- Email: [email protected]
A human lymphoid leukemia cell line with a V(D)J recombinase-mediated deletion of hprt
Mutation Research 403 Ž1998. 113–125
A human lymphoid leukemia cell line with a Vž D/ J recombinase-mediated deletion of hprt Chun-Lin Chen, Michael H. Woo, Geoffrey A.M. Neale, Rakesh M. Goorha, James C. Fuscoe, Frederick G. Behm, Susan Mathew, Mary V. Relling ) Departments of Pharmaceutical Sciences, St. Jude Children’s Research Hospital, Memphis, TN 38101, USA Department of Virology and Molecular Biology, St. Jude Children’s Research Hospital, Memphis, TN 38101, USA Department of Pathology, St. Jude Children’s Research Hospital, Memphis, TN 38101, USA Colleges of Pharmacy and Medicine, UniÕersity of Tennessee, Memphis, TN 38101, USA EnÕironmental Carcinogenesis DiÕision, National Health and EnÕironmental Effects Research Laboratory, US EnÕironmental Protection Agency, Research Triangle Park, NC 27711, USA Received 9 January 1998; revised 2 February 1998; accepted 5 February 1998
Abstract Large deletions of exons 2 and 3 of the hprt gene are the most common type of hprt mutation in lymphocytes of newborn infants, and their frequency increases in cultured human T-lymphoid cells as a result of exposure to etoposide. Sequenced PCR products for these deletions are consistent with a VŽD.J recombinase-mediated mechanism underlying their genesis. Herein, we describe the isolation and characterization of an etoposide-induced mutant CEM cell line that is clonal for a VŽD.J recombinase-mediated exon 2 q 3 deletion. Human CCRF-CEM cells were exposed to 5 m M etoposide for 4 h, selected in 6-thioguanine, and an exon 2 q 3 deletion mutant was isolated through serial limiting dilution, using a PCR-based assay for detection of the exon 2 q 3 deletion. Untreated CEM cells and cells treated with 6-thioguanine alone were similarly subcultured. The exon 2 q 3 deletion-containing line was termed SJCEM808 and had a slightly longer doubling time than the control lines, tended to clump in suspension, and was characterized by cell membrane blebbing. Compared to the parent line, SJCEM808 had similar cytogenetic abnormalities, lower CD2, CD1, and CD10 expression, and negligible RAG-1 expression. However, RAG-1 expression was down-regulated in some untreated parental subclones following similar subculturing. The sequence of the exon 2 q 3 deletion mutation exhibited nucleotide insertions, and the breakpoints were adjacent to heptamer signal recognition sequences in intact hprt, consistent with a VŽD.J recombinasemediated mechanism underlying its genesis. There were no MLL gene or interlocus T-cell receptor ŽTCR. rearrangements. These results indicate that non-homologous recombination following etoposide treatment is neither necessarily accompanied by other large DNA rearrangements nor simply a pre-lethal event, and this cell line may serve as a useful tool for studying illegitimate VŽD.J recombinase-mediated deletions. q 1998 Elsevier Science B.V. All rights reserved. Keywords: Etoposide; VŽD.J recombinase; Mutagenesis; Deletion mutant; CCRF-CEM; Leukemia; RAG-1
) Corresponding author. Pharmaceutical Sciences, St. Jude Children’s Research Hospital, 332 N. Lauderdale, Memphis, TN 38101-0318, USA. Tel.: q1-901-495-2348; fax: q1-901-525-6869; E-mail: [email protected]
0027-5107r98r$19.00 q 1998 Elsevier Science B.V. All rights reserved. PII: S 0 0 2 7 - 5 1 0 7 Ž 9 8 . 0 0 0 6 2 - 1
114
C.-L. Chen et al.r Mutation Research 403 (1998) 113–125
1. Introduction VŽD.J recombinase is a well-characterized sitespecific mammalian recombinase system w1,2x. The genes RAG-1 and RAG-2 confer VŽD.J recombinase activity to otherwise negative cells w3x and have homology to topoisomerases w4,5x, one of which Žtopoisomerase II. is the pharmacologic target of the epipodophyllotoxin anticancer drug, etoposide. Characteristic signature sequence motifs indicative of inappropriate VŽD.J recombinase activity are evident in various translocation breakpoints associated with lymphoid malignancies, involving as many as 25% of de novo T-cell leukemias and lymphomas w6x, the majority of B-cell follicular lymphomas w7x, and in the tŽ4;11. and less clearly in the tŽ9;11. translocation involving the MLL gene w8,9x. Other sequencespecific recombinational motifs have also been noted in acute leukemias characterized by MLL translocations. Thus, illegitimate recombination is a key mechanism in carcinogenesis, and is likely involved in etoposide-induced secondary leukemias w10–13x. Although there is no standard test for mammalian illegitimate DNA recombination, illegitimate VŽD.J recombinase activity has been implicated in a high percentage of hprt mutations in lymphocytes isolated from normal newborns w14x, and was detected at a higher frequency in a single patient treated with etoposide w15x. We demonstrated w16x that a single 4-h treatment with etoposide significantly increased the frequency of hprt exon 2 q 3 deletions in human leukemic CCRF-CEM cells in a concentration-dependent fashion, as early as 4 h after treatment. We hypothesized that this etoposide-induced illegitimate DNA recombination was not necessarily a lethal event, because an equally cytotoxic but longer exposure to etoposide resulted in a lower ultimate Žat 6 days. frequency of hprt exon 2 q 3 deletions w17x. This hypothesis is supported by the expansion of human lymphocytes containing the hprt exon 2 q 3 deletion w14,18,19x. Herein, we demonstrate that the hprt exon 2 q 3 deletions generated by etoposide can be maintained stably in progeny indefinitely, because cells containing the exon 2 q 3 deletion could be cloned by simple limiting dilution techniques. Moreover, the deletion was not associated with other cytogenetic changes or with other VŽD.J recombinase-mediated
non-homologous recombination. This is the first report of a cell line which is clonal for a VŽD.J recombinase-mediated deletion in hprt. 2. Materials and methods 2.1. Materials All hprt primers were synthesized by the Center for Biotechnology at St Jude Children’s Research Hospital. Nucleotide numbering for the hprt gene was as given by Edwards et al. w20x. Tris base, 2-mercaptoethanol, Ficoll 400, placental genomic DNA, and other related chemicals were purchased from Sigma ŽSt. Louis, MO.. Nucleotides ŽdATP, dTTP, dCTP, dGTP. were purchased from Pharmacia Biotech ŽPiscataway, NJ.. Taq DNA polymerase, HindIII, and Pst I were obtained from Promega ŽMadison, WI.. The 123-bp DNA ladder and BamHI were from Life Technologies ŽGaithersburg, MD.. Calf thymus DNA, used as the standard for quantitation of DNA, and the TKO 100 mini-fluorometer were from Hoefer Scientific Instruments ŽSan Francisco, CA.. PCR water from Promega was used in all PCR reactions. The Qiagen Blood and Cell Culture DNA kits Žused for extraction of DNA for Southern blotting., and QIAmp kits Žused for DNA for exon 2 q 3 deletion assay. were obtained from Qiagen ŽSanta Clarita, CA.. Quick Spine G-25 Sephadex columns for purification of radiolabelled DNA were purchased from Boehringer Mannheim ŽIndianapolis, IN.. Cell culture medium ŽRPMI 1640. and Lglutamine were from BioWhittaker ŽWalkersville, MD.. The hprt cDNA ŽATCC 57057. Ž; 900 bp. was purchased from American Type Culture Collection ŽRockville, MD.; MLL cDNA was provided by Dr. James Downing ŽSt. Jude Children’s Research Hospital. and encompassed exons 5–11 Ž; 800 bp. w21x. Fetal bovine serum was purchased from Hyclone ŽLogan, UT. and was heat-inactivated at 568C for 30 min before use. An OmniGene temperature cycler ŽModel TR3 SM2. ŽHybaid, Woodbridge, NJ. was used throughout the study. GeneScreen Plus w hybridization transfer membranes were from NEN Life Science Products ŽBoston, MA.; PEI cellulose was from EM Separations ŽGibbstown, NJ.. Storage phosphor screens, PhosphoImager and SpotFindere were from Molecular Dynamics ŽSunnyvale, CA..
C.-L. Chen et al.r Mutation Research 403 (1998) 113–125
2.2. Limiting dilutions of cells Human lymphoid leukemia cells ŽCCRF-CEM and its derivatives. were cultured in log phase in RPMI 1640 supplemented with 2 mM glutamine and 10%
115
fetal bovine serum at 378C in a 5% CO 2-humidified incubator. Three sets of cells were serially diluted and subcultured: untreated CEM cells, CEM cells treated with 10 m grml 6-thioguanine, and CEM cells treated with etoposide and then with 6-thio-
Table 1 Limiting dilution results Average starting number of cells cultured per aliquot Žwell. a
Number of wells found to have any deletions per number of wells submitted for assay b
Number of replicates positive for the exon 2 q 3 deletion per number of replicates assayed Žamount of DNA per PCR replicate. c
Approximate hprt exon 2 q 3 deletion frequency per cell d
10,000
1r20
; 8.67 = 10y5
4000
1r36
1250 250
1r19 1r78
50
1r98
1
1r407
0.04
1r341
0.04
12r12
4r4 Žat 1 m g. 4r4 Žat 0.5 m g. 1r4 Žat 0.1 m g. 2r2 Žat 0.4 m g. 2r2 Žat 0.2 m g. 1r2 Žat 0.007 m g. 2r2 Žat 0.006 m g. 2r2 Žat 0.003 m g. 1r2 Žat 0.001 m g. 1r2 Žat 0.002 m g. 2r2 Žat 0.001 m g. 1r2 Žat 0.0005 m g. 1r2 Žat 0.00025 m g. 4r4 Žat 800 pg. 1r4 Žat 160 pg. 3r4 Žat 80 pg. 0r4 Žat 40 pg. 20r20 Žat 20 pg. 48r50 Žat 10 pg. 11r20 Žat 6 pg. 3r10 Žat 5 pg. duplicate samples from each well at approximately 50 pg; all tested positive
a
; 1.67 = 10y4 ; 6.06 = 10y4 ; 4.33 = 10y3
; 1.7 = 10y2
; 0.1
;1
not estimated
This is the number of cells which were cultured in each of multiple aliquots. In general, the rough estimate from the previous dilution’s deletion frequency predicted that one of 10 ‘wells’ should have been positive following this dilution. In practice, more than 10 aliquots had to be checked to find a positive aliquot. Cells were subcultured until an adequate number were available for DNA extraction and subsequent analysis by PCR. b This is the number of wells Žor aliquots. found to have exon 2 q 3 deletions followed by the number of wells from which DNA was submitted for the exon 2 q 3 deletion assay. The amount of DNA used per PCR replicate at this screening stage was generally equal to the largest amount listed in the third column Žsee table note c.. As expected, only at the final ‘dilution’ step were all aliquots checked positive for the exon 2 q 3 deletion. c When the screening step indicated a positive aliquot of cells, multiple PCR replicates were assayed using DNA from that same aliquot Žusually at more than one amount of DNA per PCR replicate. to obtain a rough estimate of the deletion frequency at that level of cell dilution. Examples of the number of replicates positive for the exon 2 q 3 deletion out of the total number of replicates assayed at specific DNA amounts per PCR replicate are indicated Žnot all data are shown.. The estimate of the deletion frequency was used to decide on the number of cellsrwell to be aliquoted at the following dilution step. d Using Poisson statistics w16x and making rough assumptions of the proportion of null replicates Ž P0 . when all replicates at a given DNA concentration were positive Že.g., P0 assumed to be s 0.01 if 2 of 2 positive., the ‘average’ approximate hprt exon 2 q 3 deletion frequency per cell was estimated, only for purposes of deciding on cell dilutions for the next step.
116
C.-L. Chen et al.r Mutation Research 403 (1998) 113–125
guanine. For the last group, CEM cells in log phase were treated with 5 m M etoposide for 4 h, and subcultured in drug-free medium for 6 days, which caused an increased frequency of the exon 2 q 3 deletion as described w16x. The treated cells were partially enriched for the exon 2 q 3 deletion by selection for hprt mutants, using 10 m grml 6-thioguanine throughout the remainder of the limiting dilution experiments. The dilution steps required to clone a pure population of cells are outlined in Table 1.
GTGCTAAGTGCTAGAGTTACGGC, nucleotide 2658–2680., 50 mM Tris–HCl ŽpH 9.5., 15 mM ŽNH 4 . 2 SO4 , 2 mM MgCl 2 , 1.5 U of Taq polymerase, 0.25 mM each dNTP, and varying amounts of DNA, in a total of 50 m l. After layering 50 m l of mineral oil, the reactions were denatured at 948C for 4 min and then submitted to 35 cycles of denaturation Ž948C, 1 min., annealing Ž558C, 45 s., and elongation Ž728C, 1 min., and final extension at 728C for 7 min. Using these conditions, intact hprt Žwithout exon 2 q 3 deletion. yielded a fragment of 586 bp.
