Antisense ATM gene therapy: a strategy to increase the ... - Nature

Gene Therapy (2000) 7, 852–858  2000 Macmillan Publishers Ltd All rights reserved 0969-7128/00 $15.00 www.nature.com/gt

ACQUIRED DISEASES

RESEARCH ARTICLE

Antisense ATM gene therapy: a strategy to increase the radiosensitivity of human tumors C Guha1, U Guha1, S Tribius1, A Alfieri2, D Casper3, P Chakravarty1, W Mellado4, TK Pandita4 and B Vikram1 Departments of 1Radiation Oncology and 3Neurological Surgery of the Albert Einstein College of Medicine, Montefiore; 2Beth Israel Medical Centers, New York, NY; and 4Center for Radiological Research, Columbia University, NY, USA

Atm, the gene mutated in ataxia-telangiectasia (AT) patients, is an essential component of the signal transduction pathway that responds to DNA damage due to ionizing radiation (IR). We attenuated ATM protein expression in human glioblastoma cells by expressing antisense RNA to a functional domain of the atm gene. While ATM expression decreased,

constitutive expression of p53 and p21 increased. Irradiated ATM-attenuated cells failed to induce p53, demonstrated radioresistant DNA synthesis, and increased radiosensitivity. Antisense-ATM gene therapy in conjunction with radiation therapy may provide a novel strategy for the treatment of cancer. Gene Therapy (2000) 7, 852–858.

Keywords: antisense gene therapy; ATM; radiation therapy; glioblastoma multiforme

Introduction Ataxia telangiectasia (AT) is an autosomal recessive disease with a pleiotropic phenotype that includes exquisite radiosensitivity.1,2 The mechanism by which the atm (AT mutated) gene product is involved in cellular responses to ionizing radiation (IR) is currently under investigation. Cells from AT patients display defects in several cell cycle checkpoints, which normally allow the repair of IRdamaged DNA before replication and chromosomal segregation.3,4 The failure to arrest DNA synthesis in S-phase upon irradiation (radioresistant DNA synthesis) is a hallmark of AT cells.5 ATM is a member of a family of large proteins showing sequence homology to the catalytic domain of phosphatidylinositol-3 (PI-3) kinase, implying a role in signal transduction.6 The presence of a leucine zipper domain in ATM also suggests dimerization or interactions with additional proteins.6 Overexpression of the leucine zipper domain as a truncated ATM protein in a human colorectal carcinoma cell line functioned as a dominant negative ATM mutant and increased radiosensitivity.7 Radiosensitivity was complemented by the overexpression of the truncated ATM carboxy terminal PI-3 kinase, demonstrating that this domain is essential for cell survival after irradiation.7 Doses as low as 4 Gy are lethal to transgenic mice with mutant forms of the mouse atm homolog.8 Fibroblasts from ATM knockout mice also demonstrate increased radiosensitivity in homozygotes, and cells with the heterozygous genotype demonstrate an intermediate phenotype.9 Following irradiation, AT cells exhibit defective/delayed induction of p53 and p21 genes, with inability to arrest

Correspondence: C Guha, Department of Radiation Oncology, 111 East 210 Street, Bronx, NY 10467, USA Received 25 May 1999; accepted 19 January 2000

the cell cycle at G1.4,10 Recent studies have indicated that ATM acts as a protein kinase that is activated by IR and subsequently phosphorylates p53 on serine-15.11,12 In AT cells, this post-irradiation activation and phosphorylation of the p53 protein is reduced, which might explain the increased vulnerability of AT cells to IR-induced DNA damage. Additionally, the deletion of atm radiosensitizes even p53-deficient bone marrow and mouse embryonic fibroblast cells.13 Thus, ATM appears to be an essential component of the p53-dependent, as well as, p53-independent signal transduction machineries responsible for cell survival following IR. Although AT patients and Atm-null mice are extremely radiosensitive, it is becoming increasingly apparent that hypersensitivity to IR associated with Atm dysfunction is cell-type specific and has a diverse relationship with the status of the p53 gene expression. For example, IR-induced apoptosis is Atm and p53dependent in developing CNS,14 Atm and p53-independent in lungs, and atm-independent and p53-dependent in thymus.15 We proposed that attenuation of ATM function in human cancer cells would increase their sensitivity to IR irrespective of their status of the p53 gene mutation. Glioblastoma multiforme (GBM) is one of the most lethal brain tumors. It is rarely completely resected because of ill-defined margins. Delivering high enough doses of radiotherapy to eradicate the tumor is also impossible without seriously damaging the adjacent brain. We postulated that if it were possible to increase the radiosensitivity of GBMs, we might achieve tumor control at lower doses, which would not damage the surrounding brain. In order to test this hypothesis, we first cloned the sequence for the PI-3 kinase domain of the ATM gene in an antisense orientation into a retroviral expression vector MFG, and then transfected it into the human glioblastoma multiforme cell line U-87 MG. In this report we demonstrate that antisense ATM-trans-

