In anti HIV therapy, usually at least two nucleoside analogue reverse transcriptase inhibitors (NRTI), such as zidovudine (ZDV), zalcitabine (ddC), didanosine (ddI), lamivudine (3TC) or stavudine (d4T) are combined with a protease inhibitor or with a non-nucleoside analogue reverse transcriptase inhibitor (NNRTI), such as efavirenz (EFV). With prolonged exposure to antiretroviral drugs, several long-term side effects have been associated with the use of NRTI and their ability to inhibit polymerase-γ the enzyme responsible for replication of mitochondrial (mt) DNA . mtDNA depletion in the liver may manifest as microvesicular steatosis, steatohepatitis or organ failure [2–4]. Long-term use of NRTI is also associated with lactic acidosis, another manifestation of mitochondrial respiratory chain dysfunction [2,4–6].
Although it has been suggested from data collected in vitro, in animals and in humans that NRTI are able to cause mtDNA depletion and respiratory chain dysfunction [2,4,7–9], the interaction between two NRTI on mitochondria has not been assessed. In order to address this question, we incubated the human hepatoma HepG2 cell line with single NRTI and with combinations of NRTI as they are typically prescribed to HIV patients in antiretroviral therapy. We measured cell growth, lactate production, intracellular lipids, mtDNA, and the expression of the mtDNA-encoded cytochrome c oxidase subunit II (COX II). In order to assess long-term toxicity, the cells were exposed to the antiviral agents over a period of 25 days.
Materials and methods
The human hepatoma HepG2 cell line was provided by the American Type Culture Collection (ATCC HB-8065). Cell culture flasks (75 cm2) were from Becton Dickinson (Beiersdorf, Germany) and 10% foetal bovine serum was from PAA Laboratories (Linz, Germany). ZDV, ddI, d4T, were from Sigma (Deisenhofen, Germany), and 3TC, ddC and EFV were kindly provided by Glaxo Smith Kline (Stevenage, UK), Roche (Basle, Switzerland) and DuPont, respectively. The effects of abacavir, the sixth currently licensed NRTI, were not assessed because this drug was not released by the manufacturer. Oil-Red was from Sigma and L-lactate was determined using a kit from Roche Diagnostics. Pepstatin A was from Boehringer Mannheim (Mannheim, Germany). The anti-COX II, anti-COX IV and anti-mouse monoclonal antibodies were from Molecular Probes (Leiden, Netherlands) and Gibco (Karlsruhe, Germany).
HepG2 cells were propagated at 37°C and 5% CO2 in Dulbecco's modified Eagle's medium, containing 4.5 g/l glucose and 110 mg/l pyruvate, supplemented with 10% foetal bovine serum, 50 U/ml streptomycin, 50 U/l penicillin and 0.25 amphotericin B. A total of 2.7 × 106 HepG2 cells were seeded during logarithmic growth at day 1. At days 5, 10, 15, 20 and 25, the cells were harvested. Viable cells were counted by the Trypan blue exclusion method in a Neubauer chamber. Then, 2.7 × 106 cells were reseeded in new flasks. Medium was renewed on the days of trypsinization and on every third day after seeding.
NRTI were added in concentrations corresponding to the steady-state peak plasma levels (Cmax) in humans during HIV therapy, e.g., 7.0 μM ZDV, 3.6 μM d4T, 177 nM ddC, 8.3 μM 3TC and 11.8 μM ddI (product data sheets). In addition, concentrations corresponding to one-third and to 10 times the Cmax were tested. With the exception of ZDV [which was dissolved in medium containing 2.5% dimethyl sulfoxide (DMSO)] all NRTI were dissolved in medium. For the HepG2 cells exposed to ZDV, the final DMSO concentration was 0.0025–0.00025%. Controls were incubated in medium without NRTI, or with EFV at Cmax (12.4 μM). The EFV concentration probably represents multiplicities of those in human plasma, because it was not adjusted for the high protein binding of this NNRTI.
Intracellular lipid droplets were determined by Oil-Red-O staining. Cells grown on coverslips were stained in Oil-Red-O for 30 min, counterstained with Meyer's hematoxylin (8 min), washed in water and embedded in glycerol–gelatine. Slides from incubations in medium without any added NRTI, from HepG2 cells with ddC 59 nM and with ddC 177 nM at day 20 were chosen as standards. Subsequently all slides were compared to these standards by a blinded investigator. When day 20 slides could not be used because of early cell death, earlier slides were assessed.