2.3. Assay of the hprt exon 2 q 3 deletion 2.5. Multiplex PCR for hprt exons DNA was used in PCR experiments directly without dialysis. A hemi-nested polymerase chain reaction ŽPCR. was performed using modifications of previously described procedures w14,15,22x to measure the frequency of the hprt exon 2 q 3 deletion. All conditions were as reported w16,17x, except that a new antisense primer, A23030, was used in place of A106. The first round of the PCR reaction was carried out by using sense primer A107 Ž5X-CAGTTTCCCGGGTTCGG., which anneals at nucleotides 1835–1851, and antisense primer A23030 Ž5XCCTGTCTATGGTCTCGATTCA., which anneals at nucleotides 23010–23030. The second round of the PCR reaction was performed by using sense primer A115 Ž5X-GTGCGATGGTGAGGTTCT., which anneals at nucleotides 2094–2111, and antisense A23030. DNA from cells with hprt exon 2 q 3 deletion Ždeletion of about 20 kb nucleotides. yields an 886 bp fragment, while DNA without hprt exon 2 q 3 deletion cannot be amplified under these conditions. The frequency of the exon 2 q 3 hprt deletion mutants was estimated using Poisson statistics as previously described w16x. 2.4. Detection of intact hprt without the exon 2 q 3 deletion In order to confirm that the exon 2 q 3 deletioncontaining clone was not contaminated with other hprt mutants, DNA was submitted to a PCR for intron 1, which would amplify only in cells with an intact intron 1 Ži.e., without an exon 2 q 3 deletion.. The PCR reaction contained 0.1 m M each of primer A115 and an antisense primer A2658 Ž5X-
To confirm that the exon 2 q 3 deletion was the only large hprt deletion, a multiplex PCR assay using 16 primers to amplify all nine exons of hprt was performed as previously reported w23,24x. Approximately 150 ng of DNA was used in each PCR and 30 cycles were performed. 2.6. Southern blotting for hprt deletions and for MLL rearrangements A total of 10 m g of cellular genomic DNA was digested overnight Žseparately. with PstI or HindIII Žfor hprt . or BamHI or SstI Žfor MLL., electrophoresed in 0.8% agarose, and transferred to nylon membranes under alkali solution Ž0.4 N NaOH and 0.6 N NaCl.. The membranes were pre-hybridized using a commercially available solution ŽGibco BRL, Grand Island, NY. for 4 h, hybridized overnight at 658C with hprt or MLL probes labelled via random primer amplification ŽRediPrimer kit, Amersham, Arlington Heights, IL. and washed at 658C with 2 = SSC solution for 15 min, 2 = SSC q 0.1% SDS for 30 min, and then 0.2 = SSC for 15 min. Phosphorimaging was used to detect hybridized DNA fragments. 2.7. Sequencing of hprt exons 2 q 3 deletion mutants PCR products were purified with the QIAquick Gel Extraction Kit, and sequenced using the cycle sequencing reaction employing fluorescence-tagged dye terminator ŽPRISM, Applied Biosystems. at the Center of Biotechnology of St. Jude Children’s Re-
C.-L. Chen et al.r Mutation Research 403 (1998) 113–125
search Hospital. Sequence information was analyzed using the University of Wisconsin Genetics Computer Group software package ŽMadison, WI.. 2.8. The hprt actiÕity The hprt activity was measured using a modification of the protocol of Zimm et al. w25x, measuring the conversion of radiolabelled hypoxanthine to inosine monophosphate ŽIMP.. The assay medium contained 0.1 M glycine buffer, pH 10.0, 1 mM phosphoribosyl pyrophosphate, 5 mM MgCl 2 , 0.15 mM 14 C-hypoxanthine, and 20 m l of cell lysate Žat least 5 = 10 3 cellsrm l. in a total volume of 100 m l. After centrifugation at 10,000 rpm for 30 s, the sample was incubated at room temperature for 15 min. The reaction was stopped by placing samples on ice and adding 5 m l of 0.25 M EDTA. The standard curve was made by spiking the above reaction medium on ice with 10 m l of 14 C IMP at varying concentrations and 5 m l of 0.25 M EDTA. A total of 10 m l of standards and samples were spotted on PEI cellulose, dried overnight, washed thrice in 1 mM ammonium bicarbonate and once in distilled water, allowed to dry, and counted in 20 ml of scintillation fluid. 2.9. Northern blot analysis for RAG gene expression Northern analysis was performed using probes for RAG-1 and RAG-2 as described w26x using 20 m g of total RNA. Quantitation of autoradiographs was done using a Kodak Bioimage densitometer and associated software. 2.10. Cell surface markers Cells were washed with PBS and incubated with 0.1% EDTA for 10 min at 378C. The cells were separated mechanically by gentle pipetting and then stained using a standard indirect immunofluorescence method w27x by incubation with a panel of monoclonal antibodies to lymphoid-associated antigens for 30 min at 48C, washed, incubated with FŽabX . 2 goat anti-mouse conjugated to fluoroscein isothiocyanate, and analyzed by flow cytometry ŽCoulter Elite, Miami Lakes, FL.. Studies included isotype matched monoclonal antibody and CD45 positive controls. Primary antibodies ŽBecton-Dick-
117
inson, San Jose, CA. included those directed against lymphoid-associated ŽCD19, 20, 21, 22, 7, 5, 2, 1, 3, 4, 8., myeloid-associated ŽCD13, 15, 33, 11b, 11c, 14., and non-specific ŽCD34, HLA-DR, and CD10. cell-surface antigens. 2.11. Chromosomal analysis Cytogenetic analysis was done according to standard methods by using a trypsin–Giemsa or a quinacrine fluorescence-banding technique. Aberrant karyotypes were described according to the International System for Human Cytogenetic Nomenclature w28x. Fluorescence in situ hybridization ŽFISH. was carried out using the MLL probe on 11q23 ŽOncor, Gaithersburg.. The MLL probe was labelled with digoxigenin and pre-mixed with blocking DNA and hybrisol VII Ž50% formamide, 2 = SSC.. The slides with fixed cells were treated in RNaser2= SSC for 1 h at 378C. Slides were denatured in 70% formamider2= SSC, pH 7.0 at 708C for 2 min and dehydrated immediately in cold Žy208C. 70, 80, and then 100% ethanol. Probe Ž10 m lrslide. was prewarmed at 378C and placed on the target DNA, covered with glass, sealed with rubber cement, and incubated in a humidified chamber overnight at 378C. Slides were washed three times 5 min each in 50% formamide 2 = SSC ŽpH 7.0. at 43–458C, followed by 2 = SSC washes for 10 min. Slides were rinsed in phosphate buffered detergent ŽPBD., and were immunocytochemically-stained using rhodaminelabelled anti-digoxigenin, rabbit-anti-sheep antibody, and rhodamine-labelled anti-rabbit antibody, with each stain separated by three PBD washes. Counterstaining was with 4,6-diamidino-2-phenylindole. A minimum of 200 interphase nuclei with intact morphology was screened using an Olympus fluorescence microscope with a triple band filter ŽChroma Technology, Brattleboro, VT.. 2.12. Electron microscopy Cell pellets were fixed in 2% glutaraldehyde in 0.1 M cacodylate buffer ŽpH 7.4. followed by postfixation with 1% osmium tetroxide in the same buffer. Samples were dehydrated in graded series of ethanols and embedded in Spurr’s low viscosity embedding medium. Ultrathin sections were cut with a
118
C.-L. Chen et al.r Mutation Research 403 (1998) 113–125
diamond knife on a Sorvall MT 6000 ultramicrotome and stained with Reynold’s lead citrate and uranyl acetate. The sections were examined in a JEOL 1200 EX II microscope operated at 60 kV. 2.13. T-cell receptor (TCR) gene rearrangement and interlocus recombination assay TCR-b rearrangements in DNA from parental CEM cells, CEM cells treated with 6-thioguanine, SJCEM962, and SJCEM808 were assessed by Southern blot as previously described w29x. Interlocus recombination between TCR Vg and TCR Jb was analyzed as described w30x.
3. Results 3.1. Limiting dilution results Parental CEM cells had no detectable hprt exon 2 q 3 deletion Ždeletion frequency - 2.5 = 10y8 .. After a single treatment with etoposide and subculture for 6 days, the hprt exon 2 q 3 deletion frequency was significantly increased to 1.67 = 10y7 w16x. Further selection for hprt mutants in 6-thioguanine increased the frequency to 1.11 = 10y6 . Limiting dilution experiments were carried out, using the PCR assay for the exon 2 q 3 deletion to identify positive subpopulations for further division and testing, until a pure population of cells exhibiting the hprt exon 2 q 3 deletion was obtained. The dilution steps are summarized in Table 1.
3.2. Purity of the clone At a density of 0.04 cells platedrwell, one positive clone Žtermed SJCEM808. was detected and was further subcloned at the same density of 0.04 cellsrwell in a 96-well plate. This clone ŽSJCEM808. was expanded, and its frequency of the exon 2 q 3 deletion was found to be similar to that predicted using Poisson statistics. The assumptions were that each PCR reaction was able to yield a positive PCR product if the deletion was present even at a density as low as 1 copy per PCR replicate, that each cell contained 6.25 pg DNA, and that every cell present carried the exon 2 q 3 deletion. For example, three out of 10 replicates Ž30%. were found to be positive for the exon 2 q 3 deletion using 5 pg DNA from SJCEM808 per PCR reaction, and the Poisson’s distribution predicts 55% of replicates at this dilution should be positive; 20 of 20 replicates Ž100%. were found to be positive using 20 pg DNA from SJCEM808 per PCR reaction, and the Poisson’s distribution predicts 96% of replicates at this dilution to be positive ŽTable 1.. These results suggest that all cells in this clone have the hprt exon 2 q 3 deletion. Using identical PCR conditions, there was no exon 2 q 3 deletion PCR product detected after testing at least 20 replicates, using 2.5 m g DNA per replicate, from the parental CEM cells, CEM cells treated with 6-thioguanine, and a sister subclone ŽSJCEM962. of CEM cells Žisolated along with SJCEM808 at the initial 0.04 cellsrwell dilution step. which were also treated with etoposide, 6-thioguanine, and stepwise limiting dilutions ŽFig. 1.. Thus, the appropriate
Fig. 1. Ethidium bromide-stained agarose gel of electrophoresed PCR products for the exon 2 q 3 deletion of hprt. From left to right are five replicates of PCR products using 2.5 m g DNA from the parental CEM cells; 2.5 m g DNA from the parental CEM cells treated with 6-thioguanine ŽCEMq TG.; 2.5 m g DNA from a sister subclone Ž962. which was isolated concurrently with SJCEM808 following treatment of CEM cells with etoposide, 6-thioguanine, and stepwise limiting dilutions; a molecular weight marker lane; and five replicates each of only 100 or 10 pg DNA per PCR from the SJCEM808 cell line Ž808.. The exon 2 q 3 deletion was present in five of five replicates at 100 pg and four of five replicates at 10 pg per PCR reaction. Replacement of water for DNA in PCR reactions yielded no PCR products.
C.-L. Chen et al.r Mutation Research 403 (1998) 113–125
119
Fig. 2. Ethidium bromide-stained agarose gel of electrophoresed PCR products for intact intron 1 of hprt Žindicating the lack of an exon 2 q 3 deletion.. From left to right, duplicates of progressively smaller amounts of DNA per PCR replicate from parental CEM cells were tested; PCR products for intron 1 are barely visible at 1 ng DNA per replicate, but easily detectable at 5 ng. Detectable PCR products are evident using 500 and 67 ng of DNA from CEM cells treated with 6-thioguanine ŽCEMq TG., from 67 ng DNA from an early selection passage of CEM cells treated with 6-thioguanine and etoposide ŽCV6., and from 67 ng DNA from two sister clones Ž150 and 616. of SJCEM808. However, no PCR product is evident using 67 ng Žshown. or even 2.5 m g Žnot shown. DNA from SJCEM808.
control lines exhibited hprt exon 2 q 3 deletion frequencies less than 1.28 = 10y7 . In a PCR reaction designed to amplify hprt only in cells with intact exons 2 and 3, the assay was sensitive to 5 ng, i.e., 5 ng DNA from parental CEM cells reliably yielded a PCR product in an ethidium bromide-stained agarose gel ŽFig. 2.. Using 67 ng DNA, detectable PCR products indicating intact ex-
ons 2 and 3 were observed when the source was parental CEM cells, CEM cells treated with 6-thioguanine, early selection passages of CEM cells treated with 6-thioguanine and etoposide ŽCV6., and two sister clones of SJCEM808, SJCEM150 and 616 ŽFig. 2.. However, no PCR product for intact exons 2 and 3 was detected from SJCEM808 even using as much as 2.5 m g DNA per PCR replicate.
