Introduction

The HIV-associated lipodystrophy is made up of fat redistribution, consisting of peripheral fat loss or truncal fat accumulation; a mixed form, which is the combination of these is also observed. The HIV-related lipodystrophic syndrome (HIV-LDS) combines lipodystrophies with metabolic disturbances like insulin resistance and hyperlipidemia. The frequency of these disorders observed in several cross-sectional studies is highly variable.1,2,3,4,5,6,7 Diabetes is rare with a prevalence of 1–7%, but insulin resistance is more frequent. This syndrome has been reported shortly after the initiation of highly active antiretroviral therapy (HAART), but the mechanism remains unknown. In the prospective studies implemented in patients under HAART, the incidence of lipodystrophy has been reported to be as high as 9% after 12 months or 17% after 18 months of follow-up.8,9

We have previously reported that HIV-LDS prevails in children in a similar pattern as in adults, although all features are less pronounced before puberty.10 In the same cross-sectional study, clinical lipodystrophies were observed in one third of the HIV-infected children. These lipodystrophic children were insulin resistant in comparison to the non-lipodystrophic patients when insulin resistance was assessed by insulin circulating levels during OGTT in patients with normal glucose tolerance.

The metabolic activity itself of the adipose tissue in HIV infection has been poorly documented. It is therefore not known whether the morphological changes, and more specifically hypertrophy, are associated with specific metabolic dysfunctions. The microdialysis technique enables a local in vivo approach to follow the actual interstitial metabolic profile.11,12

The aim of our study was to assess in situ the insulin sensitivity of the lipohypertrophic subcutaneous abdominal adipose tissue using the microdialysis technique in HIV-infected children by measuring the anti-lipolytic effect of insulin on glycerol release. Since we have previously shown that peripheral insulin resistance was observed under OGTT in HIV-infected children with lipodystrophy, we selected the same test in the present study. To distinguish metabolic changes in adipose tissue caused by HIV infection from that due to adipose tissue hypertrophy alone (obesity), we compared the results with those obtained in a group of obese, non-HIV-infected children, who were sex- and age-matched with the HIV-lipohypertrophic patients.

Subjects and methods

Subjects

A total of 20 patients participated in the study (aged of 6–8 y): 14 HIV-infected children and six non HIV-infected obese children. HIV-infected patients were included irrespective of their treatment and of metabolic disturbances. Nearly all of them were treated by HAART, including at least two nucleoside reverse transcriptase inhibitors and in most cases one protease inhibitor. HIV infection was generally well controlled as reflected by the CD4+ cell counts and by the viral mRNA levels (Table 1).

Table 1 Characteristics of the patients
Full size table

Non-inclusion criteria in all groups were age <6 y, oral contraception, corticotherapy, acute infection and known defect of blood coagulation. An additional exclusion criterion was applied to HIV-infected children in whom partial or generalized lipoatrophy was confirmed by skinfold thickness values below the third percentile of the distribution for age and gender in four different areas.13,14

The HIV-infected patients were selected into two groups with respect to the presence or not of subcutaneous abdominal hypertrophy defined as a supra-iliac skinfold thickness >97th percentile of the distribution for gender and age. Six children had abdominal hypertrophy (HIV/LH+ group). The remaining children (n=8) did not show any lipodystrophy (HIV/LH− group). A group of obese, non HIV-infected children (BMI >25 kg/m2) was used for comparison of metabolic functions of the adipose tissue. Abdominal hypertrophy, as defined above, was observed in all obese children. Gender distribution and frequency of children in prepuberty/puberty were similar in the three groups.

Children were included after informed parental consent and personal assent. The study was approved by the Ethical Committee of the St-Louis Medical School at the University of Paris.

Study protocol

Clinical data collection

Patients were admitted to the Clinical Investigation Unit at the Robert Debre hospital. Body height and weight were measured and body mass indexes (BMI) were calculated and read as Z-scores of the distribution for age and gender.15 Puberty was quoted using Tanner stages.16 Skinfold thicknessess (biceps, triceps, supra-iliac, supra-scapular) were measured twice on the left side of the body by a harpender skinfold calliper (Chambly-Medical, Chambly, France) at a 5 min interval in each site. The two individual values for each site were averaged and were expressed as a percentile of the normal distribution for gender and age according to the French growth standard curves.13,14 Trunk-to-arm ratio was calculated and expressed as a Z-score for gender and age. Percentage body fat mass was calculated after bioelectrical impedance analysis (BIA101Q, RJL Systems, Clintontownship, MI, USA).

Biological parameters

A standard oral glucose tolerance test (OGTT; 1.75 g glucose/kg body weight up to 75 g glucose) was performed after a 12 h overnight fast following the WHO recommendations. Plasma glucose and insulin levels were measured every 30 min until 120 min. Fasting serum lipids and glycerol levels were also measured.

