Uridine Abrogates Mitochondrial Toxicity Related to Nucleoside Analogue Reverse Transcriptase Inhibitors in Hepg2 Cells

Home , Purine analogue

Antiviral Therapy 8:463-470 Uridine abrogates mitochondrial toxicity related to nucleoside analogue reverse transcriptase inhibitors in HepG2 cells Ulrich A Walker1*, Nils Venhoff1, Eva C Koch2, Manfred Olschewski3, Josef Schneider2 and Bernhard Setzer1

1Department of Rheumatology and Clinical Immunology; 2Department of Virology, Institute for Medical Microbiology and Hygiene; and 3Department of Medical Biometry and Statistics, Medizinische Universitätsklinik, Freiburg, Germany

*Corresponding author: Tel: +49 761 270 3401; Fax: +49 761 270 3446; E-mail: [email protected]

Objective: To assess in vitro if uridine may be suitable to about 65% of NRTI-unexposed control cells. This effect prevent or treat mitochondrial toxicity related to nucleo- was dose-dependent, with a maximum at 200 µM of side analogue reverse transcriptase inhibitors (NRTIs). uridine. Uridine also rapidly and fully restored cell func- Methods: Human HepG2-hepatocytes were exposed to tion when added to cells with established mitochondrial NRTIs with or without uridine for 25 days. Cell growth, dysfunction (zalcitabine for 15 days) despite continued lactate production, intracellular lipids, mitochondrial zalcitabine exposure. Uridine also normalized cell prolif- DNA (mtDNA) and the ratio between the respiratory eration in HepG2 cells exposed to 36 µM of stavudine chain components COX II (mtDNA-encoded) and COX IV and protected HepG2-cells exposed to 7 µM of zidovu- (nuclear-encoded) were measured. dine + 8 µM of lamivudine (pyrimidine analogues), but Results: HepG2 cells exposed to zalcitabine (177 nM) failed to improve cell function or mtDNA in cells exposed without uridine developed a severe depletion of mtDNA to 11.8 or 118 µM of didanosine (a purine analogue). (to 8% of wild-type mtDNA levels), resulting in a decline Conclusions: The pyrimidine precursor uridine may atten- of cell proliferation and COX II levels, with increased uate the mitochondrial toxicity of antiretroviral pyrimidine lactate and lipid accumulation. Uridine fully abrogated NRTIs in vitro, and its supplementation may represent a the adverse effects of zalcitabine on hepatocyte prolifer- promising strategy in the prevention or treatment of ation and normalized lactate synthesis, intracellular mitochondrial toxicities in HIV-infected patients. lipids and COX II levels by adjusting mtDNA levels to

Introduction

Highly active antiretroviral therapy (HAART), In cases of life-threatening hyperlactataemia and usually comprising nucleoside analogue reverse tran- other forms of severe mitochondrial toxicity, cessation scriptase inhibitors (NRTIs) such as zidovudine of NRTI-treatment is advised [9], but biochemical and (AZT, 3′-azido-3′-deoxythimidine), zalcitabine (ddC, clinical resolution is often of slow offset. Several ′ ′ ′ ′ 2 ,3 -dideoxycytidine), didanosine (ddI, 2 ,3 -dide- dietary supplements, such as coenzyme Q10 oxyinosine), lamivudine (3TC, 2′,3′-dideoxy- (ubiquinone), vitamins or L-carnitine have been advo- 3′-thiacytidine) or stavudine (d4T, 2′,3′-didehydro- cated to improve electron flux, prevent 2′,3′-deoxythymidine), has resulted in a significant radical-mediated respiratory chain damage or to decrease of HIV-associated morbidity and mortality. provide substrates for the respiratory chain [10]. In However, several side effects have been associated vitro models of NRTI-toxicity and isolated HIV cases with the long-term use of such NRTIs and their have lent some support for this recommendation, as ability to inhibit polymerase-γ, which replicates mito- some restoration of mitochondrial damage (such as in chondrial DNA (mtDNA) [1–3]. Decreased mtDNA organelle ultrastructure, markers of oxidative stress or levels result in decreased synthesis of mtDNA- intracellular steatosis) were noted [11,12]. However, encoded respiratory chain subunits and ultimately in no clinical effectiveness has been demonstrated in a defect in oxidative phosphorylation [4]. Many controlled studies of patients with inherited mtDNA organs are thought to be involved [1,2]. The mito- mutations [10]. L-Carnitine, thiamine, riboflavin and chondrial toxicity of NRTIs in the liver has been radical scavengers also failed to demonstrate efficacy in associated with steatosis, steatohepatitis and acute reducing NRTI-related mitochondrial toxicity of liver failure, and may contribute to lactic acidosis in cultured human hepatocytes [13]. Furthermore, many some patients [5–8]. supplements have either a poor pharmacokinetic

