Susan H. MacPhail

Radiation Sensitivity, H2AX Phosphorylation, and Kinetics of Repair of DNA Strand Breaks in Irradiated Cervical Cancer Cell Lines

Six human cervical cancer cell lines [five human papillomavirus (HPV) positive, one HPV negative] for induction and rejoining of DNA strand breaks and for kinetics of formation and loss of serine 139 phosphorylated histone H2AX (γH2AX). X-rays induced the same level of DNA breakage for all cell lines. By 8 hours after 20 Gy, <2% of the initial single-strand breaks remained and no double-strand breaks could be detected. In contrast, 24 hours after irradiation, γH2AX representing up to 30% of the initial signal still present. SW756 cells showed almost four times higher background levels of γH2AX and no residual γH2AX compared with the most radiosensitive HPV-negative C33A cells that showed the lowest background and retained 30% of the maximum level of γH2AX. Radiation sensitivity, measured as clonogenic-surviving fraction after 2 Gy, was correlated with the fraction of γH2AX remaining 24 hours after irradiation. A substantial correlation with γH2AX loss half-time measured over the first 4 hours was seen only when cervical cell lines were included in a larger series of p53-deficient cell lines. Interestingly, p53 wild-type cell lines consistently showed faster γH2AX loss half-times than p53-deficient cell lines. We conclude that cell line-dependent differences in loss of γH2AX after irradiation are related in part to intrinsic radiosensitivity. The possibility that the presence of γH2AX foci may not always signify the presence of a physical break, notably in some tumor cell lines, is also supported by these results.

Cell Cycle-Dependent Expression of Phosphorylated Histone H2AX: Reduced Expression in Unirradiated but not X-Irradiated G1-Phase Cells

Abstract MacPhail, S. H., Banáth, J. P., Yu, Y. T., Chu, E. and Olive, P. L. Cell Cycle-Dependent Expression of Phosphorylated Histone H2AX: Reduced Expression in Unirradiated but not X-Irradiated G1-Phase Cells. Radiat. Res. 159, 759–767 (2003). Exposure of cells to ionizing radiation causes phosphorylation of histone H2AX at sites flanking DNA double-strand breaks. Detection of phosphorylated H2AX (γH2AX) by antibody binding has been used as a method to identify double-strand breaks. Although generally performed by observing microscopic foci within cells, flow cytometry offers the advantage of measuring changes in γH2AX intensity in relation to cell cycle position. The importance of cell cycle position on the levels of endogenous and radiation-induced γH2AX was examined in cell lines that varied in DNA content, cell cycle distribution, and kinase activity. Bivariate analysis of γH2AX expression relative to DNA content and synchronization by centrifugal elutriation were used to measure cell cycle-specific expression of γH2AX. With the exception of xrs5 cells, γH2AX level was approximately 3 times lower in unirradiated G1-phase cells than S- and G2-phase cells, and the slope of the G1-phase dose–response curve was 2.8 times larger than the slope for S-phase cells. Cell cycle differences were confirmed using immunoblotting, indicating that reduced antibody accessibility in intact cells was not responsible for the reduced antibody binding in G1-phase cells. Early apoptotic cells could be easily identified on flow histograms as a population with 5–10-fold higher levels of γH2AX, although high expression was not maintained in apoptotic cells by 24 h. We conclude that expression of γH2AX is associated with DNA replication in unirradiated cells and that this reduces the sensitivity for detecting radiation-induced double-strand breaks in S- and G2-phase cells.

Residual γH2AX foci as an indication of lethal DNA lesions

BackgroundEvidence suggests that tumor cells exposed to some DNA damaging agents are more likely to die if they retain microscopically visible γH2AX foci that are known to mark sites of double-strand breaks. This appears to be true even after exposure to the alkylating agent MNNG that does not cause direct double-strand breaks but does produce γH2AX foci when damaged DNA undergoes replication.MethodsTo examine this predictive ability further, SiHa human cervical carcinoma cells were exposed to 8 DNA damaging drugs (camptothecin, cisplatin, doxorubicin, etoposide, hydrogen peroxide, MNNG, temozolomide, and tirapazamine) and the fraction of cells that retained γH2AX foci 24 hours after a 30 or 60 min treatment was compared with the fraction of cells that lost clonogenicity. To determine if cells with residual repair foci are the cells that die, SiHa cervical cancer cells were stably transfected with a RAD51-GFP construct and live cell analysis was used to follow the fate of irradiated cells with RAD51-GFP foci.ResultsFor all drugs regardless of their mechanism of interaction with DNA, close to a 1:1 correlation was observed between clonogenic surviving fraction and the fraction of cells that retained γH2AX foci 24 hours after treatment. Initial studies established that the fraction of cells that retained RAD51 foci after irradiation was similar to the fraction of cells that retained γH2AX foci and subsequently lost clonogenicity. Tracking individual irradiated live cells confirmed that SiHa cells with RAD51-GFP foci 24 hours after irradiation were more likely to die.ConclusionRetention of DNA damage-induced γH2AX foci appears to be indicative of lethal DNA damage so that it may be possible to predict tumor cell killing by a wide variety of DNA damaging agents simply by scoring the fraction of cells that retain γH2AX foci.