Antisense ATM gene therapy C Guha et al

853

fected U-87 MG cells exhibited decreased ATM protein expression and higher constitutive expression of p53 and p21 proteins. Following irradiation, they displayed features of AT cells, such as defective p53 protein induction, radioresistant DNA synthesis, aberrant G2 checkpoint control and increased radiosensitivity.

Results Construction of an ATM antisense vector and transfection into U-87 MG cells In order to increase the intrinsic radiosensitivity of GBM cells, we proposed to attenuate ATM protein expression by expressing antisense RNA targeting the PI-3 kinase domain of the ATM. An MFG retroviral vector (Figure 1a) expressing the antisense-ATM RNA was permanently transfected into the GBM cell line U-87 MG to generate U-87␣ATM cells. Control cells were transfected with the neomycin-resistant vector, pRc-CMV, alone (U87cneo). Because a truncated PI-3 kinase domain of the ATM protein can be radioprotective,7 we avoided using the sense Atm insert as our negative control. The expression of antisense ATM RNA in U-87␣ATM cells was verified by RT-PCR with MFG and ATM-specific primers (Figure 1b; arrows). Figure 1c demonstrates that a single 430 bp fragment could be amplified from U87␣ATM cells, but not from control U-87cneo cells. Attenuation of ATM protein expression Immunoblotting, performed to compare the expression levels of ATM protein in U-87cneo and U-87␣ATM cells, revealed a band in U-87cneo cells, whose migration was similar to the bands in the atm wild-type lymphoblastic cell line GM02184D (LB(wt)) and HeLa cells, and agreed with the expected size (370 kDa) of the ATM protein (Figure 2). In contrast, ATM expression was absent in extracts prepared from the U-87␣ATM cells and the lymphoblastic cell line GM01526D (LB(AT)) that was generated from an AT patient. Blots processed, under identical conditions, without primary antibody did not reveal any bands (data not shown). Constitutive and radiation-induced p53 and p21 expression AT cells display defective and/or delayed increase in p53 and p53-dependent p21 protein expression following IR.4,10 Levels of p53 and p21 proteins were, therefore, assessed by immunoblotting in U87-␣ATM cells to determine whether the attenuation of ATM protein expression resulted in defective induction of the p53-dependent pathways. In the unirradiated U-87␣ATM cells, the basal levels of p53 and p21 proteins were increased by twoand 10-fold, respectively, compared with the U-87 parental and U-87cneo cells (Figure 3). Four hours after irradiation (10 Gy) of U-87 parent and cneo cells, p53 and p21 protein levels increased by two-fold and ⭓3-fold, respectively. The p53 induction was absent in U-87␣ATM cells, in contrast, p21 levels were increased by 1.3- to 2.0-fold. Analysis of radioresistant DNA synthesis A variety of dysfunctional phenotypes have been noted in AT cells, but one universal characteristic is radioresistant DNA synthesis,5 defined as the failure of AT cells to

Figure 1 Expression of antisense ATM RNA in U87 cells. (a) Diagram of MFG-␣ATM. A 281 bp BamHI–NcoI fragment (8710–8991 bp) of the ATM cDNA, containing the PI-3 kinase domain (8753–8815 bp), was subcloned from the pGEM-ATM14 plasmid into the MFG vector in the antisense orientation to generate MFG-␣ATM. (b) Oligonucleotide primers for RT-PCR, depicted as MFGU1, upstream primer from MFG vector sequence (5′-CCGCCGACACCAGACTAA-3′) and ATML1, downstream primer from the ATM PI-3 kinase domain (5′-GGTTTGAG AAGCGATTGG-3′). (c) RT-PCR amplification of the antisense ATM RNA in U87 cells. Total RNA, digested for 30 min at 37°C with RNasefree DNase (Boehringer Mannheim), followed by its inactivation (95°C; 5 min), was analyzed by RT-PCR in presence or absence of Superscript RT. The PCR products were electrophoresed on an ethidium bromidestained 15% agarose gel. Note the presence of a 430 bp fragment in the U87-␣ATM (clone 10) lane and MFG-␣ATM14 plasmid lane, and the absence of the band in control lanes (U87-cneo). Gene Therapy