profile (coenzyme Q10) or are less likely to be very concentration of 200 µM. Control HepG2 cells were effective in patients with mtDNA depletion, compared incubated in medium without any NRTI or with to mtDNA mutations [10,13]. uridine alone. In a search for better alternatives, we hypothesized that uridine supplementation may be used as a novel Cellular steatosis approach to prevent or treat NRTI-related mitochon- Intracellular lipid droplets were determined by Oil- drial toxicity, because fibroblasts, depleted of mtDNA Red-O-staining [4]. by long-term exposure to the polymerase-γ inhibitor ethidium bromide, became more dependent upon Lactate production uridine for growth [14]. Uridine also improved the 1.5 ml of culture fluid was collected immediately prior survival and neurite outgrowth of neuronal cells to cell trypsinization and L-lactate was determined exposed to ddC [15], and reversed the toxicity of AZT enzymatically in an automated analyser to bone marrow progenitors [16]. Therefore, we eval- (Roche/Hitachi 917) according to the manufacturers uated if uridine may be protective against instructions. Production of L-lactate was calculated in mitochondrial toxicity in human liver (HepG2) cells mmol per 105 cells [4]. exposed to NRTIs. Quantification of mtDNA Materials and methods mtDNA content was densitometrically quantified by Southern blot using Scion-image (Scion Corporation) Materials as previously described [4,17]. 5 µg of cellular DNA The human hepatoma HepG2 cell line was provided by was digested with PvuII, electrophoresed on agarose the American Type Culture Collection (ATCC HB- and blotted to a nylon membrane. mtDNA was 8065). Cell culture flasks (75 cm3) were purchased probed with a 12.9 kbp, random-prime digoxigenin- from Becton Dickinson and 10% fetal bovine serum labelled fragment, spanning nucleotide positions 3470 from PAA Laboratories, Linz, Austria. AZT, ddI, d4T, and 16379 of human mtDNA. Nuclear DNA (nDNA) uridine and Oil-Red were obtained from Sigma. was detected simultaneously with a second probe GlaxoSmithKline and Roche kindly provided 3TC and directed against 18S ribosomal DNA. The mtDNA and ddC. Pepstatin A was purchased from Boehringer nDNA signals were visualized with an alkaline-phos- Mannheim (Germany). Anti-COX II and anti-COX IV phatase conjugated, anti-digoxigenin monoclonal monoclonal antibodies were from Molecular Probes. antibody (Boehringer Mannheim, Germany) according Abacavir was unfortunately not released for tests on to the manufacturers instructions. mtDNA was HepG2 cells by its manufacturer. normalized for nDNA content by calculating the mtDNA/nDNA ratio. Cell culture ° HepG2 cells were propagated at 37 C and 5% CO2 in Quantification of the mtDNA-encoded COX II Dulbecco’s Modified Eagle Medium (DMEM), respiratory chain subunit containing 4.5 g/l glucose and 110 mg/l pyruvate, The subunit II of cytochrome c-oxidase (COX II) is supplemented with 10% fetal bovine serum, 50 U/ml encoded by mtDNA, whereas the subunit IV of streptomycin, 50 U/l penicillin and 250 µg/l ampho- cytochrome c-oxidase (COX IV) is encoded by tericin B. 2.7×106 HepG2 cells were seeded during nDNA. COX II was quantified by immunoblot logarithmic growth at day 1 and harvested at days 5, densitometry and normalized to the signal of a simul- 10, 15, 20 and 25 when viable cells were counted, taneously used antibody against COX IV as and 2.7×106 cells were replated in new flasks. described in detail elsewhere [4]. The ratios between Medium was renewed at the harvest and on every the COX II and the COX IV signals were calculated third day after plating. and the results expressed in percent of the control NRTIs with or without uridine were added in group mean. concentrations corresponding to the steady state peak plasma levels (Cmax) in humans during HIV therapy, for Statistics example, 7.0 µM of AZT, 177 nM of ddC and 8.3 µM The absolute values at each time point of the cell 5 of 3TC (product data sheets). For d4T, 10-fold Cmax count, of the lactate measurements per 10 cells, of the concentrations (36 µM) were used, because Cmax mtDNA/nDNA ratio and of the COX II/COX IV ratio concentrations did not display discernable mitochon- were compared by repeated measures analysis of vari- drial toxicity. For ddI, two concentrations were tested ance (ANOVA) [4] using the SAS statistical program

