Defective DNA Repair in Cells With Human T-Cell Leukemia/Bovine Leukemia Viruses: Role of tax Gene

JNCI Journal of the National Cancer Institute, Jun 1999

BACKGROUND: Human T-cell leukemia virus (HTLV)/bovine leukemia virus (BLV) group retroviruses, which cause hematopoietic cancers, encode a unique protein, Tax, involved in the transformation of infected cells. Our purpose was to determine whether the mechanism by which Tax protein induces transformation in HTLV- or BLV-infected cells involves DNA damage. METHODS: We used a micronucleus assay to measure chromosomal damage and alkali denaturation analysis to test host-cell DNA integrity in cells infected with HTLV, BLV, or simian T-lymphotropic virus or in cells transfected with the tax gene of HTLV or BLV. Controls included uninfected cells and cells infected with other oncogenic retroviruses or oncogenic DNA viruses. We used a plasmid reactivation assay to examine whether the damage might be due to the inhibition of DNA repair. To ascertain which of several repair pathways might be inhibited, chemical methods were used to selectively introduce lesions repaired by specific pathways into the reporter plasmid. RESULTS: The presence of Tax was associated with DNA damage. HTLV- or BLV-infected or tax-transfected cells showed normal ability to repair damage induced by deoxyribonuclease I or psoralen but markedly decreased ability to repair damage induced by UV light, quercetin, or hydrogen peroxide. CONCLUSIONS: These data suggest that the DNA repair pathway most inhibited by Tax is base-excision repair of oxidative damage. To our knowledge, this is the first report demonstrating inhibition of DNA repair by any retrovirus and suggests that this inhibition of DNA repair may contribute to the mechanism of cell transformation by the HTLV/BLV group of viruses.

A PDF file should load here. If you do not see its contents the file may be temporarily unavailable at the journal website or you do not have a PDF plug-in installed and enabled in your browser.

Alternatively, you can download the file locally and open with any standalone PDF reader:

https://jnci.oxfordjournals.org/content/91/11/933.full.pdf

Defective DNA Repair in Cells With Human T-Cell Leukemia/Bovine Leukemia Viruses: Role of tax Gene

Sean M. Philpott 0 1 2 Gertrude C. Buehring ) 0 1 2 0 Oxford University Press 1 Affiliation of authors: Program in Infectious Diseases, School of Public Health, University of California, Berkeley. eases, School of Public Health, University of California, Berkeley , Berkeley, CA 94720 ( 2 Journal of the National Cancer Institute , Vol. 91, No. 11, June 2, 1999 Background: Human T-cell leukemia virus (HTLV)/bovine leukemia virus (BLV) group retroviruses, which cause hematopoietic cancers, encode a unique protein, Tax, involved in the transformation of infected cells. Our purpose was to determine whether the mechanism by which Tax protein induces transformation in HTLV- or BLV-infected cells involves DNA damage. Methods: We used a micronucleus assay to measure chromosomal damage and alkali denaturation analysis to test host-cell DNA integrity in cells infected with HTLV, BLV, or simian T-lymphotropic virus or in cells transfected with the tax gene of HTLV or BLV. Controls included uninfected cells and cells infected with other oncogenic retroviruses or oncogenic DNA viruses. We used a plasmid reactivation assay to examine whether the damage might be due to the inhibition of DNA repair. To ascertain which of several repair pathways might be inhibited, chemical methods were used to selectively introduce lesions repaired by specific pathways into the reporter plasmid. Results: The presence of Tax was associated with DNA damage. HTLV- or BLV-infected or tax-transfected cells showed normal ability to repair damage induced by deoxyribonuclease I or psoralen but markedly decreased ability to repair damage induced by UV light, quercetin, or hydrogen peroxide. Conclusions: These data suggest that the DNA repair pathway most inhibited by Tax is base-excision repair of oxidative damage. To our knowledge, this is the first report demonstrating inhibition of DNA repair by any retrovirus and suggests that this inhibition of DNA repair may contribute to the mechanism of cell transformation by the HTLV/BLV group of viruses. [J Natl Cancer Inst 1999;91:933-42] - Human T-cell leukemia virus types 1 and 2 (HTLV-1 and HTLV-2, respectively), bovine leukemia virus (BLV), and simian T-lymphotropic virus (STLV) belong to the HTLV/BLV group of oncogenic retroviruses. These viruses are etiologically linked to a variety of hematopoietic malignancies (1). Most oncogenic retroviruses induce cellular transformation through one of two mechanisms (2). Acutely transforming retroviruses, such as Rous sarcoma virus (RSV), carry a transduced cellular proto-oncogene, giving the virus transcriptional control over the oncogene. Expression of these genes at inappropriate times, at unacceptable levels, or in unsuitable cell types causes rapid induction of tumors through dysfunction of biochemical pathways regulating cellular growth. The poorly transforming nonacute retroviruses, such as avian leukosis virus, induce malignancies by insertional mutagenesis. Proviral DNA integrates close to cellular proto-oncogenes, stimulating excess or inappropriate transcription of neighboring genes and disrupting control of growth processes. Malignancies occur infrequently and only after long latent periods. HTLV/BLV group viruses do not appear to use either of these methods. These viruses contain a transforming gene, tax, not homologous to cellular proto-oncogenes, and there is no preferred site of integration. The molecular mechanism by which the encoded protein Tax induces cellular transformation is not known, but most studies have focused on its interaction with cellular proto-oncogenes and tumor suppressor genes (3). Tax has been shown to increase expression of cellular genes