854

dine incorporation after 4 and 8 Gy in the U-87 parental (68.0 ± 10.2% and 63.8 ± 2.1%, respectively) and U-87cneo cell lines (73.0 ± 10.0 % and 72.5 ± 3.6%, respectively). The rate of thymidine incorporation was unchanged in the U87␣ATM cells (98.4 ± 7.6% at 4 Gy and 94.1 ± 7.1% at 8 Gy). The S-phase fraction of these cell lines was also evaluated following a single dose of irradiation (6 Gy) by flow cytometry. A reduction in the number of BrdU-positive cells was observed in the U-87 parental and U-87cneo cell lines, indicating a transitional phase block in DNA synthesis (Figure 4a and b). In contrast, U87␣ATM cells continued to traverse this block in DNA synthesis and the number of BrdU-positive S-phase cells remained nearly constant before and after irradiation.

Figure 2 Attenuation of ATM protein expression in U-87␣ATM cells. Immunoblots (75 ␮g per lane) of cell extracts isolated from U-87cneo, U87␣ATM (clone 10), lymphoblastic cell lines GM02184D (LB(wt), wildtype atm), and GM01526D (LB(AT), mutated atm) and HeLa cells. Arrow represents position of thyroglobulin, MW 330 000. Note the absence of the ATM-specific band in the U-87␣ATM and LB(AT) cell lines. Immunoblots are a representation of two separate experiments.

Radiation sensitivity A clonogenic assay was performed and colony formation was measured following irradiation, to determine whether attenuation of ATM resulted in a radiosensitive phenotype. After IR, colonies which demonstrated clonogenic capacity were counted and the surviving fractions were calculated. The clonogenic survival curves of a representative U-87␣ATM clone (U-87 ␣ATM10) along with the control cell lines are shown in Figure 5a. Two separate anti-ATM-transfected clones (clone 2 and 3) were irradiated with a clinically relevant dose of 2 Gy to determine the survival fraction (SF2). The survival fractions after exposure to a clinically relevant dose of 2 Gy (SF2) of U-87, U-87cneo, U-87␣ATM2, U-87␣ATM3 and U87␣ATM10 were 0.54, 0.58 and 0.32, 0.25 and 0.28, respectively (Figure 5b). This indicates a substantially increased radiosensitivity of U-87-␣ATM cells irrespective of any clonal variation that might have resulted from the selection of the antisense ATM clones. Table 1 shows that the U-87 parental cells exhibited one and two log kills at 5.1 and 9.6 Gy, respectively, the U-87cneo cells after 4.6 and 8.1 Gy, respectively, and the U87␣ATM cells after 3.4 and 6.4 Gy, respectively. Thus, attenuation of the ATM protein resulted in a dose modification factor of 1.5 in U-87␣ATM cells.

Discussion We have demonstrated that antisense ATM-transfected U-87 cells exhibit many features of AT cells. AT cells proFigure 3 Immunoblot analysis of p53 and p21 proteins. Immunoblots (20 ␮g per lane) of cell lysates from U-87 parental, cneo and ␣ATM (clone 10) cell lines, after no irradiation (0), and 4 h following irradiation with 10 Gy (10), were probed with a mixture of mouse monoclonal antibodies to beta actin (Ab-1) and p53 (Ab-3) in the same blocking solution. For detection of p21 protein, the filters were stripped after incubation in stripping buffer, containing mercaptoethanol, tris (pH 6.8) and SDS for 30 min at 50°C, according to the manufacturer’s protocol (Amersham) and were then reprobed with a 2.5 ␮g/ml of mouse monoclonal antibody to p21. Note that actin levels are approximately equal for all lanes, indicating equivalent protein concentrations in the extracts. Immunoblots are a representation of three separate experiments.

halt DNA replication after exposure to IR. AT cells continue to incorporate nucleotides into DNA on a damaged template, indicating a defective S-phase checkpoint. Therefore, the integrity of the S-phase checkpoint was assessed by (1) a thymidine incorporation assay after cells were irradiated at 4 and 8 Gy; and (2) a flow cytometry assay after labeling DNA synthesizing cells with BrdU. There was significant (P ⬍ 0.005) reduction in 3H-thymiGene Therapy

Table 1 Radiobiological parameters of U-87 parental, cneo and ␣ATM cells Alpha (␣) Beta (␤) SF2 ±s.e.