(11.8 µM corresponding to Cmax, and 118 µM). Unless (vs. 6.12). otherwise indicated, uridine was used at a standard

Results Uridine prevents ddC-mediated cytotoxicity When uridine (200 µM) was added at day 1 to the Mitochondrial toxicity of ddC on HepG2 cells medium containing ddC, the cells did not deteriorate HepG2 control cells (without any NRTI) proliferated with respect to cell function (growth and lactate rapidly and had multiplied by an average factor of 4.9 production) or COX II/COX IV ratio. This represented within each 5 day interval, resulting in 13.2×106 cells a significant difference, compared to ddC medium (SD ±1.8×106). ddC was chosen in the first set of without uridine (Figure 1, Table 1). Uridine appeared experiments, because this substance is probably the to maintain normal cell function under ddC, as no strongest inhibitor of mtDNA synthesis in HepG2 cells significant difference was noted, compared to among the currently licensed NRTIs [4,18]. As previ- untreated controls. mtDNA levels were also signifi- ously described [4], ddC caused a time-dependent cantly elevated in cells treated with ddC plus uridine, inhibition of HepG2 growth (Table 1, Figure 1). At day compared to cells treated without uridine. However, 20, the cells not only displayed severe changes in size uridine did not fully normalize mtDNA, as mtDNA and shape, but also showed increased intracellular undulated at levels of around 65% of those of normal lipids (Figure 2). Lactic acid, the product of compen- HepG2 cells. Another control, consisting of cells satory glycolysis that results from an inhibition of supplemented with uridine alone (without ddC), did mitochondrial oxidative phosphorylation, increased neither reveal any unspecific benefits of uridine, nor progressively to 350% of control values after 25 days. any toxicity (Table 1). The Oil-Red stain demon- These changes were paralleled by a decline in the rela- strated that the profound effects of ddC on cell tive COX II content of the respiratory chain, but morphology and on the accumulation of intracellular preceded by a steep and rapid decline of mtDNA/cell, lipids appeared to be fully prevented by uridine suggesting that mtDNA depletion is central in the onset (Figure 2). of the cytotoxicity.

Table 1. Cell proliferation, lactate production, mtDNA/nDNA and COX II/COX IV ratios expressed as a percentage of control HepG2 cells (mean ±SD)

Cell count (% of control) Lactate (% of control)

Day 5 Day 10 Day 15 Day 20 Day 25 Day 5 Day 10 Day 15 Day 20 Day 25

Control 100 ±14 100 ±10 100 ±13 100 ±7 100 ±9 100 ±16 100 ±18 100 ±20 100 ±15 100 ±17 Uridine 79 ±4 101 ±34 97 ±24 120 ±11 85 ±13 § 118 ±6 115 ±46 121 ±23 87 ±8 109 ±29 § ddC 85 ±13 76 ±5 58 ±16 38 ±14 21 ±6 † 125 ±25 137 ±32 216 ±96 295 ±29 350 ±29 † ddC + uridine 85 ±14 123 ±4 99 ±18 90 ±9 92 ±7 *§ 98 ±12 90 ±1 99 ±7 117 ±12 92 ±18 *§ d4T 87 ±25 80 ±15 58 ±7 67 ±6 46 ±13 † 120 ±41 140 ±67 171 ±59 149 ±39 138 ±95 § d4T + uridine 101 ±4 102 ±22 101 ±9 114 ±11 109 ±25 *§ 91 ±10 104 ±5 111 ±6 91 ±21 70 ±17 § ddI 82 ±18 56 ±8 28 ±5 33 ±30 24 ±28 † 125 ±24 209 ±41 322 ±93 350 ±122 673 ±359 † ddI + uridine 86 ±5 63 ±16 36 ±18 35 ±4 25 ±6 108 ±21 203 ±75 282 ±105 328 ±131 433 ±177 AZT + 3TC 66 ±7 53 ±20 31 ±24 22 ±30 15 ±21 † 120 ±32 140 ±25 158 ±9 153 ±11 189 ±34 † AZT + 3TC + uridine 64 ±3 83 ±16 63 ±1 86 ±6 73 ±16 *§ 134 ±8 108 ±17 128 ±27 99 ±13 118 ±53 *§

mtDNA (% of control) COX (% of control)

