It is well established that nucleoside analogue reverse transcriptase inhibitors (NRTI) can deplete mitochondrial DNA (mtDNA) in cell culture by inhibiting DNA polymerase gamma, the enzyme primarily responsible for the synthesis of mtDNA [1,2]. Rare cases of NRTI-associated lactic acidosis were described since early in the HIV epidemic, but more recently, a milder form of this syndrome: ‘symptomatic hyperlactatemia’ has been reported to occur much more frequently, with an incidence of 25.6 cases/1000 person–years among stavudine users .
Recently, the association of NRTI-associated hyperlactatemia and lipoatrophy (peripheral fat wasting) led to the hypothesis that mitochondria may be playing a key role in these body fat changes, possibly through release of apoptosis mediators that in turn lead to peripheral fat loss [4–7]. At the same time, increased awareness of the lactic acidosis syndrome and of its less severe form, symptomatic hyperlactatemia led to a heightened concern about the long- term toxic effects of the NRTIs. Clinicians are currently relying on measurements of venous lactate levels for the diagnosis of mitochondrial toxicity, but lactate levels are non-specific and insensitive for the detection of early mitochondrial dysfunction [8,9].
Tissue biopsy remains the ‘gold standard’ for the evaluation of mitochondrial dysfunction, but invasive procedures are not suitable for use in large clinical studies. Therefore, in this study, the potential for mitochondrial toxicity of antiretrovirals was evaluated by mtDNA isolation from peripheral leukocytes, with the performance of mtDNA depletion and extensive mutation analysis. Depletion testing of mtDNA in peripheral blood mononuclear cells (PBMC) has been successfully performed in the past .
Patients and methods
Ten blood DNA samples were used as controls (age 3–53 years; median 29 years), eight of which were drawn from patients who were affected by/carriers of phenylketonuria disease (Table 1). Sixteen blood samples were drawn from HIV-infected patients. Twelve specimens were collected from 10 HIV-infected, NRTI-treated patients (age 32–73 years; median 47 years) with lipoatrophy; in addition, one of these patients had serious hyperlactatemia and had three specimens collected at about 4 week-intervals during his acute symptoms and while he remained on antiretrovirals. Four specimens were collected from four HIV-infected patients (age 28–39 years; median 35 years) who were naive to all antiretrovirals. Lipoatrophy was defined as self-reporting by the patients of fat wasting in the face, buttocks or limbs, and was confirmed by the investigator. Hyperlactatemia was defined as a non-exercise venous lactate of more than twice the limit of normal, i.e. lactate > 4.8 mmol/l, drawn as per the recommendations of the AIDS Clinical Trials Group . Written informed consent was obtained from each patient prior to enrollment and this study was approved by the Institutional Review Board at the University Hospitals of Cleveland, Cleveland, Ohio.
DNA was isolated from peripheral blood leukocytes by the salting out method . Common point mutations were analyzed by multiplex polymerase chain reaction (PCR) amplification of the relevant segments of the mitochondrial genome followed by dot blot and hybridization with allele specific oligonucleotide (ASO) probes. The most common mtDNA point mutations: A3243G, T3271C, A8344G, T8356C, T8993G, T8993C, G8363A, G11778A, G3460A, G14459A, and T14484C, were analyzed [13,14]. Large deletions, DNA rearrangements, and mtDNA depletion were analyzed by Southern blot followed by hybridization with mitochondrial DNA and nuclear 18S rDNA probes . Specifically, the mitochondrial DNA was digested with Hind III and Eag I restriction endonucleases followed by Southern analysis. Deletions or DNA rearrangements were identified by abnormal banding patterns on restriction fragment length polymorphism (RFLP) analysis. Mitochondrial DNA depletion was determined by comparing the signal intensity of mtDNA with that of 18S rDNA, a nuclear gene. Depletion would reduce the ratio of signal intensity of mtDNA to that of 18S rDNA . Screening for unknown point mutations in the entire mitochondrial genome, was accomplished by temporal temperature gradient gel electrophoresis (TTGE) analysis of 32 overlapping regions of the mtDNA followed by sequencing of the abnormal DNA fragments . Selected primer sequences and TTGE conditions have been presented . Deletions of mtDNA were analyzed by PCR with three primer pairs 467F/4508R, 4013F/8380R, and 8278F/14499R. These primer pairs will amplify the mtDNA only if there is a large enough DNA deletion between the primers permitting the completion of the PCR-based amplification.