U-87 U-87cneo U-87-␣ATM

0.4325 0.405 0.615

0.0048 0.54 0.055 0.019 0.58 0.038 0.0136 0.28 0.073

Surviving Dose fraction (Gy) modifying factor 0.1 0.01 (DMF) 5.1 4.6 3.4

9.6 8.1 6.4

1 1.15 1.5

Radiobiological parameters were derived from clonogenic survival experiments after analysis by a computer-assisted program (Norman Albright Fit v2.4) which calculates the best fit curves according to the linear quadratic model by iterative weighted least squares regression analysis of all data points, estimating the covariance of the survival curve parameters and their corresponding confidence limits. ␣ and ␤ are the initial and final slopes of the linear quadratic survival curves.


ceed through the cell cycle despite DNA damage, and exhibit dysfunctional cell cycle checkpoint regulation and radioresistant DNA synthesis.5,16 Like AT cells, the U87␣ATM cells seemed to lack the control mechanism(s) that reduce the rate of DNA synthesis after irradiation (Figure 4) and exhibited increased radiosensitivity (Figure 5). It has been shown that apoptosis is minimally induced (⬍1%) in U-87 cells after exposure to 4–6 Gy of RT.17,18 In our studies, the apoptotic index at varying time intervals (1, 3 and 5 days) after RT (10 and 30 Gy) was ⭐1% for the parental and cneo U-87 cells, while U87␣ATM cells exhibited ⭐8% apoptotic cells (data not shown). In addition, irradiated U-87␣ATM cells showed aberrant nuclear morphological changes with accumulation of giant mononuclear cells with higher ploidy after 3–5 days after RT (manuscript in preparation). Whether this was caused by dysfunctional cell cycle checkpoints, defective induction of the p53 protein, or both, needs to be examined further in light of recent investigations into the pathway by which ATM acts. The p53 protein is a transcription factor that is activated in response to genotoxic stress such as IR. It is critical in producing cell cycle arrest by inducing the expression of proteins such as p21, and for eliminating lethally damaged cells by inducing apoptosis.19–21 ATM is implicated in the signaling pathways that mediate increased p53 expression4 and in the activation of p53 after IR.11,12,22 We found that attenuation of ATM protein expression in human GBM cells resulted in a failure to induce p53 protein and decreased the induction of p21 protein after IR. Similarly, a p53-independent induction of p21 protein was seen in the thymus of atm-deficient transgenic mice after exposure to IR.15 The constitutive levels of both p53 and p21 proteins were increased in U87␣ATM cells. This could imply that the attenuation of ATM protein contributes to decreased fidelity of the genomic repair pathways, resulting in genomic instability and enhanced basal levels of p53 and p21 proteins. In comparison to the U-87 parental cneo cells, U-87␣ATM cells demonstrated slower growth in culture (cell doubling time 36 h versus ⭐26 h, respectively) and increased serum-dependence (15% FBS, data not shown). These cellular proliferative defects were also reported in mouse embryonic fibroblasts9,23 and astrocytes24 from Atm−/− mice, where constitutive elevation of p21 protein was associated with slower cell growth. A truncated leucine zipper domain of the ATM protein has been shown to function as a dominant negative ATM mutant. Overexpression of this domain has been shown to increase radiosensitivity of colon cancer cells.7 Strategies using dominant negative atm gene mutants in regions such as the leucine zipper domain could have bystander effects on the function of other leucine zipper proteins. In contrast, the present strategy using antisense vectors will directly target nucleotide homologies and should be more specific to the atm gene. The mechanism of antisense atm-mediated attenuation of the ATM protein expression could be mediated by inhibition of the atm gene transcription. In addition, the selection of the PI-3 kinase domain of the Atm gene for antisense RNA expression could also contribute to this effect as this domain has been implicated in radioprotection.7 The intrinsic radiosensitivity of human tumors (measured as SF2, the fraction of tumor cells surviving after a clinically relevant 2 Gy dose) has been shown to

