- Email: [email protected]
Genotoxicity and mitochondrial damage in human lymphocytic cells chronically exposed to 3′-azido-2′,3′-dideoxythymidine
Mutation Research 390 Ž1997. 223–231
Genotoxicity and mitochondrial damage in human lymphocytic cells chronically exposed to 3X-azido-2X ,3X-dideoxythymidine Ram P. Agarwal a , Ofelia A. Olivero b
b,)
a Medical Oncology, UniÕersity of Miami School of Medicine, Miami, FL 33136, USA Laboratory of Cellular Carcinogenesis and Tumor Promotion, National Cancer Institute, NIH, Bethesda, MD 20892, USA
Received 9 May 1996; revised 6 December 1996; accepted 10 December 1996
Abstract AZT Ž3X-azido-2X ,3X-dideoxythymidine., the first nucleoside analog approved for the treatment of AIDS Žacquired immunodeficiency syndrome., induces significant toxic effects in humans exposed to therapeutic doses. As an inhibitor of the HIV-1 Žhuman immunodeficiency virus 1. reverse transcriptase, AZT blocks the incorporation of nucleotides into the host’s newly synthesized DNA. Incorporation of AZT into mammalian DNA as well as specific localization of the drug into telomeric DNA, has been previously documented by immunohistochemistry. As with other nucleoside analogs, AZT has affinity for polymerase-g, the enzyme responsible for the replication of mitochondrial DNA. In order to examine the mechanisms of toxic events induced by long-term AZT exposure, human T-lymphocytic H 9 cells were cultured with 25 mM AZT for 7 months. In the resulting H 9-AZT cells, incorporation of AZT into DNA was demonstrated by radioimmunoassay and immunohistochemistry, chromosomal aberrations and micronuclei were scored and intracellular lipid distribution was determined. Two pmol of AZT per microgram of DNA were detected by radioimmunoassay in H 9-AZT cells. Control cells showed negative values in the radioimmunoassay. Cytogenetic observations on H 9-AZT cells showed an increase in chromosomal aberrations and nuclear fragmentation when compared with unexposed H 9 cells. Electron microscopy revealed mitochondrial damage and an elevated accumulation of neutral intracellular lipid deposits probably as a consequence of a distortion in the b-oxidation of fatty acids normally carried out by this organelle. The toxicities explored here suggest that the mechanisms of AZT induced cytotoxicity in bone marrow of the patients chronically exposed to the drug in vivo may involve both chromosomal and mitochondrial DNA damage. Keywords: Mitochondrium; Mitochondrial DNA; Nucleoside analog
)
Corresponding author. National Cancer Institute, NIH, Laboratory of Cellular Carcinogenesis and Tumor Promotion, Bldg 37, Rm 3B 12, 37 Convent Drive MSC 4255, Bethesda, MD 208924255, USA. Tel.: q1 Ž301. 496-8898; Fax: q1 Ž301. 496-8709; E-mail: [email protected]
1. Introduction In 1987 the thymidine analog 3X-azido-2X ,3X-dideoxythymidine ŽAZT. was approved by the US Food and Drug Administration as the first drug to be used in the therapy of AIDS. Since then, a number of
1383-5718r97r$17.00 Copyright q 1997 Elsevier Science B.V. All rights reserved. PII S 1 3 8 3 - 5 7 1 8 Ž 9 7 . 0 0 0 1 4 - 1
224
R.P. Agarwal, O. A. OliÕeror Mutation Research 390 (1997) 223–231
toxic effects have been described as side effects associated with the use of the drug w1x. Inhibition of reverse transcriptase and DNA chain termination are the two mechanisms implicated in the therapeutic action of the analog w2,3x. Several studies have shown the ability of AZT to incorporate into mammalian DNA in vitro, w4–7x. Also, preferential incorporation of AZT into telomeric DNA of CHO cells has been described w8,9x. Chronic exposure of a human lymphocytic cell line to AZT has previously been shown to increase thymidine kinase activity and AZTmonophosphate efflux w10,11x as well as to decrease thymidine metabolism w12x. There is some evidence that nucleoside analogs have some specific effects in mitochondria. For example, 2X ,3X-dideoxycytidine ŽddC., 2X ,3X-dideoxyinosine ŽddI., and 2X ,3X-dideoxy2X ,3X-didehydrothymidine Žd4T. have been shown to induce mitochondrial impairment by reducing the amount of mtDNA w13x. The production of lactic acid as a result of the inhibition of mitochondrial oxidative phosphorylation has been reported with ddC, 2X ,3X-dideoxyadenosine ŽddA., ddI, 2X ,3X-dideoxyguanosine ŽddG., and their 2X-b-fluoro analogs, and this has been linked to mitochondrial toxicity w14x. In contrast, AZT does not inhibit mtDNA synthesis in human lymphoid CEM cells w15x or neuronal PC12 cells w16x, possibly because AZT triphosphate is a much weaker inhibitor of polymerase-g than other dideoxynucleotide triphosphates. AZT triphosphate, for example, has been shown to be approximately 500 times less potent than ddC triphosphate in inducing mitochondrial toxicity w17x. Nonetheless, mitochondrial myopathies, including ragged red fibers and paracrystalline inclusions in mitochondria, have been observed during prolonged AZT treatment of individuals infected with HIV-1 w18,19x. In an attempt to explore the long-term effects of AZT exposure, genotoxicity and mitochondrial damage were evaluated in a population of H 9 human lymphocytic cells chronically exposed to AZT. 2. Materials and methods 2.1. Cell culture and DNA isolation H 9 cells were cultured in RPMI ŽGibco, Grand Island, NY. supplemented with 10% fetal bovine
serum ŽFBS; Gibco.. H 9-AZT cells chronically grown in medium containing 25 mM AZT for 7 months were described earlier w10x. 2.2. Incorporation of AZT into genomic DNA of H9 - A ZT cells Triplicate pellets of H 9 and H 9-AZT cells were processed to obtain nuclear DNA by CsCl density gradient centrifugation w20x. After dialysis at 48C to remove the CsCl, the DNA was quantitated by spectrophotometry at A 260 nm and AZT incorporation was determined by radioimmunoassay ŽRIA. in order to determine the amount of AZT incorporated into genomic DNA of H 9-AZT cells. The RIA specific for AZT was performed as already described using an anti-AZT antibody ŽSigma, St. Louis, MO. diluted 1:5000 and w 3 HxAZT Ž20 Cirmmol, Moravek Biochemicals Inc., Brea, CA. as tracer and unlabeled AZT for the standard curve w7x. DNA from H 9-AZT and H 9 cells was adjusted to a concentration of 30 mgrml, and 0.1 ml was assayed by RIA. Either AZT standards, unknown H 9-AZT-DNA samples or H 9DNA with no AZT incorporation were incubated with the antibody for 90 min at 378C. The w 3 HxAZT tracer and goat anti-rabbit IgG ŽICN ImmunoBiologicals, Costa Mesa, CA. were then added and the mixture incubated for 25 min. Centrifugation of the sample for 15 min at 1500 = g separated bound from unbound tracer. The unbound tracer Žsupernatant. was decanted. The remaining pellets were dissolved in 200 ml of 100 mM NaOH and counted in a liquid scintillation counter. The standard curve 50% inhibition in the RIA was at 1.78 " 0.36 pmol AZT Žmean " SD, n s 60. and the lower limit of assay sensitivity when the biological samples were assayed was approximately 0.33 " 0.05 pmol AZT Žmean " SD, n s 20.. 2.3. Analysis of chromosomal aberrations One hour before harvesting, 1 mg of colchicine per ml of medium was added to the cultures. The cells were processed to prepare chromosome spreads by trypsinization, hypotonic treatment with 75 mM KCl, fixation with Carnoy’s solution, and air drying w21x. Chromosomal aberrations Ždicentrics, breaks, micronuclei, fragments, gaps and rings. were scored in 50 metaphases from H 9 and 50 H 9-AZT cells stained with Giemsa stain.