Fig. 3. Ethidium bromide-stained agarose gel of the electrophoresed products of multiplex PCR amplification of hprt exons Žpositions indicated along right side of gel., using DNA from parental CEM and from individual sister clones Ž974–983, 951, 972, and 973. and four replicates of DNA from SJCEM808 cell line. All exons were present for all of the clones examined, except for SJCEM808, which clearly lacked both exons 2 and 3.
120
C.-L. Chen et al.r Mutation Research 403 (1998) 113–125
Fig. 4. Sequence of the PCR product following amplification for the hprt exon 2 q 3 deletion product Žmutation C. w16x, using DNA from SJCEM808. The corresponding sequences of wild type intron 1 and intron 3 and the sequence for SJCEM808 are indicated across the top, bottom, and middle of the figure, respectively. The breakpoints in mutation C were consistent with the inappropriate action of VŽD.J recombinase in their genesis, as evidenced by their proximity to slightly mismatched heptamer and nonamer recognition signal sequences Žunderlined. and the presence of N-nucleotide insertions Žitalicized..
Multiplex PCR for hprt exons confirmed that SJCEM808 had deletions of exons 2 and 3, with the seven other exons intact. Parental CEM cells, CEM cells treated with 6-thioguanine, and other sister clones of SJCEM808 had intact hprt exons ŽFig. 3.. RFLP analysis of the same four cell lines with PstI or HindIII digestion and hybridization to the hprt cDNA probe was also consistent with only SJCEM808 having the hprt exon 2 q 3 deletion Ždata not shown.. 3.3. Identity of the hprt exon 2 q 3 deletion in clone SJCEM808 PCR products were sequenced at several steps throughout the limiting-dilution experiments, and all revealed the same sequence ŽFig. 4.. The breakpoints in the 886-bp PCR product were consistent with the inappropriate action of VŽD.J recombinase in the genesis of this mutation, as evidenced by the proximity to slightly mismatched heptamer and nonamer recognition signal sequences and the presence of N-nucleotide insertions. Not surprisingly, this exact mutation was one of several sequences observed 1 week following etoposide treatment w16,17x; it was also sequenced at several intermediate dilution steps prior to the single-cell cloning of SJCEM808. 3.4. The hprt actiÕity Parental CEM cells had an hprt activity of 4.32 nmolrhr10 6 cells, whereas CEM cells treated with 6-thioguanine, SJCEM962, and SJCEM808 had no detectable hprt activity Ž- 0.04 nmolrhr10 6 cells..
Fig. 5. Transmission electron micrographs showing the morphology of the parental CEM cells ŽA. and the mutant SJCEM808 cells ŽB..
C.-L. Chen et al.r Mutation Research 403 (1998) 113–125
3.5. Growth characteristics
121
3.7. T-cell receptor gene status and RAG gene expression
The parental CEM cells grew in suspension with a doubling time of 19 h. SJCEM808 had a slightly slower doubling time of 24.2 h, and tended to grow in adherent clumps, which was consistent with the prominent cell surface protrusions detected with electron microscopy ŽFig. 5.. Morphologically, the parental CEM cells were spherical with a large nucleus and sparse cytoplasm typical of T-lymphocytes; the mutant SJCEM808 cells had a similar morphology except that they possessed many more microvilli than the parental CEM cells ŽFig. 5.. 3.6. Immunologic cell surface markers The SJCEM808 clone differed from parental CEM cells, 6-thioguanine-treated CEM cells, and SJCEM962 cells in that it expressed no detectable CD2 or CD1 Žboth T-cell markers. and lower levels of CD10 antigen ŽcALLa. ŽTable 2.. All B-cell markers ŽCD19, CD22, CD20, and CD21. were negative in all cell lines tested.
There were no differences in TCR b chain gene rearrangements among SJCEM808 cells and parental CEM cells, 6-thioguanine-treated CEM cells, or the sister clone SJCEM962, indicating that all cell lines appeared clonal relative to their TCR rearrangements Ždata not shown.. In addition, no interlocus rearrangements between TCR Vg and Jb segments were detected Ždata not shown., indicating that aberrant recombination of endogenous Žalthough still inappropriate. target genes of VŽD.J recombinase do not necessarily accompany VŽD.J recombinase-mediated deletions of hprt. Interestingly, RAG-1 ŽFig. 6. and RAG-2 Ždata not shown. expression in SJCEM808 cells was negligible and was low in the sister clones SJCEM962, 981, and 982. RAG gene expression was much lower in these clones relative to the parental CEM cells and 6-thioguanine-treated CEM cells that had not been serially diluted and subcloned as had the derivative SJCEM808 and SJCEM962 lines ŽFig. 6.. To test whether RAG gene expression
Table 2 Percentage of cells positively reacting with antibodies directed to the indicated immunophenotypic markers Phenotypic markers
Parental CEM cells
CEMq TG
SJCEM962 subclone
SJCEM808 subclone
Nonspecific markers CD45 leukocyte Common Ag HLA-DR CD34 CD10 ŽcALLa.
99 0 4 25
99 0 9 74
99 0 0 21
99 0 0 4
T-cell markers CD7 CD5 CD2 CD1 CD3 CD4 CD8
98 98 73 24 12 79 32
98 99 61 19 2 97 5
98 98 98 14 15 97 0
99 99 0 0 5 66 0
Myelo r monocytic markers CD33 0 CD13 0 CD15 18 CD11b 0 CD11c 0 CD14 0
0 3 61 2 2 0
0 0 99 1 0 0
0 0 91 2 0 0
122
C.-L. Chen et al.r Mutation Research 403 (1998) 113–125
Fig. 6. Northern blot of RAG-1 and b-actin expression in parental CEM cells, 6-thioguanine-treated CEM cells ŽCEMqTG., the derivative SJCEM962 Ž962., SJCEM808 Ž808. lines, and several other sister clones Ž981, 982, 983, 974, and 975. isolated following etoposide treatment and prolonged subculturing and dilution. The ratio of RAG-1 to b-actin expression, set at 100 for CEM cells, is depicted below the b-actin signal for each lane.
was reduced following an extended period of laboratory culture and no etoposide exposure, the parental CEM and 6-thioguanine-treated CEM cells were serially diluted and eventually subcloned from single cells, using the identical dilution scheme as had been applied to the cells that eventually gave rise to SJCEM808 and SJCEM962 lines. RNA from multiple subclones was then tested for RAG-1 and RAG-2 expression. Of 10 untreated CEM subclones, six had high RAG expression, two had intermediate expression, and two had negligible expression. Of 10 6thioguanine-treated CEM subclones, three had high, four had intermediate, two had low, and one had negligible RAG expression. Of 11 etoposide q6thioguanine-treated subclones, none had high RAG expression, two had intermediate expression, five had low expression Žincluding SJCEM962., and four had negligible expression Žincluding SJCEM808.. 3.8. Cytogenetics and MLL status There were no significant gross chromosomal alterations in the SJCEM808 line compared to parental CEM cells. The parental CEM cells displayed the previously described w31x pericentric inversion 9 and accompanying deletion of 9p and extra chromosome 20 w32x, and an 11q deletion. All derivatives retained the inversion 9. The 6-thioguanine-treated CEM cells had the extra chromosome 20; the 11q deletion was
not apparent. The SJCEM962 subclone exhibited a terminal deletion of 2p and both chromosomes 19 were in derivative form. The SJCEM808 subclone was cytogenetically similar to the parental CEM cells, with invŽ9.Žqh., 11q y , and q20. RFLP analysis using the MLL cDNA probe and BamHI and Sst I digestion revealed no detectable rearrangements of MLL in the parental CEM cells, the 6-thioguanine-treated CEM cells, the SJCEM962, or the SJCEM808 clones. Lack of MLL rearrangement was also supported by FISH, which demonstrated only the two hybridization signals expected in cells lacking rearrangement of the gene.
4. Discussion The hprt sequence in the cell line derived herein displayed hallmark characteristics w2x indicative of a VŽD.J-recombinase-mediated coding joint. The breakpoints bracketing the approximately 20-kb deletion were just 5X and 3X of slightly mismatched heptamer and nonamerrheptamer recognition signal sequences, respectively, with the signal sequences and intervening exons 2 and 3 spliced out. In addition, the mutant cell line contained four non-templated N-nucleotides inserted at the junction of intron 1 and intron 3 sequence. We previously showed that such VŽD.J recombinase-mediated deletions in hprt as a result of etoposide exposure were concentration-dependent w16,17x. It has been suggested that etoposide-induced non-homologous recombination is inseparable from its cytotoxic effects, such that inappropriate DNA rearrangement would constitute a collateral, pre-lethal mutation. However, the dissociation between etoposide-induced exon 2 q 3 deletions and cytotoxicity observed with more prolonged etoposide exposure w17x, and the fact that the deletion frequency appeared to plateau Žrather than decrease. between 2 and 6 days following short-term etoposide exposure w16x, suggests that etoposide’s ability to induce exon 2 q 3 deletions is not necessarily linked with lethal DNA rearrangements. Moreover, the same type of exon 2 q 3 deletion has been reported to occur in human T-lymphocytes which have been cloned and expanded ex vivo w14,18,19x. Herein, we have cloned a pure population of human lymphoid cells exhibiting a unique hprt exon 2 q 3 deletion
C.-L. Chen et al.r Mutation Research 403 (1998) 113–125
that arose following etoposide treatment. Our results demonstrate that this deletion of over 20 kb from hprt is not a lethal event, and therefore may represent a stable and unique indicator of non-homologous DNA recombination. The reason this is an important point is that etoposide-induced acute myeloid leukemia presumably occurs through similar types of site-specific non-homologous recombination w10–13x in a hematopoietic cell which survives to be transformed and immortalized. Although it has been shown previously w33,34x that etoposide can cause large gene rearrangements acutely, we have shown that those rearrangements demonstrate site-specificity and can survive in cell progeny indefinitely. Expression of RAG-1 and RAG-2 is necessary to confer VŽD.J recombinase activity w1,3,35x. We chose CEM cells as the model system in which to evaluate etoposide’s capacity to induce illegitimate VŽD.J recombination because CEM cells are known to express RAG-1 and RAG-2 and to be capable of carrying out VŽD.J recombination w26,36x. Interestingly, clone SJCEM808, with the exon 2 q 3 deletion, exhibited a complete lack of RAG-1 expression. However, it has been suggested that RAG-1 expression may decrease in cells undergoing prolonged laboratory culture w2,37x. Our data confirm that loss of RAG-1 expression in culture is possible: in subclones of serially diluted untreated parental CEM cells that underwent prolonged laboratory culture, RAG-1 expression was decreased in 40% of subclones. Interestingly, expression was normal in 60% of clones, demonstrating the heterogeneity of RAG expression within this commonly used cell line. These cells were not treated with etoposide, and did not exhibit the hprt exon 2 q 3 deletion. Thus, the loss of RAG gene expression in CEM clones isolated after treatment with etoposide is likely due to the prolonged isolation and culture conditions. We did not find that etoposide altered the expression of RAG genes acutely: RAG-1 expression remained high and unaltered Žby Northern blot analysis. in CEM cells over the 72 h immediately following 4 h of etoposide treatment Ždata not shown.. Therefore, although RAG gene expression may be necessary to induce illegitimate VŽD.J recombination, this expression can subsequently be lost in cells carrying the hprt exon 2 q 3 deletion. Interlocus recombination between TCR genes has
123
been proposed as an indicator of non-homologous VŽD.J recombination in mammalian cells in vitro and in vivo w30,38x. Our analysis demonstrated no interlocus recombination in these cells, SJCEM808, that clearly had undergone VŽD.J recombinase-mediated deletions of an illegitimate DNA substrate, hprt. This is not surprising, however, in that it seems highly unlikely that two unrelated targets for illegitimate VŽD.J recombination, TCR genes and hprt, would undergo rearrangement in the exact same cell. If the two events are independent, the likelihood of observing the two rearrangements in the same cell line would be on the order of 1 in 10 11 Žassuming a frequency of 10y4 for interlocus TCR rearrangement and 10y7 for hprt exon 2 q 3 deletions.. Likewise, the MLL gene Ža putative target for etoposide-induced non-homologous recombination. was not rearranged in the clone selected for the hprt exon 2 q 3 deletion. However, our findings do not rule out the possibility that MLL or other targets are rearranged as a result of etoposide exposure in cells which do not Žor only coincidentally. delete exons 2 q 3 of hprt. In fact, such MLL rearrangement does likely occur as a result of etoposide exposure in some percentage of treated cells w39x. The fact that the hprt exon 2 q 3 deletions were not associated with other genetic chromosomal rearrangements in the affected cell, or with other illegitimate VŽD.J recombinasemediated rearrangements, may have importance in assessing the mechanisms underlying leukemogenesis caused by etoposide. Our findings are supportive of etoposide being able to induce targeted, sitespecific gene rearrangements in a non-lethal fashion, rather than only promiscuous mutagenic effects. Whether there is any functional significance to the hprt exon 2 q 3 deletion, other than loss of hprt activity, is unknown. The SJCEM808 cells exhibited slower and more adherent growth characteristics and a greater degree of cell surface projections than the parental cells or than ‘sister clones’ of hprt deficient cells that did not exhibit the exon 2 q 3 deletion. The abundance of microvilli on the cell surface may have contributed to the cell–cell interactions responsible for the adherent phenotype observed in these mutant cells. The SJCEM808 clone also lost expression of some cell surface antigens associated with T-cell immunophenotype and had decreased expression of cALLa compared to parental cells.