Day 5 Day 10 Day 15 Day 20 Day 25 Day 5 Day 10 Day 15 Day 20 Day 25

Control 100 ±12 100 ±17 100 ±20 100 ±18 100 ±18 100 ±26 100 ±38 100 ±43 100 ±23 100 ±19 Uridine 98 ±7 101 ±4 102 ±1 100 ±3 106 ±4 § 97 ±0 84 ±7 81 ±3 96 ±2 95 ±1 § ddC 27 ±14 24 ±17 14 ±6 11 ±6 8 ±3 † 72 ±11 48 ±14 30 ±21 11 ±9 8 ±5 † ddC + uridine 76 ±40 62 ±24 65 ±21 53 ±15 71 ±33 *§ 108 ±25 100 ±32 92 ±19 110 ±32 91 ±11 *§ d4T 115 ±4 86 ±21 69 ±2 61 ±8 68 ±20 § 94 ±10 83 ±7 79 ±2 76 ±6 70 ±9 § d4T + uridine 92 ±10 93 ±2 79 ±5 86 ±4 102 ±7 § 94 ±2 74 ±8 81 ±16 92 ±29 86 ±15 § ddI 29 ±13 16 ±8 21 ±20 15 ±10 11 ±14 † 74 ±7 38 ±13 19 ±21 18 ±16 13 ±15 † ddI + uridine 47 ±20 28 ±13 32 ±16 20 ±7 16 ±10 102 ±5 56 ±34 43 ±35 30 ±25 18 ±23 AZT + 3TC 97 ±18 99 ±2 105 ±5 90 ±1 84 ±2 § 87 ±6 86 ±24 71 ±2 93 ±52 83 ±8 § AZT + 3TC + uridine 99 ± 31 90 ± 4 101 ± 10 84 ± 7 106 ± 7 § 91 ± 8 81 ±32 63 ±1 106 ±21 107 ±13 §

The SD represents two or three independent repeats (except for the controls, for which nine assays were performed). † Significant (P<0.05) effects of NRTIs compared to controls; *statistical difference between the absence and presence of uridine; § no statistical difference compared to controls.

Antiviral Therapy 8:5 465 UA Walker et al.

Figure 1. Preventive effect of uridine on ddC-related mitochondrial toxicity

400 120 Cell number Lactate

100 300 80

60 200

40 Control 100 20 ddC ddC + Uri 0 0

140 mtDNA/nDNA 140 COX II/COX IV

% of control mean 120 120 100 100 80 80 60 60 40 40 20 20 0 0 0255 10 15 20 0255 10 15 20

Days of treatment

Uridine (200 µM) maintained virtually normal levels of mtDNA-encoded respiratory chain protein, cell growth and lactate and attenuated mtDNA depletion by ddC (177 nM). Symbols and bars represent means ±SD.

Figure 2. Oil-Red stain of HepG2 cells at day 20

AB C

Cells without ddC or uridine proliferated rapidly and showed only scarce intracellular lipids (A). Cells incubated with ddC (177 nM) were decreased in number, enlarged and had increased amounts of intracellular lipids (B). Uridine (200 µM) prevented the ddC-related changes in cell morphology and steatosis (C). Scale bar: 20 µM.

Uridine treats ddC-mediated cytotoxicity also improved lactate production, as well as the In a second set of experiments, the action of uridine mtDNA/nDNA and COX II/COX IV ratios. A similar was tested on cells that had already acquired severe recovery was observed in cells exposed to medium mtDNA depletion and respiratory chain dysfunction. without ddC. Cells were split after 15 days of exposure to ddC and 2.7×106 cells replated in three new flasks. Medium Uridine dose-response with neither of ddC or uridine was added to the first, Experiments with ddC and different uridine concentra- medium containing both ddC and uridine to the second, tions (2, 20, 50 and 200 µM) indicated an effect of and uridine alone to the third flask (Figure 3). In this uridine on all parameters possibly starting at concen- ‘therapeutic’ setting, the addition of uridine to ddC led trations of 50 µM, although we did not perform to apparently normal cell growth within 10 days and enough repeats in order to make a statistically sound