R.P. Agarwal, O. A. OliÕeror Mutation Research 390 (1997) 223–231
2.4. Immunohistochemical staining for AZT incorporation into DNA Anti-AZT antibody was reconstituted in 1 ml of distilled water, dialyzed against distilled water overnight at 48C, and preabsorbed as previously described w7x. Cell suspensions were centrifuged onto slides in a Shandon cytospin centrifuge ŽShandon Southern Instruments, Inc. Sewickley, PA. for 5 min at 400 rpm. Prior to the addition of antibody the cells were incubated for 30 min at 378C with a 1:10 dilution of normal goat serum. The absorbed antiAZT antibody was then diluted with 100 mg of thymidine per mg antibody and added to the slides for a 3-hour incubation at 378C. A secondary antirabbit-rhodamine conjugated IgG antibody diluted 1:50 was applied at room temperature for an additional 30 min. Hoechst 33258 Ž2X-w4-hydroxyphenylx5-w4-methyl-1-piperazinylx-2,5X-bi-1 H-benzimidazole; Sigma. was used at a 1:60 000 dilution to stain DNA. The slides were air dried and mounted with Pro-Texx mounting medium ŽAmerican Scientific Products, McGaw Park, IL.. Slides with H 9-AZT-treated cells were digested with DNase for 1 h prior to the incubation with anti-AZT antibody as previously described w8x. The absence of signal confirmed the specificity of the DNA staining.
225
again in 1% osmium tetroxide for 1 h, washed and dehydrated in graded dilutions of ethanol. The dehydrated cells were embedded in Spurr’s resin, sectioned by Porter-Blum MT-2 ultramicrotome and viewed and photographed in a Jeol-CX100 electron microscope at =24 000. In order to localize lipid droplets H 9-AZT cells were collected and fixed for 60 min in glutaraldehyde in 0.1 M cacodylate buffer, pH 7.3. After a postfixation treatment with 1% phosphate-buffered osmium tetroxide at 48C, and dehydration in alcohol, the cells were embedded in Spurr epoxy resin at 608C. Ultra-thin sections were cut and mounted on 200 mesh copper grids. Preparations stained with uranyl acetate and lead citrate were viewed and photographed as described above. 2.7. Lactic acid determination H 9 , H 9-AZT and H 9 acutely treated with 25 mM AZT Žlabeled H 9-AZTaq. were incubated at 378C. The cell number was determined and the culture medium was collected by centrifugation at 24, 48 and 72 h of incubation. Lactic acid concentration in the medium was determined using a lactic acid kit ŽSigma.. 3. Results
2.5. Staining for intracellular lipids
3.1. AZT-DNA incorporation
The red-oil-O technique, specific to visualize neutral intracellular lipids, was applied to standard cytospin preparations. Briefly, the preparations were fixed in 10% formalin for 1 h, washed in tap water for 20 min and stained with a solution of red-oil-O in propylene glycol for 1 h. After an additional washing, the cells were counterstained with hematoxylin for 5 min w22x. The stain is frequently used to determine the amount of lipid droplets accumulated in the cytoplasm. Furthermore, lipid accumulation was confirmed by electron microscopy.
Three micrograms of DNA isolated from three different populations of H 9-AZT cells were analyzed
2.6. Electron microscopy The cells were harvested by centrifugation, washed 3 times with PBS and fixed overnight in glutaraldehyde Ž2.5%. plus formalin Ž2%. in isotonic phosphate buffer, pH 7.3. The fixative was rinsed out with phosphate buffer and the cells were fixed
Fig. 1. Bar graph indicating comparison of chromosomal aberrations in 50 metaphases of H 9 and H 9-AZT cells. Significance is indicated by asterisks.
226
R.P. Agarwal, O. A. OliÕeror Mutation Research 390 (1997) 223–231
Table 1 Summary of genomic and mitochondrial alterations in H 9-AZT cells Test
Endpoint
H 9-AZT
H9
Figure
RIA
AZT-DNA incorporation
2.0 pmol AZTrmg DNA
0.0 pmol AZTrmg DNA
NS
Chromosomal aberrations
gaps rings dicentrics micronuclei breaks fragments
1 4 2 10 11 8
1 1 0 1 3 3
Fig. 1 Fig. 1 Fig. 1 Figs. 1 and 2 Fig. 1 Fig. 1
total
36
9
Fig. 1
Mitochondrial metabolism
intracellular lipids ŽLM. intracellular lipids ŽEM.
17.0 " 1.4 abundant
1.8 " 0.7 sparse
Fig. 3bFig. 5 Fig. 3c
Mitochondrial appearance
cristae size morphology
lost swollen abnormal
present normal normal
Fig. 4a Fig. 4b Fig. 4c
a
a
Average lipid droplets" SEM Ž n s 15..
Fig. 2. Nuclei and micronuclei from H 9-AZT cells. a: Giemsa stain. b: Hoechst stain for DNA. c,d: anti-AZT antibody, and secondary anti-rabbit rhodamine-conjugated antibody.
R.P. Agarwal, O. A. OliÕeror Mutation Research 390 (1997) 223–231
by RIA in three separate assays. The results showed an incorporation of AZT equivalent to 2 pmol AZTrmg of DNA. DNA from H 9 cells was used as a control and was assayed in parallel in the same RIA. No incorporation of AZT in DNA from H 9 cells in three independent assays was detected. 3.2. Cytogenetic analysis Fifty metaphases of both H 9 and H 9-AZT cells were analyzed to score chromosomal aberrations ŽFig. 1, Table 1.. Dicentrics, breaks, micronuclei, fragments, gaps and rings were scored. The total of 36 aberrations for H 9-AZT cells, was 4-fold higher than in the control H 9 cells Ž p - 0.05.. Two of the most dramatically increased aberrations in H 9-AZT cells were breaks Ž3 in H 9 cells to 9 in H 9-AZT cells, Fig. 1, Table 1. and fragments Ž3 in H 9 cells to 8 in H 9-AZT cells, Fig. 1, Table 1.. These elevated frequencies could be explained as a consequence of
227
AZT being a chain terminator. The occurrence of rings, on the other hand, could be seen as a result of a disturbance in the telomeres, already described as a preferential site for AZT incorporation w8,9x, and the consequent adhesiveness of the chromosomal ends. Because micronuclei were the most evident increased aberration Žfrom 2 in H 9 cells to 10 in H 9-AZT cells, Fig. 1, Table 1., further studies with anti-AZT antibodies and immunostaining were performed to determine the presence of AZT in the micronuclei. 3.3. Immunostaining Micronuclei, seen as small round fragments close to interphase nuclei, were analyzed in interphase by specific immunostaining to detect AZT-DNA incorporation. All of the 50 micronuclei scored were positive for AZT staining. Fig. 2a–d shows micronuclei of H 9-AZT cells stained by Giemsa ŽFig. 2a.,
Fig. 3. Cytospin preparations of H 9 Ža. and H 9-AZT Žb. cells stained by the red-oil-O technique to visualize lipid droplets Žb, arrowheads.. c: electron microscopy of H 9-AZT cell, arrows indicate lipid droplets. =24 000.
228
R.P. Agarwal, O. A. OliÕeror Mutation Research 390 (1997) 223–231
Hoechst ŽFig. 2b. and anti-AZT antibody and antirabbit rhodamine-conjugated antibody ŽFig. 2c,d.. Red stain Žpositive for AZT. in Fig. 2c,d is the result of recognition of AZT by the anti-AZT antibody, and the secondary rhodamine-conjugated antibody. 3.4. Staining for intracellular lipids Neutral intracellular lipids were identified as an indirect measurement of mitochondrial function w23x. Abnormal structure of the mitochondria has been linked with impaired organelle function mainly due to reduced amount of DNA and cytochromes. As a consequence, the long-chain fatty acids are not effectively utilized by the mitochondria and accumulate in the cytoplasm where they are identified as lipid droplets by the red-O-oil technique w23x. Pictures of cytospin preparations of H 9 and H 9-AZT cells stained by red-oil-O technique were compared. An evident
increase in the amount of lipid droplets was noticed in the H 9-AZT cell preparations ŽFig. 3b.. Lipid droplets were counted in 15 individual cells from H 9 and H 9-AZT lines. The average of that score was of 1.8 dropletrcell for H 9 and of 17.0 dropletrcell for H 9-AZT cells. ŽFig. 5, Table 1.. A further observation of cells through electron microscopy confirmed the lipid droplet increase in H 9-AZT cells ŽFig. 3c.. 3.5. Electron microscopy Electron microscopic examination of H 9 and H 9-AZT cells showed evidence of significant mitochondrial defects in cells chronically exposed to AZT. Fig. 4 illustrates some of the mitochondrial defects detected in H 9-AZT cells. Loss of cristae ŽFig. 4a., swollen mitochondria ŽFig. 4b. and abnormal mitochondria ŽFig. 4c. were documented.