Statistical analysis was performed using the Wilcoxon rank-sum test.
All blood samples were drawn between May and July 2000. The race, age, sex, duration of HIV infection, duration of NRTI therapy, type of antiretrovirals used, blood lactate, and other clinical abnormalities of each subject are listed in Table 1. All four NRTI-naive patients were known to be infected with HIV for only a short period of time (≤ 4 months). The 10 patients with lipoatrophy were on NRTI therapy for a median of 69 months. HIV infection had been diagnosed at a median of 89 months prior to study entry, versus a median of 3 months for the treatment-naive patients. Of note: one patient (patient 15) with lipoatrophy had a peak lactate of 10.8 (normal ≤ 2.4 mmol/l) at the time of blood collection, and was symptomatic with severe fatigue, abdominal pain, recent 20 lb weight loss, otherwise unexplained abnormal liver transaminases and a wide anion gap metabolic acidosis.
The ratio of mtDNA to 18S rDNA from quantitative analysis is listed in Table 1 for each sample analyzed. The mean ratio was 2.74 (SD, 0.42), 2.76 (SD, 0.83), and 3.41 (SD, 0.69) for healthy control, NRTI-naive, and NRTI-treated groups, respectively. Pair-wise comparison using Wilcoxon rank-sum test revealed that the difference in mtDNA/18S rDNA ratio between the NRTI-naive and NRTI-treated groups, and between healthy controls and NRTI-naive group was not statistically significant (P = 0.21 and 0.73, respectively). When the NRTI-treated group was compared with the healthy control group, there was significant increase in mtDNA in the NRTI-treated group (P = 0.019). Common pathogenic mtDNA point mutations, deletions, and DNA rearrangements were not detected in any of the samples. TTGE scanning of the entire mitochondrial genome for unknown mutations revealed more than 20 nucleotide substitutions. Most of them represent benign homoplasmic polymorphisms and reflect the variations among normal individuals. Among the nucleotide substitutions discovered, 16 of them have been reported as polymorphisms in the MITOMAP database [18,19], and 16 of them have not been reported. Three of the polymorphisms actually involve amino acid changes. These missense polymorphisms are: a threonine to methionine (C3992T) change in ND1 of patient 17, a threonine to alanine change in cytochrome b (A15326G) of several patients, and a tyrosine to histidine change in the cytochrome c oxidase subunit I (T7389C) of patient 14. It will require more extensive studies to determine whether these polymorphisms are associated with the HIV and/or NRTI-induced mitochondrial dysfunction. Ten of the 16 novel nucleotide changes were found in messenger RNA regions but none of them result in amino acid change. The remaining nucleotide changes are in tRNA and rRNA regions. The only heteroplasmic changes were observed in the WANCY region of patient 21 with a C to T change at nucleotide position (np)5586 and -1 position tRNA ala, and an A to G change at np5823 of tRNA Cys.
Table 2 summarizes the results of TTGE and sequencing analysis of the HIV-infected patients. It should be pointed out that although silent mutations do not result in amino acid changes, there are differences in the usage frequency of different codons. Most remarkable is the ACA to ACG change (A4026G) in ND1 of patient 21. The relative codon usage frequency is 133 to 10 for ACA and ACG respectively. Thus, although both ACA and ACG would encode for threonine, the efficiency of incorporating threonine at this amino acid position may have been reduced with ACG in place, thus affecting overall translational efficiency.
