Introduction

Combined highly active antiretroviral therapy (HAART), comprising protease and reverse transcriptase inhibitors, is effective in preventing sequelae of human immunodeficiency virus (HIV) infection. However, some patients taking HAART develop a lipodystrophic syndrome, with loss of subcutaneous fat and accumulation of fat in the abdomen and nape of the neck. This syndrome is associated with hypertriglyceridaemia and insulin resistance [1]. The mechanisms underlying HAART-associated lipodystrophy (HAL) are unknown. The distribution of fat accumulation in HAL is similar to that induced by increased plasma glucocorticoid levels in Cushing’s syndrome. However, circulating cortisol concentrations are not elevated in HAL patients [2].

Recent research has highlighted that glucocorticoid action within adipose tissue is regulated independently of circulating cortisol concentrations by the intra-adipose regeneration of cortisol from its inactive metabolite cortisone, catalysed by the enzyme 11β-hydroxysteroid dehydrogenase type 1 (11β-HSD1) [3, 4]. Transgenic overexpression of 11β-HSD1 in adipose tissue in mice causes visceral obesity, diabetes and dyslipidaemia [5], while 11β-HSD1 knock-out mice are protected from obesity and stress-related hyperglycaemia [6]. 11β-HSD1 is up-regulated by inflammatory cytokines [7].

Here, we test the hypothesis that increased adipose tissue 11β-HSD1 is associated with lipodystrophy and its metabolic manifestations in a previously described cohort of HIV-seropositive patients taking HAART [8, 9].

Subjects and methods

All participants were attending an HIV outpatient clinic, had been taking HAART for at least 18 months, and gave written consent for the study, which was approved by the local ethics committee. We studied 5 women and 25 men who reported lipodystrophic symptoms and had clinical features of lipodystrophy confirmed by one investigator. As control subjects we studied four women and nine men without symptoms or clinical features of HAL. None of the women was postmenopausal and the age distribution was almost identical between the groups (38±4 vs 37±6 years, lipodystrophic vs non-lipodystrophic subjects). Patients with diabetes mellitus were excluded.

Body fat and its distribution were assessed by body mass index, waist-to-hip ratio, bioimpedance and cross-sectional abdominal magnetic resonance imaging (MRI) scans. Blood was obtained at 08.00 hours after an overnight fast for glucose, insulin, lipid and C-reactive protein analyses. Thereafter, a needle aspiration biopsy of abdominal subcutaneous fat was obtained, from which RNA was extracted and mRNA species quantified by real-time polymerase chain reaction in 40 participants (13 non-lipodystrophic) [8, 9, 10]. The mRNA results are expressed relative to mRNA for β2-microglobulin. Early-morning urine was collected from 35 (10 non-lipodystrophic) participants for analysis of cortisol and its metabolites by gas chromatography/mass spectrometry [10].

The Mann–Whitney U test was used to compare differences between the groups. Correlations were calculated using Spearman’s rank correlation coefficient. All data are given as means ± SD. A p value of less than 0.05 was considered statistically significant.

Results

As described previously [8, 9], the combinations of HAART drugs and severity or duration of HIV-1 infection did not differ between HAL and non-HAL patients. The most common non-HIV medications being taken were co-trimoxazole for prophylaxis for Pneumocystis carinii pneumonia (20% vs 15%, lipodystrophic vs non-lipodystrophic group) and pravastatin (10% vs 8%, respectively). Use of other drugs was infrequent (0–1 patients in either group used acyclovir, cetirizine, loperamide, mirtazapine, omeprazole) with no difference between the groups.

Patients with HAL had higher intra-abdominal fat volume, waist-to-hip ratio, fasting serum insulin, triglycerides and C-reactive protein, and lower HDL-cholesterol (Table 1). There were no differences in blood pressure, body mass index or total % body fat, but the abdominal subcutaneous fat volume was reduced in HAL patients. In subcutaneous adipose tissue biopsies, patients with HAL had markedly higher mRNA levels for 11β-HSD1 (Table 1). In all subjects, adipose tissue 11β-HSD1 mRNA correlated positively with visceral fat volume, waist-to-hip ratio, body mass index, fasting serum insulin, triglycerides (Fig. 1a), total cholesterol and C-reactive protein, and correlated inversely with HDL-cholesterol. However, adipose tissue 11β-HSD1 mRNA did not correlate with subcutaneous fat volume or with total % body fat (Table 1).