involved in the regulation of lymphocyte growth, including interleukin 2 and its receptor (49), and to interact with putative tumor suppressor proteins like p53 (1012) and Int-6 (13,14). Alternatively, we have found that the Tax protein of HTLV/ BLV group viruses imbues infected cells with a mutator phenotype. Cancers develop through a series of well-defined steps, changing progressively from normal cells into premalignant cells and then into localized tumors and metastatic lesions. Even after exposure to strong carcinogens, the mutation rate in normal cells is too low to account for the myriad of alterations that accumulates during tumor progression. Several researchers (15,16) have argued that the earliest step in oncogenesis is a change that increases the cellular mutation rate. When compared with the mutation rate seen in primary cultures of normal human cells, the mutation rate in transformed cells is up to 1000-fold higher (17). This higher mutation rate may be due to decreased replication fidelity or inhibition of DNA damage repair. Many hereditary disorders involving defective DNA repair are characterized by a higher frequency of certain types of cancers. For example, patients with xeroderma pigmentosum (XP) lack at least one of the enzymes responsible for repairing UV radiation-induced DNA damage. XP patients exhibit 100-fold greater susceptibility to skin cancers (18). Some cancers induced by DNA viruses are also suspected of involving deficient DNA repair because of frequently observed chromosomal instability in infected cells (1923). Chromosomal abnormalities are also common in HTLV-1and BLV-infected leukemia cells (24,25) and may mirror disease severity, suggesting that these changes are important events in the progression to cancer (26). Cells immortalized in vitro with HTLV-1 exhibit similar chromosomal aberrations and enhanced sensitivity to genotoxic chemicals, as measured by increased frequency of micronuclei following treatment with the chemicals (27,28). Finally, expression of human b-polymerase (an enzyme involved in DNA excision repair) is decreased in HTLV-1-infected cells, suggesting that HTLV-1 might inhibit DNA repair in infected cells (29). In this article, we examine the hypothesis that one mechanism by which HTLV/BLV group retroviruses may transform infected cells is by inducing genomic instability, most likely by inhibiting cellular repair of spontaneous DNA damage. MATERIALS AND METHODS Cell Lines and Cell Culture essential medium (Life Technologies, Inc.); and Hep3B line in Eagles minimum essential medium (Life Technologies, Inc.) with 0.1 mM nonessential amino acids and 1.0 mM sodium pyruvate. All media were supplemented with 100 mg/mL streptomycin (Pfizer, New York, NY), 2 mM L-glutamine, 10 mg/mL insulin, 100 U/mL penicillin, 50 U/mL polymyxin B, and 5%10% fetal bovine serum (FBS) (all from Sigma Chemical Co., St. Louis, MO). All cell lines were grown at 37 C in a moist atmosphere of 5% CO2. Culture fluids were changed once or twice per week. Monolayer cells were passaged at confluence. Suspension cells were passaged to maintain a density of 0.51.0 106 cells/mL. Plasmids and Plasmid Propagation The 38 cell lines used for this study are summarized in Table 1. Cell lines were maintained as follows: most of the monolayer lines in Dulbeccos modified high glucose Eagles medium (Life Technologies, Inc. [GIBCO/BRL], Gaithersburg, MD); lymphoid-derived suspension lines in RPMI-1640 medium (Life Technologies, Inc.); Chinese hamster ovary (CHO) lines in alpha-modified minimal Six plasmids or modifications thereof were used. Plasmid pRSV-tax1 contains the tax gene of HTLV-1, and plasmid pRSV-tax2 encodes HTLV-2 tax, both under the control of the RSV promoter. Plasmid pBLV-tax contains BLV tax under the control of the simian virus 40 (SV40) promoter (43). Plasmid pK30 contains the whole HTLV-1 provirus (44). Plasmid pRSV-CAT contains the Cell type (M/S)* Virus contained or genetic defect *M 4 monolayer; S 4 suspension. Other virus abbreviations are as follows: BPV 4 bovine papillomavirus; EBV 4 Epstein-Barr virus; FeLV 4 feline leukemia virus; HBV 4 hepatitis B virus; HPV 4 human papillomavirus; MLV 4 mouse leukemia virus; MMTV 4 mouse mammary tumor virus; MPMV 4 Mason-Pfizer monkey virus; MSV 4 mouse sarcoma virus; RSV 4 Rous sarcoma virus; STLV 4 simian T-lymphotropic virus; SV40 4 simian virus 40. Chinese hamster ovary (M) Chinese hamster ovary (M) Chinese hamster ovary (M) Bovine mammary epithelium (M) Bovine mammary epithelium (M) Bovine mammary epithelium (M) Mouse mammary epithelium (M) Human mammary epithelium (M) Human mammary epithelium (M) Human T lymphocyte (S) Human promyelocyte (S) Bat lung epithelium (M) Bat lung epithelium (M) Bovine mammary epithelium (M) Bovine mammary epithelium (M) Bovine mammary epithelium (M) Bovine mammary epithelium (M) Bovine mammary epithelium (M) Lamb kidney fibroblast (M) Baboon T lymphocyte (S) Human T lymphocyte (S) Human T lymphocyte (S) Human T lymphocyte (S) Human T lymphocyte (S) Mouse mammary epithelium (M) Human B lymphocyte (S) Human liver (M) Human uterine cervix (M) Human lung fibroblast (M) CAT (chloramphenicol acetyltransferase) reporter gene under the control of the RSV promoter (45). Plasmid pBluescribe (Stratagene, La Jolla, CA) contains the gene for b-galactosidase. Vector-only deletion mutants were made of the taxcontaining plasmids by cleaving out the tax insert with EcoRI and converting the plasmid back to closed circular form by use of standard methods (46). Deletion mutants were confirmed for the absence of the tax genes by their size, as determined by agarose gel electrophoresis and by negative polymerase chain reaction (PCR) results by use of primers for regions within the respective tax genes. All plasmids contain an ampicillin-resistance gene for selection during propagation and purification. Plasmids were propagated according to standard methods (46) in XL1-Blue strain