855

Figure 4 Radioresistant DNA synthesis in U-87, U-87cneo and U87␣ATM cells. (a) Representative flow cytometry scatter plots of cells exposed to 0 Gy or 6 Gy of irradiation and incubated with BrdU, plotted as increasing fluorescence of propidium iodide (X axis) versus increasing FITC fluorescence obtained with an anti-BrdU-FITC conjugated antibody to detect BrdU incorporation (Y axis). Gating for S-phase cells is shown in the box in the panels. (b) Graph based on the primary data from (a) representing changes in the percentage of S-phase cells after irradiation relative to unirradiated cells (four independent FACS experiments). The percentage change of S-phase cells in U-87␣ATM (clone 10) cells after irradiation was significantly different from U-87 parental (P ⬍ 0.0005) and U-87cneo (P = 0.006) samples by Student’s t test.

Gene Therapy


856

tissues. Finally, the use of retrovirus-based gene therapy vectors would target cycling tumor cells and spare the noncycling brain parenchyma. Other attempts to increase the radiosensitivity of human tumors have explored the use of ‘suicide’ gene therapy approaches, using the viral thymidine kinase (HSVtk) gene,29 the cytosine deaminase/thymidine kinase fusion genes,30 or expressing tumor necrosis factor alpha (TNF-␣) under the control of a radiation-inducible promoter (Egr-1).31 Long-term effects of such suicide gene therapy approaches to the nervous system are not known. However, one recent report demonstrated that adenovirus gene therapy with HSVtk resulted in chronic brain inflammation because of generalized expression of the suicide gene and the presence of a diffusible toxic product. Thus, the absence of a diffusible product in our approach may improve the therapeutic ratio. Moreover, by targeting an intrinsic radioprotective signal transduction pathway, we eliminated an important problem that is common to most suicide gene therapies, ie inadequate prodrug delivery. Thereby, antisense-ATM gene therapy could serve as a paradigm for strategies targeting the intrinsic survival mechanisms of cancer cells after irradiation.

Materials and methods

Figure 5 Attenuation of ATM protein expression increases the radiation sensitivity of U-87␣ATM cells. (a) U-87 parental, U-87cneo and U-87␣ATM (clone 10) cells were exposed to 0, 2, 4, 6 and 8 Gy of IR and then incubated for 2–3 weeks before fixation, staining and colony counting. Clonogenic assays were performed in triplicate. Mean colony numbers of irradiated cells relative to plating efficiencies and unirradiated clonogenic numbers are plotted and the s.e. are shown as error bars. Curves were generated by a best-fit analysis to the linear quadratic model using a computer program (N Albright Fit v2.4) for the Macintosh. (b) Survival fraction of U-87 cells after 2 Gy (SF2). SF2 was determined in three separate clones of U-87␣ATM cells (clones 2, 3 and 10). The SF2s in U-87␣ATM clones, clone 2 (0.32), 3 (0.25) and 10 (0.28) were significantly lower (P ⬍ 0.005) than that of U-87 parental (0.54) and cneo (0.58) cells by Student’s t test.

be an independent prognostic factor following RT for cancers of the cervix25,26 and head and neck.27 In another study, the mean SF2 of tumor cell lines from GBM patients (0.50) was found to be significantly higher than that of tumor cell lines from patients with anaplastic astrocytoma (0.34), who survive much longer after RT.28 Our experiments demonstrated that the SF2 of U-87 cells was reduced approximately 50% (from 0.54 to 0.28) by attenuation of the ATM protein with antisense atm gene therapy. Recent studies with Atm-null mice showed that Atm−/− neurons are resistant to IR-induced apoptosis,14 while the survival of Atm−/− astrocytes are similar to the wild-type astrocytes.24 Our data suggests that in contrast to Atm-null neurons and astrocytes, disruption of Atm in glial tumors confers radiosensitivity. Thus, antisense ATM gene therapy may preferentially radiosensitize glial tumors, without affecting the adjacent normal brain Gene Therapy