Figure 3. Uridine in the treatment of established mitochondrial toxicity

Cell number 200 Lactate 120 ddC+, Uri- 180 + + ddC , Uri 160 100 ddC-, Uri+ ddC-, Uri- 140

80 120

100 60 80

mtDNA/nDNA COX II/COX IV 120

% of control mean 100 100

80 80 60

40 60 20

0 40 0255 10 15 20 0255 10 15 20

Days of treatment

HepG2 cells were incubated with ddC (177 nM) for 15 days (open triangles) and then reseeded in three flasks (2.7×106 cells each). Flasks were then incubated for further 10 days with ddC only (open triangles), uridine only (filled circles) or with uridine (200 µM) plus ddC (filled circles). conclusion. Uridine in concentrations of 2 and 20 tended to improve these parameters. In the AZT/3TC µM did not appear to have discernible effects (data combination, the addition of uridine from day 1 not shown). improved cell growth and lactate, as these parameters were statistically normalized (Table 1, Figure 4). Mitochondrial toxicity of other NRTIs In contrast to the benefits of uridine for cells To compare the toxicities of other NRTIs and their exposed to d4T or AZT/3TC, uridine failed to provide combinations with previously described effects [4] we protection against the mitochondrial toxicity mediated included ddI, d4T and the AZT/3TC mixture into our by 118 µM (Table 1) and 11.8 µM (not shown) of ddI. studies. The latter was chosen due to an elevated effect of the mixture in comparison to its individual compo- Discussion nents. In our system, ddI reduces mtDNA levels, the COX II/COX IV ratio and cell proliferation, and We first examined the effect of uridine on the long- increases lactate (Table 1). The effects of d4T appear to term mitochondrial toxicity of ddC and recorded be similar to those of ddI, but quantitatively less mtDNA, mtDNA-encoded respiratory chain protein pronounced and at times only at the border of statistical and cell function (lactate production, intracellular significance. The combination of AZT and 3TC leads to lipids and cell proliferation). Cell function virtually a not yet fully characterized mitochondrial toxicity by normalized with uridine, which contrasts with the lack inhibiting cell growth and increasing lactate production of improvement by a multivitamin cocktail (riboflavin, without specifically affecting mtDNA levels and thiamine and vitamin C) and of L-carnitine in the same mtDNA-encoded respiratory chain components. model [13]. We found that in the presence of 177 nM of ddC, uridine appeared to adjust mtDNA levels to Effect of uridine on other NRTIs about 65% of NRTI-naive HepG2-cells. Other investi- d4T reduced HepG2 cell growth and this statistically gators have studied cells harbouring heteroplasmic significant effect was prevented by the addition of mtDNA mutations and found that wild-type mtDNA uridine (200 µM) to the medium. Lactate, mtDNA levels in the order of 20% provided protection from content and COX II/COX IV ratio were adversely (but respiratory chain dysfunction [14,19]. Maintenance of not significantly) affected by d4T, but uridine also mtDNA above this threshold in our ddC experiments

Antiviral Therapy 8:5 467 UA Walker et al.

Figure 4. Uridine (200 µM) in the prevention of toxic effects of AZT (7 µM) combined with 3TC (8.3 µM)