of Escherichia coli, and plasmid DNA was purified by use of the Wizard(tm) Maxiprep kit (Promega Corp., Madison, WI). Cell Transfections The established H9 T-lymphocytic cell line was transfected with the DNA of various plasmids by electroporation. Plasmid (10 mg) was added to cells suspended at a concentration of 1 107/mL in 0.3 mL of RPMI-1640 medium supplemented with 10 mM dextrose (Sigma Chemical Co.) and 1 mM dithiothreitol (Fisher Scientific Co., Santa Clara, CA). After 10 minutes on ice, the mixture was subjected to a single electrical pulse of 250 V, 960 mF from a GenePulser Transfection apparatus equipped with a Capacitance Extender unit (Bio-Rad Laboratories, Richmond, CA) and maintained on ice another 10 minutes. The mixture was then diluted into 5 mL of medium supplemented with antimicrobials and 15% FBS and incubated at 37 C for 48 hours before analysis. Efficient expression of the transfected tax genes was verified by reverse transcriptionPCR analysis (Fig. 1). Cells transfected with pBluescribe or vector alone (pRSV-tax1, pRSV-tax2, and pBLV-tax with the inserted tax genes removed) were used as experimental controls. PCR and Reverse TranscriptionPCR These assays were performed to confirm expression of tax genes in transfected cells and the absence of tax in plasmid deletion mutants. Assays were done according to standard methods (46). Primers used in these assays were synthesized by Operon (Alameda, CA). Primer sequences are detailed in Table 2. Micronucleus Assays These assays were undertaken to detect micronuclei, small perinuclear bodies containing either chromosomal fragments or entire chromosomes that have failed to attach to the mitotic spindle during cell division. Micronuclei are considered to be a marker of gross chromosomal damage. Thirty-six cell lines were assayed (AA8, UV41, EM9, C72, M23, M26, C127I, MCF7, T47D, H9, HL-60, Tb1Lu, Bat2Cl6, FLK, C37Cl1, C72/BLV, C761, M26/BLV, M26/BLV-Cl1, C8166-45, MT2, MT4, clone 19, FeLV-3281, JLSV5, L691D, GR, CMMT, CE/RSV, KC, XC, ID13, Raji, Hep3B, HeLa, and SV40-infected IMR-90 cells). Cultured cells in the logarithmic phase of growth were treated overnight with 3 mg/mL cytochalasin-B added to the culture medium in order to arrest cells in the process of mitosis, the phase during which micronuclei appear. Cells were washed twice Fig. 1. Reverse transcription polymerase chain reaction (PCR) products confirming tax expression in transfected cells. Total genomic RNA was collected from transfected H9 cells (human T-cell line) or control MT2 cell line (human T-cell leukemia virus type 1 [HTLV-1]infected T-cell line), reverse transcribed, and subjected to nested PCR by use of tax-specific oligonucleotide primers. PCR products were subjected to electrophoresis on a 3% NuSieve (FMC BioProducts) agarose gel. The presence of an amplified band of the appropriate size indicates expression of the tax gene. From left to right: lane 1 4 base-pair marker (HinfI- restricted fX174 DNA); lane 2 4 product from reverse tanscription mixture with no added RNA; lane 3 4 product from MT2 cell RNA; lane 4 4 product from untransfected H9 cell RNA; lane 5 4 product from H9 cells transfected with bovine leukemia virus tax; lane 6 4 product from H9 cells transfected with HTLV-1 tax; lane 7 4 product from H9 cells transfected with HTLV-2 tax; lane 8 4 product from H9 cells transfected with whole HTLV-1. Table 2. Oligonucleotides used for polymerase chain reaction amplification to confirm expression of tax genes in transfected cells and absence of tax in plasmid deletion mutants Sequence, 58 to 38 as synthesized CTTCGGGATACATTACCTGA AATTGGCATTGGTAGGGCTG ATATCACCATCGATGCCTGG TTCACTTTCCCCCTTCGAGC TTATCAGCCCACTTCCCAGG TGGTGGGCAAACAGTCTTCG CGGATACCCAGTCTACGTGT GAGCCGATAACGCGTCCATC Nucleotide Amplification product size, bp 406 *Virus abbreviations are as follows: BLV 4 bovine leukemia virus; HTLV 4 human T-cell leukemia virus. bp 4 base pairs. with ice-cold Dulbeccos phosphate-buffered saline (DPBS) (Life Technologies, Inc.), placed in hypotonic solution (0.75% sodium citrate) for 10 minutes to swell cells, then fixed in absolute methanol. Monolayer cells had been grown directly on glass slides, whereas suspension cells were harvested onto glass slides after fixation with the use of a Cytospin-II cytocentrifuge (Shandon Inc., Pittsburgh, PA) at 600 rpm for 6 minutes at room temperature. All cell preparations were dried and stained for 5 minutes with 0.01 mg/mL ethidium bromide. Approximately 1000 cells per slide were examined for the presence of micronuclei under green epifluorescent illumination (450490 nm), and the number of cells with or without micronuclei was scored. Structures were identified as micronuclei if they were 1) similar in shape and staining to, but smaller than, the normal nucleus, 2) nonrefractile, and 3) not linked to the nucleus by a nucleoplasmic bridge (47). Scoring was blind; i.e., all cell cultures were prepared by one of us (G. C. Buehring) and given to the other (S. M. Philpott) as coded specimens for scoring. The kinetochore assay was performed as described by Majone et al. (27) to determine if the micronuclei observed were due to whole-chromosome loss or to chromosome fragmentation. The presence of a kinetochore (centromere) within a micronucleus suggests that a whole chromosome is present, whereas the absence of a kinetochore suggests that only a chromosome fragment is present. After methanol fixation on glass slides, cells were incubated with antikinetochore monoclonal antibody (Chemicon, Temecula, CA) diluted 1 : 1000 in DPBS, then in biotinylated horse anti-mouse immunoglobulin G (Vector Laboratories, Inc., Burlingame, CA) diluted 1 : 1000 in DPBS, and finally in 5 mg/mL fluoresceinated streptavidin (Vector Laboratories, Inc.). Each incubation was done for 1 hour at room temperature; between each step, slides were rinsed three times with DPBS. After these incubations, cells were stained with propidium