Mitochondrial DNA depletion and morphologic changes in adipocytes associated with nucleoside reverse transcriptase inhibitor therapy
AIDS 2003; 17(9):1329-1338
David Nolan; Emma Hammond; Annalise Martin; Louise Taylor; Susan Herrmann; Elizabeth McKinnon; Cecily Metcalf a; Bruce Latham a; Simon Mallal. From the Centre for Clinical Immunology and Biomedical Statistics, Royal Perth Hospital and Murdoch University, and aDepartment of Pathology, Royal Perth Hospital, Western Australia.
Background: Nucleoside analogue reverse transcriptase inhibitor (NRTI) therapy provides sufficient conditions for progressive subcutaneous fat wasting in HIV-infected patients. As NRTI-induced host toxicity is proposed to involve cellular mitochondrial DNA (mtDNA) depletion, determinants of cellular mtDNA copy number and mitochondrial mass in adipocyte samples from NRTI-treated HIV-infected patients and antiretroviral-naive controls were investigated. Adipose
tissue morphology was also assessed.
Methods: Subcutaneous fat samples were obtained from NRTI-treated, HIV-infected patients (n = 21), antiretroviral therapy-naive HIV-infected controls (n = 11), and HIV-seronegative controls (n = 6). Non-adipocytes were removed by collagenase digestion. Adipocyte mtDNA copies/cell was measured using a real time PCR-based assay, and adipocyte mitochondrial protein content was also measured. Light and electron microscopy were performed on tissue samples.
Findings: Adipocyte mtDNA copies/cell values were similar (P = 0.56) in HIV seronegative and HIV-infected control groups. NRTI treatment was associated with reduced adipocyte mtDNA copies/cell, representing mean mtDNA depletion in NRTI-treated individuals of 77.7% compared with the mean value for the HIV-infected control group (P < 0001). Additionally, significant differences were found in adipocyte mtDNA copies/cell between patients receiving stavudine (n= 12, mean mtDNA depletion 87.1%) and zidovudine (n = 9, mean mtDNA depletion 52.1%) (P < 0.001). Adipocyte mitochondrial mass was increased in the stavudine group only (mean increase 289%, P < 0.01).
Interpretation: NRTI therapy is associated with mtDNA depletion and mitochondrial proliferation in adipocytes, consistent with the hypothesis that NRTI-induced mtDNA depletion contributes to the pathogenesis of subcutaneous fat wasting. Morphologic assessment also supports a role for NRTI therapy in inducing adipocyte metabolic dysfunction and cell death.
Nucleoside analogue reverse transcriptase inhibitors (NRTI) were the first therapeutic agents to demonstrate clinical efficacy as antiretroviral therapy for HIV infection, and continue to be utilized in contemporary highly active
antiretroviral therapy (HAART) regimens that may combine three NRTI, or two NRTI drugs with either HIV protease inhibitors (PI) or non-nucleoside reverse transcriptase inhibitors (NNRTI). The use of drugs from this class appears to
be an important determinant of the durability and effectiveness of these HAART regimens, so it is likely that NRTI therapy will continue to be used for the foreseeable future. In this context, it is recognized that while the use of
HAART has substantially reduced the risk of progressive immune deficiency induced by uncontrolled HIV infection, these benefits may be offset by an increased burden of long-term antiretroviral drug toxicity, including those mediated specifically by NRTI therapy.
The antiretroviral activity of NRTI drugs is determined by their ability to inhibit the RNA-dependent DNA polymerase activity of HIV reverse transcriptase. Hence, NRTI drugs compete with endogenous nucleic acids for incorporation by
DNA polymerase, and cause premature termination of viral DNA chain elongation when incorporated. Adverse effects of NRTI drugs also appear to be mediated by their effects on host DNA polymerase activity. Mitochondrial
DNA (mtDNA) polymerase gamma appears particularly susceptible to inhibition, as this unique polymerase - indispensable for mtDNA synthesis - lacks the ability shared by nuclear DNA (nDNA) polymerases to effectively discriminate against NRTI in favour of endogenous nucleic acids. NRTI-induced inhibition of mtDNA synthesis is proposed to induce depletion of cellular mtDNA copy number, ultimately leading to cellular toxicity when mitochondrial
function is compromised to the extent that bioenergetic needs cannot be met. Within the NRTI class, the capacity of a select drug to induce mtDNA depletion is determined by a number of factors, including the efficiency with which the
relevant nucleoside (or nucleotide) analogue is converted to its active form, and its ability to gain entry to the mitochondrial compartment within the cell. The principal determinant, however, is its affinity for DNA polymerase
gamma. It is notable that the relative antiretroviral activity of specific drugs within this class do not predict their effects on mtDNA synthesis, indicating that inhibition of viral and host DNA polymerases act as independent processes.