124
C.-L. Chen et al.r Mutation Research 403 (1998) 113–125
VŽD.J recombinase-mediated DNA rearrangement acting on unnatural substrates may be an important biomarker of undesired recombinogenesis caused by anticancer drugs and other xenobiotics. Assessing such recombination in an endogenous gene substrate Že.g., hprt ., as opposed to a non-integrated vector substrate, permits an assessment of non-homologous recombination in normal cells Žsuch as circulating lymphocytes. on a transcribed gene DNA target in its complex chromatin milieu. Because transcription, chromatin interactions, and integration status have all been hypothesized to be important determinants of DNA cleavage and recombination w2,10,40–42x, VŽD.J recombinase-mediated rearrangement of hprt is being evaluated as an indicator of non-homologous DNA recombination w 16,17,19,22,43 x . SJCEM808 can serve as an useful source of positive control for PCR reactions aimed at detection of the VŽD.J recombinase-mediated deletion of exons 2 and 3 in hprt.
Acknowledgements We thank Ms. Pamela McGill, Ms. Natasha Lenchik, Ms. Ya Qin Chu, Ms. Eve Su, Ms. Margaret Needham, Ms. Susan Ragsdale, and the Cytogenetics, Electron Microscopy, and Molecular Resource Center shared resources of SJCRH for their excellent technical assistance. This article has been reviewed by the USEPA and approved for publication. Approval does not signify that the contents necessarily reflect the views or policies of the Agency, nor does mention of trade names or commercial products constitute endorsement or recommendation for use. This work is supported by Cancer Center CORE Grant CA-21765, NCI CA-51001, by a Center of Excellence grant from the State of Tennessee, and American Lebanese Syrian Associated Charities ŽALSAC..
References w1x S.M. Lewis, G.E. Wu, The origins of VŽD.J recombination, Cell 88 Ž1997. 159–162. w2x M. Gellert, Molecular analysis of VŽD.J recombination, Annu. Rev. Genet. 26 Ž1992. 425–446.
w3x M.A. Oettinger, D.G. Schatz, C. Gorka, D. Baltimore, RAG-1 and RAG-2, adjacent genes that synergistically activate VŽD.J recombination, Science 248 Ž1990. 1517–1523. w4x S. Kallenbach, T. Brinkmann, F. Rougeon, Rag-1: a topoisomerase?, Int. Immunol. 5 Ž1992. 231–232. w5x J.C. Wang, P.R. Caron, R.A. Kim, The role of DNA topoisomerases in recombination and genome stability: a doubleedged sword, Cell 62 Ž1990. 403–406. w6x L. Brown, Q. Cheng, M.J. Siciliano, W. Crist, G. Buchanan, R. Baer, Site-specific recombination of the tal-1 gene is a common occurrence in human T-cell leukemia, EMBO J. 9 Ž1990. 3343–3351. w7x B. Tycko, J. Sklar, Chromosomal translocations in lymphoid neoplasia: a reappraisal of the recombinase model, Cancer Cells 2 Ž1990. 1–8. w8x Y. Gu, G. Cimino, H. Alder, T. Nakamura, R. Prasad, O. Canaani, D.T. Moir, C. Jones, P.C. Nowell, C.M. Croce, E. Canaani, The Ž4;11.Žq21;q23. chromosome translocations in acute leukemias involve the VŽD.J recombinase, Proc. Natl. Acad. Sci. U.S.A. 89 Ž1992. 10464–10468. w9x M. Negrini, C.A. Felix, C. Martin, B.J. Lange, T. Nakamura, E. Canaani, C.M. Croce, Potential topoisomerase II DNA-binding sites at the breakpoints of a tŽ9;11. chromosome translocation in acute myeloid leukemia, Cancer Res. 53 Ž1993. 4489–4492. w10x P.L. Broeker, H.G. Super, M.J. Thirman, H. Pomykala, Y. Yonebayashi, S. Tanabe, N. Zeleznik-Le, J.D. Rowley, Distribution of 11q23 breakpoints within the MLL breakpoint cluster region in de novo acute leukemia and in treatment-related acute myeloid leukemia: correlation with scaffold attachment regions and topoisomerase II consensus binding sites, Blood 87 Ž1996. 1912–1922. w11x C.A. Felix, B.J. Lange, M.R. Hosler, J. Fertala, M. Bjornsti, Chromosome band 11q23 translocation breakpoints are DNA topoisomerase II cleavage sites, Cancer Res. 55 Ž1995. 4287–4292. w12x M.A. Smith, L. Rubinstein, R.S. Ungerleider, Therapy-related acute myeloid leukemia following treatment with epipodophyllotoxins: estimating the risks, Med. Pediatr. Oncol. 23 Ž1995. 86–98. w13x J.R. Downing, A.T. Look, MLL fusion genes in the 11q23 acute leukemias, Cancer Treat. Res. 84 Ž1996. 73–92. w14x J.C. Fuscoe, L.J. Zimmerman, M.J. Lippert, J.A. Nicklas, J.P. O’Neill, R.J. Albertini, VŽD.J recombinase-like activity mediates hprt gene deletion in human fetal T-lymphocytes, Cancer Res. 51 Ž1991. 6001–6005. w15x J.C. Fuscoe, L.J. Zimmerman, K. Harrington-Brock, L. Burnette, M.M. Moore, J.A. Nicklas, J.P. O’Neill, R.J. Albertini, VŽD.J recombinase-mediated deletion of the hprt gene in T-lymphocytes from adult humans, Mutat. Res. 283 Ž1992. 13–20. w16x C-L. Chen, J.C. Fuscoe, Q. Liu, M.V. Relling, Etoposide causes illegitimate VŽD.J recombination in human lymphoid leukemic cells, Blood 89 Ž1996. 1701–1707. w17x C-L. Chen, J.C. Fuscoe, Q. Liu, C.-H. Pui, H.H. Mahmoud, M.V. Relling, Relationship between cytotoxicity and sitespecific DNA recombination after in vitro exposure of
C.-L. Chen et al.r Mutation Research 403 (1998) 113–125
w18x
w19x
w20x
w21x
w22x
w23x
w24x
w25x
w26x
w27x
w28x w29x
w30x
leukemia cells to etoposide, J. Natl. Cancer Inst. 88 Ž1996. 1840–1847. M.J. Mcginniss, J.A. Nicklas, R.J. Albertini, Molecular analyses of in vivo hprt mutations in human T-lymphocytes: IV. Studies in newborns, Environ. Mol. Mutagen. 14 Ž1989. 229–237. B.A. Finette, T. Poseno, R.J. Albertini, VŽD.J recombinasemediated hprt mutations in peripheral blood lymphocytes of normal children, Cancer Res. 56 Ž1996. 1405–1412. A. Edwards, H. Voss, P. Rice, A. Civitello, J. Stegemann, C. Schwager, J. Zimmerman, H. Erfle, C.T. Caskey, W. Ansorge, Automated DNA sequencing of the human hprt locus, Genomics 6 Ž1990. 593–608. F.G. Behm, S.C. Raimondi, J.L. Frestedt, Q. Liu, W.M. Crist, J.R. Downing, G.K. Rivera, J.H. Kersey, C-H. Pui, Rearrangement of the MLL gene confers a poor prognosis in childhood acute lymphoblastic leukemia, regardless of presenting age, Blood 87 Ž1996. 2870–2877. J.C. Fuscoe, L.K. Vira, D.D. Collard, M.M. Moore, Quantification of hprt gene deletions mediated by illegitimate VŽD.J recombination in peripheral blood cells of humans, Environ. Mol. Mutagen. 29 Ž1997. 28–35. R.A. Gibbs, P.N. Nguyen, A. Edwards, A.B. Civitello, C.T. Caskey, Multiplex DNA deletion detection and exon sequencing of the hypoxanthine phosphoribosyltransferase gene in Lesch–Nyhan families, Genomics 7 Ž1990. 235–244. J.C. Fuscoe, L.J. Zimmerman, K. Harrington-Brock, M.M. Moore, Large deletions are tolerated at the hprt locus of in vivo derived human T-lymphocytes, Mutat. Res. 283 Ž1992. 255–262. S. Zimm, G. Reaman, R.F. Murphy, D.G. Poplack, Biochemical parameters of mercaptopurine activity in patients with acute lymphoblastic leukemia, Cancer Res. 46 Ž1986. 1495– 1498. G.A.M. Neale, T.J. Fitzgerald, R.M. Goorha, Expression of the VŽD.J recombinase gene RAG-1 is tightly regulated and involves both transcriptional and post-transcriptional control, Mol. Immunol. 29 Ž1992. 1457–1466. F.G. Behm, S.C. Raimondi, M.J. Schell, A.T. Look, G.K. Rivera, C-H. Pui, Lack of CD45 antigen on blast cells in childhood acute lymphoblastic leukemia is associated with chromosomal hyperdiploidy and other favorable prognostic features wsee commentsx, Blood 79 Ž1992. 1011–1016. F. Mitelman, ISCN95: An International System For Human Cytogenetic Nomenclature, S. Karger, Basel, 1995. T.J. Fitzgerald, G.A. Neale, S.C. Raimondi, R.M. Goorha, Rhom-2 expression does not always correlate with abnormalities on chromosome 11 at band p13 in T-cell acute lymphoblastic leukemia, Blood 80 Ž1992. 3189–3197. S. Lipkowitz, V.F. Garry, I.R. Kirsch, Interlocus V–J recom-
w31x
w32x
w33x
w34x
w35x
w36x
w37x
w38x
w39x
w40x
w41x
w42x
w43x
125
bination measures genomic instability in agriculture workers at risk for lymphoid malignancies, Proc. Natl. Acad. Sci. U.S.A. 89 Ž1992. 5301–5305. D.E. Moore, K. Weise, R. Zawydiwski, E.B. Thompson, The karyotype of the glucocorticoid-sensitive, lymphoblastic human T-cell line CCRF-CEM shows a unique deleted and inverted chromosome 9, Cancer Genet. Cytogenet. 14 Ž1985. 89–94. R. Zawydiwski, J.M. Harmon, E.B. Thompson, Glucocorticoid-resistant human acute lymphoblastic leukemic cell line with functional receptor, Cancer Res. 43 Ž1983. 3865–3873. J. Maraschin, B. Dutrillaux, A. Aurias, Chromosome aberrations induced by etoposide are not random, Int. J. Cancer 46 Ž1990. 808–812. N.A. Berger, S. Chatterjee, J.A. Schmotzer, S.R. Helms, Etoposide-induced gene alterations: potential contribution to cell death, Proc. Natl. Acad. Sci. U.S.A. 33 Ž1991. 8740– 8742. J.F. McBlane, D.C. van Gent, D.A. Ramsden, C. Romeo, C.A. Cuomo, M. Gellert, M.A. Oettinger, Cleavage at a VŽD.J recombination signal requires only RAG1 and RAG2 proteins and occurs in two steps, Cell 83 Ž1995. 387–395. J.F. Riou, L. Grondard, O. Petitgenet, M. Abitbol, F. Lavelle, Altered topoisomerase I activity and recombination activating gene expression in a human leukemia cell line resistant to doxorubicin, Biochem. Pharmacol. 46 Ž1993. 851–861. F.W. Alt, G. Rathbun, E. Oltz, G. Taccioli, Y. Shinkai, Function and control of recombination-activating gene activity, Ann. New York Acad. Sci. 651 Ž1992. 277–295. J.M. Abdallah, D.P. Lombardi, I.R. Kirsch, Genetic instability in patients with Hodgkin’s disease undergoing chemotherapy, J. Clin. Invest. 96 Ž1995. 2744–2747. P.D. Aplan, D.S. Chervinsky, M. Stanulla, W.C. Burhans, Site-specific DNA cleavage within the MLL breakpoint cluster region induced by topoisomerase II inhibitors, Blood 87 Ž1996. 2649–2658. M.T. Howard, M.P. Lee, T.S. Hsieh, J.D. Griffith, Drosophila topoisomerase II-DNA interactions are affected by DNA structure, J. Mol. Biol. 217 Ž1991. 53–62. A. Udvardy, P. Schedl, M. Sander, T.S. Hsieh, Topoisomerase II cleavage in chromatin, J. Mol. Biol. 191 Ž1986. 231–246. D.G. Schatz, M.A. Oettinger, M.S. Schlissel, VŽD.J recombination: molecular biology and regulation, Annu. Rev. Immunol. 10 Ž1992. 359–383. W.G. McGregor, V.M. Maher, J.J. McCormick, Kinds and locations of mutations induced in the hypoxanthine-guanine phosphoribosyltransferase gene of human T-lymphocytes by 1-nitrosopyrene, including those caused by VŽD.J recombinase, Cancer Res. 54 Ž1994. 4207–4213.