The microdialysis technique

The method was similar as described by Lafontan et al17 After a local anaesthesia, a microdialysis catheter. (30×0.5 mm, 20 000 Da molecular weight cut-off) was inserted at both sizes of the umbilicus in the abdominal subcutaneous tissue at maximum 8 cm lateral to the umbilicus. The nylon tubing inlets of the catheters were connected to a microinjection pump. A Ringer solution supplemented with 4.0 mmol/l glucose was perfused through the catheters. After equilibration a recovery procedure was performed. Briefly, the catheters were perfused at three successive flow-rates. (3.0, 1.5 and 0.5 μl/min, respectively) and the glycerol concentrations were determined in the dialysate at the end of each perfusion period. (30 min). These concentrations were plotted against the perfusion rates. Regression analysis was used to calculate the glycerol concentration at ‘zero flow'. The recovery rate of the catheter (%) at 0.5 μl/min was calculated by the ratio (glycerol concentration adipose tissue at 0.5 μl/min)/(interstitial glycerol concentration at `zero flow')×100. The recovery rates were >90% in all children. The perfusion rate was at 0.5 μl/min during the test.

Dialysate samples were collected for glycerol concentration measurements at 15 minutes interval during OGTT.

Analytical methods

Plasma and dialysate glycerol levels were measured by an automatic analyzer specifically designed for microdialysis samples (CMA 600 Microdialysis Analyser, Solna, Sweden). Glycerol was measured in the microdialysis analyzer by an optical system. The limit of detection was 10 μmol/l and CV was 3.7% at 12 μmol. Glucose was measured by a glucose oxydase method.

Serum insulin concentrations were measured using a double-antibody radio-immuno-assay (ERIA Diagnostics Pasteur, Paris, France). Cross-reactivity with proinsulin and the derived metabolites was less than 1%.

Plasma was assayed for HIV-RNA by a quantitative reverse transcriptase polymerase chain reaction assay (Amplicor Monitor, Roche-Diagnostic, Meylan, France). The lower limit of quantification was 200 copies per milliliter. CD4+ lymphocytes were quantified by flow cytometry.

Statistical analyses

All data were entered and analyzed on the SAS software (SAS Institute, Cary, NC, USA). Results are presented as mean±1 s.d. in the text and in the tables and for more clarity as mean±1 s.e. on the figures. Comparisons among the study groups for continuous variables were performed using non-parametric tests: a Kruskal–Wallis test for the three groups and a Mann–Whitney U-test for two groups. Values for insulin concentrations were log-transformed prior to analysis. Comparisons between the groups were performed taking into account an independent effect of BMI and gender using a GLM analysis.

Qualitative variables were compared using a χ2 or a Fisher exact test. The level of statistical significance was set at 0.05.

Results

Data on the study population are presented in Table 1. There were no statistical differences in CD4+ cell count, in viral load and in distribution of HIV- therapy between HIV/LH+ and HIV/LH− children. Anthropometric parameters were significantly different between the three groups (except for trunk to arm ratio), being consistently lower in the HIV/LH− group. The HIV/LH− children presented with BMI and skinfold thickness in the normal range for age and gender. The HIV/LH+ children showed BMI in the normal range but increased central skinfold thickness; moreover, the ratio of central to peripheral skinfold thickness was increased attesting for an abnormal fat distribution for age and gender.

Fasting cholesterol and triglycerides plasma levels were not significantly different between the groups. Plasma glycerol at fast was higher, although not significantly, in obese vs HIV-infected children (P=0.06). Results of the OGTT are presented in Figure 1. All patients had normal glucose tolerance. Serum insulin concentrations at fast were higher, although not significantly, in HIV/LH+ group than in HIV/LH− children. (6.8±4.3 vs 3.4±2.0 mU/l; P= 0.09). A significant difference was observed in serum insulin level between obese children and HIV/LH− children at fast. (7.3±3.2 vs 3.4±2.0 mU/l; P=0.04). There were no statistical differences between the HIV/LH+ and the obese children at fast (P=0.6). The insulin area under the curve (IAUC) was significantly different between the three groups (P=0.04). The difference remained significant even after adjustment for gender, BMI or percentage fat mass. It was significantly higher in the obese vs HIV/LH− groups. (8769±5429 vs 3618±2222 mU min l−1; P=0.02) and higher, although not significantly, in the HIV/LH+ vs HIV/LH− groups (8161±4552 vs 3618±2222 mU min l−1; P=0.06) and not different in obese vs HIV/LH+ groups (8769±5429 vs 8161±4552 mU min l−1; P=0.9).

Figure 1
figure 1

Time course of plasma glucose and insulin during OGTT. Data are presented as mean±s.e. Symbols: HIV/LH+ group (λ), HIV/LH− group (O) and obese group (σ). Results for the comparisons show the statistical significance between the groups independent of BMI (*P<0.05).

Full size image

The time-course of glycerol concentration in the dialysate of the subcutaneous abdominal adipose tissue during OGTT is presented in Figure 2. No statistical differences were found between the three groups. Mean glycerol level at baseline was higher in obese vs HIV/LH+ children (P=0.2) and significantly higher in obese than in the HIV/LH− group (P=0.03).