It has been proposed that this model of NRTI toxicity - commonly referred to as the 'pol-[gamma]' hypothesis - is relevant to the pathogenesis of the 'lipodystrophy syndrome', a highly prevalent clinical entity among HAART recipients in which progressive loss of subcutaneous fat in the peripheries and face is a prominent feature. NRTI therapy alone provides sufficient conditions for the development of subcutaneous fat wasting, and is an independent risk factor for the development of fat wasting in individuals treated with HAART regimens. Moreover, data from observational studies and clinical trials indicate differences in relative risk of lipodystrophy onset associated with select drugs within the NRTI class. Hence, a comprehensive model of lipodystrophy pathogenesis must be able to account for the contribution of NRTI therapy, and provide for the minimal conditions under which fat wasting can develop; that is, in the setting of NRTI therapy alone.
We sought to address the hypothesis that adipocyte-specific mtDNA depletion contributes to the pathogenesis of NRTI-associated subcutaneous fat wasting, as proposed by Brinkman and colleagues. In this study, subcutaneous
fat biopsies have been performed using an excisional technique to avoid contamination with non-adipose tissue, and adipocytes were subsequently purified from subcutaneous fat samples through the use of collagenase digestion. Adipocyte mtDNA quantity was measured using a precise real-time PCR quantification assay. Additionally, histological and ultrastructural features of adipose tissue samples were assessed, as well as mitochondrial protein content as a marker of mitochondrial organelle mass. Results of this cross-sectional study indicate an association between NRTI therapy and mtDNA depletion as well as mitochondrial organellar proliferation in vivo, of sufficient severity to provide a plausible mechanism for adipose tissue toxicity in the context of antiretroviral therapy-associated lipodystrophy.
Subcutaneous fat wasting
Measures of adipocyte mtDNA copies/cell and mitochondrial protein content were not significantly different between antiretroviral-treated patients with clinically apparent fat wasting at the time of biopsy, and those without (P = 0.44 and P = 0.88 respectively; post-switch cases excluded from analysis). Given the highly significant differences in these variables between stavudine- and zidovudine-treated patients, informative cases in these analyses may be those treated with stavudine who had no evidence of subcutaneous fat wasting at the time of biopsy, and zidovudine-treated patients with clinical fat wasting at the time of biopsy. In the stavudine-treated group, results from the three patients without fat wasting were: 49 copies/cell (5 months stavudine/didanosine/efavirenz), 98 copies/ cell (20 months
stavudine/lamivudine/nelfinavir) and 50 copies/cell (22 months stavudine/lamivudine/nelfinavir). Over a 12-month follow-up period, the first two of these patients developed clinically apparent fat wasting. In the zidovudine group, results from patients with fat wasting at the time of biopsy were: 156 copies/cell (30 months zidovudine/lamivudine/nevirapine), and 427 copies/cell (49 months zidovudine/lamivudine/indinavir). No
zidovudine-treated patient developed fat wasting subsequent to the biopsy, over 12 months of follow-up.
The major finding of this study is that NRTI therapy is associated with adipocyte mtDNA depletion and mitochondrial proliferation. Additionally, morphologic assessment of adipose tissue indicates that subcutaneous fat wasting is associated with adipose cell loss associated with lipogranuloma formation, and characteristic mitochondrial proliferation in the remaining adipocytes. None of these outcomes were dependent on the presence of HIV PI in antiretroviral therapy regimens. Highly significant differences in adipocyte mtDNA depletion were also noted between the zidovudine- and stavudine-treated groups, which were not abrogated after excluding the possible confounding effects of previous NRTI therapy, and of choice of concurrent NRTI within the HAART regimen. This does not exclude the possibility that differences between didanosine and lamivudine may exist - rather, that in this study the observed stavudine effect could not be attributed to an effect of the second NRTI in the regimen.
