We present a procedure for the comet assay, a gel electrophoresis-based method that can be used to measure DNA damage in individual eukaryotic cells. It is versatile, relatively simple to perform and sensitive. Although most investigations make use of its ability to measure DNA single-strand breaks, modifications to the method allow detection of DNA double-strand breaks, cross-links, base damage and apoptotic nuclei. The limit of sensitivity is approximately 50 strand breaks per diploid mammalian cell. DNA damage and its repair in single-cell suspensions prepared from yeast, protozoa, plants, invertebrates and mammals can also be studied using this assay. Originally developed to measure variation in DNA damage and repair capacity within a population of mammalian cells, applications of the comet assay now range from human and sentinel animal biomonitoring (e.g., DNA damage in earthworms crawling through toxic waste sites) to measurement of DNA damage in specific genomic sequences. This protocol can be completed in fewer than 24 h.

No abstract

A method for measuring DNA damage to individual cells, based on the technique of microelectrophoresis, was described by Ostling and Johanson in 1984 (Biochem. Biophys. Res. Commun. 123, 291-298). Cells embedded in agarose are lysed, subjected briefly to an electric field, stained with a fluorescent DNA-binding stain, and viewed using a fluorescence microscope. Broken DNA migrates farther in the electric field, and the cell then resembles a "comet" with a brightly fluorescent head and a tail region which increases as damage increases. We have used video image analysis to define appropriate "features" of the comet as a measure of DNA damage, and have quantified damage and repair by ionizing radiation. The assay was optimized for lysing solution, lysing time, electrophoresis time, and propidium iodide concentration using Chinese hamster V79 cells. To assess heterogeneity of response of normal versus malignant cells, damage to both tumor cells and normal cells within mouse SCC-VII tumors was assessed. Tumor cells were separated from macrophages using a cell-sorting method based on differential binding of FITC-conjugated goat anti-mouse IgG. The "tail moment", the product of the amount of DNA in the tail and the mean distance of migration in the tail, was the most informative feature of the comet image. Tumor and normal cells showed significant heterogeneity in damage produced by ionizing radiation, although the average amount of damage increased linearly with dose (0-15 Gy) and suggested similar net radiosensitivities for the two cell types. Similarly, DNA repair rate was not significantly different for tumor and normal cells, and most of the cells had repaired the damage by 30 min following exposure to 15 Gy. The heterogeneity in response did not appear to be a result of differences in response through the cell cycle.

GammaH2AX can be detected with excellent sensitivity using both flow and image analysis. The rate of gammaH2AX loss may be an important factor in the response of cells to ionizing radiation, with more rapid loss and less retention associated with more radioresistant 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 (␥H2AX). 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, ␥H2AX representing up to 30% of the initial signal still present. SW756 cells showed almost four times higher background levels of ␥H2AX and no residual ␥H2AX compared with the most radiosensitive HPV-negative C33A cells that showed the lowest background and retained 30% of the maximum level of ␥H2AX. Radiation sensitivity, measured as clonogenic-surviving fraction after 2 Gy, was correlated with the fraction of ␥H2AX remaining 24 hours after irradiation. A substantial correlation with ␥H2AX 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 ␥H2AX loss half-times than p53-deficient cell lines. We conclude that cell line-dependent differences in loss of ␥H2AX after irradiation are related in part to intrinsic radiosensitivity. The possibility that the presence of ␥H2AX foci may not always signify the presence of a physical break, notably in some tumor cell lines, is also supported by these results.