Figure 2
figure 2

Time course of glycerol in the dialysate during OGTT. Data are presented as mean±s.e. Symbols: HIV/LH+ group (λ), HIV/LH− group (O) and obese group (σ).

Full size image

The inhibition of glycerol release, expressed as the relative change in glycerol level over baseline (glycerol T120−T0 /T0), was not statistically different between the three groups (−49±14, −38±14 and −51±17% in the obese, HIV/LH+ and HIV/LH− groups, respectively; P= 0.3). The inhibition of glycerol release corrected for the level of circulating insulin during the OGTT (expressed by IAUC) was very similar in HIV/LH+ and obese groups (−6±5.10−3vs −7±5.10−3 l mU−1 min−1; P=0.4) and four fold less than in the HIV/LH− group (−24±25.10−3 l mU−1 min−1; P=0.02 for both comparisons between HIV/LH+ vs HIV/LH− and obese vs HIV/LH− groups).

Discussion

To our knowledge this is the first time that the metabolic activity of the adipose tissue of the HIV-associated lipodystrophies has been investigated in vivo. The microdialysis technique makes it possible to measure in situ the actual metabolic changes in the adipose tissue. Here we used the stimulation by insulin under OGTT and insulin sensitivity of the adipose tissue was assessed by inhibition of glycerol release in lipodystrophic areas in HIV-infected children.

The HIV/LH+ group presented with mild morphological changes. Indeed, there was a hypertrophy of the adipose tissue in the abdominal area according to our inclusion criteria. Otherwise most HIV/LH+ children had BMI calculated in the normal range for age and gender. In contrast, supra-iliac skinfolds were measured two-fold thicker in the obese children than in the HIV-infected children with central hypertrophy. However, the degree of peripheral insulin resistance, assessed by IAUC together with normal glucose tolerance, was similar in the HIV/LH+ children and in obese children. This insulin resistance in the HIV-lipodystrophic children is not surprising since we have already reported elevated circulating insulin levels in HIV-infected children with lipodystrophy. Indeed it has been shown that the first stages of puberty induce insulin resistance18 and that is the reason why the children in each study group were matched for gender and pubertal stage; therefore, differences in the level of insulin sensitivity between the lipodystrophic and not lipodystrophic children could not be explained by the effect of puberty. Moreover, mean BMI (expressed for age and gender) was lower in the HIV-infected children and could not therefore explain a similar degree of insulin resistance as in obese children. There was no statistical and independent effect of BMI on the comparisons of circulating insulin levels between these two groups. In contrast, in the non-lipodystrophic children normal glucose tolerance and low insulin levels were observed. It is therefore very unlikely that the peripheral insulin resistance observed in the HIV-infected children with lipohypertrophy were related to the HIV-infection itself.

The mechanism of HIV-LDS remains controversial and unclear. Several investigators have demonstrated that protease inhibitors (PIs) inhibit in vitro the preadipocyte differentiation.19,20 This could explain to some extent the process of lipodystrophy and the insulin resistance. It has been shown that PIs are able to inhibit in vitro the glucose transporter activity of GLUT-4 in an adipocyte-derived cell line, providing direct evidence for the mechanism of insulin resistance at the level of the adipocyte associated with the use of PI in HIV infection.21,22,23 One of this report also shows that nelfinavir is able to increase lipolysis in the same cell line.22 Here, we provide another evidence of insulin resistance at the level of the adipose tissue. This observation is in keeping with the reports on the decrease of the whole peripheral glucose uptake in patients receiving PIs.24,25

It is well known that obesity is a risk factor for the development of insulin resistance even in populations with no increased genetic risk for type 2 diabetes.26 If the resistance to the anti-lipolytic effect of insulin was due to the hypertrophy of the abdominal adipose tissue itself in the HIV-infected children it would be anticipated a lesser degree of insulin resistance than in the obese children in whom the abdominal wall was much thicker.

The in situ insulin sensitivity of the subcutaneous abdominal adipose tissue was similar in HIV/LH+ and in obese children whereas the metabolic activity of the adipose tissue at fast (dialysate glycerol adipose tissue T0) was slightly lower in HIV/LH+ than in obese children. Although no information is available on the morphometry of the adipose tissue in lipodystrophies, it has been shown that the size of abdominal adipose cells interferes with the lipolytic rate.27,28 It would be interesting to test whether the local stimulation of lipolysis is impaired in lipodystrophic patients in comparison to obese subjects.12

In conclusion, our data argue in favor of insulin resistance in the adipose tissue in lipohypertophies associated with HIV infection. This adipose tissue showed a high degree of insulin resistance not anticipated from the modest anthropometric changes. This would suggest that adipocyte lesions contribute to the metabolic disturbances observed in the lipodystrophic syndrome associated with HIV infection.