patients with lipoatrophy with truncal fat accumulation (the mixed group) (n = 12); patients with lipoatrophy without truncal fat accumulation (the lipoatrophic group) (n =15); and patients without LD (HIV controls) (n=15). All patients were on a stable and effective nucleoside analogue-based HAART with no changes in antiretroviral therapy during the preceding 8 weeks. The combinations were as follows: 10 patients were receiving PIs and 4 nonnucleoside analogues in the mixed group; 10 patients were receiving PIs and 4 nonnucleoside analogues in the lipoatrophic group; in the HIV control group, 9 patients were receiving PIs and 5 nonnucleoside analogues; and 12 age-matched HIV-negative healthy men served as controls.

None of the HIV-infected patients had signs of ongoing infections or fasting glucose >7 mM.

Peripheral blood samples were obtained after an overnight fasting. Measurements of total cholesterol (mM), high-density lipoprotein (HDL) cholesterol (mM), low-density lipoprotein (LDL) cholesterol (mM), triglycerides (mM), plasma glucose (mM), and insulin (pM) were determined immediately using routine methods.

CD4 cell counts were calculated by flow cytometry and HIV RNA copies were measured by the Amplicor HIV Monitor (Roche Molecular Systems, Branchburg, NJ) (lower limit of detection: 20 copies/mL).

Cytokines IL-18, IL-6, and TNF-a were measured in plasma.

Adiponectin was determined by a radioimmunoassay kit (LINCO Research, Inc., St. , MO). Threshold of detection was 1 ng/mL. Insulin resistance was assessed from fasting plasma insulin and glucose by using the homeostasis model (HOMA-IR) as described previously.

Body Composition Analysis

Fat and fat-free tissue masses for the whole body, trunk, and extremities were measured using dual-energy x-ray absorptiometry (DXA) scanner, Norland XR 36 (version 3.94; Norland Corp., Fort Atkinson, WI). All measurements were done in a single laboratory. Whole-body and regional fat measurements (trunk and extremity) were determined.