Control AZT + 3TC AZT + 3TC + Uridine 250 120 Cell number Lactate

100 200 80

60 150

% of control mean 100 20

0 0 0255 10 15 20 0255 10 15 20

Days of treatment

Symbols and bars represent means ±SD. could therefore explain why uridine completely increasing the relative amounts of NRTIs, which normalized COX composition and cell function; but then compete more efficiently at polymerase-γ. We how may uridine affect mtDNA levels? hypothesize that such a ‘vicious circle’ can be The d-drug NRTIs (ddC, d4T and ddI) inhibit disrupted by supplying uridine as an exogenous polymerase-γ by competing for their natural mito- source of pyrimidine precursors, and that this chondrial nucleotide counterparts and by causing represents the mechanism for the beneficial effects chain termination after their incorporation into the of uridine on mtDNA depletion and its conse- nascent mtDNA strand. The onset of severe mtDNA quences to cell function. It is also possible, that depletion and the associated decline in the synthesis uridine metabolites compete for the pyrimidine of mtDNA-encoded respiratory chain subunits affects NRTIs at kinases responsible for the intramito- several metabolic pathways: first, ATP can no longer chondrial activation or for transporters responsible be synthesized efficiently through oxidative phospho- for their import [3]. rylation and glycolysis has to be relied upon; second, Both hypotheses need not be mutually exclusive and the NADH/NAD+ ratio increases, which may are also supported by the observation that uridine did contribute to lactate synthesis and inhibit key not abrogate the toxicity of ddI, presumably due to the enzymes of β-oxidation, resulting in the intracellular inability of its pyrimidine derivatives to compete with accumulation of triglycerides; third, an intact respira- ddI as a purine. With d4T, another pyrimidine d-drug, tory chain function is also necessary for the de novo reduced cell counts were observed despite only a synthesis of all cellular pyrimidines, because an effi- moderate mtDNA depletion. It is unclear to us whether cient electron flux, especially through complexes III this represents an effect of statistics or may indicate and IV, is essential for the activity of dihydroorotate non-mtDNA effects that contribute to d4T toxicity. dehydrogenase (DHODH; E.C. 1.3.99.11) [20]. Uridine also appeared to improve the d4T-mediated DHODH synthesizes orotate from which uridine cell growth inhibition. monophosphate and all (mitochondrial and nuclear) We also noted a positive effect of uridine in pyrimidines are derived. Cells with a defect in the preventing the poorly understood toxicity of the respiratory chain, therefore, become auxotrophic for AZT/3TC combination. Strategies aimed at increasing uridine. The fact that uridine provides precursors cellular pyrimidine precursors have previously also necessary for DNA synthesis at mitosis has been used been promising in models of AZT-induced anaemia to explain its positive effect on the proliferation of and leukopenia [16,21]. AZT may bind intramito- cells harbouring dysfunctional mtDNA [14]. This chondrially to adenylate kinase, an enzyme involved in mechanism, however, does not yet explain why ATP formation, and inhibit the mitochondrial ADP- uridine maintained normal lactate and improved ATP translocator or nucleoside diphosphate kinase [3]. mtDNA in ddC-treated cells. We suggest that respira- It is conceivable, that uridine-derived pyrimidines may tory chain inhibition due to NRTI exposure also also compete with AZT for one or several of these depletes intramitochondrial pyrimidines, thereby enzymes [16].