iodide in an antifade solution (Sigma Chemical Co.). Cells were scored for micronuclei as described above. When a micronucleus was located, the presence or absence of a kinetochore spot in each micronucleus was determined by switching from green to blue illumination (510560 nm). The number of kinetochore-positive micronuclei, the number of kinetochore-negative micronuclei, and the total number of cells containing at least one micronucleus were all recorded separately. These samples were not coded. Apoptosis Assays Apoptotic cell death was measured in cells cultured in 10% FBS-containing maintenance medium, described above in section entitled Cell Lines and Cell Culture, as well as under conditions of induced apoptosis, namely, serum withdrawal or the addition of the chemotherapeutic drugs doxorubicin or dexamethasone at a final concentration of 10 mM. Apoptotic cell death in HTLV-1infected cells (C8166-45, MT2, and MT4), BLV-infected cells (Bat2Cl6), STLVinfected cells (KIA), virally transfected cells (H9 cells transfected with the tax-containing plasmids or whole HTLV-1-containing plasmid), and uninfected cells (H9 and Tb1Lu) was measured three ways as described in detail previously (48). Briefly, trypan blue dye exclusion to detect viable cells was performed daily by counting and plotting viable cells versus time. Apoptosis before and after serum withdrawal was evaluated by electrophoretic analysis of genomic DNA. Extraction of DNA was done by digestion of a DPBS-rinsed cell pellet in 20 mL of 50 mM TrisHCl (pH 8.0), 10 mM ethylenediamine-tetraacetic acid (EDTA), 0.5% (wt/vol) sodium dodecyl sulfate, and proteinase K (Promega Corp.) (0.5 mg/mL) for 1 hour at 50 C and then 1 hour with 0.1 mg/mL ribonuclease A (RNase A) (Boehringer-Mannheim, Indianapolis, IN) added. We heated samples to 70 C and added 10 mL of loading solution (10 mM EDTA, 1% [wt/vol] SeaKem low-gelling-temperature agarose [FMC BioProducts, Rockland, ME], 0.25% [wt/vol] bromphenol blue, and 40% [wt/vol] sucrose). Samples were subjected to electrophoresis through a 3% agarose gel for 3 hours at 60 V in TrisBorateEDTA buffer (1.1 M Tris and 90 mM boric acid [pH 8.4]). Cells were considered to have undergone apoptotic death if they exhibited a characteristic ladder pattern of genomic DNA degradation. Finally, apoptosis was measured as the amount of fragmented DNA in spent culture media after cells were incubated 024 hours in the presence or absence of camptothecin, doxorubicin, or dexamethasone, all anticancer drugs that induce apoptosis in actively dividing cells. A commercially available enzyme-linked immunosorbent assay (ELISA) kit to measure the amount of fragmented DNA released from apoptotic cells (Cellular DNA Fragmentation ELISA; Boehringer-Mannheim) was used according to the manufacturers directions. Optical density of the final ELISA reaction was measured after different durations of drug exposure. Transfected cells were not subjected to apoptosis studies until 48 hours after transfection, so that those cells injured or killed during transfection could be removed before the assays were begun. Alkali Denaturation Assay This assay was used to measure the relative frequency of DNA strand breaks in 29 cell lines (AA8, UV41, EM9, C127I, MCF7, T47D, H9, Tb1Lu, Bat2Cl6, C72/BLV, C761, M26/BLV, C8166-45, MT2, KIA, FeLV-3281, JLSV5, L691D, GR, CMMT, F81, CE/RSV, KC, XC, ID13, Raji, Hep3B, HeLa, and SV40-infected IMR-90 cells). Procedures for performing this type of assay have been described in detail by other investigators (49,50). The analysis was blind; i.e., the person performing the analysis (S. M. Philpott) received coded cells from the person growing the cell cultures (G. C. Buehring). Washed cells were suspended at a concentration of 2.05.0 105 cells/mL in 5 mL of ice-cold DPBS and transferred as 500-mL aliquots into nine prechilled 12 75-mm borosilicate glass tubes. The first three tubes (labeled A) represented the control of undenatured double-stranded DNA. To these was added 1.0 mL of a fresh 1 : 1 mixture of 0.1 N NaOH (denaturing solution) and 0.1 M KH2PO4 (stop buffer) (final pH 7.4); then we added 500 mL of bisbenzamide buffer (indicator) (0.16% N-laroylsarcosine (Sigma Chemical Co.), 0.2 M K4P2O7, 0.04 M EDTA, and 1.0 mg/mL bisbenzamide [Hoechst 33258; Sigma Chemical Co.] [pH 7.4]). Bisbenzamide is a fluorescent dye that binds to double-stranded but not to singlestranded DNA. The next three tubes (labeled B) represented the experimental values. Denaturing solution was added, and the cellular DNA was allowed to unwind for exactly 10 minutes under chilled (4 C), light-free conditions before stop buffer and indicator were added. This set of tubes contained DNA in partially denatured form. The tubes labeled C represented another control, the maximum unwinding that the DNA is capable of attaining. The DNA in these tubes was allowed to unwind for 2 hours under chilled, light-free conditions before the reaction was stopped. Fluorescence was measured with a fluorometer (excitation, 365 nm; emission, 465 nm; narrow band pass). The fraction (F) of double-stranded DNA remaining after 10 minutes of alkaline denaturation was then calculated as follows: F 4 (mean fluorescence tubes B mean fluorescence tubes C)/(mean fluorescence tubes A mean fluorescence tubes C). Plasmid Reactivation Assay This assay, used to measure rates of repair of induced DNA damage, has been described by others (51,52). Plasmid pRSV-CAT harboring the CAT reporter gene under control of the RSV promoter was used throughout. For UV damage, 50100 mL of plasmid DNA (50 mg/mL in TE buffer [10 mM TrisHCl at pH 8.0 and 1 mM EDTA]) per well of a sterile 24-well tissue culture plate was exposed to an unfiltered UV lamp for 1530 seconds at a distance of 1 inch (approximately 23 J/m2 per second). For the preparation of deoxyribonuclease (DNase)-nicked DNA, plasmid DNA was diluted (50 mg/mL in an ice-cold solution containing 50 mM