A human lymphoid leukemia cell line with a Vž D/ J recombinase-mediated deletion of hprt Chun-Lin Chen, Michael H. Woo, Geoffrey A.M. Neale, Rakesh M. Goorha, James C. Fuscoe, Frederick G. Behm, Susan Mathew, Mary V. Relling ) Departments of Pharmaceutical Sciences, St. Jude Children’s Research Hospital, Memphis, TN 38101, USA Department of Virology and Molecular Biology, St. Jude Children’s Research Hospital, Memphis, TN 38101, USA Department of Pathology, St. Jude Children’s Research Hospital, Memphis, TN 38101, USA Colleges of Pharmacy and Medicine, UniÕersity of Tennessee, Memphis, TN 38101, USA EnÕironmental Carcinogenesis DiÕision, National Health and EnÕironmental Effects Research Laboratory, US EnÕironmental Protection Agency, Research Triangle Park, NC 27711, USA Received 9 January 1998; revised 2 February 1998; accepted 5 February 1998
Abstract Large deletions of exons 2 and 3 of the hprt gene are the most common type of hprt mutation in lymphocytes of newborn infants, and their frequency increases in cultured human T-lymphoid cells as a result of exposure to etoposide. Sequenced PCR products for these deletions are consistent with a VŽD.J recombinase-mediated mechanism underlying their genesis. Herein, we describe the isolation and characterization of an etoposide-induced mutant CEM cell line that is clonal for a VŽD.J recombinase-mediated exon 2 q 3 deletion. Human CCRF-CEM cells were exposed to 5 m M etoposide for 4 h, selected in 6-thioguanine, and an exon 2 q 3 deletion mutant was isolated through serial limiting dilution, using a PCR-based assay for detection of the exon 2 q 3 deletion. Untreated CEM cells and cells treated with 6-thioguanine alone were similarly subcultured. The exon 2 q 3 deletion-containing line was termed SJCEM808 and had a slightly longer doubling time than the control lines, tended to clump in suspension, and was characterized by cell membrane blebbing. Compared to the parent line, SJCEM808 had similar cytogenetic abnormalities, lower CD2, CD1, and CD10 expression, and negligible RAG-1 expression. However, RAG-1 expression was down-regulated in some untreated parental subclones following similar subculturing. The sequence of the exon 2 q 3 deletion mutation exhibited nucleotide insertions, and the breakpoints were adjacent to heptamer signal recognition sequences in intact hprt, consistent with a VŽD.J recombinasemediated mechanism underlying its genesis. There were no MLL gene or interlocus T-cell receptor ŽTCR. rearrangements. These results indicate that non-homologous recombination following etoposide treatment is neither necessarily accompanied by other large DNA rearrangements nor simply a pre-lethal event, and this cell line may serve as a useful tool for studying illegitimate VŽD.J recombinase-mediated deletions. q 1998 Elsevier Science B.V. All rights reserved. Keywords: Etoposide; VŽD.J recombinase; Mutagenesis; Deletion mutant; CCRF-CEM; Leukemia; RAG-1
) Corresponding author. Pharmaceutical Sciences, St. Jude Children’s Research Hospital, 332 N. Lauderdale, Memphis, TN 38101-0318, USA. Tel.: q1-901-495-2348; fax: q1-901-525-6869; E-mail: [email protected]
0027-5107r98r$19.00 q 1998 Elsevier Science B.V. All rights reserved. PII: S 0 0 2 7 - 5 1 0 7 Ž 9 8 . 0 0 0 6 2 - 1
114
C.-L. Chen et al.r Mutation Research 403 (1998) 113–125
1. Introduction VŽD.J recombinase is a well-characterized sitespecific mammalian recombinase system w1,2x. The genes RAG-1 and RAG-2 confer VŽD.J recombinase activity to otherwise negative cells w3x and have homology to topoisomerases w4,5x, one of which Žtopoisomerase II. is the pharmacologic target of the epipodophyllotoxin anticancer drug, etoposide. Characteristic signature sequence motifs indicative of inappropriate VŽD.J recombinase activity are evident in various translocation breakpoints associated with lymphoid malignancies, involving as many as 25% of de novo T-cell leukemias and lymphomas w6x, the majority of B-cell follicular lymphomas w7x, and in the tŽ4;11. and less clearly in the tŽ9;11. translocation involving the MLL gene w8,9x. Other sequencespecific recombinational motifs have also been noted in acute leukemias characterized by MLL translocations. Thus, illegitimate recombination is a key mechanism in carcinogenesis, and is likely involved in etoposide-induced secondary leukemias w10–13x. Although there is no standard test for mammalian illegitimate DNA recombination, illegitimate VŽD.J recombinase activity has been implicated in a high percentage of hprt mutations in lymphocytes isolated from normal newborns w14x, and was detected at a higher frequency in a single patient treated with etoposide w15x. We demonstrated w16x that a single 4-h treatment with etoposide significantly increased the frequency of hprt exon 2 q 3 deletions in human leukemic CCRF-CEM cells in a concentration-dependent fashion, as early as 4 h after treatment. We hypothesized that this etoposide-induced illegitimate DNA recombination was not necessarily a lethal event, because an equally cytotoxic but longer exposure to etoposide resulted in a lower ultimate Žat 6 days. frequency of hprt exon 2 q 3 deletions w17x. This hypothesis is supported by the expansion of human lymphocytes containing the hprt exon 2 q 3 deletion w14,18,19x. Herein, we demonstrate that the hprt exon 2 q 3 deletions generated by etoposide can be maintained stably in progeny indefinitely, because cells containing the exon 2 q 3 deletion could be cloned by simple limiting dilution techniques. Moreover, the deletion was not associated with other cytogenetic changes or with other VŽD.J recombinase-mediated
non-homologous recombination. This is the first report of a cell line which is clonal for a VŽD.J recombinase-mediated deletion in hprt. 2. Materials and methods 2.1. Materials All hprt primers were synthesized by the Center for Biotechnology at St Jude Children’s Research Hospital. Nucleotide numbering for the hprt gene was as given by Edwards et al. w20x. Tris base, 2-mercaptoethanol, Ficoll 400, placental genomic DNA, and other related chemicals were purchased from Sigma ŽSt. Louis, MO.. Nucleotides ŽdATP, dTTP, dCTP, dGTP. were purchased from Pharmacia Biotech ŽPiscataway, NJ.. Taq DNA polymerase, HindIII, and Pst I were obtained from Promega ŽMadison, WI.. The 123-bp DNA ladder and BamHI were from Life Technologies ŽGaithersburg, MD.. Calf thymus DNA, used as the standard for quantitation of DNA, and the TKO 100 mini-fluorometer were from Hoefer Scientific Instruments ŽSan Francisco, CA.. PCR water from Promega was used in all PCR reactions. The Qiagen Blood and Cell Culture DNA kits Žused for extraction of DNA for Southern blotting., and QIAmp kits Žused for DNA for exon 2 q 3 deletion assay. were obtained from Qiagen ŽSanta Clarita, CA.. Quick Spine G-25 Sephadex columns for purification of radiolabelled DNA were purchased from Boehringer Mannheim ŽIndianapolis, IN.. Cell culture medium ŽRPMI 1640. and Lglutamine were from BioWhittaker ŽWalkersville, MD.. The hprt cDNA ŽATCC 57057. Ž; 900 bp. was purchased from American Type Culture Collection ŽRockville, MD.; MLL cDNA was provided by Dr. James Downing ŽSt. Jude Children’s Research Hospital. and encompassed exons 5–11 Ž; 800 bp. w21x. Fetal bovine serum was purchased from Hyclone ŽLogan, UT. and was heat-inactivated at 568C for 30 min before use. An OmniGene temperature cycler ŽModel TR3 SM2. ŽHybaid, Woodbridge, NJ. was used throughout the study. GeneScreen Plus w hybridization transfer membranes were from NEN Life Science Products ŽBoston, MA.; PEI cellulose was from EM Separations ŽGibbstown, NJ.. Storage phosphor screens, PhosphoImager and SpotFindere were from Molecular Dynamics ŽSunnyvale, CA..
C.-L. Chen et al.r Mutation Research 403 (1998) 113–125
2.2. Limiting dilutions of cells Human lymphoid leukemia cells ŽCCRF-CEM and its derivatives. were cultured in log phase in RPMI 1640 supplemented with 2 mM glutamine and 10%
115
fetal bovine serum at 378C in a 5% CO 2-humidified incubator. Three sets of cells were serially diluted and subcultured: untreated CEM cells, CEM cells treated with 10 m grml 6-thioguanine, and CEM cells treated with etoposide and then with 6-thio-
Table 1 Limiting dilution results Average starting number of cells cultured per aliquot Žwell. a
Number of wells found to have any deletions per number of wells submitted for assay b
Number of replicates positive for the exon 2 q 3 deletion per number of replicates assayed Žamount of DNA per PCR replicate. c
Approximate hprt exon 2 q 3 deletion frequency per cell d
10,000
1r20
; 8.67 = 10y5
4000
1r36
1250 250
1r19 1r78
50
1r98
1
1r407
0.04
1r341
0.04
12r12
4r4 Žat 1 m g. 4r4 Žat 0.5 m g. 1r4 Žat 0.1 m g. 2r2 Žat 0.4 m g. 2r2 Žat 0.2 m g. 1r2 Žat 0.007 m g. 2r2 Žat 0.006 m g. 2r2 Žat 0.003 m g. 1r2 Žat 0.001 m g. 1r2 Žat 0.002 m g. 2r2 Žat 0.001 m g. 1r2 Žat 0.0005 m g. 1r2 Žat 0.00025 m g. 4r4 Žat 800 pg. 1r4 Žat 160 pg. 3r4 Žat 80 pg. 0r4 Žat 40 pg. 20r20 Žat 20 pg. 48r50 Žat 10 pg. 11r20 Žat 6 pg. 3r10 Žat 5 pg. duplicate samples from each well at approximately 50 pg; all tested positive
a
; 1.67 = 10y4 ; 6.06 = 10y4 ; 4.33 = 10y3
; 1.7 = 10y2
; 0.1
;1
not estimated
This is the number of cells which were cultured in each of multiple aliquots. In general, the rough estimate from the previous dilution’s deletion frequency predicted that one of 10 ‘wells’ should have been positive following this dilution. In practice, more than 10 aliquots had to be checked to find a positive aliquot. Cells were subcultured until an adequate number were available for DNA extraction and subsequent analysis by PCR. b This is the number of wells Žor aliquots. found to have exon 2 q 3 deletions followed by the number of wells from which DNA was submitted for the exon 2 q 3 deletion assay. The amount of DNA used per PCR replicate at this screening stage was generally equal to the largest amount listed in the third column Žsee table note c.. As expected, only at the final ‘dilution’ step were all aliquots checked positive for the exon 2 q 3 deletion. c When the screening step indicated a positive aliquot of cells, multiple PCR replicates were assayed using DNA from that same aliquot Žusually at more than one amount of DNA per PCR replicate. to obtain a rough estimate of the deletion frequency at that level of cell dilution. Examples of the number of replicates positive for the exon 2 q 3 deletion out of the total number of replicates assayed at specific DNA amounts per PCR replicate are indicated Žnot all data are shown.. The estimate of the deletion frequency was used to decide on the number of cellsrwell to be aliquoted at the following dilution step. d Using Poisson statistics w16x and making rough assumptions of the proportion of null replicates Ž P0 . when all replicates at a given DNA concentration were positive Že.g., P0 assumed to be s 0.01 if 2 of 2 positive., the ‘average’ approximate hprt exon 2 q 3 deletion frequency per cell was estimated, only for purposes of deciding on cell dilutions for the next step.
116
C.-L. Chen et al.r Mutation Research 403 (1998) 113–125
guanine. For the last group, CEM cells in log phase were treated with 5 m M etoposide for 4 h, and subcultured in drug-free medium for 6 days, which caused an increased frequency of the exon 2 q 3 deletion as described w16x. The treated cells were partially enriched for the exon 2 q 3 deletion by selection for hprt mutants, using 10 m grml 6-thioguanine throughout the remainder of the limiting dilution experiments. The dilution steps required to clone a pure population of cells are outlined in Table 1.
GTGCTAAGTGCTAGAGTTACGGC, nucleotide 2658–2680., 50 mM Tris–HCl ŽpH 9.5., 15 mM ŽNH 4 . 2 SO4 , 2 mM MgCl 2 , 1.5 U of Taq polymerase, 0.25 mM each dNTP, and varying amounts of DNA, in a total of 50 m l. After layering 50 m l of mineral oil, the reactions were denatured at 948C for 4 min and then submitted to 35 cycles of denaturation Ž948C, 1 min., annealing Ž558C, 45 s., and elongation Ž728C, 1 min., and final extension at 728C for 7 min. Using these conditions, intact hprt Žwithout exon 2 q 3 deletion. yielded a fragment of 586 bp.