We failed to show evidence of mtDNA depletion in the leukocytes of a group of patients with moderate to severe lipoatrophy, including one case of lactic acidosis. In fact, the mtDNA/18S rDNA was higher in the NRTI-treated patients than in the healthy controls, but there was no significant difference between the NRTI-naive and -treated patients. The increase in mtDNA in the NRTI-treated patients could be due to mitochondrial proliferation in order to offset the unfavorable functional condition in the presence of NRTI drugs. Proliferation of defective mitochondria has been reported in patients with mitochondrial disorders . Thus, NRTI-induced mitochondrial toxicity may not solely be a consequence of broad effects of NRTI on mitochondrial DNA synthesis. Another possible explanation for the lack of mtDNA depletion in the peripheral leukocytes is the tissue specificity of NRTI-induced mitochondrial dysfunction, translating to specific tissue involvement, which could also be different and dependent on the type of NRTI used. Tissue-specific mtDNA depletions have been observed within a single patient [21,22]. Mariotti et al. found mtDNA depletion in muscle but not in leukocytes or fibroblasts of one patient . Chariot et al. found mtDNA depletion in skeletal muscle and liver, but not myocardium or kidney of another patient . These results support the hypothesis of tissue specificity.
Our results are consistent with those of Shikuma et al., who failed to show mtDNA depletion in PBMC in a group of patients with lipoatrophy, despite showing mtDNA depletion in adipose tissue . We extend this observation to patients with hyperlactatemia. Similarly, Blanche et al. studied eight children who had biochemical evidence of mitochondrial dysfunction in muscle tissues but could not demonstrate mtDNA depletion or mtDNA mutations in either lymphocytes or in muscle . Davison et al.  studied the mtDNA levels in the brain of HIV-positive patients after zidovudine therapy and found that the dementia and encephalopathy of HIV patients were not associated with depletion of mtDNA. The importance of mitochondrial DNA depletion in the pathogenesis of NRTI-related mitochondrial dysfunction is not clear. Mechanisms of NRTI-induced mitochondrial dysfunction may be linked to uncommon mutations, or may be independent of mtDNA damage. For instance, the lactic acidosis syndrome could be the result of a selective mitochondrial dysfunction in the liver, and lipodystrophy could represent a localized mitochondrial dysfunction in adipocytes. In this regard, Masini and colleagues described a dual mechanism of zidovudine-induced mitochondrial dysfunction in an animal model ; one of these mechanisms was independent of mtDNA damage, and was rather due to a direct effect of zidovudine on the respiratory chain .
In inherited mitochondrial myopathies, mtDNA deletions are often not detected in lymphocytes, while present in muscle tissues . In contrast, point mutations in these disorders can usually be detected in the blood, although analysis of skeletal muscle provides the most reliable and sensitive assay for these mutations [27,28]. Thus, failure to show mtDNA deletions in the peripheral leukocytes does not necessarily rule out a role for such mtDNA mutations in the pathogenesis of NRTI-induced mitochondrial dysfunction, but rather emphasizes the need for targeted tissue sampling of affected tissues, muscle, fat and liver, to assess the role of mitochondrial DNA abnormalities in the pathogenesis of AIDS-associated lipoatrophy and hyperlactatemia.
On the other hand, we were able to find several point mutations/variations in peripheral blood cell mitochondrial DNA samples from several patients. These variations were detected in some antiretroviral-naive patients, as well as in patients with lipoatrophy and long-term NRTI therapy. In the past, investigators looked at mtDNA deletions, depletions, and some specific known mutations, but not the entire mtDNA genome to correlate with toxicity. Conceivably, these mtDNA variations have subtle effects on respiratory chain activity in healthy individuals and these defects may be augmented in the presence of HIV infection or NRTI treatment to result in sufficient dysfunction to result in clinical disease. For example in patient 21, the translational efficiency may be further reduced by the addition of NRTIs, such that mitochondrial toxicity becomes fully manifested only after the introduction of NRTIs. Since the production of mature, functional mitochondrial rRNA, tRNA and mRNA is dependent upon endonuclease cleavage of properly folded primary transcripts, any nucleotide change can potentially affect secondary structure, thus, interfering with RNA processing. Therefore, a seemingly silent nucleotide change may drastically impair the maturation process of RNA that may be complicated by NRTI therapy. This is why one possible hypothesis would be that NRTI may not directly cause mtDNA damage, but may affect mitochondrial respiratory chain function  and that this may be potentiated in HIV disease, perhaps as a result of direct toxicities of HIV proteins such as vpr . Further studies of the frequencies of each reported and novel mtDNA polymorphism/mutation in ethnically defined populations and in HIV-infected patients will be required to ascertain if these sequences play any role in the development of mitochondrial-related disorders in HIV disease.