TrisHCl [pH 7.8], 5 mM b-mercaptoethanol, 5 mM MgCl2, 50 mg/mL BSA, and 0.5 Kunitz unit/mL DNase I [BoehringerMannheim]) and incubated on ice for 15 minutes. For the induction of psoralen damage, plasmid DNA was diluted to 50 mg/mL in 500 mL of sterile ice-cold water and then mixed with 20 mL of 4.6 mM psoralen (furo[2,3-g]coumarin) in ethanol, incubated at 25 C for 60 minutes, and exposed for 60 seconds to light from a mercury sun lamp to trigger the psoralenDNA cross-linking reaction. For the introduction of damage with quercetin, plasmid DNA was diluted to 50 mg/mL in sterile ice-cold water, mixed with an equal volume of a solution of 0.4 mM cupric chloride, 0.4 mM quercetin, and 20 mM TrisHCl (pH 8.0), and then incubated at 37 C for 60 minutes. Plasmid DNA was oxidized by dilution to 50 mg/mL in 3% (vol/vol) H2O2 and incubation at room temperature for 30 minutes. Undamaged control plasmids were treated with the vehicle solutions without exposure to the damaging agents. After all treatments, the damaged DNA was purified by ethanol precipitation and resuspended in TE buffer to a final concentration of 500 mg/mL. To verify the extent of damage, we subjected treated DNA to electrophoresis on a 1.5% agarose gel and compared it with undamaged DNA (Fig. 2, A, B, and C). Target cells were transfected with 25 mg of damaged or undamaged pRSV-CAT plasmid DNA by electroporation and were incubated at 37 C for approximately 40 hours, the time of maximal enzyme expression in transfected lymphocytes (51). Harvested cells were resuspended in 250 mM TrisHCl (pH 7.8) and lysed by freeze-thawing three times, the total protein was determined (BCA protein assay kit; Pierce Chemical Co., Rockford, IL), and DNA repair rates were measured as a function of CAT enzyme activity quantitatively determined by use of a commercially available CAT-specific ELISA kit (CAT-ELISA; Boehringer-Mannheim). Statistical Analysis The data presented in Fig. 3, A and B (number of micronuclei and proportion of single-stranded DNA), were analyzed by three group comparison methods (53): one that assumes normal distribution of each group being compared (Scheffes method) and two that are nonparametric (KruskalWallis method and the median counterpart of Fishers exact test). The nonparametric analyses were performed with the use of both Monte Carlo and asymptotic approaches. BeFig. 2. Gel electrophoresis in 1.5% SeaKem agarose to determine extent of plasmid (pRSV-CAT) damage by various agents. For all treatments, no detectable undamaged (supercoiled) DNA remained. The base-pair markers are from HindIIIrestricted l DNA. A) From left to right: lane 1 4 untreated plasmid; lane 2 4 quercetin-treated plasmid; lane 3 4 EcoRI-restricted plasmid (as a marker for linearization); lane 4 4 base-pair markers. B) From left to right: lane 1 4 base-pair markers; lane 2 4 untreated plasmid; lane 3 4 deoxyribonuclease I-treated plasmid. C) From left to right: lane 1 4 base-pair markers; lane 2 4 untreated plasmid; lane 3 4 hydrogen peroxide-treated plasmid; lane 4 4 psoralen-treated plasmid; lane 5 4 UV light-treated plasmid. Symbol and abbreviation key: 4 HTLV-1-infected human T-cell lines (C8166-45, MT2, and MT4) or H9 human T cells transfected with the HTLV-1 tax gene in plasmid pRSV-tax1; 4 HTLV-2-infected T-cell line (clone 19) or H9 cells transfected with the HTLV-2 tax gene in plasmid pRSV-tax2; 4 STLV (simian T-cell lymphotropic virus)-infected baboon T-cell line (KIA); (simian virus 40)-infected IMR-90 (human lung fibroblast) line; 4 CHO (Chinese hamster ovary) cell line AA8 (normal for DNA repair mechanisms); 4 CHO cell lines UV41 (defective in nucleotide excision repair) and EM9 (defective in DNA strand break repair); mammary epithelial cells negative for BLV infection (C72, M23, and M26); 4 H9 cells transfected with plasmid pRSV-tax1 with the HTLV-1 tax gene deleted; 4 H9 cells transfected with plasmid pRSV-tax2 with the HTLV-2 tax gene deleted; 4 H9 cells transfected with plasmid pBLV-tax with the BLV tax gene deleted; 4 H9 cells transfected with BlueScribe plasmid. Fig. 3. Comparison of markers of DNA damage in cell lines infected with various oncogenic viruses or transfected with tax from human T-cell leukemia virus (HTLV) or bovine leukemia virus (BLV). A) Frequency of micronuclei. Each data point represents the mean number of micronuclei per 1000 cells (duplicate experiments) for each cell line. B) Proportion of DNA in singlestranded conformation. Each data point represents the mean value of triplicate experiments. For both plots, each data point refers to a different cell line. The mean value for a given group is represented by the horizontal bar. Group means were compared by use of ScheffJs multiple comparison procedure, the Kruskal Wallis method, and the median counterpart of Fishers exact test. Cells infected or transfected with HTLV/BLV group viruses had a statistically significantly higher mean micronucleus frequency (A) and proportion of single-stranded DNA (B) [(1 F), where F is the proportion of double-stranded DNA; see section entitled Alkali Denaturation Assay in Materials and Methods] than uninfected cells, cells infected with other oncogenic viruses, cells transfected with the plasmid vectors with the tax gene deleted, and cells transfected with an irrelevant plasmid (pBlueScribe) not containing a tax gene (P<.003 for all comparisons). cause our experimental design took a fixed-effects approach, the inferences apply only to the cell lines that we used. Data in Fig. 4 (plasmid reactivation assays) were analyzed by the Dunnett t test (53). This is a multiple comparison method that uses a pooled standard deviation and allows the comparison of multiple values to a single control. The minimum significance level was P<.05. All tests were two-sided. RESULTS Frequency of Micronuclei in HTLV- and BLV-Infected/ Transfected Cells Initially, we investigated the ability of HTLV/BLV group viruses to induce