Radiosensitization by the Histone Deacetylase Inhibitor PCI-24781

Purpose: PCI-24781 is a novel broad spectrum histone deacetylase inhibitor that is currently in phase I clinical trials. The ability of PCI-24781 to act as a radiation sensitizer and the mechanisms of radiosensitization were examined. Experimental Design: Exponentially growing human SiHa cervical and WiDr colon carcinoma cells were exposed to 0.1 to 10 μmol/L PCI-24781 in vitro for 2 to 20 h before irradiation and 0 to 4 h after irradiation. Single cells and sorted populations were analyzed for histone acetylation, H2AX phosphorylation, cell cycle distribution, apoptotic fraction, and clonogenic survival. Results: PCI-24781 treatment for 4 h increased histone H3 acetylation and produced a modest increase in γH2AX but negligible cell killing or radiosensitization. Treatment for 24 h resulted in up to 80% cell kill and depletion of cells in S phase. Toxicity reached maximum levels at a drug concentration of ∼1 μmol/L, and cells in G1 phase at the end of treatment were preferentially spared. A similar dose-modifying factor (DMF0.1 = 1.5) was observed for SiHa cells exposed for 24 h at 0.1 to 3 μmol/L, and more radioresistant WiDr cells showed less sensitization (DMF0.1 = 1.2). Limited radiosensitization and less killing were observed in noncycling human fibroblasts. Cell sorting experiments confirmed that depletion of S-phase cells was not a major mechanism of radiosensitization and that inner noncycling cells of SiHa spheroids could be sensitized by nontoxic doses. PCI-24781 pretreatment increased the fraction of cells with γH2AX foci 24 h after irradation but did not affect the initial rate of loss of radiation-induced γH2AX or the rate of rejoining of DNA double-strand breaks. Conclusions: PCI-24781 shows promise as a radiosensitizing agent that may compromise the accuracy of repair of radiation damage.

Higher-Order Chromatin Structure-Dependent Repair of DNA Double-Strand Breaks: Involvement of the V(D)J Recombination Double-Strand Break Repair Pathway

Repair of DNA double-strand breaks (DSBs) is linked to the V(D)J recombination pathway through investigations of radiation-sensitive mutants. Here we report a possible association between the distribution of DSBs within higher-order chromatin structures and this pathway. Both murine severe combined immunodeficient (SCID) and Chinese hamster XR-1 cells exhibit defective DNA DSB repair and defective V(D)J recombination. The DSB repair defect is not complete, with only a subset of slowly repairing lesions affected by the mutations in these cell lines. We used a modified neutral filter elution procedure which retained elements of higher-order chromatin structures, namely nuclear matrix-DNA interactions. X-ray-induced DSBs that occurred as multiples within looped DNA structures were nonrepairable in SCID and XR-1 cells. In contrast, these lesions were repaired in radioresistant wild-type cells. Cell lines complemented with human DNA containing the respective complementing genes (XRCC7 and XRCC4) showed an increased rate of DSB repair. These results agree with previous findings with xrs5 cells (a member of the XRCC5 group). Xrs5 cells are defective for the Ku p80 subunit of the V(D)J recombination complex and show repair and V(D)J recombination defects similar to those of SCID and XR-1 cells.

Higher-Order Chromatin Structure-Dependent Repair of DNA Double-Strand Breaks: Factors Affecting Elution of DNA from Nucleoids

The nuclear matrix is increasingly identified with the processing of DNA damage. Previous work has suggested that association of DNA with the matrix can influence the repair of DNA double-strand breaks (DSBs) and the sensitivity of mammalian cells to ionizing radiation. By selectively examining DSBs that occur as multiples (multiple DSBs) within looped DNA structures, we have identified a subset of DSBs that repair with slow kinetics through the V(D)J recombination-associated DSB repair pathway. Enrichment of S-phase populations by centrifugal elutriation and selective examination of nascent DNA by pulse-labeling were used to demonstrate that elution of DNA from nucleoids is retarded by the presence of replicating DNA. Previously, application of a Poisson-based model of induction of multiple DSBs and DNA elution to a panel of mammalian cell lines indicated that the size of the looped chromatin domains varied between cell lines. The data presented here explain the range in domain sizes between cells as the result of differences in the percentage of cells actively replicating their DNA. Correction of the model to account for S-phase populations results in a looped domain size of 2.9 Mbp independent of cell type. Single-cell gel electrophoresis of nucleoids provides additional evidence for such sized structures. Stabilization of DNA to elution during S phase does not permit repair of DSBs in the DSB repair mutants xrs5 and St.SCID, both defective for the DSB repair pathway associated with V(D)J recombination.