Fig. 4. Mitochondrial appearance visualized by electron microscopy in H 9-AZT cells. a: loss of internal cristae. b: swollen mitochondria. c: abnormal mitochondria. d: control H 9 cells. =24 000.
R.P. Agarwal, O. A. OliÕeror Mutation Research 390 (1997) 223–231 Table 2 Effect of chronic and acute AZT treatment of H 9 cells on lactic acid production 24 h
48 h
72 h
H 9 Ž ns 3. 72.7"28.1 b 123.9"33.2 114.1" 9.8 H 9-AZT Ž ns 3. 69.7"19.2 140.8"38.3 145.9"21.8 H 9-AZTaq Ž ns 3. a 111.5"49.8 190.6"28.2 167.8"60.7 a b
H 9-AZTaq cells treated acutely with 25 mM AZT. mg lactaterdl=10 6 cells.
3.6. Lactic acid determination The lactic acid content expressed as mgrdl per 10 6 cells, in the medium of AZT-acutely treated cells ŽH 9AZTaq . are presented in Table 2. The values in H 9-AZT Žpreviously chronically exposed to AZT. did not differ significantly from those of control cells ŽH 9 .. However, the cells that were treated acutely with AZT at 25 mM expressed 47–53% more lactic acid than the control cells.
4. Discussion Mitochondrial and chromosomal damage has been studied in a population of a human T-cell line derivative, H 9 , chronically exposed to AZT ŽH 9-AZT cells.. This study clearly demonstrates incorporation of AZT into genomic DNA, chromosomal aberrations, nuclear fragmentation, mitochondrial damage and altered lipid metabolism in the chronically AZT exposed cells. AZT has several undesired side effects in humans w1x. Some of them, like bone marrow aplasia, could be caused by incorporation of AZT into human DNA with a disturbance in both genomic and mitochondrial DNA replication. Previous studies reported incorporation of AZT into mammalian DNA of different species w4,5x in spite of the low affinity of AZT-5X-triphosphate for mammalian DNA polymerases w4,24–26x. Furthermore, preferential localization of AZT into telomeric DNA of Chinese hamster ovary ŽCHO. cells was documented by immunofluorescence with anti-AZT antibodies w8x. The same study also showed a disturbance in the normal development of telomeres probed indirectly by an increase in the number of chromosome bridges scored
229
in anaphase–telophase of CHO cells w8x. In that report, the authors speculated about the possibility that AZT might inhibit telomerase, the enzyme responsible for the elongation of the telomeres w21,27,28x. Chromosomal aberrations were previously studied in vivo in bone marrow of mice exposed to AZT for 28 days, revealing a 4-fold increase as compared with untreated mice w7x. Here, we report for the first time an integral study of both genomic and mitochondrial AZT-induced toxicities in the same cellular system under chronic exposure. Chromosomal aberrations – mainly fragments and breaks – increased 4-fold in H 9-AZT as compared with untreated H 9 cells ŽFig. 1, Table 1., probably as a result of incorporation of AZT into genomic DNA and subsequent chain termination. Rings, another chromosomal aberration that was increased in AZT chronically exposed cells, could be the result of chromosome stickiness by incorporation of AZT into the telomeres w8,9x. Nuclear fragmentation, probably as a result of AZT incorporation, was one of the aberrations that increased most dramatically ŽFig. 1, Table 1.. This is in agreement with earlier reports that demonstrated the ability of AZT, among other nucleoside analogs, to induce micronuclei in different cellular systems w29,30x. Immunohistochemical analysis of the micronuclei with anti-AZT antibodies confirmed that those nuclear fragments were indeed positive for AZT incorporated into the DNA ŽFig. 2.. A DNase pretreatment of exposed cells prior to anti-AZT incubation revealed no staining, confirming that the AZT signal was specific for AZT-DNA. An important body of information exists about degenerative effects induced by AZT into mitochondria in human myotubes w18,31,32x as well as destructive changes in the myofibrils, accumulation of lipid droplets, and enzymatic defects in the respiratory chain system w23,33x; however, there is no evidence of chromosomal aberrations or preferential binding studies in muscle cells. In addition to the genotoxic effects induced by AZT in H 9-AZT chronically exposed cells, mitochondrial damage was evaluated here by electron microscopy ŽFig. 4.. Degenerative lesions such as loss of cristae, swollen mitochondria and abnormal mitochondria ŽFig. 4a–c., lesions already described in myotubes of AZT induced myopathies w18,32x, have been observed in the cellular system studied here. We also report an in-
230
R.P. Agarwal, O. A. OliÕeror Mutation Research 390 (1997) 223–231
fect of AZT in human populations exposed to therapeutic doses of the drug.
Acknowledgements The authors are very thankful to Ms. Jun He for her technical assistance. This work was partially supported by the Department of Health and Human Services Grant NIAID AI 29155. We would like to thank Margaret Taylor for editorial assistance.
Fig. 5. Lipid droplets accumulated in the cytoplasm of H 9 and H 9-AZT cells. Each dot represents the amount of lipid droplets scored in one cell. A total of 15 cells were scored and the average of the observations is shown. A statistical significance is shown by an asterisk.
crease in cytoplasmic accumulation of lipid droplets ŽFig. 3bFig. 5, Table 1., indirect evidence of the impaired ability of the mitochondria to utilize longchain fatty acids w23x. As an adaptive cellular response, such a mitochondrial deficiency is expected to induce stimulation of glycolysis resulting in increased lactic acid secretion. Failure to observe any increase in lactate in H 9-AZT cells, compared to acutely treated cells, suggests that during chronic treatment, despite impaired mitochondria, the cells had adapted to grow under these conditions. Since the use of nucleoside analogs is widespread in AIDS and hepatitis B therapy, a careful study of their potential clastogenic ability should be taken into account as already suggested w34x. It is not clear whether nucleoside analog-induced toxicity is a consequence of a decrease in mtDNA w13,15,35x or of mutations impairing normal mitochondrial function. Nonetheless, mtDNA incorporation of nucleoside analog drugs remains a sensitive measure of drug toxicity. A careful investigation of human liver toxicity reported in patients receiving AZT might indicate whether or not chronic AZT exposure could induce liver failure by mechanisms similar to those seen with the nucleoside analog 5-FIAU Ž2X-fluoro-2-deoxyarabinofuranosyl-5-iodocytidine. w36x. Studies of chronic exposure in different systems are needed to address the potential carcinogenic ef-