chromosomal abnormalities by use of micronucleus formation as an indicator of DNA damage (47). Replicate values were obtained at different times for 36 cell lines. The mean number of micronuclei for each line was used for subsequent statistical manipulations (Fig. 3, A). Established cell lines infected with HTLV/BLV group viruses had statistically significantly higher frequencies of micronuclei per 1000 cells than uninfected cells or established cell lines infected with other types of oncogenic viruses (Scheffes method, P<.0001; both KruskalWallis test and Fishers exact test, P<.001). When assayed 48 hours after transfection, cell lines transfected with Taxencoding plasmids gave similar results (Fig. 3, A). Cells transfected with the tax-deleted plasmid vectors did not show an increase in micronucleus frequency. Role of Apoptosis and Chromosome Loss on Increased Frequency of Micronuclei Although micronuclei are generally used as a biomarker for DNA damage in genotoxic studies, micronuclei and micronucleus-like structures can arise through several mechanisms: apoptosis, whole-chromosome loss, or chromosomal breakage (47). We first examined the possibility that the increased rate of formation of micronuclei seen in infected cells was due to apoptosis. When analyzed by trypan blue dye exclusion, agarose gel electrophoresis, or cell death immunoassay, there was no significant difference between infected or transfected cells and uninfected or untransfected cells in the degree of apoptosis demFig. 4. Plasmid reactivation analysis of cellular DNA repair. Cells were transfected with normal CAT (chloramphenicol acetyltransferase) plasmid DNA or the same plasmid damaged by UV light, deoxyribonuclease I (DNase I), psoralen, quercetin, or hydrogen peroxide. DNA repair rates were measured as the ratio of CAT activity in extracts from cells transfected with a damaged plasmid to CAT activity in extracts from cells transfected with an undamaged plasmid. All experiments were run in triplicate. Chinese hamster ovary (CHO) cell line EM9, deficient in DNA strand break repair, was unable to repair quercetin and hydrogen peroxide damage (A), and control CHO cell line UV41, defective in nucleotide-excision repair, was unable to repair psoralen damage (A). Normal CHO cell line AA8 was able to repair plasmids with all types of damage (A). Repair rates of plasmids damaged by UV light, quercetin, or hydrogen peroxide were statistically significantly inhibited in cells transfected with the tax gene (B) or infected with human T-cell leukemia virus (HTLV) (C) and bovine leukemia virus (BLV) (D) (two-sided P<.01, Dunnett t test for all comparisons). Repair rates of damage induced by DNase I or psoralen were not inhibited (B, C, and D). onstrated. Furthermore, there was no increased sensitivity to various apoptosis-inducing treatments (serum starvation and addition of camptothecin) (data not shown) (48). We also eliminated the possibility that the increased rate of formation of micronuclei seen in infected cells was due to virusinduced mitotic loss of whole chromosomes. A modified micronucleus assay using an anti-kinetochore antibody was carried out to distinguish micronuclei with a kinetochore (marker for a whole chromosome) from micronuclei without (27,28). We saw no increase in the frequency of kinetochore-positive micronuclei in infected or tax-transfected cells (data not shown) (48). Frequency of DNA Strand Breaks in HTLV- and BLV-Infected/Transfected Cells To confirm that the micronuclei seen in cells infected with HTLV/BLV group viruses arose from DNA-damaging events, we used an alkali denaturation assay to detect the relative frequency of DNA strand breaks in various infected and uninfected cell lines. Commonly used in radiobiology (50), this method is based on the observation that the unwinding rate of DNA molecules in a mild alkali solution increases with prior exposure to ionizing agents. Since DNA strand breaks are believed to be responsible for this increase, the rate of unwinding can be used as a surrogate measure of strand breakage. This method has been used to study the effect of acute herpesvirus infection on the integrity of host cell DNA (49), and our experiments represent an application of this assay to persistently infected, transformed cells. The greater the rate of decrease of fluorescence after alkali treatment, the greater the presumed amount of singlestrandedness and DNA damage (or DNA repair intermediates) present. The mean of triplicate values for each cell line was used for subsequent statistical manipulations (Fig. 3, B). Cells infected with HTLV/BLV group viruses showed a statistically significant decrease in fluorescence (increase in singlestrandedness due to DNA damage) when compared with uninfected cell lines or with cells infected with other oncogenic viruses, including retroviruses known to transform cells by insertional mutagenesis or a viral oncogene (Scheffes method, P<.0001; KruskalWallis test and Fishers exact test, Monte Carlo approach, P<.001; Fishers exact test, asymptotic approach, P<.003). In vitro transfection experiments using Taxencoding plasmids gave similar results (Fig. 3, B). Cells transfected with tax-deleted plasmid vectors did not show a statistically significant difference from untransfected cells. Repair of DNA Damage Caused by UV Light, DNase, Psoralen, Quercetin, or Hydrogen Peroxide in HTLV- and BLV-Infected/Transfected Cells The micronucleus and alkali denaturation results raised the question of whether HTLV/BLV group viruses themselves cause DNA damage or whether they inhibit repair of spontaneous damage. To test the theory that Tax inhibited normal cellular DNA repair processes, a modified plasmid reactivation assay was used to measure repair rates in uninfected, infected, and tax-transfected cells. Cells will repair lesions in exogenous DNA, so it is possible to measure repair of damaged viral or plasmid DNA transfected into cells in