In summary, we were unable to show evidence of mitochondrial DNA depletion or deletion in the peripheral leukocytes of HIV-infected patients with lipoatrophy or hyperlactatemia, including a case of fulminant lactic acidosis. This is the first report that includes an extensive analysis of the entire mitochondrial genome in this clinical syndrome. This analysis identified several novel mutations/polymorphisms in patients’ samples. The presence of such mutations/polymorphisms in cells of antiretroviral-therapy-naive patients suggests that NRTI therapy may collaborate with pre-existent subtle mitochondrial variations, which can be the result of individual mtDNA variations or of HIV infection itself. These variations may be clinically silent in health but may result in subtle impairments in oxidative phosphorylation activity. Conceivably, these subtle variations in mtDNA accumulate over time and eventually lead to clinically demonstrable mitochondrial dysfunction. Furthermore, NRTI may unmask silent mtDNA variations through impairment of mitochondrial function and promote the occurrence of this disorder.
1. Chen C, Vasquez-Padua M, Cheng Y. Effect of anti-immunodeficiency virus nucleoside analogs on mitochondrial DNA and its implication for delayed toxicity. Mol Pharmacol 1991, 39: 625–628.
2. Medina DJ, Tsai CH, Hsiung GD, Cheng YC. Comparison of mitochondrial morphology, mitochondrial DNA content, and cell viability in cultured cells treated with three anti-human immunodeficiency virus dideoxynucleosides. Antimicrob Agents Chemother 1994, 38: 1824–1828.
3. Lonergan, JT, Havlir D, Barber E, Matthews WC. Incidence and outcome of hyperlactatemia associated with clinical manifestations in HIV-infected adults receiving NRTI-containing regimens. Eighth Conference on Retroviruses and Opportunistic Infections
. Chicago, January–February 2001 [abstract 624],
4. Mallal SA, John M, Moore CB, James IR. et al
. Contribution of nucleoside analogue reverse transcriptase inhibitors to subcutaneous fat wasting in patients with HIV infection
2000, 14: 1309–1316.
5. Brinkman K, Smeitink JA, Romijn, Reiss P. Mitochondrial toxicity induced by nucleoside analogue reverse-transcriptase inhibitors is a key factor in the pathogenesis of antiretroviral therapy–related lipodystrophy
. Lancet 1999, 354: 1112–1115.
6. Cai J, Yang J, Jones DP. Mitochondrial control of apoptosis: the role of cytochrome c
. Biochim Biophys Acta 1998, 1366: 139–149.
7. Susin SA, Zamzami N, Kroemer G. Mitochondria
as regulators of apoptosis: doubt no more. Biochim Biophys Acta-Bioenerg 1998, 1366: 151–156.
8. Jackson MJ, Schaefer JA, Johnson MA. et al
. Presentation and clinical investigation of mitochondrial respiratory chain disease: A study of 51 patients. Brain 1995, 118: 339–357.
9. Sherratt EJ, Thomas AW, Alcolado JC. Mitochondrial DNA defects: a widening clinical spectrum of disorders. Clin Sci 1997, 92: 225–235.