2.3. Assay of the hprt exon 2 q 3 deletion 2.5. Multiplex PCR for hprt exons DNA was used in PCR experiments directly without dialysis. A hemi-nested polymerase chain reaction ŽPCR. was performed using modifications of previously described procedures w14,15,22x to measure the frequency of the hprt exon 2 q 3 deletion. All conditions were as reported w16,17x, except that a new antisense primer, A23030, was used in place of A106. The first round of the PCR reaction was carried out by using sense primer A107 Ž5X-CAGTTTCCCGGGTTCGG., which anneals at nucleotides 1835–1851, and antisense primer A23030 Ž5XCCTGTCTATGGTCTCGATTCA., which anneals at nucleotides 23010–23030. The second round of the PCR reaction was performed by using sense primer A115 Ž5X-GTGCGATGGTGAGGTTCT., which anneals at nucleotides 2094–2111, and antisense A23030. DNA from cells with hprt exon 2 q 3 deletion Ždeletion of about 20 kb nucleotides. yields an 886 bp fragment, while DNA without hprt exon 2 q 3 deletion cannot be amplified under these conditions. The frequency of the exon 2 q 3 hprt deletion mutants was estimated using Poisson statistics as previously described w16x. 2.4. Detection of intact hprt without the exon 2 q 3 deletion In order to confirm that the exon 2 q 3 deletioncontaining clone was not contaminated with other hprt mutants, DNA was submitted to a PCR for intron 1, which would amplify only in cells with an intact intron 1 Ži.e., without an exon 2 q 3 deletion.. The PCR reaction contained 0.1 m M each of primer A115 and an antisense primer A2658 Ž5X-
To confirm that the exon 2 q 3 deletion was the only large hprt deletion, a multiplex PCR assay using 16 primers to amplify all nine exons of hprt was performed as previously reported w23,24x. Approximately 150 ng of DNA was used in each PCR and 30 cycles were performed. 2.6. Southern blotting for hprt deletions and for MLL rearrangements A total of 10 m g of cellular genomic DNA was digested overnight Žseparately. with PstI or HindIII Žfor hprt . or BamHI or SstI Žfor MLL., electrophoresed in 0.8% agarose, and transferred to nylon membranes under alkali solution Ž0.4 N NaOH and 0.6 N NaCl.. The membranes were pre-hybridized using a commercially available solution ŽGibco BRL, Grand Island, NY. for 4 h, hybridized overnight at 658C with hprt or MLL probes labelled via random primer amplification ŽRediPrimer kit, Amersham, Arlington Heights, IL. and washed at 658C with 2 = SSC solution for 15 min, 2 = SSC q 0.1% SDS for 30 min, and then 0.2 = SSC for 15 min. Phosphorimaging was used to detect hybridized DNA fragments. 2.7. Sequencing of hprt exons 2 q 3 deletion mutants PCR products were purified with the QIAquick Gel Extraction Kit, and sequenced using the cycle sequencing reaction employing fluorescence-tagged dye terminator ŽPRISM, Applied Biosystems. at the Center of Biotechnology of St. Jude Children’s Re-
C.-L. Chen et al.r Mutation Research 403 (1998) 113–125
search Hospital. Sequence information was analyzed using the University of Wisconsin Genetics Computer Group software package ŽMadison, WI.. 2.8. The hprt actiÕity The hprt activity was measured using a modification of the protocol of Zimm et al. w25x, measuring the conversion of radiolabelled hypoxanthine to inosine monophosphate ŽIMP.. The assay medium contained 0.1 M glycine buffer, pH 10.0, 1 mM phosphoribosyl pyrophosphate, 5 mM MgCl 2 , 0.15 mM 14 C-hypoxanthine, and 20 m l of cell lysate Žat least 5 = 10 3 cellsrm l. in a total volume of 100 m l. After centrifugation at 10,000 rpm for 30 s, the sample was incubated at room temperature for 15 min. The reaction was stopped by placing samples on ice and adding 5 m l of 0.25 M EDTA. The standard curve was made by spiking the above reaction medium on ice with 10 m l of 14 C IMP at varying concentrations and 5 m l of 0.25 M EDTA. A total of 10 m l of standards and samples were spotted on PEI cellulose, dried overnight, washed thrice in 1 mM ammonium bicarbonate and once in distilled water, allowed to dry, and counted in 20 ml of scintillation fluid. 2.9. Northern blot analysis for RAG gene expression Northern analysis was performed using probes for RAG-1 and RAG-2 as described w26x using 20 m g of total RNA. Quantitation of autoradiographs was done using a Kodak Bioimage densitometer and associated software. 2.10. Cell surface markers Cells were washed with PBS and incubated with 0.1% EDTA for 10 min at 378C. The cells were separated mechanically by gentle pipetting and then stained using a standard indirect immunofluorescence method w27x by incubation with a panel of monoclonal antibodies to lymphoid-associated antigens for 30 min at 48C, washed, incubated with FŽabX . 2 goat anti-mouse conjugated to fluoroscein isothiocyanate, and analyzed by flow cytometry ŽCoulter Elite, Miami Lakes, FL.. Studies included isotype matched monoclonal antibody and CD45 positive controls. Primary antibodies ŽBecton-Dick-
117
inson, San Jose, CA. included those directed against lymphoid-associated ŽCD19, 20, 21, 22, 7, 5, 2, 1, 3, 4, 8., myeloid-associated ŽCD13, 15, 33, 11b, 11c, 14., and non-specific ŽCD34, HLA-DR, and CD10. cell-surface antigens. 2.11. Chromosomal analysis Cytogenetic analysis was done according to standard methods by using a trypsin–Giemsa or a quinacrine fluorescence-banding technique. Aberrant karyotypes were described according to the International System for Human Cytogenetic Nomenclature w28x. Fluorescence in situ hybridization ŽFISH. was carried out using the MLL probe on 11q23 ŽOncor, Gaithersburg.. The MLL probe was labelled with digoxigenin and pre-mixed with blocking DNA and hybrisol VII Ž50% formamide, 2 = SSC.. The slides with fixed cells were treated in RNaser2= SSC for 1 h at 378C. Slides were denatured in 70% formamider2= SSC, pH 7.0 at 708C for 2 min and dehydrated immediately in cold Žy208C. 70, 80, and then 100% ethanol. Probe Ž10 m lrslide. was prewarmed at 378C and placed on the target DNA, covered with glass, sealed with rubber cement, and incubated in a humidified chamber overnight at 378C. Slides were washed three times 5 min each in 50% formamide 2 = SSC ŽpH 7.0. at 43–458C, followed by 2 = SSC washes for 10 min. Slides were rinsed in phosphate buffered detergent ŽPBD., and were immunocytochemically-stained using rhodaminelabelled anti-digoxigenin, rabbit-anti-sheep antibody, and rhodamine-labelled anti-rabbit antibody, with each stain separated by three PBD washes. Counterstaining was with 4,6-diamidino-2-phenylindole. A minimum of 200 interphase nuclei with intact morphology was screened using an Olympus fluorescence microscope with a triple band filter ŽChroma Technology, Brattleboro, VT.. 2.12. Electron microscopy Cell pellets were fixed in 2% glutaraldehyde in 0.1 M cacodylate buffer ŽpH 7.4. followed by postfixation with 1% osmium tetroxide in the same buffer. Samples were dehydrated in graded series of ethanols and embedded in Spurr’s low viscosity embedding medium. Ultrathin sections were cut with a
118
C.-L. Chen et al.r Mutation Research 403 (1998) 113–125
diamond knife on a Sorvall MT 6000 ultramicrotome and stained with Reynold’s lead citrate and uranyl acetate. The sections were examined in a JEOL 1200 EX II microscope operated at 60 kV. 2.13. T-cell receptor (TCR) gene rearrangement and interlocus recombination assay TCR-b rearrangements in DNA from parental CEM cells, CEM cells treated with 6-thioguanine, SJCEM962, and SJCEM808 were assessed by Southern blot as previously described w29x. Interlocus recombination between TCR Vg and TCR Jb was analyzed as described w30x.
3. Results 3.1. Limiting dilution results Parental CEM cells had no detectable hprt exon 2 q 3 deletion Ždeletion frequency - 2.5 = 10y8 .. After a single treatment with etoposide and subculture for 6 days, the hprt exon 2 q 3 deletion frequency was significantly increased to 1.67 = 10y7 w16x. Further selection for hprt mutants in 6-thioguanine increased the frequency to 1.11 = 10y6 . Limiting dilution experiments were carried out, using the PCR assay for the exon 2 q 3 deletion to identify positive subpopulations for further division and testing, until a pure population of cells exhibiting the hprt exon 2 q 3 deletion was obtained. The dilution steps are summarized in Table 1.
3.2. Purity of the clone At a density of 0.04 cells platedrwell, one positive clone Žtermed SJCEM808. was detected and was further subcloned at the same density of 0.04 cellsrwell in a 96-well plate. This clone ŽSJCEM808. was expanded, and its frequency of the exon 2 q 3 deletion was found to be similar to that predicted using Poisson statistics. The assumptions were that each PCR reaction was able to yield a positive PCR product if the deletion was present even at a density as low as 1 copy per PCR replicate, that each cell contained 6.25 pg DNA, and that every cell present carried the exon 2 q 3 deletion. For example, three out of 10 replicates Ž30%. were found to be positive for the exon 2 q 3 deletion using 5 pg DNA from SJCEM808 per PCR reaction, and the Poisson’s distribution predicts 55% of replicates at this dilution should be positive; 20 of 20 replicates Ž100%. were found to be positive using 20 pg DNA from SJCEM808 per PCR reaction, and the Poisson’s distribution predicts 96% of replicates at this dilution to be positive ŽTable 1.. These results suggest that all cells in this clone have the hprt exon 2 q 3 deletion. Using identical PCR conditions, there was no exon 2 q 3 deletion PCR product detected after testing at least 20 replicates, using 2.5 m g DNA per replicate, from the parental CEM cells, CEM cells treated with 6-thioguanine, and a sister subclone ŽSJCEM962. of CEM cells Žisolated along with SJCEM808 at the initial 0.04 cellsrwell dilution step. which were also treated with etoposide, 6-thioguanine, and stepwise limiting dilutions ŽFig. 1.. Thus, the appropriate
Fig. 1. Ethidium bromide-stained agarose gel of electrophoresed PCR products for the exon 2 q 3 deletion of hprt. From left to right are five replicates of PCR products using 2.5 m g DNA from the parental CEM cells; 2.5 m g DNA from the parental CEM cells treated with 6-thioguanine ŽCEMq TG.; 2.5 m g DNA from a sister subclone Ž962. which was isolated concurrently with SJCEM808 following treatment of CEM cells with etoposide, 6-thioguanine, and stepwise limiting dilutions; a molecular weight marker lane; and five replicates each of only 100 or 10 pg DNA per PCR from the SJCEM808 cell line Ž808.. The exon 2 q 3 deletion was present in five of five replicates at 100 pg and four of five replicates at 10 pg per PCR reaction. Replacement of water for DNA in PCR reactions yielded no PCR products.
C.-L. Chen et al.r Mutation Research 403 (1998) 113–125
119
Fig. 2. Ethidium bromide-stained agarose gel of electrophoresed PCR products for intact intron 1 of hprt Žindicating the lack of an exon 2 q 3 deletion.. From left to right, duplicates of progressively smaller amounts of DNA per PCR replicate from parental CEM cells were tested; PCR products for intron 1 are barely visible at 1 ng DNA per replicate, but easily detectable at 5 ng. Detectable PCR products are evident using 500 and 67 ng of DNA from CEM cells treated with 6-thioguanine ŽCEMq TG., from 67 ng DNA from an early selection passage of CEM cells treated with 6-thioguanine and etoposide ŽCV6., and from 67 ng DNA from two sister clones Ž150 and 616. of SJCEM808. However, no PCR product is evident using 67 ng Žshown. or even 2.5 m g Žnot shown. DNA from SJCEM808.
control lines exhibited hprt exon 2 q 3 deletion frequencies less than 1.28 = 10y7 . In a PCR reaction designed to amplify hprt only in cells with intact exons 2 and 3, the assay was sensitive to 5 ng, i.e., 5 ng DNA from parental CEM cells reliably yielded a PCR product in an ethidium bromide-stained agarose gel ŽFig. 2.. Using 67 ng DNA, detectable PCR products indicating intact ex-
ons 2 and 3 were observed when the source was parental CEM cells, CEM cells treated with 6-thioguanine, early selection passages of CEM cells treated with 6-thioguanine and etoposide ŽCV6., and two sister clones of SJCEM808, SJCEM150 and 616 ŽFig. 2.. However, no PCR product for intact exons 2 and 3 was detected from SJCEM808 even using as much as 2.5 m g DNA per PCR replicate.
Fig. 3. Ethidium bromide-stained agarose gel of the electrophoresed products of multiplex PCR amplification of hprt exons Žpositions indicated along right side of gel., using DNA from parental CEM and from individual sister clones Ž974–983, 951, 972, and 973. and four replicates of DNA from SJCEM808 cell line. All exons were present for all of the clones examined, except for SJCEM808, which clearly lacked both exons 2 and 3.