References w1x Yarchoan, R. and S. Broder Ž1987. Preliminary results on the use of dideoxynucleosides in the therapy of AIDS, in: R.M. Chanock, R.A. Lerner, F. Brown and H.S. Ginsberg ŽEds.., Vaccine 87, Cold Spring Harbor Press, New York, pp. 214–224. w2x P.A. Furman, J.A. Fyfe, M.H. St Clair, K. Weinhold, J.L. Rideout, G.A. Freeman, S.N. Lehrman, D.P. Bolognesi, S. X Broder, H. Mitsuya, D.A. Barry, Phosphorylation of 3 -azidoX X 3 -deoxythymidine and selective interaction of the 5 -triphosphate with human immunodeficiency virus reverse transcriptase, Proc. Natl. Acad. Sci. USA 83 Ž1986. 8333–8337. w3x Y.C. Cheng, G.E. Dutschman, K.F. Bastow, M.G. Sarngadharan, R.Y.C. Ting, Human immunodeficiency virus reverse transcriptase. General properties and its interactions with nucleoside triphosphate analogs, J. Biol. Chem. 262 Ž1987. 2187–2189. w4x M.A. Vazquez-Padua, M.C. Starnes, Y.C. Cheng, IncorporaX X tion of 3 -azido-3 -deoxythymidine into cellular DNA and its removal in a human leukemic cell line, Cancer Commun. 2 Ž1990. 55–62. w5x V.I. Avramis, W. Markson, R.L. Jackson, E. Gomperts, Biochemical pharmacology of zidovudine in human Tlymphoblastoid cells ŽCEM., Aids 3 Ž1989. 417–422. w6x J.P. Sommadossi, R. Carlisle, Z. Zhou, Cellular pharmacolX X ogy of 3 -azido-3 -deoxythymidine with evidence of incorporation into DNA of human bone marrow cells, Mol. Pharmacol. 36 Ž1989. 9–14. w7x O.A. Olivero, F.A. Beland, M.C. Poirier, Immunofluorescent X X X localization and quantitation of 3 -azido-2 ,3 -dideoxythymidine ŽAZT. incorporated into chromosomal DNA of human, hamster and mouse cell lines, Int. J. Oncol. 4 Ž1994. 49–54. w8x O.A. Olivero, M.C. Poirier, Preferential incorporation of X X X 3 -azido-2 ,3 -dideoxythymidine ŽAZT. into telomeric DNA and Z-DNA-containing regions of chinese hamster ovary cells, Mol. Carcinog. 8 Ž1993. 81–88. w9x D.E. Gomez, A. Kassim, O.A. Olivero, Preferential incorpoX X X ration of 3 -Azido-2 ,3 -dideoxythymidine ŽAZT. in telomeric sequences of CHO cells, Int. J. Oncol. 7 Ž1995. 1057–1060. w10x R.P. Agarwal, 3X-Azido-2X ,3X-dideoxythymidine induces
R.P. Agarwal, O. A. OliÕeror Mutation Research 390 (1997) 223–231
w11x
w12x
w13x
w14x
w15x
w16x
w17x
w18x
w19x
w20x
w21x w22x
thymidine kinase activity in vitro, Int. J. Purine Pyrimidine Res. 2 Ž1991. 1–3. R.P. Agarwal, J. Padmanabhan, J. He, Kinetics of accumulaX X X tion and efflux of 3 -azido-3 -deoxythymidine-5 -monophosphate in H 9 and H 9-AZT cells, Cell. Pharmacol. 1 Ž1994. 121–125. R.P. Agarwal, A.M. Mian, Thymidine and zidovudine metabolism in chronically zidovudine-exposed cells in vitro, Biochem. Pharmacol. 42 Ž1991. 905–911. D.J. Medina, C.H. Tsai, G.D. Hsiung, Y.C. Cheng, Comparison of mitochondrial morphology, mitochondrial DNA content, and cell viability in cultured cells treated with three anti-human immunodeficiency virus dideoxynucleosides, Antimicrob. Agents Chemother. 38 Ž1994. 1824–1828. C.H. Tsai, S.L. Doong, D.G. Johns, J.S. Driscoll, Y.C. Cheng, Effect of anti-HIV 2X-beta-fluoro-2X ,3X-dideoxynucleoside analogs on the cellular content of mitochondrial DNA and on lactate production, Biochem. Pharmacol. 48 Ž1994. 1477–1481. C.H. Chen, M. Vazquez-Padua, Y.C. Cheng, Effect of antihuman immunodeficiency virus nucleoside analogs on mitochondrial DNA and its implication for delayed toxicity, Mol. Pharmacol. 39 Ž1991. 625–628. S.A. Keilbaugh, W.H. Prusoff, M.V. Simpson, The PC12 cell as a model for studies of the mechanism of induction of peripheral neuropathy by anti-HIV-1 dideoxynucleoside analogs, Biochem. Pharmacol. 42 Ž1991. 5–8. J.L. Martin, C.E. Brown, N. Matthews-Davis, J.E. Reardon, Effects of antiviral nucleoside analogs on human DNA polymerases and mitochondrial DNA synthesis, Antimicrob. Agents Chemother. 38 Ž1995. 2743–2749. M.C. Dalakas, I. Illa, G.H. Pezeshkpour, J.P. Laukaitis, B. Cohen, J.L. Griffin, Mitochondrial myopathy caused by long-term zidovudine therapy, New Engl. J. Med. 322 Ž1990. 1098–1105. C. Mhiri, M. Baudrimont, G. Bonne, C. Geny, F. Degoul, C. Marsac, E. Roullet, R. Gherardi, Zidovudine myopathy: a distinctive disorder associated with mitochondrial dysfunction, Ann. Neurol. 29 Ž1991. 606–614. Flamm, W.G., M.L. Birnstiel and P.M.B. Walker Ž1967. Preparation and fractionation and isolation of single strands of DNA by ultracentrifugation in fixed-angle rotors, in: G.D. Birnie and S.M. Fox ŽEds.., Subcellular Components: Preparation and Fractionation, Butterworths, London, pp. 126–154. C.W. Greider, Telomeres, telomerase and senescence, Bioessays 12 Ž1990. 363–369. Coolidge, B.J. and R.M. Howard Ž1979. Animal Histology Procedures, U.S. Dept. of Health, Education and Welfare, Bethesda, MD.
231
w23x M.C. Dalakas, M.E. Leon-Monzon, I. Bernardini, W.A. Gahl, C.A. Jay, Zidovudine-induced mitochondrial myopathy is associated with muscle carnitine deficiency and lipid storage, Ann. Neurol. 35 Ž1994. 482–487. w24x E.L. White, W.B. Parker, L.J. Macy, S.C. Shaddix, G. McCaleb, J.A. Secrist, R. Vince, W.M. Shannon, Comparison of the effect of Carbovir, AZT, and dideoxynucleoside triphosphates on the activity of human immunodeficiency virus reverse transcriptase and selected human polymerases, Biochem. Biophys. Res. Commun. 161 Ž1989. 393–398. w25x P. Huang, D. Farquhar, W. Plunkett, Selective action of 3X-azido-3X-deoxythymidine 5X-triphosphate on viral reverse transcriptases and human DNA polymerases, J. Biol. Chem. 265 Ž1990. 11914–11918. w26x W. Nickel, S. Austermann, G. Bialek, F. Grosse, Interactions of azidothymidine triphosphate with the cellular DNA polymerases alpha, delta, and epsilon and with DNA primase, J. Biol. Chem. 267 Ž1992. 848–854. w27x E.H. Blackburn, Telomeres: structure and synthesis, J. Biol. Chem. 265 Ž1990. 5919–5921. w28x D. Shippen-Lentz, E.H. Blackburn, Functional evidence for an RNA template in telomerase, Science 247 Ž1990. 546–552. w29x M.D. Phillips, B. Nascimbeni, R.R. Tice, M.D. Shelby, Induction of micronuclei in mouse bone marrow cells: an evaluation of nucleoside analogues used in the treatment of AIDS, Environ. Mol. Mutagen. 18 Ž1991. 168–183. w30x M. Gonzalez Cid, I. Larripa, Genotoxic activity of azidothymidine ŽAZT. in in vitro systems, Mutation Res. 321 Ž1994. 113–118. w31x G.H. Pezeshkpour, I. Illa, M.C. Dalakas, Ultrastructural characteristics and DNA immunocytochemistry in human immunodeficiency virus and zidovudine-associated myopathies, Hum. Pathol. 22 Ž1991. 1281–1288. w32x B.S. Peters, J. Winer, D.N. Landon, A. Stotter, A.J. Pinching, Mitochondrial myopathy associated with chronic zidovudine therapy in AIDS, Q. J. Med. 86 Ž1993. 5–15. w33x M.C. Semino-Mora, M.E. Leon-Monzon, M.C. Dalakas, Effect of L-carnitine on the zidovudine-induced destruction of human myotubes. Part I: L-carnitine prevents the myotoxicity of AZT in vitro, Lab. Invest. 71 Ž1994. 102–112. w34x M.D. Shelby, The cytogenetic effects of AZT, Environ. Mol. Mutagen. 24 Ž1994. 148–151. w35x C.H. Chen, Y.C. Cheng, Delayed cytotoxicity and selective loss of mitochondrial DNA in cells treated with the anti-huX X man immunodeficiency virus compound 2 ,3 -dideoxycytidine, J. Biol. Chem. 264 Ž1989. 11934–11937. w36x W.B. Parker, Y.C. Cheng, Mitochondrial toxicity of antiviral nucleoside analogs, J. NIH Res. 6 Ž1994. 57–61.