Supernatant (1.5 ml) was collected immediately prior to trypsinization and L-lactate was determined enzymatically in an automated analyser (Roche/Hitachi 917) according to the manufacturer's instructions. Production of L-lactate was calculated in mmol/l per 105 cells.
Quantification of the mtDNA-encoded COX II respiratory chain subunit
Cellular protein (100 μg) was diluted (1 : 1) in 10 μl sample buffer [50 mM Tris–HCl pH 6.8, 12% glycerol, 4% sodium dodecyl sulphate (SDS), 1 μg/ml Pepstatin A, 0.01% Bromophenol blue] and boiled for 3 min. Proteins were separated by electrophoresis on an 11% polyacrylamide gel, containing 0.1% SDS and electroblotted onto nitrocellulose sheets. The nitrocellulose sheets were blocked with 10% non-fat dry milk and incubated overnight in blocking buffer containing a 1 : 100 dilution of both anti-human COX II and anti-human COX IV mouse monoclonal antibodies. Washes, incubation with an alkaline phosphatase- conjugated goat secondary anti-mouse antibody (diluted 1 : 200), and development with nitro Blue tetrazolium–5-bromo-4-chloroindol-2-yl phosphate were performed according to standard procedures. The intensities of the signals were quantified by scanning densitometry, using Scion-image. The intensity of the COX II signal was normalized to the signal of the nuclear (n) DNA-encoded (COX IV) subunit and calculated as the COX II : COX IV ratio .
Quantification of mtDNA
mtDNA content was quantified using a Southern blot technique [9–11]. Genomic DNA was extracted and digested with PvuII. Five μg of the digest were electrophoresed on a 0.8% agarose gel and blotted to a nylon membrane. mtDNA was probed with a 12.9 kilo base pair, random prime digoxigenin-labelled fragment, spanning nucleotide positions 3470 and 16379 of human mtDNA. In order to obtain an internal standard, nDNA was probed simultaneously with a second digoxigenin-labelled probe, directed against the multi-copy 18S ribosomal DNA gene. The mtDNA and nDNA signals were visualized with an anti-digoxigenin alkaline phosphatase conjugated monoclonal antibody (Boehringer Mannheim) according to the manufacturer's instructions. The signals were quantified by scanning densitometry using Scion-image (Scion Corporation, Frederick, ML, USA). mtDNA content was calculated as the ratio between mtDNA and nDNA signals. The Southern blots were also screened for the presence of large-scale mtDNA deletions.
For each NRTI, the absolute values of the cell count, lactate and mtDNA : nDNA ratio measurements at each time point were tested by repeated measures analysis of variance (ANOVA) for concentration and time dependency. NRTI combinations were tested for between-subjects effects using the general linear models procedure of the SAS statistical programme (version 6.12). Due to a relatively large interblot variation in the COXII : COXIV immunoblots (inter-run SD of about 30% within controls), this parameter was not statistically tested. The results from the COXII : COXIV ratio determinations therefore represent calculations relative to same-day controls tested on the same immunoblot.
To detect possible differences between the individual NRTI with regard to mitochondrial toxicity, the HepG2 cells were incubated with single NRTI. The results are given in Tables 1 and 2.
The HepG2 cells proliferated rapidly in medium without any added anti-HIV agent: at the time of each harvest, the 2.7 × 106 cells seeded had multiplied on average by a factor of 4.6, resulting in 12.5 × 106 cells (SD ± 1.3 × 106). Cell growth was inhibited by ddC, ZDV, ddI and d4T. Cytotoxicity declined in the order ddC > ZDV > ddI > d4T, whereas 3TC and EFV did not cause inhibition of cell growth.
The cytotoxic effects of ddC, ZDV, ddI and d4T were dose dependent (P < 0.001 for each). At concentrations corresponding to peak plasma concentrations in human anti-HIV therapy, the toxicity was substantially reduced compared to the 10-fold Cmax concentrations.
The changes in cell growth mediated by ddC, ZDV, ddI and d4T were also time dependent (P < 0.001 for each). Inhibition of cell growth was not discernible at day 5 at most concentrations, but was observed only with prolonged exposure.