120
C.-L. Chen et al.r Mutation Research 403 (1998) 113–125
Fig. 4. Sequence of the PCR product following amplification for the hprt exon 2 q 3 deletion product Žmutation C. w16x, using DNA from SJCEM808. The corresponding sequences of wild type intron 1 and intron 3 and the sequence for SJCEM808 are indicated across the top, bottom, and middle of the figure, respectively. The breakpoints in mutation C were consistent with the inappropriate action of VŽD.J recombinase in their genesis, as evidenced by their proximity to slightly mismatched heptamer and nonamer recognition signal sequences Žunderlined. and the presence of N-nucleotide insertions Žitalicized..
Multiplex PCR for hprt exons confirmed that SJCEM808 had deletions of exons 2 and 3, with the seven other exons intact. Parental CEM cells, CEM cells treated with 6-thioguanine, and other sister clones of SJCEM808 had intact hprt exons ŽFig. 3.. RFLP analysis of the same four cell lines with PstI or HindIII digestion and hybridization to the hprt cDNA probe was also consistent with only SJCEM808 having the hprt exon 2 q 3 deletion Ždata not shown.. 3.3. Identity of the hprt exon 2 q 3 deletion in clone SJCEM808 PCR products were sequenced at several steps throughout the limiting-dilution experiments, and all revealed the same sequence ŽFig. 4.. The breakpoints in the 886-bp PCR product were consistent with the inappropriate action of VŽD.J recombinase in the genesis of this mutation, as evidenced by the proximity to slightly mismatched heptamer and nonamer recognition signal sequences and the presence of N-nucleotide insertions. Not surprisingly, this exact mutation was one of several sequences observed 1 week following etoposide treatment w16,17x; it was also sequenced at several intermediate dilution steps prior to the single-cell cloning of SJCEM808. 3.4. The hprt actiÕity Parental CEM cells had an hprt activity of 4.32 nmolrhr10 6 cells, whereas CEM cells treated with 6-thioguanine, SJCEM962, and SJCEM808 had no detectable hprt activity Ž- 0.04 nmolrhr10 6 cells..
Fig. 5. Transmission electron micrographs showing the morphology of the parental CEM cells ŽA. and the mutant SJCEM808 cells ŽB..
C.-L. Chen et al.r Mutation Research 403 (1998) 113–125
3.5. Growth characteristics
121
3.7. T-cell receptor gene status and RAG gene expression
The parental CEM cells grew in suspension with a doubling time of 19 h. SJCEM808 had a slightly slower doubling time of 24.2 h, and tended to grow in adherent clumps, which was consistent with the prominent cell surface protrusions detected with electron microscopy ŽFig. 5.. Morphologically, the parental CEM cells were spherical with a large nucleus and sparse cytoplasm typical of T-lymphocytes; the mutant SJCEM808 cells had a similar morphology except that they possessed many more microvilli than the parental CEM cells ŽFig. 5.. 3.6. Immunologic cell surface markers The SJCEM808 clone differed from parental CEM cells, 6-thioguanine-treated CEM cells, and SJCEM962 cells in that it expressed no detectable CD2 or CD1 Žboth T-cell markers. and lower levels of CD10 antigen ŽcALLa. ŽTable 2.. All B-cell markers ŽCD19, CD22, CD20, and CD21. were negative in all cell lines tested.
There were no differences in TCR b chain gene rearrangements among SJCEM808 cells and parental CEM cells, 6-thioguanine-treated CEM cells, or the sister clone SJCEM962, indicating that all cell lines appeared clonal relative to their TCR rearrangements Ždata not shown.. In addition, no interlocus rearrangements between TCR Vg and Jb segments were detected Ždata not shown., indicating that aberrant recombination of endogenous Žalthough still inappropriate. target genes of VŽD.J recombinase do not necessarily accompany VŽD.J recombinase-mediated deletions of hprt. Interestingly, RAG-1 ŽFig. 6. and RAG-2 Ždata not shown. expression in SJCEM808 cells was negligible and was low in the sister clones SJCEM962, 981, and 982. RAG gene expression was much lower in these clones relative to the parental CEM cells and 6-thioguanine-treated CEM cells that had not been serially diluted and subcloned as had the derivative SJCEM808 and SJCEM962 lines ŽFig. 6.. To test whether RAG gene expression
Table 2 Percentage of cells positively reacting with antibodies directed to the indicated immunophenotypic markers Phenotypic markers
Parental CEM cells
CEMq TG
SJCEM962 subclone
SJCEM808 subclone
Nonspecific markers CD45 leukocyte Common Ag HLA-DR CD34 CD10 ŽcALLa.
99 0 4 25
99 0 9 74
99 0 0 21
99 0 0 4
T-cell markers CD7 CD5 CD2 CD1 CD3 CD4 CD8
98 98 73 24 12 79 32
98 99 61 19 2 97 5
98 98 98 14 15 97 0
99 99 0 0 5 66 0
Myelo r monocytic markers CD33 0 CD13 0 CD15 18 CD11b 0 CD11c 0 CD14 0
0 3 61 2 2 0
0 0 99 1 0 0
0 0 91 2 0 0
122
C.-L. Chen et al.r Mutation Research 403 (1998) 113–125
Fig. 6. Northern blot of RAG-1 and b-actin expression in parental CEM cells, 6-thioguanine-treated CEM cells ŽCEMqTG., the derivative SJCEM962 Ž962., SJCEM808 Ž808. lines, and several other sister clones Ž981, 982, 983, 974, and 975. isolated following etoposide treatment and prolonged subculturing and dilution. The ratio of RAG-1 to b-actin expression, set at 100 for CEM cells, is depicted below the b-actin signal for each lane.
was reduced following an extended period of laboratory culture and no etoposide exposure, the parental CEM and 6-thioguanine-treated CEM cells were serially diluted and eventually subcloned from single cells, using the identical dilution scheme as had been applied to the cells that eventually gave rise to SJCEM808 and SJCEM962 lines. RNA from multiple subclones was then tested for RAG-1 and RAG-2 expression. Of 10 untreated CEM subclones, six had high RAG expression, two had intermediate expression, and two had negligible expression. Of 10 6thioguanine-treated CEM subclones, three had high, four had intermediate, two had low, and one had negligible RAG expression. Of 11 etoposide q6thioguanine-treated subclones, none had high RAG expression, two had intermediate expression, five had low expression Žincluding SJCEM962., and four had negligible expression Žincluding SJCEM808.. 3.8. Cytogenetics and MLL status There were no significant gross chromosomal alterations in the SJCEM808 line compared to parental CEM cells. The parental CEM cells displayed the previously described w31x pericentric inversion 9 and accompanying deletion of 9p and extra chromosome 20 w32x, and an 11q deletion. All derivatives retained the inversion 9. The 6-thioguanine-treated CEM cells had the extra chromosome 20; the 11q deletion was
not apparent. The SJCEM962 subclone exhibited a terminal deletion of 2p and both chromosomes 19 were in derivative form. The SJCEM808 subclone was cytogenetically similar to the parental CEM cells, with invŽ9.Žqh., 11q y , and q20. RFLP analysis using the MLL cDNA probe and BamHI and Sst I digestion revealed no detectable rearrangements of MLL in the parental CEM cells, the 6-thioguanine-treated CEM cells, the SJCEM962, or the SJCEM808 clones. Lack of MLL rearrangement was also supported by FISH, which demonstrated only the two hybridization signals expected in cells lacking rearrangement of the gene.
4. Discussion The hprt sequence in the cell line derived herein displayed hallmark characteristics w2x indicative of a VŽD.J-recombinase-mediated coding joint. The breakpoints bracketing the approximately 20-kb deletion were just 5X and 3X of slightly mismatched heptamer and nonamerrheptamer recognition signal sequences, respectively, with the signal sequences and intervening exons 2 and 3 spliced out. In addition, the mutant cell line contained four non-templated N-nucleotides inserted at the junction of intron 1 and intron 3 sequence. We previously showed that such VŽD.J recombinase-mediated deletions in hprt as a result of etoposide exposure were concentration-dependent w16,17x. It has been suggested that etoposide-induced non-homologous recombination is inseparable from its cytotoxic effects, such that inappropriate DNA rearrangement would constitute a collateral, pre-lethal mutation. However, the dissociation between etoposide-induced exon 2 q 3 deletions and cytotoxicity observed with more prolonged etoposide exposure w17x, and the fact that the deletion frequency appeared to plateau Žrather than decrease. between 2 and 6 days following short-term etoposide exposure w16x, suggests that etoposide’s ability to induce exon 2 q 3 deletions is not necessarily linked with lethal DNA rearrangements. Moreover, the same type of exon 2 q 3 deletion has been reported to occur in human T-lymphocytes which have been cloned and expanded ex vivo w14,18,19x. Herein, we have cloned a pure population of human lymphoid cells exhibiting a unique hprt exon 2 q 3 deletion
C.-L. Chen et al.r Mutation Research 403 (1998) 113–125
that arose following etoposide treatment. Our results demonstrate that this deletion of over 20 kb from hprt is not a lethal event, and therefore may represent a stable and unique indicator of non-homologous DNA recombination. The reason this is an important point is that etoposide-induced acute myeloid leukemia presumably occurs through similar types of site-specific non-homologous recombination w10–13x in a hematopoietic cell which survives to be transformed and immortalized. Although it has been shown previously w33,34x that etoposide can cause large gene rearrangements acutely, we have shown that those rearrangements demonstrate site-specificity and can survive in cell progeny indefinitely. Expression of RAG-1 and RAG-2 is necessary to confer VŽD.J recombinase activity w1,3,35x. We chose CEM cells as the model system in which to evaluate etoposide’s capacity to induce illegitimate VŽD.J recombination because CEM cells are known to express RAG-1 and RAG-2 and to be capable of carrying out VŽD.J recombination w26,36x. Interestingly, clone SJCEM808, with the exon 2 q 3 deletion, exhibited a complete lack of RAG-1 expression. However, it has been suggested that RAG-1 expression may decrease in cells undergoing prolonged laboratory culture w2,37x. Our data confirm that loss of RAG-1 expression in culture is possible: in subclones of serially diluted untreated parental CEM cells that underwent prolonged laboratory culture, RAG-1 expression was decreased in 40% of subclones. Interestingly, expression was normal in 60% of clones, demonstrating the heterogeneity of RAG expression within this commonly used cell line. These cells were not treated with etoposide, and did not exhibit the hprt exon 2 q 3 deletion. Thus, the loss of RAG gene expression in CEM clones isolated after treatment with etoposide is likely due to the prolonged isolation and culture conditions. We did not find that etoposide altered the expression of RAG genes acutely: RAG-1 expression remained high and unaltered Žby Northern blot analysis. in CEM cells over the 72 h immediately following 4 h of etoposide treatment Ždata not shown.. Therefore, although RAG gene expression may be necessary to induce illegitimate VŽD.J recombination, this expression can subsequently be lost in cells carrying the hprt exon 2 q 3 deletion. Interlocus recombination between TCR genes has
123
been proposed as an indicator of non-homologous VŽD.J recombination in mammalian cells in vitro and in vivo w30,38x. Our analysis demonstrated no interlocus recombination in these cells, SJCEM808, that clearly had undergone VŽD.J recombinase-mediated deletions of an illegitimate DNA substrate, hprt. This is not surprising, however, in that it seems highly unlikely that two unrelated targets for illegitimate VŽD.J recombination, TCR genes and hprt, would undergo rearrangement in the exact same cell. If the two events are independent, the likelihood of observing the two rearrangements in the same cell line would be on the order of 1 in 10 11 Žassuming a frequency of 10y4 for interlocus TCR rearrangement and 10y7 for hprt exon 2 q 3 deletions.. Likewise, the MLL gene Ža putative target for etoposide-induced non-homologous recombination. was not rearranged in the clone selected for the hprt exon 2 q 3 deletion. However, our findings do not rule out the possibility that MLL or other targets are rearranged as a result of etoposide exposure in cells which do not Žor only coincidentally. delete exons 2 q 3 of hprt. In fact, such MLL rearrangement does likely occur as a result of etoposide exposure in some percentage of treated cells w39x. The fact that the hprt exon 2 q 3 deletions were not associated with other genetic chromosomal rearrangements in the affected cell, or with other illegitimate VŽD.J recombinasemediated rearrangements, may have importance in assessing the mechanisms underlying leukemogenesis caused by etoposide. Our findings are supportive of etoposide being able to induce targeted, sitespecific gene rearrangements in a non-lethal fashion, rather than only promiscuous mutagenic effects. Whether there is any functional significance to the hprt exon 2 q 3 deletion, other than loss of hprt activity, is unknown. The SJCEM808 cells exhibited slower and more adherent growth characteristics and a greater degree of cell surface projections than the parental cells or than ‘sister clones’ of hprt deficient cells that did not exhibit the exon 2 q 3 deletion. The abundance of microvilli on the cell surface may have contributed to the cell–cell interactions responsible for the adherent phenotype observed in these mutant cells. The SJCEM808 clone also lost expression of some cell surface antigens associated with T-cell immunophenotype and had decreased expression of cALLa compared to parental cells.