Genotoxicity and mitochondrial damage in human lymphocytic cells chronically exposed to 3X-azido-2X ,3X-dideoxythymidine Ram P. Agarwal a , Ofelia A. Olivero b
b,)
a Medical Oncology, UniÕersity of Miami School of Medicine, Miami, FL 33136, USA Laboratory of Cellular Carcinogenesis and Tumor Promotion, National Cancer Institute, NIH, Bethesda, MD 20892, USA
Received 9 May 1996; revised 6 December 1996; accepted 10 December 1996
Abstract AZT Ž3X-azido-2X ,3X-dideoxythymidine., the first nucleoside analog approved for the treatment of AIDS Žacquired immunodeficiency syndrome., induces significant toxic effects in humans exposed to therapeutic doses. As an inhibitor of the HIV-1 Žhuman immunodeficiency virus 1. reverse transcriptase, AZT blocks the incorporation of nucleotides into the host’s newly synthesized DNA. Incorporation of AZT into mammalian DNA as well as specific localization of the drug into telomeric DNA, has been previously documented by immunohistochemistry. As with other nucleoside analogs, AZT has affinity for polymerase-g, the enzyme responsible for the replication of mitochondrial DNA. In order to examine the mechanisms of toxic events induced by long-term AZT exposure, human T-lymphocytic H 9 cells were cultured with 25 mM AZT for 7 months. In the resulting H 9-AZT cells, incorporation of AZT into DNA was demonstrated by radioimmunoassay and immunohistochemistry, chromosomal aberrations and micronuclei were scored and intracellular lipid distribution was determined. Two pmol of AZT per microgram of DNA were detected by radioimmunoassay in H 9-AZT cells. Control cells showed negative values in the radioimmunoassay. Cytogenetic observations on H 9-AZT cells showed an increase in chromosomal aberrations and nuclear fragmentation when compared with unexposed H 9 cells. Electron microscopy revealed mitochondrial damage and an elevated accumulation of neutral intracellular lipid deposits probably as a consequence of a distortion in the b-oxidation of fatty acids normally carried out by this organelle. The toxicities explored here suggest that the mechanisms of AZT induced cytotoxicity in bone marrow of the patients chronically exposed to the drug in vivo may involve both chromosomal and mitochondrial DNA damage. Keywords: Mitochondrium; Mitochondrial DNA; Nucleoside analog
)
Corresponding author. National Cancer Institute, NIH, Laboratory of Cellular Carcinogenesis and Tumor Promotion, Bldg 37, Rm 3B 12, 37 Convent Drive MSC 4255, Bethesda, MD 208924255, USA. Tel.: q1 Ž301. 496-8898; Fax: q1 Ž301. 496-8709; E-mail: [email protected]
1. Introduction In 1987 the thymidine analog 3X-azido-2X ,3X-dideoxythymidine ŽAZT. was approved by the US Food and Drug Administration as the first drug to be used in the therapy of AIDS. Since then, a number of
1383-5718r97r$17.00 Copyright q 1997 Elsevier Science B.V. All rights reserved. PII S 1 3 8 3 - 5 7 1 8 Ž 9 7 . 0 0 0 1 4 - 1
224
R.P. Agarwal, O. A. OliÕeror Mutation Research 390 (1997) 223–231
toxic effects have been described as side effects associated with the use of the drug w1x. Inhibition of reverse transcriptase and DNA chain termination are the two mechanisms implicated in the therapeutic action of the analog w2,3x. Several studies have shown the ability of AZT to incorporate into mammalian DNA in vitro, w4–7x. Also, preferential incorporation of AZT into telomeric DNA of CHO cells has been described w8,9x. Chronic exposure of a human lymphocytic cell line to AZT has previously been shown to increase thymidine kinase activity and AZTmonophosphate efflux w10,11x as well as to decrease thymidine metabolism w12x. There is some evidence that nucleoside analogs have some specific effects in mitochondria. For example, 2X ,3X-dideoxycytidine ŽddC., 2X ,3X-dideoxyinosine ŽddI., and 2X ,3X-dideoxy2X ,3X-didehydrothymidine Žd4T. have been shown to induce mitochondrial impairment by reducing the amount of mtDNA w13x. The production of lactic acid as a result of the inhibition of mitochondrial oxidative phosphorylation has been reported with ddC, 2X ,3X-dideoxyadenosine ŽddA., ddI, 2X ,3X-dideoxyguanosine ŽddG., and their 2X-b-fluoro analogs, and this has been linked to mitochondrial toxicity w14x. In contrast, AZT does not inhibit mtDNA synthesis in human lymphoid CEM cells w15x or neuronal PC12 cells w16x, possibly because AZT triphosphate is a much weaker inhibitor of polymerase-g than other dideoxynucleotide triphosphates. AZT triphosphate, for example, has been shown to be approximately 500 times less potent than ddC triphosphate in inducing mitochondrial toxicity w17x. Nonetheless, mitochondrial myopathies, including ragged red fibers and paracrystalline inclusions in mitochondria, have been observed during prolonged AZT treatment of individuals infected with HIV-1 w18,19x. In an attempt to explore the long-term effects of AZT exposure, genotoxicity and mitochondrial damage were evaluated in a population of H 9 human lymphocytic cells chronically exposed to AZT. 2. Materials and methods 2.1. Cell culture and DNA isolation H 9 cells were cultured in RPMI ŽGibco, Grand Island, NY. supplemented with 10% fetal bovine
serum ŽFBS; Gibco.. H 9-AZT cells chronically grown in medium containing 25 mM AZT for 7 months were described earlier w10x. 2.2. Incorporation of AZT into genomic DNA of H9 - A ZT cells Triplicate pellets of H 9 and H 9-AZT cells were processed to obtain nuclear DNA by CsCl density gradient centrifugation w20x. After dialysis at 48C to remove the CsCl, the DNA was quantitated by spectrophotometry at A 260 nm and AZT incorporation was determined by radioimmunoassay ŽRIA. in order to determine the amount of AZT incorporated into genomic DNA of H 9-AZT cells. The RIA specific for AZT was performed as already described using an anti-AZT antibody ŽSigma, St. Louis, MO. diluted 1:5000 and w 3 HxAZT Ž20 Cirmmol, Moravek Biochemicals Inc., Brea, CA. as tracer and unlabeled AZT for the standard curve w7x. DNA from H 9-AZT and H 9 cells was adjusted to a concentration of 30 mgrml, and 0.1 ml was assayed by RIA. Either AZT standards, unknown H 9-AZT-DNA samples or H 9DNA with no AZT incorporation were incubated with the antibody for 90 min at 378C. The w 3 HxAZT tracer and goat anti-rabbit IgG ŽICN ImmunoBiologicals, Costa Mesa, CA. were then added and the mixture incubated for 25 min. Centrifugation of the sample for 15 min at 1500 = g separated bound from unbound tracer. The unbound tracer Žsupernatant. was decanted. The remaining pellets were dissolved in 200 ml of 100 mM NaOH and counted in a liquid scintillation counter. The standard curve 50% inhibition in the RIA was at 1.78 " 0.36 pmol AZT Žmean " SD, n s 60. and the lower limit of assay sensitivity when the biological samples were assayed was approximately 0.33 " 0.05 pmol AZT Žmean " SD, n s 20.. 2.3. Analysis of chromosomal aberrations One hour before harvesting, 1 mg of colchicine per ml of medium was added to the cultures. The cells were processed to prepare chromosome spreads by trypsinization, hypotonic treatment with 75 mM KCl, fixation with Carnoy’s solution, and air drying w21x. Chromosomal aberrations Ždicentrics, breaks, micronuclei, fragments, gaps and rings. were scored in 50 metaphases from H 9 and 50 H 9-AZT cells stained with Giemsa stain.