10. Hao Zhang, Cooney D, Sreenath A. et al
. Quantitation of mitochondrial DNA in human lymphoblasts by a competitive polymerase chain reaction method: application to the study of inhibitors of mitochondrial DNA content. Mol Pharmacol 1994, 46: 1063–1069.
12. Lahiri D, Nurnberger J Jr. A rapid non-enzymatic method for the preparation of HMW DNA from blood for RFLP studies. Nucl Acids Res 1991, 19: 5444.5444.
13. Wong LJC, Senadheera D. Direct detection of multiple point mutations in mitochondrial DNA. Clin Chem 1997, 43: 1857–1861.
14. Liang MH, Wong LJC. Yield of mtDNA mutation analysis in 2000 patients. Am J Med Genet 1998, 77: 395–400.
15. Vu TH, Sciacco M, Tanji K. et al
. Clinical manifestations of mitochondrial DNA depletion. Neurology 1998, 50: 1783–1790.
16. Chen TJ, Boles RG, Wong L. Detection of mitochondrial DNA mutations by temporal temperature gradient gel electrophoresis. Clin Chem 1999, 45: 1162–1167.
17. Liang MH, Wang JJ, Chen TJ, Fan W, Wong LJ. Novel mtDNA mutations detected by TTGE. Am J Hum Genet 1999, 65 (suppl) : A458.A458.
18. Kogelnik AM, Lott MY, Brown MD, Navathe SB, Wallace DC. MITOMAP: a human mitochondrial genome database. Nucl Acids Res 1996, 24: 177–179.
19. Kogelnik AM, Lott MY, Brown MD, Navathe SB, Wallace DC. MITOMAP: Shuman mitochondrial genome database – 1998 update. Nucl Acids Res 1998, 26: 112–115.
20. Oldfors A, Holme E, Tulinius M, Larsson NG. Tissue distribution and disease manifestations of the tRNA(Lys) A←G(8344) mitochondrial DNA mutation in a case of myoclonus epilepsy and ragged red fibres. Acta Neuropathol (Berl) 1995, 90: 328–333.
21. Chariot P, Drogou I, de Lacroix-Smania I. et al
. Zidovudine-induced mitochondrial disorder with massive liver steatosis, myopathy, lactic acidosis
, and mitochondrial DNA depletion. J Hepatol 1999, 30: 156–160.
22. Mariotti C, Uziel G, Carrara F. et al
. Early-onset encephalomyopathy associated with tissue-specific mitochondrial DNA depletion: a morphological, biochemical and molecular-genetic study. J Neurol 1995, 242: 547–556.
23. Shikuma C, Hu N, Milne C, Shiramizu B. Subcutaneous adipose tissue mitochondrial DNA analysis from individuals with HAART-associated lipodystrophy
2001, 15: 1801–1809.
24. Blanche S, Tardieu M, Rustin P. et al
. Persistent mitochondrial dysfunction and perinatal exposure to antiretroviral nucleoside analogues. Lancet 1999, 354: 1084–1089.
25. Davison FD, Sweeney BJ, Scaravilli F. Mitochondrial DNA levels in the brain of HIV-positive patients after zidovudine therapy. J Neurol 1996, 243: 648–651.
26. Masini A, Scotti C, Calligaro A. et al
. Zidovudine-induced experimental myopathy: dual mechanism of mitochondrial damage. J Neurol Sci 1999, 166: 31–140.
27. Larsson NG, Clayton DA. Molecular genetic aspects of human mitochondrial disorders. Annu Rev Genet 1995, 29: 151–178.
28. Kiechle FL, Kaul KL, Farkas DH. Mitochondrial disorders. Methods of specimen selection for diagnostic molecular pathology.
Arch Pathol Lab Med 1996, 120: 597–603.
29. Macreadie IG, Thorburn DR, Kirby DM, Castelli LA, de Rosario NL, Azad AA. HIV-1 protein Vpr causes gross mitochondrial dysfunction in the yeast Saccharomyces cerevisiae
. FEBS Lett 1997, 410: 145–149.