Construction of MFG-␣ATM vector The MFG retroviral vector system was used to express antisense ATM RNA, in which Moloney murine leukemia virus long terminal repeat sequences generate both a full length viral RNA (for encapsidation into virus particles) and a subgenomic mRNA expressing the inserted sequences.32 A 281 bp BamHI–NcoI fragment (8710– 8991 bp) of the ATM cDNA, containing the PI-3 kinase domain (8753–8815 bp), was subcloned from the pGEMATM14 plasmid (a gift from Dr Tej Pandita, Center for Radiological Research, Columbia University, NY, USA) into the MFG vector in the antisense orientation to generate MFG-␣ATM (Figure 1a). Cell culture The U87 MG cell line33 and transfectants were cultured in high glucose-DMEM with 15% heat inactivated fetal bovine serum (FBS), and 1% penicillin/streptomycin (P/S). The lymphoblastic cell lines GM02184D and GM01526D, from normal and AT patients respectively (NIGMS Genetic Mutant Cell Repository, Camden, NJ, USA), were cultured in RPMI with 10% FBS and 1% P/S. HeLa cells were grown in MEM with 10% FBS (all media and supplements were purchased from Life Technologies, Grand Island, NY, USA). Generation of antisense ATM RNA-expressing U87 cells U87 cells were cotransfected with MFG-␣ATM and pRcCMV, a plasmid expressing the neomycin resistance gene (Invitrogen, Carlsbad, CA, USA), in a molar ratio of 10:1, using the TransIT polyamine, LT-1, as per the manufacturer’s protocol (Panvera, Madison, WI, USA). Stable neomycin-resistant transfectants were selected in the presence of genticin (500 ␮g/ml). Total RNA was extracted from pRcCMV-transfected (U-87cneo), and MFG␣ATM14-transfected (U-87␣ATM) cells using the Qiagen RNeasy kit (Qiagen, Santa Clarita, CA, USA). Antisense


ATM RNA expression was confirmed by reverse transcriptase polymerase chain reaction (RT-PCR) analysis using a PCR kit (Gibco/BRL, Gaithersburg, MD, USA) according to the manufacturer’s instructions. Three clones (clone 2, 3 and 10), expressing antisense ATM RNA, were randomly selected for SF2 assays. Extensive studies were performed with one of these clones (clone 10).

Electrophoresis and immunoblotting Immunoblotting was performed using standard procedures.34 Protein concentrations in cell lysates were determined using the BioRad protein assay (BioRad Laboratories, Richmond, CA, USA). For ATM expression, 75 ␮g of cell lysates were loaded on SDS 5% polyacrylamide gel (SDS-PAGE), electrophoresed and transferred to a polyvinyldifluoride (PVDF) membrane (Millipore, Bedford, MA, USA) in a Transblot apparatus (BioRad). The filters were probed with a polyclonal rabbit antiATM antibody, Ab-3 (2 ␮g/ml) (Oncogene Science, Cambridge, MA, USA), and were then incubated with horseradish peroxidase-conjugated anti-rabbit IgG (Amersham, Arlington Heights, IL, USA), followed by development with enhanced chemiluminescence (ECL plus; Amersham). Immunoblotting for p53, p21, and actin was performed in a similar fashion, except that 20 ␮g of cell lysates were separated on a 12.5% SDS-PAGE. The primary antibodies for p53, p21 and beta actin were murine monoclonal antibodies (Oncogene Science), and the secondary antibodies were horseradish peroxidaseconjugated anti-mouse IgM and IgG1 antibodies (Oncogene). The immunoblots were scanned by UMAX Astra600S scannner and the analysis of area and density of bands on autoradiograms was performed using the video densitometry software program IMAGE (NIH IMAGE version 1.49). IMAGE analysis of the bands was represented as peaks and the area under the curve was arbitrarily given units of measurement for comparison. For comparison, the individual band intensities were normalized to actin intensity for each sample and expressed as fold increase over controls. Radiation procedure Irradiation was delivered to cells plated in large T-flasks and multiwell plates with a 6 MV linear accelerator (Clinac 6–100; Varian Oncology Systems, Palo Alto, CA, USA) using a single field and 1.5 cm superflab bolus, with a 40 × 40 cm field size and a dose rate of 2.5 Gy/min. Smaller T-flasks were irradiated in a rotating vertical column using a Cesium137 IBL 437c irradiator (Campagnie, Oris Industrie, Canada) at a dose rate of 9.1 Gy/min. Clonogenic assay Cell lines were plated in triplicate at limiting dilutions into 60 mm culture dishes for 16 h and then irradiated (0–8 Gy) and incubated for 2–3 weeks at 37°C. Before counting the colonies, the culture medium was decanted and the cells were fixed in 10% formalin, rinsed, and stained with 1.0% crystal violet in citric acid. Plating efficiencies for the three cell lines were similar and ranged from 20 to 41% between successive experiments. Analysis of radioresistant DNA synthesis Two methods were used. (1) Thymidine assay: the effect of radiation on DNA synthesis was determined as pre-