R.P. Agarwal, O. A. OliÕeror Mutation Research 390 (1997) 223–231
2.4. Immunohistochemical staining for AZT incorporation into DNA Anti-AZT antibody was reconstituted in 1 ml of distilled water, dialyzed against distilled water overnight at 48C, and preabsorbed as previously described w7x. Cell suspensions were centrifuged onto slides in a Shandon cytospin centrifuge ŽShandon Southern Instruments, Inc. Sewickley, PA. for 5 min at 400 rpm. Prior to the addition of antibody the cells were incubated for 30 min at 378C with a 1:10 dilution of normal goat serum. The absorbed antiAZT antibody was then diluted with 100 mg of thymidine per mg antibody and added to the slides for a 3-hour incubation at 378C. A secondary antirabbit-rhodamine conjugated IgG antibody diluted 1:50 was applied at room temperature for an additional 30 min. Hoechst 33258 Ž2X-w4-hydroxyphenylx5-w4-methyl-1-piperazinylx-2,5X-bi-1 H-benzimidazole; Sigma. was used at a 1:60 000 dilution to stain DNA. The slides were air dried and mounted with Pro-Texx mounting medium ŽAmerican Scientific Products, McGaw Park, IL.. Slides with H 9-AZT-treated cells were digested with DNase for 1 h prior to the incubation with anti-AZT antibody as previously described w8x. The absence of signal confirmed the specificity of the DNA staining.
225
again in 1% osmium tetroxide for 1 h, washed and dehydrated in graded dilutions of ethanol. The dehydrated cells were embedded in Spurr’s resin, sectioned by Porter-Blum MT-2 ultramicrotome and viewed and photographed in a Jeol-CX100 electron microscope at =24 000. In order to localize lipid droplets H 9-AZT cells were collected and fixed for 60 min in glutaraldehyde in 0.1 M cacodylate buffer, pH 7.3. After a postfixation treatment with 1% phosphate-buffered osmium tetroxide at 48C, and dehydration in alcohol, the cells were embedded in Spurr epoxy resin at 608C. Ultra-thin sections were cut and mounted on 200 mesh copper grids. Preparations stained with uranyl acetate and lead citrate were viewed and photographed as described above. 2.7. Lactic acid determination H 9 , H 9-AZT and H 9 acutely treated with 25 mM AZT Žlabeled H 9-AZTaq. were incubated at 378C. The cell number was determined and the culture medium was collected by centrifugation at 24, 48 and 72 h of incubation. Lactic acid concentration in the medium was determined using a lactic acid kit ŽSigma.. 3. Results
2.5. Staining for intracellular lipids
3.1. AZT-DNA incorporation
The red-oil-O technique, specific to visualize neutral intracellular lipids, was applied to standard cytospin preparations. Briefly, the preparations were fixed in 10% formalin for 1 h, washed in tap water for 20 min and stained with a solution of red-oil-O in propylene glycol for 1 h. After an additional washing, the cells were counterstained with hematoxylin for 5 min w22x. The stain is frequently used to determine the amount of lipid droplets accumulated in the cytoplasm. Furthermore, lipid accumulation was confirmed by electron microscopy.
Three micrograms of DNA isolated from three different populations of H 9-AZT cells were analyzed
2.6. Electron microscopy The cells were harvested by centrifugation, washed 3 times with PBS and fixed overnight in glutaraldehyde Ž2.5%. plus formalin Ž2%. in isotonic phosphate buffer, pH 7.3. The fixative was rinsed out with phosphate buffer and the cells were fixed
Fig. 1. Bar graph indicating comparison of chromosomal aberrations in 50 metaphases of H 9 and H 9-AZT cells. Significance is indicated by asterisks.
226
R.P. Agarwal, O. A. OliÕeror Mutation Research 390 (1997) 223–231
Table 1 Summary of genomic and mitochondrial alterations in H 9-AZT cells Test
Endpoint
H 9-AZT
H9
Figure
RIA
AZT-DNA incorporation
2.0 pmol AZTrmg DNA
0.0 pmol AZTrmg DNA
NS
Chromosomal aberrations
gaps rings dicentrics micronuclei breaks fragments
1 4 2 10 11 8
1 1 0 1 3 3
Fig. 1 Fig. 1 Fig. 1 Figs. 1 and 2 Fig. 1 Fig. 1
total
36
9
Fig. 1
Mitochondrial metabolism
intracellular lipids ŽLM. intracellular lipids ŽEM.
17.0 " 1.4 abundant
1.8 " 0.7 sparse
Fig. 3bFig. 5 Fig. 3c
Mitochondrial appearance
cristae size morphology
lost swollen abnormal
present normal normal
Fig. 4a Fig. 4b Fig. 4c
a
a
Average lipid droplets" SEM Ž n s 15..
Fig. 2. Nuclei and micronuclei from H 9-AZT cells. a: Giemsa stain. b: Hoechst stain for DNA. c,d: anti-AZT antibody, and secondary anti-rabbit rhodamine-conjugated antibody.
R.P. Agarwal, O. A. OliÕeror Mutation Research 390 (1997) 223–231
by RIA in three separate assays. The results showed an incorporation of AZT equivalent to 2 pmol AZTrmg of DNA. DNA from H 9 cells was used as a control and was assayed in parallel in the same RIA. No incorporation of AZT in DNA from H 9 cells in three independent assays was detected. 3.2. Cytogenetic analysis Fifty metaphases of both H 9 and H 9-AZT cells were analyzed to score chromosomal aberrations ŽFig. 1, Table 1.. Dicentrics, breaks, micronuclei, fragments, gaps and rings were scored. The total of 36 aberrations for H 9-AZT cells, was 4-fold higher than in the control H 9 cells Ž p - 0.05.. Two of the most dramatically increased aberrations in H 9-AZT cells were breaks Ž3 in H 9 cells to 9 in H 9-AZT cells, Fig. 1, Table 1. and fragments Ž3 in H 9 cells to 8 in H 9-AZT cells, Fig. 1, Table 1.. These elevated frequencies could be explained as a consequence of
227
AZT being a chain terminator. The occurrence of rings, on the other hand, could be seen as a result of a disturbance in the telomeres, already described as a preferential site for AZT incorporation w8,9x, and the consequent adhesiveness of the chromosomal ends. Because micronuclei were the most evident increased aberration Žfrom 2 in H 9 cells to 10 in H 9-AZT cells, Fig. 1, Table 1., further studies with anti-AZT antibodies and immunostaining were performed to determine the presence of AZT in the micronuclei. 3.3. Immunostaining Micronuclei, seen as small round fragments close to interphase nuclei, were analyzed in interphase by specific immunostaining to detect AZT-DNA incorporation. All of the 50 micronuclei scored were positive for AZT staining. Fig. 2a–d shows micronuclei of H 9-AZT cells stained by Giemsa ŽFig. 2a.,
Fig. 3. Cytospin preparations of H 9 Ža. and H 9-AZT Žb. cells stained by the red-oil-O technique to visualize lipid droplets Žb, arrowheads.. c: electron microscopy of H 9-AZT cell, arrows indicate lipid droplets. =24 000.
228
R.P. Agarwal, O. A. OliÕeror Mutation Research 390 (1997) 223–231
Hoechst ŽFig. 2b. and anti-AZT antibody and antirabbit rhodamine-conjugated antibody ŽFig. 2c,d.. Red stain Žpositive for AZT. in Fig. 2c,d is the result of recognition of AZT by the anti-AZT antibody, and the secondary rhodamine-conjugated antibody. 3.4. Staining for intracellular lipids Neutral intracellular lipids were identified as an indirect measurement of mitochondrial function w23x. Abnormal structure of the mitochondria has been linked with impaired organelle function mainly due to reduced amount of DNA and cytochromes. As a consequence, the long-chain fatty acids are not effectively utilized by the mitochondria and accumulate in the cytoplasm where they are identified as lipid droplets by the red-O-oil technique w23x. Pictures of cytospin preparations of H 9 and H 9-AZT cells stained by red-oil-O technique were compared. An evident
increase in the amount of lipid droplets was noticed in the H 9-AZT cell preparations ŽFig. 3b.. Lipid droplets were counted in 15 individual cells from H 9 and H 9-AZT lines. The average of that score was of 1.8 dropletrcell for H 9 and of 17.0 dropletrcell for H 9-AZT cells. ŽFig. 5, Table 1.. A further observation of cells through electron microscopy confirmed the lipid droplet increase in H 9-AZT cells ŽFig. 3c.. 3.5. Electron microscopy Electron microscopic examination of H 9 and H 9-AZT cells showed evidence of significant mitochondrial defects in cells chronically exposed to AZT. Fig. 4 illustrates some of the mitochondrial defects detected in H 9-AZT cells. Loss of cristae ŽFig. 4a., swollen mitochondria ŽFig. 4b. and abnormal mitochondria ŽFig. 4c. were documented.
