NRTI induced mitochondrial toxicity as a mechanism for HAART related lipodystrophy: fact or fiction?

One of the most debilitating side-effects of long-term highly active antiretroviral therapy (HAART) is the physical alteration of body composition, also known as lipodystrophy syndrome. This syndrome has yet to be clearly defined, but is generally agreed to include physical alterations such as peripheral fat wasting in the face, buttocks and limbs, abnormal fat accumulation in the central abdomen or subcutaneously (‘buffalo hump’), and also features indicating disturbed metabolism, such as insulin resistance with overt diabetes and dyslipidaemia [1]. Although initially linked to the use of protease inhibitors (PIs), lipodystrophy is also observed in PI naive patients [2-4]. Unfortunately, the aetiology of the syndrome is unknown at present and it is unclear whether all the features even share a common pathophysiology. One of the proposed mechanisms is nucleoside reverse transcriptase inhibitor (NRTI) induced mitochondrial toxicity [5, 6]. Although this remains a hypothesis, several recent observations have emerged that support this mechanism. HAART related lipodystrophy closely resembles multiple symmetrical lipomatosis (MSL), a syndrome associated with mutations in mitochondrial DNA (mtDNA) and dysfunction of the mitochondrial respiratory chain [7-9]. Poly neuropathy often accompanies MSL, thus demonstrating mitochondrial dysfunction in nonadipocyte tissue. In HAART related lipodystrophy several features of NRTI mitochondrial toxicity, such as polyneuropathy, liver steatosis and hyperlactataemia, occur concomitantly [4]. In the Western Australian cohort, NRTI use represents an independent risk factor for lipodystrophy in general and fat wasting in particular [10]. In the randomized Prometheus study (which had an NRTI sparing arm), the use of an NRTI, stavudine (d4T), was a precipitating factor for the development of lipoatrophy [11]. Within the NRTI family, d4T is associated with lipodystrophy particularly strongly [2, 10]. This was recently further demonstrated by the randomized NOVAVIR Trial [12]. In this trial, zidovudine (ZDV) was compared with d4T on a lamivudine (3TC)/indinavir (IDV) backbone. The chief drawbacks of this study were that it was not blinded and patients were not HAART naive at randomization. Patients had received prior therapy with ZDV, didanosine (ddI) or zalcitabine (ddC). Patients with lipoatrophy, unlike patients with fat accumulation, have several metabolic abnormalities consistent with mitochondrial toxicity [13]. All these observations strongly indicate the role of NRTIs (in particular d4T) in inducing HAART related lipoatrophy, although none of the studies show direct evidence for mitochondrial toxicity. Other observations have recently shown that mitochondria does, indeed, appear to be involved in lipodystrophy syndrome. The mitochondria within adipocytes from subjects with lipodystrophy showed ultrastructural abnormalities, such as marked variation in size and shape, a disturbed architecture of the cristae and inclusion bodies [14-16]. The cytoplasm of the adipocytes accumulates clusters of lipid droplets, similar to microvesicular changes observed in the liver and muscle tissue after exposure to mitochondrial toxins [17]. Furthermore, genetic abnormalities in mtDNA have been detected in lipoatrophic regions from the buttocks. mtDNA was reduced by 44% in subcutaneous adipose tissue of HIV-infected subjects treated with NRTIs, compared to NRTI naive HIV-infected subjects. Also, lipoatrophic fat had a lower mtDNA content than fat from patients without clinical signs of lipoatrophy [16]. Decreased amounts of mtDNA in subcutaneous adipose tissue were also suggested by a second study using a semiquantitative polymerase chain reaction (PCR) [18, 19]. This technique, however, may be imprecise in this setting; also, only two patients with lipoatrophy were tested. One patient with fat redistribution and multiple deletions of mtDNA in subcutaneous adipose tissue was described [19]. These deletions may represent coexisting MSL and may not be secondary to NRTI use, as other investigators did not describe deletions in a larger group of patients [16] even when using more sensitive techniques [18]. Since NRTI induced mitochondrial toxicity was suggested as a key factor in the induction of HAART related lipodystrophy, several studies have demonstrated that NRTIs do indeed play an important role in this syndrome and also that mitochondria in the adipose tissue appear to be affected. These observations support the mitochondrial hypothesis for lipodystrophy in general and lipoatrophy in particular. However, further investigations are necessary to reproduce the depletion of mtDNA in subcutaneous adipose tissue, to delineate the affected cell compartment and to point out possible differences between regions of fat loss and fat accumulation. It is also necessary to demonstrate any biochemical defect(s) in adipocytes' respiratory chain and elucidate how these defects might lead to the phenotypic alterations in adipose tissue before the hypothesis is conclusively proven.