Hypertonic Saline Enhances Expression of Phosphorylated Histone H2AX after Irradiation

Abstract Reitsema, T. J., Banáth, J. P., MacPhail, S. H. and Olive,P. L. Hypertonic Saline Enhances Expression of Phosphorylated Histone H2AX after Irradiation. Radiat. Res. 161, 402– 408 (2004). Phosphorylation of histone H2AX at serine 139 occurs at sites surrounding DNA double-strand breaks, producing discrete spots called “foci” that are visible with a microscope after antibody staining. This modification is believed to create changes in chromatin structure and assemble various repair proteins at sites of DNA damage. To examine the role of chromatin structure, human SiHa cells were exposed to hypertonic salt solutions that are known to condense chromatin and sensitize cells to chromosome damage and killing by ionizing radiation. Postirradiation incubation in 0.5 M Na+ increased γH2AX expression about fourfold as measured by flow cytometry and immunoblotting, and loss of γH2AX was inhibited in the presence of high salt. Focus size rather than the number of radiation-induced γH2AX foci was also increased about fourfold. When high-salt treatment was delayed for 1 h after irradiation, effects on focus size and retention were reduced. The increase in focus size was associated with a decrease in the rate of rejoining of double-strand breaks as measured using the neutral comet assay. We conclude that γH2AX expression after irradiation is sensitive to salt-induced changes in chromatin structure during focus formation, and that a large focus size may be an indication of a reduced ability to repair DNA damage.

RPA foci are associated with cell death after irradiation

Abstract MacPhail, S. H. and Olive, P. L. RPA Foci are Associated with Cell Death after Irradiation. Radiat. Res. 155, 672–679 (2001). Complexes containing replication protein A (RPA) were observed in human TK6 and WIL-2NS lymphoblast cells and SiHa cervical carcinoma cells exposed to 250 kV X rays. Image analysis of individual cells with fluorescence-tagged anti-RPA antibodies was used to measure numbers of discrete foci per cell. RPA foci formed in S-phase cells in response to radiation doses as low as 0.5 Gy, and the number of foci/nucleus was linearly related to dose up to 50 Gy. The maximum number of cells with foci occurred 4–8 h after exposure to 4 Gy, and subsequently declined. However, the number of RPA foci per nucleus (in those cells with foci) reached a maximum after 2–4 h. Apoptotic nuclei from irradiated TK6 and WIL-2NS cells initially contained foci, but these were lost as degradation continued. Radiation-induced micronuclei in SiHa cells were greatly enriched for RPA foci, and cells with nuclei without foci often contained micronuclei with multiple RPA foci. In SiHa cells examined up to 7 days after 4 Gy, RPA foci reappeared in one or more cells in up to 90% of the surviving colonies, and some cells contained 150 or more distinct foci. Reappearance of these complexes could be indicative of radiation-induced genomic instability. These results are consistent with the idea that RPA foci observed several hours after irradiation represent irreparable lesions and as such might be useful in identifying radiosensitive cells.

Growth of V79 cells as xenograft tumors promotes multicellular resistance but does not increase spontaneous or radiation-induced mutant frequency

Abstract Banáth, J. P., Sinnott, L., Larrivée, B., MacPhail, S. H. and Olive, P. L. Growth of V79 Cells as Xenograft Tumors Promotes Multicellular Resistance but does not Increase Spontaneous or Radiation-Induced Mutant Frequency. Radiat. Res. 164, 733–744 (2005). A Chinese hamster V79 xenograft model was developed to determine whether cells subjected to a hypoxic tumor microenvironment would be more likely to undergo mutation at the HPRT locus. V79-171b cells stably transfected with VEGF and EGFP were grown subcutaneously in immunodeficient NOD/ SCID mice. V79-VE tumors were characterized for host cell infiltration, doubling time, hypoxic fraction, vascular perfusion, and response to ionizing radiation. When irradiated in vitro, the mutant frequency for a given surviving fraction did not differ for cells grown in vivo or in vitro. Similar results were obtained using HCT116 human colorectal carcinoma cells grown as xenografts. However, V79-VE cells grown as xenografts were significantly more resistant to killing than monolayers. The background mutant frequency and the radiation-induced mutant frequency did not differ for tumor cells close to or distant from blood vessels. Similarly, tumor cells from well-perfused regions showed the same rate of strand break rejoining and the same rate of loss of phosphorylated histone H2AX as cells sorted from poorly perfused regions. Therefore, deleterious effects of the tumor microenvironment on DNA repair efficiency or mutation induction could not be demonstrated in these tumors. Rather, development of multicellular resistance in V79-VE tumors acted to reduce mutant frequency for a given dose of radiation.