3. Kakuda TN. Pharmacology of nucleoside and nucleotide The effect of exogenous uridine on mtDNA has reverse transcriptase inhibitor-induced mitochondrial toxi- important potential implications: firstly, HIV patients city. Clinical Therapeutics 2000; 22:685–708. suffering from NRTI-related liver toxicity may benefit 4. Walker UA, Setzer B & Venhoff N. Increased long-term mitochondrial toxicity in combinations of nucleoside from strategies aimed at increasing uridine in hepato- analogue reverse-transcriptase inhibitors. AIDS 2002; cytes. This may also be true for NRTI toxicities in 16:2165–2173. non-hepatic tissues, although other tissues may have 5. Lambert JS, Seidlin M, Reichman RC, Plank CS, Laverty M, Morse GD, Knupp C, McLaren C, Pettinelli C & different requirements for natural pyrimidines, vary in Valentine FT. 2′,3′-dideoxyinosine (ddI) in patients with the the activation and intraorganellar transport of pyrimi- acquired immunodeficiency syndrome or AIDS-related complex. A Phase I trial. New England Journal of dine analogues, or differentially depend on respiratory Medicine 1990; 322:1333–1340. chain function [1,2,17,22]; secondly, uridine supplemen- 6. Morris AA, Taanman JW, Blake J, Cooper JM, Lake BD, tation may also be useful in improving the therapeutic Malone M, Love S, Clayton PT, Leonard JV & Schapira AH. Liver failure associated with mitochondrial DNA index of NRTIs and thereby allow higher dosing of these depletion. Journal of Hepatology 1998; 28:556–563. antiretroviral drugs, which could overcome HIV resis- 7. Chariot P, Drogou I, de Lacroix-Szmania I, Eliezer-Vanerot tance; thirdly, it may be possible, that some NRTIs also MC, Chazaud B, Lombes A, Schaeffer A & Zafrani ES. Zidovudine-induced mitochondrial disorder with massive indirectly inhibit DHODH in lymphocytes, which could liver steatosis, myopathy, lactic acidosis, and mitochondrial result in a seemingly paradoxical immunsuppressive DNA depletion. Journal of Hepatology 1999; 30:156–160. effect [23,24]. 8. Gerard Y, Maulin L, Yazdanpanah Y, Tribonniere X, Amiel C, Maurage CA, Robin S, Sablonniere B, Dhennain C & Pharmacokinetic studies of uridine have been Mouton Y. Symptomatic hyperlactataemia: an emerging conducted in order to improve the therapeutic index complication of antiretroviral therapy. AIDS 2000; 14:2723–2730. of 5-fluorouracil. Parental administration of uridine 9. Brinkman K, Vrouenraets S, Kauffman R, Weigel H & resulted in plasma levels far above those required to Frissen J. Treatment of nucleoside reverse transcriptase rescue HepG2 cells, for example, in the millimolar inhibitor-induced lactic acidosis. AIDS 2000; 14:2801–2802. range, and was tolerated well by humans [25,26]. 10. Walker UA & Byrne E. The therapy of respiratory chain Oral uridine can also be administered at high doses, encephalomyopathy: a critical review of the past and with mild osmotic diarrhoea being dose-limiting current perspective. Acta Neurologica Scandinavica 1995; 92:273–280. beyond 12 g/m2 [27]. Due to the pharmacokinetic 11. de la Asuncion JG, del Olmo ML, Sastre J, Pallardo FV & properties of uridine (bioavailability of 7% and Vina J. Zidovudine (AZT) causes an oxidation of mito- serum half-life of 2 h), other sources of pyrimidine chondrial DNA in mouse liver. Hepatology 1999; 29:985–987. precursors, such as deoxycytidine [16,28,29], or 12. Fouty B, Frerman F & Reves R. Riboflavin to treat nucleo- inhibitors of uridine catabolism or excretion may be side analogue-induced lactic acidosis. Lancet 1998; envisaged [30,31]. 352:291–292. Factors limiting the use of pyrimidine precursors in 13. Venhoff N, Setzer B, Lebrecht D & Walker UA. Dietary supplements in the treatment of NRTI-related mitochon- HIV patients may be a theoretical risk of an increased drial toxicity. AIDS 2002; 16:800–802. nuclear or mitochondrial mutation rate, and compro- 14. Bodnar AG, Cooper JM, Leonard JV & Schapira AH. mised antiretroviral efficacy. Uridine metabolites might Respiratory-deficient human fibroblasts exhibiting defec- tive mitochondrial DNA replication. Biochemical Journal compete with the NRTIs at the level of HIV reverse 1995; 305:817–822. transcriptase or might interfere with the activation of 15. Keilbaugh SA, Hobbs GA & Simpson MV. Anti-human the NRTIs to the antiretrovirally active nucleotide immunodeficiency virus type 1 therapy and peripheral neuropathy: prevention of 2′,3′-dideoxycytidine toxicity in triphosphates. However, uridine precursors in molar PC12 cells, a neuronal model, by uridine and pyruvate. excess were reported not to impair the antiretroviral Molecular Pharmacology 1993; 44:702–706. activity NRTIs [16,21,32]. 16. Sommadossi JP, Carlisle R, Schinazi RF & Zhou Z. Uridine reverses the toxicity of 3′-azido-3′-deoxythymidine in normal Taken together, the data indicate that uridine attenu- human granulocyte-macrophage progenitor cells in vitro ates the mitochondrial toxicity of antiretroviral without impairment of antiretroviral activity. Antimicrobial Agents & Chemotherapy 1988; 32:997–1001. pyrimidine NRTIs in HepG2 cells in vitro and that 17. Walker UA, Bickel M, Lütke Volksbeck SI, Ketelsen UP, strategies aimed at increasing uridine could be promising Schofer H, Setzer B, Venhoff N, Rickerts V & Staszewski S. in the prevention or treatment of mitochondrial side Evidence of nucleoside analogue reverse transcriptase inhibitor-associated genetic and structural defects of mito- effects in HIV-infected patients. chondria in adipose tissue of HIV-infected patients. Journal of Acquired Immune Deficiency Syndromes 2002; 29:117–121. References 18. Birkus G, Hitchcock MJ & Cihlar T. Assessment of mitochondrial toxicity in human cells treated with tenofovir: comparison with other nucleoside reverse transcriptase 1. Lewis W & Dalakas MC. Mitochondrial toxicity of inhibitors. Antimicrobial Agents & Chemotherapy 2002; antiviral drugs. Nature Medicine 1995; 1:417–422. 46:716–723. 2. Brinkman K, Smeitink JA, Romijn JA & Reiss P. 19. Attardi G, Yoneda M & Chomyn A. Complementation and Mitochondrial toxicity induced by nucleoside analogue segregation behavior of disease-causing mitochondrial reverse transcriptase inhibitors is a key factor in the patho- DNA mutations in cellular model systems. Biochimica et genesis of antiretroviral therapy-related lipodystrophy. Biophysica Acta 1995; 1271:241–248. Lancet 1999; 354:1112–1115.