124
C.-L. Chen et al.r Mutation Research 403 (1998) 113–125
VŽD.J recombinase-mediated DNA rearrangement acting on unnatural substrates may be an important biomarker of undesired recombinogenesis caused by anticancer drugs and other xenobiotics. Assessing such recombination in an endogenous gene substrate Že.g., hprt ., as opposed to a non-integrated vector substrate, permits an assessment of non-homologous recombination in normal cells Žsuch as circulating lymphocytes. on a transcribed gene DNA target in its complex chromatin milieu. Because transcription, chromatin interactions, and integration status have all been hypothesized to be important determinants of DNA cleavage and recombination w2,10,40–42x, VŽD.J recombinase-mediated rearrangement of hprt is being evaluated as an indicator of non-homologous DNA recombination w 16,17,19,22,43 x . SJCEM808 can serve as an useful source of positive control for PCR reactions aimed at detection of the VŽD.J recombinase-mediated deletion of exons 2 and 3 in hprt.
Acknowledgements We thank Ms. Pamela McGill, Ms. Natasha Lenchik, Ms. Ya Qin Chu, Ms. Eve Su, Ms. Margaret Needham, Ms. Susan Ragsdale, and the Cytogenetics, Electron Microscopy, and Molecular Resource Center shared resources of SJCRH for their excellent technical assistance. This article has been reviewed by the USEPA and approved for publication. Approval does not signify that the contents necessarily reflect the views or policies of the Agency, nor does mention of trade names or commercial products constitute endorsement or recommendation for use. This work is supported by Cancer Center CORE Grant CA-21765, NCI CA-51001, by a Center of Excellence grant from the State of Tennessee, and American Lebanese Syrian Associated Charities ŽALSAC..
References w1x S.M. Lewis, G.E. Wu, The origins of VŽD.J recombination, Cell 88 Ž1997. 159–162. w2x M. Gellert, Molecular analysis of VŽD.J recombination, Annu. Rev. Genet. 26 Ž1992. 425–446.
w3x M.A. Oettinger, D.G. Schatz, C. Gorka, D. Baltimore, RAG-1 and RAG-2, adjacent genes that synergistically activate VŽD.J recombination, Science 248 Ž1990. 1517–1523. w4x S. Kallenbach, T. Brinkmann, F. Rougeon, Rag-1: a topoisomerase?, Int. Immunol. 5 Ž1992. 231–232. w5x J.C. Wang, P.R. Caron, R.A. Kim, The role of DNA topoisomerases in recombination and genome stability: a doubleedged sword, Cell 62 Ž1990. 403–406. w6x L. Brown, Q. Cheng, M.J. Siciliano, W. Crist, G. Buchanan, R. Baer, Site-specific recombination of the tal-1 gene is a common occurrence in human T-cell leukemia, EMBO J. 9 Ž1990. 3343–3351. w7x B. Tycko, J. Sklar, Chromosomal translocations in lymphoid neoplasia: a reappraisal of the recombinase model, Cancer Cells 2 Ž1990. 1–8. w8x Y. Gu, G. Cimino, H. Alder, T. Nakamura, R. Prasad, O. Canaani, D.T. Moir, C. Jones, P.C. Nowell, C.M. Croce, E. Canaani, The Ž4;11.Žq21;q23. chromosome translocations in acute leukemias involve the VŽD.J recombinase, Proc. Natl. Acad. Sci. U.S.A. 89 Ž1992. 10464–10468. w9x M. Negrini, C.A. Felix, C. Martin, B.J. Lange, T. Nakamura, E. Canaani, C.M. Croce, Potential topoisomerase II DNA-binding sites at the breakpoints of a tŽ9;11. chromosome translocation in acute myeloid leukemia, Cancer Res. 53 Ž1993. 4489–4492. w10x P.L. Broeker, H.G. Super, M.J. Thirman, H. Pomykala, Y. Yonebayashi, S. Tanabe, N. Zeleznik-Le, J.D. Rowley, Distribution of 11q23 breakpoints within the MLL breakpoint cluster region in de novo acute leukemia and in treatment-related acute myeloid leukemia: correlation with scaffold attachment regions and topoisomerase II consensus binding sites, Blood 87 Ž1996. 1912–1922. w11x C.A. Felix, B.J. Lange, M.R. Hosler, J. Fertala, M. Bjornsti, Chromosome band 11q23 translocation breakpoints are DNA topoisomerase II cleavage sites, Cancer Res. 55 Ž1995. 4287–4292. w12x M.A. Smith, L. Rubinstein, R.S. Ungerleider, Therapy-related acute myeloid leukemia following treatment with epipodophyllotoxins: estimating the risks, Med. Pediatr. Oncol. 23 Ž1995. 86–98. w13x J.R. Downing, A.T. Look, MLL fusion genes in the 11q23 acute leukemias, Cancer Treat. Res. 84 Ž1996. 73–92. w14x J.C. Fuscoe, L.J. Zimmerman, M.J. Lippert, J.A. Nicklas, J.P. O’Neill, R.J. Albertini, VŽD.J recombinase-like activity mediates hprt gene deletion in human fetal T-lymphocytes, Cancer Res. 51 Ž1991. 6001–6005. w15x J.C. Fuscoe, L.J. Zimmerman, K. Harrington-Brock, L. Burnette, M.M. Moore, J.A. Nicklas, J.P. O’Neill, R.J. Albertini, VŽD.J recombinase-mediated deletion of the hprt gene in T-lymphocytes from adult humans, Mutat. Res. 283 Ž1992. 13–20. w16x C-L. Chen, J.C. Fuscoe, Q. Liu, M.V. Relling, Etoposide causes illegitimate VŽD.J recombination in human lymphoid leukemic cells, Blood 89 Ž1996. 1701–1707. w17x C-L. Chen, J.C. Fuscoe, Q. Liu, C.-H. Pui, H.H. Mahmoud, M.V. Relling, Relationship between cytotoxicity and sitespecific DNA recombination after in vitro exposure of
C.-L. Chen et al.r Mutation Research 403 (1998) 113–125
w18x
w19x
w20x
w21x
w22x
w23x
w24x
w25x
w26x
w27x
w28x w29x
w30x
leukemia cells to etoposide, J. Natl. Cancer Inst. 88 Ž1996. 1840–1847. M.J. Mcginniss, J.A. Nicklas, R.J. Albertini, Molecular analyses of in vivo hprt mutations in human T-lymphocytes: IV. Studies in newborns, Environ. Mol. Mutagen. 14 Ž1989. 229–237. B.A. Finette, T. Poseno, R.J. Albertini, VŽD.J recombinasemediated hprt mutations in peripheral blood lymphocytes of normal children, Cancer Res. 56 Ž1996. 1405–1412. A. Edwards, H. Voss, P. Rice, A. Civitello, J. Stegemann, C. Schwager, J. Zimmerman, H. Erfle, C.T. Caskey, W. Ansorge, Automated DNA sequencing of the human hprt locus, Genomics 6 Ž1990. 593–608. F.G. Behm, S.C. Raimondi, J.L. Frestedt, Q. Liu, W.M. Crist, J.R. Downing, G.K. Rivera, J.H. Kersey, C-H. Pui, Rearrangement of the MLL gene confers a poor prognosis in childhood acute lymphoblastic leukemia, regardless of presenting age, Blood 87 Ž1996. 2870–2877. J.C. Fuscoe, L.K. Vira, D.D. Collard, M.M. Moore, Quantification of hprt gene deletions mediated by illegitimate VŽD.J recombination in peripheral blood cells of humans, Environ. Mol. Mutagen. 29 Ž1997. 28–35. R.A. Gibbs, P.N. Nguyen, A. Edwards, A.B. Civitello, C.T. Caskey, Multiplex DNA deletion detection and exon sequencing of the hypoxanthine phosphoribosyltransferase gene in Lesch–Nyhan families, Genomics 7 Ž1990. 235–244. J.C. Fuscoe, L.J. Zimmerman, K. Harrington-Brock, M.M. Moore, Large deletions are tolerated at the hprt locus of in vivo derived human T-lymphocytes, Mutat. Res. 283 Ž1992. 255–262. S. Zimm, G. Reaman, R.F. Murphy, D.G. Poplack, Biochemical parameters of mercaptopurine activity in patients with acute lymphoblastic leukemia, Cancer Res. 46 Ž1986. 1495– 1498. G.A.M. Neale, T.J. Fitzgerald, R.M. Goorha, Expression of the VŽD.J recombinase gene RAG-1 is tightly regulated and involves both transcriptional and post-transcriptional control, Mol. Immunol. 29 Ž1992. 1457–1466. F.G. Behm, S.C. Raimondi, M.J. Schell, A.T. Look, G.K. Rivera, C-H. Pui, Lack of CD45 antigen on blast cells in childhood acute lymphoblastic leukemia is associated with chromosomal hyperdiploidy and other favorable prognostic features wsee commentsx, Blood 79 Ž1992. 1011–1016. F. Mitelman, ISCN95: An International System For Human Cytogenetic Nomenclature, S. Karger, Basel, 1995. T.J. Fitzgerald, G.A. Neale, S.C. Raimondi, R.M. Goorha, Rhom-2 expression does not always correlate with abnormalities on chromosome 11 at band p13 in T-cell acute lymphoblastic leukemia, Blood 80 Ž1992. 3189–3197. S. Lipkowitz, V.F. Garry, I.R. Kirsch, Interlocus V–J recom-
w31x
w32x
w33x
w34x
w35x
w36x
w37x
w38x
w39x
w40x
w41x
w42x
w43x
125
bination measures genomic instability in agriculture workers at risk for lymphoid malignancies, Proc. Natl. Acad. Sci. U.S.A. 89 Ž1992. 5301–5305. D.E. Moore, K. Weise, R. Zawydiwski, E.B. Thompson, The karyotype of the glucocorticoid-sensitive, lymphoblastic human T-cell line CCRF-CEM shows a unique deleted and inverted chromosome 9, Cancer Genet. Cytogenet. 14 Ž1985. 89–94. R. Zawydiwski, J.M. Harmon, E.B. Thompson, Glucocorticoid-resistant human acute lymphoblastic leukemic cell line with functional receptor, Cancer Res. 43 Ž1983. 3865–3873. J. Maraschin, B. Dutrillaux, A. Aurias, Chromosome aberrations induced by etoposide are not random, Int. J. Cancer 46 Ž1990. 808–812. N.A. Berger, S. Chatterjee, J.A. Schmotzer, S.R. Helms, Etoposide-induced gene alterations: potential contribution to cell death, Proc. Natl. Acad. Sci. U.S.A. 33 Ž1991. 8740– 8742. J.F. McBlane, D.C. van Gent, D.A. Ramsden, C. Romeo, C.A. Cuomo, M. Gellert, M.A. Oettinger, Cleavage at a VŽD.J recombination signal requires only RAG1 and RAG2 proteins and occurs in two steps, Cell 83 Ž1995. 387–395. J.F. Riou, L. Grondard, O. Petitgenet, M. Abitbol, F. Lavelle, Altered topoisomerase I activity and recombination activating gene expression in a human leukemia cell line resistant to doxorubicin, Biochem. Pharmacol. 46 Ž1993. 851–861. F.W. Alt, G. Rathbun, E. Oltz, G. Taccioli, Y. Shinkai, Function and control of recombination-activating gene activity, Ann. New York Acad. Sci. 651 Ž1992. 277–295. J.M. Abdallah, D.P. Lombardi, I.R. Kirsch, Genetic instability in patients with Hodgkin’s disease undergoing chemotherapy, J. Clin. Invest. 96 Ž1995. 2744–2747. P.D. Aplan, D.S. Chervinsky, M. Stanulla, W.C. Burhans, Site-specific DNA cleavage within the MLL breakpoint cluster region induced by topoisomerase II inhibitors, Blood 87 Ž1996. 2649–2658. M.T. Howard, M.P. Lee, T.S. Hsieh, J.D. Griffith, Drosophila topoisomerase II-DNA interactions are affected by DNA structure, J. Mol. Biol. 217 Ž1991. 53–62. A. Udvardy, P. Schedl, M. Sander, T.S. Hsieh, Topoisomerase II cleavage in chromatin, J. Mol. Biol. 191 Ž1986. 231–246. D.G. Schatz, M.A. Oettinger, M.S. Schlissel, VŽD.J recombination: molecular biology and regulation, Annu. Rev. Immunol. 10 Ž1992. 359–383. W.G. McGregor, V.M. Maher, J.J. McCormick, Kinds and locations of mutations induced in the hypoxanthine-guanine phosphoribosyltransferase gene of human T-lymphocytes by 1-nitrosopyrene, including those caused by VŽD.J recombinase, Cancer Res. 54 Ž1994. 4207–4213.