viously described.5 In brief, cells were plated at a density of 5 × 104 cells per well into 24-well plates and were allowed to attach for 6 h. Cells were irradiated (4 and 8 Gy) and kept on ice until the medium of all wells was changed to medium containing 2 ␮Ci/ml methyl-3H-thymidine (ICN Pharmaceuticals, Irvine, CA, USA). After 15 h of incubation, the cells were harvested, washed twice with cold PBS, fixed in 70% methanol, and the cell suspension was filtered through GF/C (Whatman) filters. The filters were sequentially washed with 95 and 80% methanol, air-dried, and assayed for radioactivity in Ecoscint scintillation fluid by scintillation spectrometry (Packard Tri-Carb 4530, Downers Grove, CA, USA). (2) Bromodeoxyuridine (BrdU) assay: exponentially growing cells were irradiated (6 Gy) and 16 h later the cells were incubated for 4 h with 10 ␮m BrdU (Sigma, St Louis, MO, USA), followed by trypsinization. The cells were then pelleted, vortexed and fixed in 70% iced ethanol and stored at −20°C. The cells were later processed by resuspending in PBS and incubating with 50 ␮g/ml RNaseA at 37°C for 10 min, followed by permeabilization with 1 N HCl and 0.5% Triton X-100 for 10 min at 0°C. BrdU immunocytochemistry was performed with fluoresceinconjugated human anti-BrdU antibody (Pharmingen, San Diego, CA, USA) as per the manufacturer’s instructions. The cells were counterstained with propidium iodide (8 ␮g/ml) at 37°C for 15 min, were placed on ice and subjected immediately to flow cytometric analysis for quantification of the DNA content, and BrdU incorporation, on a FACScan (Becton Dickinson, Mountain View, CA, USA) using Cell Quest software (Becton Dickinson, San Jose, CA, USA) and WinMidi shareware (J Trotter, Scripps Clinic, La Jolla, CA, USA).

857

Acknowledgements We thank Zoya Niazova and Zuoheng Fan for technical assistance, TK Pandita for providing the pGEM-ATM14 plasmid, J Roy Chowdhury and R Kucheralapati for helpful discussions and reviewing the manuscript. This work was supported by NIH grant (NS 34746 to TKP) and by a pilot project grant from the Department of Radiation Oncology, Montefiore Medical Center, Bronx, NY.

References 1 Taylor AM et al. Ataxia telangiectasia: a human mutation with abnormal radiation sensitivity. Nature 1975; 258: 427–429. 2 Gatti RA et al. Ataxia-telangiectasia: an interdisciplinary approach to pathogenesis. Medicine 1991; 70: 99–117. 3 Beamish H, Lavin MF. Radiosensitivity in ataxia-telangiectasia: anomalies in radiation-induced cell cycle delay. Int J Radiat Biol 1994; 65: 175–184. 4 Kastan MB et al. A mammalian cell cycle checkpoint pathway utilizing p53 and GADD45 is defective in ataxia-telangiectasia. Cell 1992; 71: 587–597. 5 Painter RB, Young BR. Radiosensitivity in ataxia-telangiectasia: a new explanation. Proc Natl Acad Sci USA 1980; 77: 7315–7317. 6 Savitsky K et al. A single ataxia telangiectasia gene with a product similar to PI-3 kinase (see comments). Science 1995; 268: 1749–1753. 7 Morgan SE et al. Fragments of ATM which have dominantnegative or complementing activity. Molec Cell Biol 1997; 17: 2020–2029. 8 Barlow C et al. Atm-deficient mice: a paradigm of ataxia telangiectasia. Cell 1996; 86: 159–171. Gene Therapy