The most prominent effect on cellular lactate production was observed with ddI. ddI at doses equivalent to Cmax increased lactate production per cell by about sixfold (613%) at day 25. The respective values for ddC, ZDV and d4T were 293%, 153% and 144%. Similar to cell growth, the rise in lactate was time- (P < 0.01 for each ddI, ddC, ZDV and d4T) and dose-dependent (P < 0.01 for each ddI, ddC, ZDV and d4T). 3TC did not affect lactate at any dose.
mtDNA and COX II
In the incubations with ddC, ddI and d4T, impaired cell growth, an increase in lactate and an increase in intracellular lipids was observed and associated with a decline of mtDNA. In the HepG2 cells exposed to 177 nM ddC, mtDNA decreased by a factor of four (to 26%) by day 5, whereas the cells had multiplied by a similar factor (3.8) during the observation period. mtDNA depletion frequently preceded rapid alterations in cell number and lactate, was largely dose-dependent (P < 0.001 for ddI and ddC; P = 0.07 for d4T) and resulted in a decline of the mtDNA-encoded COX II respiratory chain subunit. In general, COX II decreased somewhat later than mtDNA. mtDNA depletion was most prominent with ddC, followed by ddI and d4T. Notably, steady-state levels of mtDNA and of COX II were not achieved by ddI and d4T after 25 days of exposure. mtDNA was reduced to 70% by 4.2 μM d4T, to 42% by 42 μM d4T, and to 2% by 11.8 μM ddI at day 30. COX II levels at day 30 declined to 46% with 4.2 μM d4T, to 48% with 42 μM d4T, and to 1% with 11.8 μM ddI. Importantly, there was no mtDNA or COX II depletion in cells exposed to ZDV, despite obvious cytotoxicity. Neither EFV, 3TC nor ZDV affected significantly mtDNA or COX II levels. mtDNA deletions were never observed using the Southern blot technique.
Oil-Red staining of HepG2 cells after 20 days’ incubation in medium without any NRTI revealed scarce intra-cytoplasmic lipid droplets. When ddC at one-third Cmax and at Cmax was added, fewer cells adhered to the coverslip. These cells also contained increased amounts of lipid droplets. The most prominent lipid accumulation was observed in cells incubated with ddC at 1.77 μM. All NRTI, except 3TC, caused some accumulation of intracellular lipids, at least at the high concentrations. EFV did not affect lipid storage.
NRTI combinations commonly used in antiviral therapy were tested, i.e., ddC–d4T, 3TC–ZDV, and ddI–d4T.
When ddC (57 nM) was combined with d4T (4.2 μM, Fig. 1a), the toxicity was enhanced with respect to cell number (P < 0.01), lactate production (P = 0.03) and mtDNA decline (P < 0.01), compared with either NRTI alone. Although not statistically tested, a clear effect on COX II expression was seen only in the combination and not with exposure to either ddC or d4T alone. When the concentration of ddC was tripled to 177 nM, the strongest toxicity was still seen in the combination; however, the differences to ddC single-treatment were less prominent than with the lower concentration of ddC (59 nM). This may suggest saturation of the toxic mechanism with higher concentrations.
Increased toxicity in the combination of 3TC (8.3 μM) with ZDV (7.1 μM) was observed with regard to cell proliferation (P = 0.04), as virtually no cell survived beyond day 15 with double-NRTI exposure (Fig. 1b). In contrast cell growth was not affected by 3TC (8.3 μM) alone and only about halved at day 25 by ZDV (7.1 μM) alone. The ZDV–3TC combination differed also from the single-drug treatments in its more rapid increase in lactate (P = 0.04). Like ZDV and 3TC alone, there was no discernable change in either mtDNA or COX II levels.
In contrast with the ddC–d4T and 3TC–ZDV combinations, the toxicity of the ddI–d4T combination in terms of all of the parameters measured did not differ significantly from the effects of ddI alone.
We tested the long-term mitochondrial toxicity of NRTI with respect to mtDNA, mtDNA-encoded respiratory chain protein and cell function (lactate production, intracellular lipids and cell proliferation) and confirm previous findings that NRTI are able to mediate a rapid decline of mtDNA [7,12,13]. Our findings suggest that the NRTI can be divided into three categories based on their ability to affect mtDNA and cell proliferation of HepG2 cells.