The severity of mtDNA depletion in adipocytes in the presence of NRTI therapy provides a plausible basis for the development of mitochondrial dysfunction and cellular pathology. mtDNA depletion attributable to stavudine- as
compared with zidovudine therapy is consistent with the results of clinical studies demonstrating increased relative risk of developing subcutaneous fat wasting associated with stavudine therapy, and is also supported by in
vitro studies of mtDNA polymerase gamma inhibition associated with these drugs. Hence, these data support the hypothesis that mtDNA depletion contributes to the pathogenesis of subcutaneous fat wasting associated with NRTI therapy, and that select drugs within the NRTI class are associated with increased relative risk of inducing adipocyte mtDNA depletion. However, establishing a causal relationship between NRTI therapy, mtDNA depletion, and clinically relevant adipose tissue toxicity requires longitudinal studies in which these parameters can be measured prior to, as well as following, the introduction of antiretroviral therapy in each individual. Additionally, further research is needed to establish the degree of correlation between NRTI-induced mtDNA depletion and mitochondrial organelle dysfunction. It
is well established that the mitochondrial genome exhibits a degree of redundancy, so that a significant loss of mtDNA through mutation or depletion may be required before tissue pathology develops. In vitro, it is estimated that this 'threshold' is attained when mtDNA is reduced by approximately 80% -a value that is consistent with the results of this study. This requires clarification in adipose tissue.
Similar studies published by Walker and Shikuma have noted significant mtDNA depletion in adipose tissue samples from patients with subcutaneous fat wasting. In this study, these observations are extended by the demonstration of NRTI-associated mtDNA depletion in adipocytes - thereby excluding confounding effects that may be attributed to the presence of stromal-vascular cell populations within adipose tissue samples. Additionally, the use of a
real-time PCR quantitation assay in this study allows for precision of measurement that has not been available previously.
These findings complement a recently published study demonstrating specific effects of HIV PI therapy on adipose tissue differentiation markers and insulin sensitivity in vivo, establishing the relevance of adipose tissue as a target
for the contributions of both NRTI and PI drug classes to the pathogenesis of lipodystrophy. This is consistent with an aetiopathogenic model in which PI and NRTI drugs make independent, as well as synergistic, contributions to
lipodystrophy pathogenesis. In this study, we were unable to identify an independent association between PI therapy and mtDNA depletion or mitochondrial proliferation, suggesting that these effects were determined primarily by NRTI use. Additionally, use of PI therapy was not required for the development of abnormal adipose tissue morphology.
The pathologic consequences of mtDNA depletion and associated mitochondrial dysfunction have generally been considered in terms of their effects on oxidative phosphorylation-the creation of cellular energy (adenosine triphosphate, ATP) within mitochondria from aerobic metabolism. In adipose tissue, however, mitochondrial energy production is predominantly harnessed to biosynthetic pathways, so that energy may be stored as triglyceride rather than expended by the adipocyte itself. Mitochondrial dysfunction in adipocytes and other lipogenic tissues appears to activate a complex regulatory response that attempts to restore bioenergetic equilibrium by increasing oxidative
metabolism and mitochondrial proliferation at the expense of energy-dependent biosynthetic reactions. Hence, while profound mitochondrial toxicity may induce cell death, as proposed by Brinkman, chronic sublethal
mitochondrial dysfunction in adipose tissue may induce adipocyte metabolic dysfunction characterized by decreased capacity for triglyceride synthesis and storage, and a reciprocal increase in substrate oxidation and mitochondrial
This phenotype is consistent with adipose tissue morphologic assessment presented in this study, and in a recent report by Lloreta and colleagues. Cellular loss was associated with the recruitment of macrophages, but the absence of other inflammatory changes suggests that apoptotic, rather than necrotic, mechanisms are involved in cell death. This is supported by studies demonstrating increased apoptosis in adipose tissue in individuals with HAART-associated lipodystrophy (assessed with TdT-mediated dUTP nick-end labelling staining) that was not ameliorated after switching from an HIV PI to NNRTI therapy while maintaining NRTI therapy. In surviving adipocytes (that is, those available for ultrastructural assessment), increased mitochondrial biogenesis may represent a compensatory response to sublethal mitochondrial dysfunction.
This cross-sectional study provides evidence that in vivo adipocyte mtDNA depletion induced by NRTI therapy may be relevant to the pathogenesis of subcutaneous fat wasting, as previously hypothesized by Brinkman and colleagues, although the establishment of causality will require longitudinal studies. It is anticipated that further elucidation of the relationships between NRTI therapy-induced mtDNA depletion and adipose tissue pathophysiology will provide an increased understanding of the cellular basis for this clinical syndrome. This may in turn provide for a directed approach to the monitoring of NRTI effects in the development and maintenance of fat wasting, based on the assessment of cellular pathology in the target tissue.