858

9 Xu Y, Baltimore D. Dual roles of ATM in the cellular response to radiation and in cell growth control (see comments). Genes Dev 1996; 10: 2401–2410. 10 Khanna KK et al. Nature of G1/S cell cycle checkpoint defect in ataxia-telangiectasia. Oncogene 1995; 11: 609–618. 11 Banin S et al. Enhanced phosphorylation of p53 by ATM in response to DNA damage. Science 1998; 281: 1674–1677. 12 Canman CE et al. Activation of the ATM kinase by ionizing radiation and phosphorylation of p53. Science 1998; 281: 1677–1679. 13 Westphal CH et al. Loss of atm radiosensitizes multiple p53 null tissues. Cancer Res 1998; 58: 5637–5639. 14 Herzog KH et al. Requirement for Atm in ionizing radiationinduced cell death in the developing central nervous system. Science 1998; 280: 1089–1091. 15 Barlow C et al. Atm selectively regulates distinct p53-dependent cell-cycle checkpoint and apoptotic pathways. Nat Genet 1997; 17: 453–456. 16 Young BR, Painter RB. Radioresistant DNA synthesis and human genetic diseases. Hum Genet 1989; 82: 113–117. 17 Haas-Kogan DA et al. p53-dependent G1 arrest and p53-independent apoptosis influence the radiobiologic response of glioblastoma. Int J Radiat Oncol Biol Phys 1996; 36: 95–103. 18 Yount GL et al. Cell cycle synchrony unmasks the influence of p53 function on radiosensitivity of human glioblastoma cells. Cancer Res 1996; 56: 500–506. 19 Kuerbitz SJ, Plunkett BS, Walsh WV, Kastan MB. Wild-type p53 is a cell cycle checkpoint determinant following irradiation. Proc Natl Acad Sci USA 1992; 89: 7491–7495. 20 Kastan MB et al. Participation of p53 protein in the cellular response to DNA damage. Cancer Res 1991; 51: 6304–6311. 21 Levine AJ. p53, the cellular gatekeeper for growth and division. Cell 1997; 88: 323–331. 22 Waterman MJ, Stavridi ES, Waterman JL, Halazonetis TD. ATMdependent activation of p53 involves dephosphorylation and association with 14–3-3 proteins. Nat Genet 1998; 19: 175–178. 23 Xu Y et al. Involvement of p53 and p21 in cellular defects and tumorigenesis in Atm−/− mice. Mol Cell Biol 1998; 18: 4385–4390.

Gene Therapy

24 Gosink EC, Chong MJ, McKinnon PJ. Ataxia telangiectasia mutated deficiency affects astrocyte growth but not radiosensitivity. Cancer Res 1999; 59: 5294–5298. 25 West CM, Davidson SE, Roberts SA, Hunter RD. Intrinsic radiosensitivity and prediction of patient response to radiotherapy for carcinoma of the cervix. Br J Cancer 1993; 68: 819– 823. 26 West CM, Davidson SE, Roberts SA, Hunter, RD. The independence of intrinsic radiosensitivity as a prognostic factor for patient response to radiotherapy of carcinoma of the cervix. Br J Cancer 1997; 76: 1184–1190. 27 Kelland LR, Steel, GG. Recovery from radiation damage in human squamous carcinoma of the cervix. Int J Radiat Biol 1989; 55: 119–127. 28 Taghian A et al. Intrinsic radiation sensitivity may not be the major determinant of the poor clinical outcome of glioblastoma multiforme. Int J Radiat Oncol Biol Phys 1993; 25: 243–249. 29 Kim JH, Kim SH, Brown SL, Freytag SO. Selective enhancement by an antiviral agent of the radiation-induced cell killing of human glioma cells transduced with HSV-tk gene. Cancer Res 1994; 54: 6053–6056. 30 Rogulski KR, Kim JH, Kim SH, Freytag SO. Glioma cells transduced with an Escherichia coli CD/HSV-1 TK fusion gene exhibit enhanced metabolic suicide and radiosensitivity. Hum Gene Ther 1997; 8: 73–85. 31 Hallahan DE et al. Spatial and temporal control of gene therapy using ionizing radiation. Nature Med 1995; 1: 786–791. 32 Dranoff G et al. Vaccination with irradiated tumor cells engineered to secrete murine granulocyte–macrophage colony-stimulating factor stimulates potent, specific, and long-lasting antitumor immunity. Proc Natl Acad Sci USA 1993; 90: 3539–3543. 33 Ponten J, Macintyre EH. Long term culture of normal and neoplastic human glia. Acta Pathol Microbiol Scand 1968; 74: 465–486. 34 Harlow E, Lane D (eds). Immunoblotting. In: Antibodies: A Laboratory Manual. Cold Spring Harbor Laboratory Press: Plainview, NY, 1988 pp 471–510.