The members of the first category are ddC, ddI and d4T, which deplete mtDNA and also have effects on cell growth, lipids, lactate and COX II levels. Within this category, the effects were time- and dose- dependent and can be graded in the order ddC > ddI > d4T. This hierarchy corresponds to the potency of the NRTI in inhibiting polymerase-γ in vitro . In our experiments the mtDNA content declined by approximately fourfold during the first 5 days of HepG2 cell exposure to ddC at 177 nM, whilst the cell number (and nDNA) had increased by a factor of 3.8. This suggests that cell division occurred almost without replication of mtDNA. Such an almost complete inhibition of polymerase-γ activity by ddC in nanomolar concentrations has been observed by other investigators also .
NRTI are prodrugs that need to be phosphorylated in the cytosol and in mitochondria prior to their interaction with DNA polymerases [14,15]. The fact that there is mtDNA depletion with ddC, ddI and d4T, strongly suggests that these NRTI have been activated and are present in the triphosphorylated form in the mitochondria. The effects on mtDNA appeared earlier and were more pronounced than the effects on cell function. This suggests that mtDNA-depletion by polymerase-γ inhibition plays an important role in mediating the toxicity of ddC, ddI and d4T, and that for these NRTI an assay for mtDNA content may be a more sensitive and earlier indicator of drug toxicity than measurement of cell viability . Polymerase-γ exists in only one isoform, is constitutively expressed and varies little among tissues . Therefore, and because the isolated inhibition of polymerase-γ is not predictive for mitochondrial toxicity , other nuclear factors are likely to modulate the toxic effects of NRTI on mitochondria. Different cells may differentially depend on oxidative phosphorylation  or may differentially express isoforms of the mitochondrial phosphorylation machinery . The importance of nuclear factors is also highlighted by the observation that in different cells (CEM cells) the hierarchy of mitochondrial toxicity was altered, e.g. d4T was more potent in reducing cell viability and mtDNA content, than ddI .
The second category is that of 3TC, which did not exert any substantial toxicity on mtDNA, COX II and all the other functional parameters. Using other systems, 3TC was found to be an only weak inhibitor of gamma-polymerase and to exert little or no mitochondrial toxicity in clinically relevant concentrations [14,18].
ZDV is the sole member of the third category, which is characterized by a reduction of cell growth, despite normal mtDNA and COX II levels. An increase in lactate was observed, and intracellular lipids were also only moderately increased. We observed enlarged mitochondria of at times gigantic size and a disrupted mitochondrial reticulum when we stained the mitochondria of cells exposed to ZDV (71 μM) with Mitotracker (Molecular Probes) . We did not formally test whether the mitochondrial cytotoxicity might represent effects of DMSO which was used as a solvent for ZDV. Although DMSO may have a variety of effects on HepG2 cells , we discard the hypothesis of DMSO cytotoxicity, because the final DMSO concentration was three to four orders of magnitude lower than in previous publications and because cytotoxicity even at concentrations in the percentage range was never observed. Finally, we believe that even if DMSO cytotoxicity was present, it would be hard to explain the delayed onset of toxicity in our observations. Despite the lack of effects on mtDNA, ZDV is a known mitochondrial toxin and numerous effects on mitochondria have been described [23–26].
An early effect on oxidative phosphorylation independent of mtDNA depletion was observed with ZDV at micromolar concentrations in murine erythroleukemic cells . In another hematopoietic cell line, ZDV has been suggested to be incorporated into nDNA and to inhibit gene expression by interfering with transcription . ZDV accumulates in the mitochondrial matrix  where it may bind to adenylate kinase , an enzyme involved in ATP formation and is in addition an inhibitor of the mitochondrial ADP–ATP translocator [25,26] and of nucleoside diphosphate kinase . Lastly, ZDV-treated mice were found to have increased levels of oxidatively damaged DNA residues, such as 7,8-dihydro-8-oxo-2′-deoxyguanosine in liver [29,30].
At first glance, our findings contrast with those of Pan-Zhou et al. , who also compared the effects of the five NRTI on HepG2 cells, but found a marked increase in lactic acid only with ZDV. Furthermore d4T did not affect the synthesis of mtDNA-encoded polypeptides. This apparent discrepancy may be reconciled by taking into account the observation period, since Pan-Zhou et al. assessed lactic acid production after 4 days, and mtDNA-encoded protein after 6 days of NRTI treatment.