Fig. 4. Mitochondrial appearance visualized by electron microscopy in H 9-AZT cells. a: loss of internal cristae. b: swollen mitochondria. c: abnormal mitochondria. d: control H 9 cells. =24 000.
R.P. Agarwal, O. A. OliÕeror Mutation Research 390 (1997) 223–231 Table 2 Effect of chronic and acute AZT treatment of H 9 cells on lactic acid production 24 h
48 h
72 h
H 9 Ž ns 3. 72.7"28.1 b 123.9"33.2 114.1" 9.8 H 9-AZT Ž ns 3. 69.7"19.2 140.8"38.3 145.9"21.8 H 9-AZTaq Ž ns 3. a 111.5"49.8 190.6"28.2 167.8"60.7 a b
H 9-AZTaq cells treated acutely with 25 mM AZT. mg lactaterdl=10 6 cells.
3.6. Lactic acid determination The lactic acid content expressed as mgrdl per 10 6 cells, in the medium of AZT-acutely treated cells ŽH 9AZTaq . are presented in Table 2. The values in H 9-AZT Žpreviously chronically exposed to AZT. did not differ significantly from those of control cells ŽH 9 .. However, the cells that were treated acutely with AZT at 25 mM expressed 47–53% more lactic acid than the control cells.
4. Discussion Mitochondrial and chromosomal damage has been studied in a population of a human T-cell line derivative, H 9 , chronically exposed to AZT ŽH 9-AZT cells.. This study clearly demonstrates incorporation of AZT into genomic DNA, chromosomal aberrations, nuclear fragmentation, mitochondrial damage and altered lipid metabolism in the chronically AZT exposed cells. AZT has several undesired side effects in humans w1x. Some of them, like bone marrow aplasia, could be caused by incorporation of AZT into human DNA with a disturbance in both genomic and mitochondrial DNA replication. Previous studies reported incorporation of AZT into mammalian DNA of different species w4,5x in spite of the low affinity of AZT-5X-triphosphate for mammalian DNA polymerases w4,24–26x. Furthermore, preferential localization of AZT into telomeric DNA of Chinese hamster ovary ŽCHO. cells was documented by immunofluorescence with anti-AZT antibodies w8x. The same study also showed a disturbance in the normal development of telomeres probed indirectly by an increase in the number of chromosome bridges scored
229
in anaphase–telophase of CHO cells w8x. In that report, the authors speculated about the possibility that AZT might inhibit telomerase, the enzyme responsible for the elongation of the telomeres w21,27,28x. Chromosomal aberrations were previously studied in vivo in bone marrow of mice exposed to AZT for 28 days, revealing a 4-fold increase as compared with untreated mice w7x. Here, we report for the first time an integral study of both genomic and mitochondrial AZT-induced toxicities in the same cellular system under chronic exposure. Chromosomal aberrations – mainly fragments and breaks – increased 4-fold in H 9-AZT as compared with untreated H 9 cells ŽFig. 1, Table 1., probably as a result of incorporation of AZT into genomic DNA and subsequent chain termination. Rings, another chromosomal aberration that was increased in AZT chronically exposed cells, could be the result of chromosome stickiness by incorporation of AZT into the telomeres w8,9x. Nuclear fragmentation, probably as a result of AZT incorporation, was one of the aberrations that increased most dramatically ŽFig. 1, Table 1.. This is in agreement with earlier reports that demonstrated the ability of AZT, among other nucleoside analogs, to induce micronuclei in different cellular systems w29,30x. Immunohistochemical analysis of the micronuclei with anti-AZT antibodies confirmed that those nuclear fragments were indeed positive for AZT incorporated into the DNA ŽFig. 2.. A DNase pretreatment of exposed cells prior to anti-AZT incubation revealed no staining, confirming that the AZT signal was specific for AZT-DNA. An important body of information exists about degenerative effects induced by AZT into mitochondria in human myotubes w18,31,32x as well as destructive changes in the myofibrils, accumulation of lipid droplets, and enzymatic defects in the respiratory chain system w23,33x; however, there is no evidence of chromosomal aberrations or preferential binding studies in muscle cells. In addition to the genotoxic effects induced by AZT in H 9-AZT chronically exposed cells, mitochondrial damage was evaluated here by electron microscopy ŽFig. 4.. Degenerative lesions such as loss of cristae, swollen mitochondria and abnormal mitochondria ŽFig. 4a–c., lesions already described in myotubes of AZT induced myopathies w18,32x, have been observed in the cellular system studied here. We also report an in-
230
R.P. Agarwal, O. A. OliÕeror Mutation Research 390 (1997) 223–231
fect of AZT in human populations exposed to therapeutic doses of the drug.
Acknowledgements The authors are very thankful to Ms. Jun He for her technical assistance. This work was partially supported by the Department of Health and Human Services Grant NIAID AI 29155. We would like to thank Margaret Taylor for editorial assistance.
Fig. 5. Lipid droplets accumulated in the cytoplasm of H 9 and H 9-AZT cells. Each dot represents the amount of lipid droplets scored in one cell. A total of 15 cells were scored and the average of the observations is shown. A statistical significance is shown by an asterisk.
crease in cytoplasmic accumulation of lipid droplets ŽFig. 3bFig. 5, Table 1., indirect evidence of the impaired ability of the mitochondria to utilize longchain fatty acids w23x. As an adaptive cellular response, such a mitochondrial deficiency is expected to induce stimulation of glycolysis resulting in increased lactic acid secretion. Failure to observe any increase in lactate in H 9-AZT cells, compared to acutely treated cells, suggests that during chronic treatment, despite impaired mitochondria, the cells had adapted to grow under these conditions. Since the use of nucleoside analogs is widespread in AIDS and hepatitis B therapy, a careful study of their potential clastogenic ability should be taken into account as already suggested w34x. It is not clear whether nucleoside analog-induced toxicity is a consequence of a decrease in mtDNA w13,15,35x or of mutations impairing normal mitochondrial function. Nonetheless, mtDNA incorporation of nucleoside analog drugs remains a sensitive measure of drug toxicity. A careful investigation of human liver toxicity reported in patients receiving AZT might indicate whether or not chronic AZT exposure could induce liver failure by mechanisms similar to those seen with the nucleoside analog 5-FIAU Ž2X-fluoro-2-deoxyarabinofuranosyl-5-iodocytidine. w36x. Studies of chronic exposure in different systems are needed to address the potential carcinogenic ef-