Table 1 Characteristics of the study subjects
Full size table
Fig. 1
figure 1

Prediction of severity of hypertriglyceridaemia by (a) adipose 11β-HSD1 mRNA (expressed relative to mRNA for β2-microglobulin) and (b) urinary cortisol : cortisone metabolites, i.e. the ratio of (5α- + 5β-tetrahydrocortisol) : tetrahydrocortisone. Filled symbols: patients with lipodystrophy; open symbols: patients without lipodystrophy. For comparisons between groups and Spearman rank correlations, see Table 1

Full size image

In urine, patients with HAL had higher ratios of cortisol : cortisone metabolites (Table 1). The urine cortisol : cortisone metabolite ratio correlated positively with fasting serum triglycerides (Fig. 1b) and inversely with subcutaneous fat volume and HDL-cholesterol (Table 1). Sex and age had no confounding effects on adipose tissue 11β-HSD1 mRNA or on urinary cortisol : cortisone metabolite ratios.

To assess the impact of increased 11β-HSD1 on intra-adipose tissue glucocorticoid signalling in HAL patients, other adipose tissue mRNAs were also quantified. Levels of 11β-HSD1 mRNA were positively correlated with mRNA levels for the glucocorticoid receptor (Spearman r=0.43, p=0.006) and for the glucocorticoid-sensitive target gene angiotensinogen (r=0.35, p=0.031). For other genes involved in regulation of adipose tissue metabolism and hormonal signalling mRNAs have been quantified previously in this cohort [8, 9]. Interestingly, 11β-HSD1 mRNA levels were correlated positively with mRNA levels for TNF-α (r=0.64, p<0.0001) and inversely with mRNA levels for leptin (r=−0.38, p=0.015) and peroxisome proliferator-activated (PPARγ) (r=−0.55, p=0.0003); leptin and cytokines up-regulate, and PPARγ agonists (thiazolidinediones) down-regulate 11β-HSD1 in cells.

Discussion

We conclude that 11β-HSD1 mRNA is increased in adipose tissue of patients with HAL and correlates with the severity of intra-abdominal fat accumulation and metabolic disturbances. The increased ratio of cortisol : cortisone metabolites in urine supports the inference that in vivo conversion of cortisone to cortisol by 11β-HSD1 is enhanced in these patients. The associated increase in mRNA for glucocorticoid receptor and angiotensinogen suggests that increased 11β-HSD1 mRNA increases intra-adipose glucocorticoid signalling and could account for the pseudo-Cushing’s characteristics of patients with HAL.

The mechanism that leads to increased adipose tissue 11β-HSD1 in HAL is unclear. In cells in culture, 11β-HSD1 is regulated by metabolic and inflammatory stimuli. Although the relevance of these observations in vivo in humans remains to be established, the associations observed by us in this study suggest that increased cytokine activity and decreased PPARγ activation may be important. Protease inhibitors, with which lipodystrophy was originally associated, have been shown to inhibit rather than enhance 11β-HSD1 activity in 3T3-L1 adipocytes [7].

Levels of 11β-HSD1 are also greater in subcutaneous adipose tissue from obese subjects than in that from non-obese subjects [10]. Notably, this is associated with increased rather than decreased leptin expression, and does not seem to be accompanied by increased glucocorticoid receptor or angiotensinogen expression [10]. The findings in HAL patients appear to more closely mirror those in Cushing’s syndrome and in mice with adipose tissue 11β-HSD1 overexpression than those in idiopathic obesity. Although not directly addressed here, a possible explanation is that glucocorticoid receptor activation is increased both in subcutaneous and in omental adipose tissue in Cushing’s syndrome, in mice overexpressing 11β-HSD1 and in HAL, whereas in idiopathic obesity glucocorticoid receptor activation is only increased in subcutaneous adipose tissue.

These observations have important clinical implications. Polymorphisms that influence 11β-HSD1 expression may underlie inter-individual differences in susceptibility to HAL and may form the basis for pharmacogenetic strategies in future. Currently, 11β-HSD1 inhibitors are being developed which, if active in adipose tissue, may offer a novel approach to prevent or treat visceral fat accumulation and insulin resistance in HAART-treated patients.