We observed increased toxicity in some NRTI combinations, namely ddC–d4T and ZDV–3TC, whereas the d4T–ddI combination did not result in increased toxicity. Interestingly where increased toxicity was observed, both combination partners were pyrimidine nucleoside analogues. Whether this finding may represent a general rule and can be extended to other pyrimidine–pyrimidine and purine–pyrimidine analogue combinations remains to be established. The mechanism for the synergistic action of NRTI on mitochondrial damage at present is not known. The fact, that an increase of mitochondrial toxicity with the ddI–d4T combination was not observed may indicate that synergistic inhibition of polymerase-γ may surprisingly not play a prominent role in the combined NRTI toxicity, although the two compounds act synergistically at HIV-reverse transcriptase . Alternatively, antagonistic effects of the ddI–d4T combination on other mechanisms (see below) may interfere with synergy on polymerase-γ. We also cannot exclude that at lower concentrations of both NRTI partners, a synergistic effect would have been unmasked.
The synthesis of mtDNA requires the uptake of deoxynucleotides into the mitochondrial matrix, a task that can be achieved by the mitochondrial deoxynucleotide carrier . Based on the inhibition characteristics of this carrier, it has been suggested that antiviral dideoxynucleosides may inhibit the uptake of the natural substrates, such as dATP. By interacting with this transporter, the antiviral nucleoside analogues, might lower the amount of available natural nucleosides to compete for polymerase-γ, thereby increasing each other's toxicity. Other sites of interaction could be drug efflux transporters, such as MRP4 , the mitochondrial ATP–ADP exchange , or the above-mentioned phosphorylation pathways that normally activate the NRTI to their respective triphosphate derivatives .
The fact that some NRTI at clinically relevant concentrations were able to decrease mtDNA beyond 25 days of exposure, suggests that studies on mitochondrial toxicity must be long-term investigations. Our finding of increased toxicity in NRTI combinations should prompt further pre-clinical research, but should also be considered in the analysis of clinical studies on mitochondrial toxicity.
1.Brinkman K, ter Hofstede HJ, Burger DM, Smeitink JA, Koopmans PP. Adverse effects of reverse transcriptase inhibitors: mitochondrial toxicity as common pathway.AIDS
2.Brivet FG, Nion I, Megarbane B, Slama A, Brivet M, Rustin P et al.Fatal lactic acidosis and liver steatosis associated with didanosine and stavudine treatment: a respiratory chain dysfunction?J Hepatol
3.Morris AA, Taanman JW, Blake J, Cooper JM, Lake BD, Malone M, et al.Liver failure associated with mitochondrial DNA depletion.J Hepatol
4.Chariot P, Drogou I, de Lacroix-Szmania I, Eliezer-Vanerot MC, Chazaud B, Lombes A, et al.Zidovudine-induced mitochondrial disorder with massive liver steatosis, myopathy, lactic acidosis, and mitochondrial DNA depletion.J Hepatol
5.John M, Moore CB, James IR, Nolan D, Upton RP, McKinnon EJ, et al.Chronic hyperlactatemia in HIV-infected patients taking antiretroviral therapy.AIDS
6.Gerard Y, Maulin L, Yazdanpanah Y, Tribonniere X, Amiel C, Maurage CA, et al.Symptomatic hyperlactataemia: an emerging complication of antiretroviral therapy.AIDS
7.Pan-Zhou XR, Cui L, Zhou XJ, Sommadossi JP, Darley-Usmar VM. Differential effects of antiretroviral nucleoside analogs on mitochondrial function in HepG2 cells.Antimicrob Agents Chemother
8.Youssef JA, Badr MZ. Disruption of mitochondrial energetics and DNA synthesis by the anti- AIDS drug dideoxyinosine.Toxicol Lett
9.Chariot P, Bonne G, Authier FJ, Marsac C, Gherardi R. Expression of cytochrome c oxidase subunits encoded by mitochondrial or nuclear DNA in the muscle of patients with zidovudine myopathy.J Neurol Sci