References w1x Yarchoan, R. and S. Broder Ž1987. Preliminary results on the use of dideoxynucleosides in the therapy of AIDS, in: R.M. Chanock, R.A. Lerner, F. Brown and H.S. Ginsberg ŽEds.., Vaccine 87, Cold Spring Harbor Press, New York, pp. 214–224. w2x P.A. Furman, J.A. Fyfe, M.H. St Clair, K. Weinhold, J.L. Rideout, G.A. Freeman, S.N. Lehrman, D.P. Bolognesi, S. X Broder, H. Mitsuya, D.A. Barry, Phosphorylation of 3 -azidoX X 3 -deoxythymidine and selective interaction of the 5 -triphosphate with human immunodeficiency virus reverse transcriptase, Proc. Natl. Acad. Sci. USA 83 Ž1986. 8333–8337. w3x Y.C. Cheng, G.E. Dutschman, K.F. Bastow, M.G. Sarngadharan, R.Y.C. Ting, Human immunodeficiency virus reverse transcriptase. General properties and its interactions with nucleoside triphosphate analogs, J. Biol. Chem. 262 Ž1987. 2187–2189. w4x M.A. Vazquez-Padua, M.C. Starnes, Y.C. Cheng, IncorporaX X tion of 3 -azido-3 -deoxythymidine into cellular DNA and its removal in a human leukemic cell line, Cancer Commun. 2 Ž1990. 55–62. w5x V.I. Avramis, W. Markson, R.L. Jackson, E. Gomperts, Biochemical pharmacology of zidovudine in human Tlymphoblastoid cells ŽCEM., Aids 3 Ž1989. 417–422. w6x J.P. Sommadossi, R. Carlisle, Z. Zhou, Cellular pharmacolX X ogy of 3 -azido-3 -deoxythymidine with evidence of incorporation into DNA of human bone marrow cells, Mol. Pharmacol. 36 Ž1989. 9–14. w7x O.A. Olivero, F.A. Beland, M.C. Poirier, Immunofluorescent X X X localization and quantitation of 3 -azido-2 ,3 -dideoxythymidine ŽAZT. incorporated into chromosomal DNA of human, hamster and mouse cell lines, Int. J. Oncol. 4 Ž1994. 49–54. w8x O.A. Olivero, M.C. Poirier, Preferential incorporation of X X X 3 -azido-2 ,3 -dideoxythymidine ŽAZT. into telomeric DNA and Z-DNA-containing regions of chinese hamster ovary cells, Mol. Carcinog. 8 Ž1993. 81–88. w9x D.E. Gomez, A. Kassim, O.A. Olivero, Preferential incorpoX X X ration of 3 -Azido-2 ,3 -dideoxythymidine ŽAZT. in telomeric sequences of CHO cells, Int. J. Oncol. 7 Ž1995. 1057–1060. w10x R.P. Agarwal, 3X-Azido-2X ,3X-dideoxythymidine induces
R.P. Agarwal, O. A. OliÕeror Mutation Research 390 (1997) 223–231
w11x
w12x
w13x
w14x
w15x
w16x
w17x
w18x
w19x
w20x
w21x w22x
thymidine kinase activity in vitro, Int. J. Purine Pyrimidine Res. 2 Ž1991. 1–3. R.P. Agarwal, J. Padmanabhan, J. He, Kinetics of accumulaX X X tion and efflux of 3 -azido-3 -deoxythymidine-5 -monophosphate in H 9 and H 9-AZT cells, Cell. Pharmacol. 1 Ž1994. 121–125. R.P. Agarwal, A.M. Mian, Thymidine and zidovudine metabolism in chronically zidovudine-exposed cells in vitro, Biochem. Pharmacol. 42 Ž1991. 905–911. D.J. Medina, C.H. Tsai, G.D. Hsiung, Y.C. Cheng, Comparison of mitochondrial morphology, mitochondrial DNA content, and cell viability in cultured cells treated with three anti-human immunodeficiency virus dideoxynucleosides, Antimicrob. Agents Chemother. 38 Ž1994. 1824–1828. C.H. Tsai, S.L. Doong, D.G. Johns, J.S. Driscoll, Y.C. Cheng, Effect of anti-HIV 2X-beta-fluoro-2X ,3X-dideoxynucleoside analogs on the cellular content of mitochondrial DNA and on lactate production, Biochem. Pharmacol. 48 Ž1994. 1477–1481. C.H. Chen, M. Vazquez-Padua, Y.C. Cheng, Effect of antihuman immunodeficiency virus nucleoside analogs on mitochondrial DNA and its implication for delayed toxicity, Mol. Pharmacol. 39 Ž1991. 625–628. S.A. Keilbaugh, W.H. Prusoff, M.V. Simpson, The PC12 cell as a model for studies of the mechanism of induction of peripheral neuropathy by anti-HIV-1 dideoxynucleoside analogs, Biochem. Pharmacol. 42 Ž1991. 5–8. J.L. Martin, C.E. Brown, N. Matthews-Davis, J.E. Reardon, Effects of antiviral nucleoside analogs on human DNA polymerases and mitochondrial DNA synthesis, Antimicrob. Agents Chemother. 38 Ž1995. 2743–2749. M.C. Dalakas, I. Illa, G.H. Pezeshkpour, J.P. Laukaitis, B. Cohen, J.L. Griffin, Mitochondrial myopathy caused by long-term zidovudine therapy, New Engl. J. Med. 322 Ž1990. 1098–1105. C. Mhiri, M. Baudrimont, G. Bonne, C. Geny, F. Degoul, C. Marsac, E. Roullet, R. Gherardi, Zidovudine myopathy: a distinctive disorder associated with mitochondrial dysfunction, Ann. Neurol. 29 Ž1991. 606–614. Flamm, W.G., M.L. Birnstiel and P.M.B. Walker Ž1967. Preparation and fractionation and isolation of single strands of DNA by ultracentrifugation in fixed-angle rotors, in: G.D. Birnie and S.M. Fox ŽEds.., Subcellular Components: Preparation and Fractionation, Butterworths, London, pp. 126–154. C.W. Greider, Telomeres, telomerase and senescence, Bioessays 12 Ž1990. 363–369. Coolidge, B.J. and R.M. Howard Ž1979. Animal Histology Procedures, U.S. Dept. of Health, Education and Welfare, Bethesda, MD.
231
w23x M.C. Dalakas, M.E. Leon-Monzon, I. Bernardini, W.A. Gahl, C.A. Jay, Zidovudine-induced mitochondrial myopathy is associated with muscle carnitine deficiency and lipid storage, Ann. Neurol. 35 Ž1994. 482–487. w24x E.L. White, W.B. Parker, L.J. Macy, S.C. Shaddix, G. McCaleb, J.A. Secrist, R. Vince, W.M. Shannon, Comparison of the effect of Carbovir, AZT, and dideoxynucleoside triphosphates on the activity of human immunodeficiency virus reverse transcriptase and selected human polymerases, Biochem. Biophys. Res. Commun. 161 Ž1989. 393–398. w25x P. Huang, D. Farquhar, W. Plunkett, Selective action of 3X-azido-3X-deoxythymidine 5X-triphosphate on viral reverse transcriptases and human DNA polymerases, J. Biol. Chem. 265 Ž1990. 11914–11918. w26x W. Nickel, S. Austermann, G. Bialek, F. Grosse, Interactions of azidothymidine triphosphate with the cellular DNA polymerases alpha, delta, and epsilon and with DNA primase, J. Biol. Chem. 267 Ž1992. 848–854. w27x E.H. Blackburn, Telomeres: structure and synthesis, J. Biol. Chem. 265 Ž1990. 5919–5921. w28x D. Shippen-Lentz, E.H. Blackburn, Functional evidence for an RNA template in telomerase, Science 247 Ž1990. 546–552. w29x M.D. Phillips, B. Nascimbeni, R.R. Tice, M.D. Shelby, Induction of micronuclei in mouse bone marrow cells: an evaluation of nucleoside analogues used in the treatment of AIDS, Environ. Mol. Mutagen. 18 Ž1991. 168–183. w30x M. Gonzalez Cid, I. Larripa, Genotoxic activity of azidothymidine ŽAZT. in in vitro systems, Mutation Res. 321 Ž1994. 113–118. w31x G.H. Pezeshkpour, I. Illa, M.C. Dalakas, Ultrastructural characteristics and DNA immunocytochemistry in human immunodeficiency virus and zidovudine-associated myopathies, Hum. Pathol. 22 Ž1991. 1281–1288. w32x B.S. Peters, J. Winer, D.N. Landon, A. Stotter, A.J. Pinching, Mitochondrial myopathy associated with chronic zidovudine therapy in AIDS, Q. J. Med. 86 Ž1993. 5–15. w33x M.C. Semino-Mora, M.E. Leon-Monzon, M.C. Dalakas, Effect of L-carnitine on the zidovudine-induced destruction of human myotubes. Part I: L-carnitine prevents the myotoxicity of AZT in vitro, Lab. Invest. 71 Ž1994. 102–112. w34x M.D. Shelby, The cytogenetic effects of AZT, Environ. Mol. Mutagen. 24 Ž1994. 148–151. w35x C.H. Chen, Y.C. Cheng, Delayed cytotoxicity and selective loss of mitochondrial DNA in cells treated with the anti-huX X man immunodeficiency virus compound 2 ,3 -dideoxycytidine, J. Biol. Chem. 264 Ž1989. 11934–11937. w36x W.B. Parker, Y.C. Cheng, Mitochondrial toxicity of antiviral nucleoside analogs, J. NIH Res. 6 Ž1994. 57–61.