order to assess inherent repair capacity (54). This assay has been used to examine the DNA repair capacity of a variety of normal and malignant cell types, as well as to demonstrate an association between diminished DNA repair proficiency and skin cancer susceptibility in individuals with XP (51,52,55,56). Gel electrophoresis of CAT reporter plasmids treated with the DNA-damaging agents indicated that, within the sensitivity of the gel assay, all of the DNA was converted from the undamaged form (supercoiled) to the damaged forms (linear, nicked, and adducted) with different mobilities (Fig. 2, A, B, and C). The results of experiments describing cellular DNA repair by plasmid reactivation analysis are described in Fig. 4, A, B, C, and D). Our initial experiments measured repair of UV light-induced DNA damage in HTLV/BLV-infected and uninfected cell lines. Following transfection with all types of damaged CAT reporter plasmids, uninfected H9 cells showed CAT enzyme activity equivalent to the level seen in cells transfected with the undamaged plasmid controls. In contrast, HTLV- and BLV-infected cell lines transfected with plasmids damaged by UV light, quercetin, and hydrogen peroxide showed a statistically significant decrease in CAT activity, indicating a diminished ability to repair these types of DNA damage (Fig. 4, C and D). There was no statistically significant decrease in their ability to repair damage induced by DNase or photoactivated psoralen (Fig. 4, C and D). The same patterns prevailed for cells transfected with plasmids containing either the BLV or the HTLV tax gene (Fig. 4, B). Cells transfected with tax-deleted plasmid vectors did not show diminished repair of DNA damaged by any of the five agents (data not shown). The mutant CHO cell line EM9, defective in DNA strand break repair, showed significantly decreased repair rates of DNA damaged by UV light, quercetin, or hydrogen peroxide as compared with the normal parental cell line AA8 (Fig. 4, A). The mutant cell line UV41, defective in nucleotideexcision repair, was unable to repair psoralen damage as compared with the normal parental cell line AA8 (Fig. 4, A). These controls indicated that our plasmid reactivation system was working correctly. DISCUSSION Our data indicating an increased frequency of micronuclei in cells infected with viruses of the HTLV/BLV group or transfected with their tax genes corroborate the work of Majone et al. (27) who used cells transfected with HTLV-1 Tax-encoding plasmids and extend it by showing, for the first time, similar results with infected cells and the transactivating Tax protein of other members of the HTLV/BLV group. In contrast to our results, the results obtained by Semmes et al. (57) did not indicate increased micronucleus frequency in HTLV-2 Taxtransfected cells. This discrepancy may be due to differences in the cell type used for transfection. We used a T-cell line, the cell type HTLV-2 normally infects, whereas Semmes et al. used intestinal epithelial cells, a cell type not known to be a target of HTLV pathogenesis. Only two non-HTLV/BLV cell lines (i.e., EM9 and Hep3B) showed significant increases in the frequencies of micronuclei and single-strandedness (decreased fluorescence) when compared with the background levels seen in uninfected cells. The EM9 CHO cell line is known to be defective in DNA strandbreak repair (58); it was included as a positive assay control. AA8, the normal parental line of EM9, and UV41, a sister clone, defective in nucleotide-excision repair (5860) did not show an increase in the frequency of formation of micronuclei or strand breakage. They were included as negative assay controls. Hep3B hepatocarcinoma cells are infected with hepatitis B virus (HBV) (30); our observations, therefore, support recent observations that HBV-infected cells are unusually susceptible to genotoxic damage. One study (20) found that the presence of HBV DNA alone increased the frequency of genetic recombination in hepatocarcinoma cells. Other researchers (6163) have reported data suggesting that the HBV X protein may be involved in inhibiting endogenous repair pathways. By using a variety of DNA-damaging agents in the plasmid reactivation assay, we sought to pinpoint the particular DNA repair pathway inhibited by the Tax protein of HTLVs and BLV. Mammalian cells use different biochemical pathways to repair different types of damage, including base-excision repair, nucleotide-excision repair, recombination, and direct reversal of damage by DNA ligase (54). Exposure of plasmid DNA to UV light can cause several types of damage. Most lesions are cyclobutane pyrimidine dimers and [64] photoproducts, but strong doses of UV radiation can also cause DNA cross-linking, strand breakage, and oxidative damage (54). To ascertain which of these repair pathways might be inhibited in Tax-transformed cells, we used chemical methods to selectively introduce specific lesions into the CAT reporter plasmid. DNase was used to introduce nicks into the reporter plasmid. These lesions are sealed by DNA ligase (64). Our data suggest that ligase-directed reversal of DNA damage is not inhibited by the Tax proteins of the HTLV/BLV group. Lesions that can be repaired by direct reversal of DNA damage are fairly rare; most lesions are repaired by processes requiring excision of damaged DNA. Photoactivated psoralen damages DNA by forming monoadducts and interstrand crosslinks; monoadducts are probably repaired by nucleotide excision, while the commonly accepted pathway for cross-link repair involves both nucleotide-excision repair and recombination (54,65). In humans, nucleotide-excision repair is the sole pathway for removing bulky adducts, such as pyrimidine dimers and psoralen adducts (66). This explains why cells from individuals with XP and other hereditary nucleotide-excision repair defects demonstrated decreased repair of psoralen damage in plasmid reactivation studies (54,67,68) similar