10.Walker UA, Bickel M, Lütke Volksbeck SI, Ketelsen UP, Schofer H, Setzer B et al.Evidence of nucleoside analogue reverse transcriptase inhibitor–associated genetic and structural defects of mitochondria in adipose tissue of HIV-infected patients
. J Acquir Immune Defic Syndr
11.Lucas GN, Amarasingha A. Sinus histiocytosis with massive lymphadenopathy [letter].Ceylon Med J
12.Chen CH, Cheng YC. Delayed cytotoxicity and selective loss of mitochondrial DNA in cells treated with the anti-human immunodeficiency virus compound 2′,3′-dideoxycytidine.J Biol Chem
13.Broder S. Pharmacodynamics of 2′,3′-dideoxycytidine: an inhibitor of human immunodeficiency virus.Am J Med
14.Kakuda TN. Pharmacology of nucleoside and nucleotide reverse transcriptase inhibitor-induced mitochondrial toxicity.Clin Ther
15.Dolce V, Fiermonte G, Runswick MJ, Palmieri F, Walker JE. The human mitochondrial deoxynucleotide carrier and its role in the toxicity of nucleoside antivirals.Proc Natl Acad Sci USA
16.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
17.Schultz RA, Swoap SJ, McDaniel LD, Zhang B, Koon EC, Garry DJ, et al.Differential expression of mitochondrial DNA replication factors in mammalian tissues.J Biol Chem
18.Martin JL, Brown CE, Matthews-Davis N, Reardon JE. Effects of antiviral nucleoside analogs on human DNA polymerases and mitochondrial DNA synthesis.Antimicrob Agents Chemother
19.Chen CH, Vazquez-Padua M, Cheng YC. Effect of anti-human immunodeficiency virus nucleoside analogs on mitochondrial DNA and its implication for delayed toxicity.Mol Pharmacol
20.Huizing M, Ruitenbeek W, Van den Heuvel LP, Dolce V, Iacobazzi V, Smeitink JA, et al.Human mitochondrial transmembrane metabolite carriers: tissue distribution and its implication for mitochondrial disorders.J Bioenerg Biomembr
21.Gilkerson RW, Margineantu DH, Capaldi RA, Selker JM. Mitochondrial DNA depletion causes morphological changes in the mitochondrial reticulum of cultured human cells.FEBS Lett
22.Paran N, Geiger B, Shaul Y. HBV infection of cell culture: evidence for multivalent and cooperative attachment.EMBO J
23.Hobbs GA, Keilbaugh SA, Rief PM, Simpson MV. Cellular targets of 3′-azido-3′-deoxythymidine: an early (non-delayed) effect on oxidative phosphorylation.Biochem Pharmacol
24.Barile M, Valenti D, Hobbs GA, Abruzzese MF, Keilbaugh SA, Passarella S, et al.Mechanisms of toxicity of 3′-azido-3′- deoxythymidine. Its interaction with adenylate kinase.Biochem Pharmacol
25.Valenti D, Barile M, Passarella S. AZT inhibition of the ADP/ATP antiport in isolated rat heart mitochondria.Int J Mol Med
26.Barile M, Valenti D, Passarella S, Quagliariello E. 3′-Azido-3′-deoxythmidine uptake into isolated rat liver mitochondria and impairment of ADP/ATP translocator.Biochem Pharmacol
27.Weidner DA, Sommadossi JP. 3′-Azido-3′-deoxythymidine inhibits globin gene transcription in butyric acid-induced K-562 human leukemia cells.Mol Pharmacol
28.Valenti D, Barile M, Quagliariello E, Passarella S. Inhibition of nucleoside diphosphate kinase in rat liver mitochondria by added 3'-azido-3'-deoxythymidine.FEBS Letters
29.de la Asuncion JG, del Olmo ML, Sastre J, Pallardo FV, Vina J. Zidovudine (AZT) causes an oxidation of mitochondrial DNA in mouse liver.Hepatology
30.Hayakawa M, Ogawa T, Sugiyama S, Tanaka M, Ozawa T. Massive conversion of guanosine to 8-hydroxy-guanosine in mouse liver mitochondrial DNA by administration of azidothymidine.Biochem Biophys Res Commun
31.Maga G, Hubscher U, Pregnolato M, Ubiali D, Gosselin G, Spadari S. Potentiation of inhibition of wild-type and mutant human immunodeficiency virus type 1 reverse transcriptases by combinations of nonnucleoside inhibitors and d- and L-(beta)-dideoxynucleoside triphosphate analogs.Antimicrob Agents Chemother
32.Schuetz JD, Connelly MC, Sun D, Paibir SG, Flynn PM, Srinivas RV, et al.MRP4: A previously unidentified factor in resistance to nucleoside-based antiviral drugs.Nature Med