Antiviral Therapy 8:5 469 UA Walker et al.

20. Löffler M, Jöckel J, Schuster G & Becker C. 27. van Groeningen C, Peters G, Nadal J, Leyva A, Gall H & Dihydroorotat-ubiquinone oxidoreductase links mitochon- Pinedo H. Phase I clinical and pharmacokinetics study of dria in the biosynthesis of pyrimidine nucleotides. orally administered uridine. Proceedings of the AACR Molecular & Cellular Biochemistry 1997; 174:125–129. 1987; 28:195. 21. Calabresi P, Falcone A, St Clair MH, Wiemann MC, Chu 28. Bhalla KN, Li GR, Grant S, Cole JT, MacLaughlin WW & SH & Darnowski JW. Benzylacyclouridine reverses Volsky DJ. The effect in vitro of 2′-deoxycytidine on the azidothymidine-induced marrow suppression without metabolism and cytotoxicity of 2′,3′-dideoxycytidine. AIDS impairment of anti-human immunodeficiency virus activity. 1990; 4:427–431. Blood 1990; 76:2210–2215. 29. Bhalla K, Birkhofer M, Li GR, Grant S, MacLaughlin W, 22. Medina DJ, Tsai CH, Hsiung GD & Cheng YC . Cole J, Graham G & Volsky DJ. 2′-Deoxycytidine protects Comparison of mitochondrial morphology, mitochondrial normal human bone marrow progenitor cells in vitro DNA content, and cell viability in cultured cells treated against the cytotoxicity of 3′-azido-3′-deoxythymidine with with three anti-human immunodeficiency virus dideoxynu- preservation of antiretroviral activity. Blood 1989; cleosides. Antimicrobial Agents & Chemotherapy 1994; 74:1923–1928. 38:1824–1828. 30. Falcone A, Darnowski JW, Ruprecht RM, Chu SH, 23. Setzer B, Schlesier M, Thomas AK & Walker UA. Brunetti I & Calabresi P. Different effect of benzylacy- Immunosuppressive effect of mtDNA-depletion by d-drugs clouridine on the toxic and therapeutic effects of in healthy human-lymphocytes. Antiviral Therapy 2003; azidothymidine in mice. Blood 1990; 76:2216–2221. 8:L43. 31. de Miranda P, Good SS, Yarchoan R, Thomas RV, Blum 24. Silva HT, Cao W, Shorthouse R & Morris RE. Mechanism MR, Myers CE & Broder S. Alteration of zidovudine phar- of action of leflunomide: in vivo uridine administration macokinetics by probenecid in patients with AIDS or reverses its inhibition of lymphocyte proliferation. AIDS-related complex. Clinical Pharmacology & Transplantation Proceedings 1996; 28:3082–3084. Therapeutics 1989; 46:494–500. 25. Leyva A, van Groeningen CJ, Kraal I, Gall H, Peters GJ, 32. Walker UA, Koch E, Venhoff N, Klinker H, Langman P, Lankelma J & Pinedo HM. Phase I and pharmacokinetic Zilly M, Schneider J & Setzer B. Uridine prevents and studies of high-dose uridine intended for rescue from 5- treats mtDNA depletion by NRTI pyrimidine analogues fluorouracil toxicity. Cancer Research 1984; and fully restores mitochondrial function. Antiviral 44:5928–5933. Therapy 2003; 8:L17. 26. van Groeningen CJ, Leyva A, Kraal I, Peters GJ & Pinedo HM. Clinical and pharmacokinetic studies of prolonged administration of high-dose uridine intended for rescue from 5-FU toxicity. Cancer Treatment Reports 1986; 70:745–750.

Received 12 May 2003; accepted 6 July 2003