to ours. Our data suggest that nucleotide-excision repair and recombinatorial repair of DNA damage are not significantly affected by BLV or HTLV Tax. The possibility exists, however, that alternative pathways could be responsible for the repair. Quercetin is a mutagenic flavonoid that damages DNA by introducing single- and double-stranded breaks (69). Although the mechanics of DNA strand break repair are still poorly understood, several studies (7072) have suggested that singlestranded breaks are repaired by base excision or nucleotide excision, whereas double-stranded breaks are repaired by homologous or illegitimate recombination. Quercetin is a mutagen only under aerobic conditions, suggesting involvement of reactive oxygen intermediates. H2O2 reacts with transition metal ions to generate hydroxyl radicals, which then react with the nitrogenous bases or the backbone-forming sugars of DNA to produce oxidized pyrimidines, oxidized purines, and singlestranded breaks (73). Free-radical quenchers inhibit quercetininduced DNA scission, thus supporting this model (69). Lesions induced by free radicals are corrected by base-excision repair. Our data on reactivation of plasmids damaged with quercetin and H2O2 suggest that the DNA repair pathway inhibited in cells infected with viruses of the HTLV/BLV group of retroviruses is normal base excision of oxidative lesions. Results showing no inhibition of psoralen-induced damage in these cells imply, but do not rule out, that nucleotide-excision repair and recombinatorial mechanisms are not affected. The transforming Tax protein characteristic of these viruses seems to be the primary determinant of the DNA repair inhibition, although further experiments (such as an in vitro acellular system to measure DNA repair in the absence and presence of recombinant Tax) are needed to confirm this observation. To our knowledge, this is the first report of inhibition of DNA repair by any retrovirus and offers a new explanation for the riddle of the mechanism of transformation by the HTLV/BLV group of retroviruses. These results open the door to future research that could focus on delineating the molecular mechanisms of Tax inhibition of DNA repair. Tax-induced decrease of b-polymerase production is one possibility for a mechanism (29). Alternatively, Tax might not directly inhibit base-excision repair. Rather, it might saturate (i.e., overwhelm) this pathway by somehow stimulating overproduction of metabolic oxidants. In future studies, it will perhaps be possible to identify exactly which Tax domain(s) might be responsible for the DNA repair inhibition that we have discovered and what cellular factors they might interact with. Most studies of retroviral oncogenesis have focused on the role of oncogenes, examining increased or aberrant expression of these genes by insertional mutagenesis or viral transduction. Although our study represents a dramatic departure from this approach, a DNA repair inhibition model does not exclude other models of Tax-induced cellular transformation. Our model simply questions whether the transformation process rests entirely on growth-regulatory mechanisms, as proposed by transactivation and tumor suppressor models. Instead, we believe that gross genomic alterations are also necessary for progression to a neoplastic state and may, in fact, be the initial event. Disturbance of oncogene and/or tumor suppressor gene function might represent a later step. Inhibition of DNA repair and interleukin 2-driven autostimulation could also work synergistically to render infected cells susceptible to multiple mutagenic changes. This model could also explain why HTLV- and BLV-induced malignancies occur in only about 1% of infected individuals and only after a long latent period. Inhibition of DNA repair might prime virally infected cells for secondary mutagenic events, but such events would be rare and would accumulate only over a long period of time. REFERENCES Present address: S. M. Philpott, Department of Infectious Disease, Wadsworth Center, New York State Department of Health, Albany. S. M. Philpott has been supported in part by the California chapter of Sigma Xi; by Public Health Service grant 8T2GM07127 from the National Institute of General Medical Sciences, National Institutes of Health, Department of Health and Human Services; and by the following University of California sources: Regents Fellowship and the School of Public Health Grossman Endowment, Wellness Fellowship, and Levine Award. We thank G. Niermann, J. Xiao, A. Kilani, S. Linn, and K. Radke for technical advice and critical comments. We also thank B. Weiser and H. Burger for reviewing manuscript drafts and M. Tarter for advice on statistics. We are grateful to G. Firestone for providing the GR cell line and the pRSV-CAT plasmid, W. Waschman for the pRSV-tax1 and pRSV-tax2 plasmids, L. Willems for the pBLV-tax plasmid, P. Duesberg for the Rous sarcoma virus-infected chick fibroblasts, T. Forte for the Hep3B cell line, M. McGrath for the M23 and M26 cell lines, and K. Radke for the Bat2Cl6 and Tb1Lu cell lines. Other cell lines were obtained from the former Naval Biosciences Laboratory (Oakland, CA) and the American Type Culture Collection (Manassas, VA). The following reagent was obtained through the AIDS Research and Reference Reagent Program, Division of Acquired Immunodeficiency Syndrome, National Institute of Allergy and Infectious Diseases, National Institutes of Health: human T-cell leukemia virus-1 K30 DNA (plasmid pK30) from T. Kindt. Manuscript received June 30, 1998; revised March 24, 1999; accepted April 5, 1999.


This is a preview of a remote PDF: https://jnci.oxfordjournals.org/content/91/11/933.full.pdf

Sean M. Philpott, Gertrude C. Buehring. Defective DNA Repair in Cells With Human T-Cell Leukemia/Bovine Leukemia Viruses: Role of tax Gene, JNCI Journal of the National Cancer Institute, 1999, 933-942, DOI: 10.1093/jnci/91.11.933