ANTICANCER RESEARCH 26: 3311-3316 (2006)
Correlations between Radiation-induced Double Strand Breaks and Cell Cycle Checkpoints in X-irradiated NIH/3T3 Fibroblasts MICKAEL J. CARIVEAUA1,2, GERHARD W. KALMUS3, ROBERTA M. JOHNKE2, RON R. ALLISON2, MARK EVANS2 and DONALD HOLBERT4
Departments of 1Physics, 3Biology and 4Biostatistics, East Carolina University, Greenville; 2Department of Radiation Oncology, The Brody School of Medicine, East Carolina University, Greenville, North Carolina, U.S.A.
Abstract. Background: To better understand the relationship between the formation of radiation-induced DNA double strand breaks (DSB) and cell cycle checkpoint activation, studies were conducted in the NIH/3T3 fibroblast cell line in order to establish correlations between the temporal appearance of Á-H2AX foci (a DSB) and the expression of the cell cycle regulatory proteins, the cyclins, and their cyclin kinase inhibitor, p21. Materials and Methods: Immunocytochemistry was used to determine the expression of cyclin E, A, B1, p21, and the generation of DSB in NIH/3T3 cells exposed to 2 or 4 Gy X-irradiation. Results and Discussion: The data suggest that the G1/S- and S-phase delay (cyclin E and cyclin A protein levels) are dependent on the dose of radiation while the G2/M (cyclin B1 protein levels) delay is dependent on the quantity of DSB sustained by the irradiated cell. One of the first responses to radiation-induced DNA double strand breaks (DSB) is the immediate and massive phosphorylation of histone H2A homologs (H2AX) mediated through activation of the ATM gene (1). The phosphorylated H2AX molecules (Á-H2AX) associate at the site of the DSB, forming microcsopic foci, with each foci representing one break in the DNA double helix (2). Recent evidence suggests that the phosphorylation of the H2AX proteins serves as a signal for the recruitment of repair enzymes (3). Histone H2AX has been implicated in the maintenance of genomic stability in response to DNA damage and is phosphorylated at a PI3KK (phosphatidyinositol-3 kinase related kinase) motif in the carboxyl (COOH) terminus minutes after DNA damage (4-7, 24). The cell’s defense against DSB formation involves an extensive repair process that can only occur during a period
Correspondence to: Mickael J. Cariveau, Ph.D., Southern Research Institute, Department of Biochemistry and Molecular Biology, 2000 9th Avenue South, Birmingham, AL 35205, U.S.A. Tel: 205581-256, e-mail:
[email protected] Key Words: DSB, cell cycle checkpoints.
0250-7005/2006 $2.00+.40
of transitional delay in cell proliferation. Better known as cell cycle, this transitional delay allows the cell time to repair the radiation-induced damage and is controlled at 3 checkpoints within the cell cycle by the interaction of specific molecules, which regulate cell cycle progression, the cyclin/CDK complexes (8-10). Cyclins function by coupling to their corresponding cyclin-dependent kinases (CDK) forming a holoenzyme referred to as a cyclin kinase, which is then responsible for regulating the transition from G1 to S, through S, and G2 to M. The G1/S transition is regulated by the cyclin E/Cdk2 complex, with cyclin E mRNA levels increasing sharply in late G1 and reaching maximum levels at the G1/S transition. The S-phase checkpoint was first identified as a point of delay by Painter and Young (11). Since that time, it has been shown that the intra S-phase checkpoint follows the same mechanism of delay as the initiation phase of G1/S with Chk2-mediated phosphorylation of Cdc25A, leading to its ubiquitinated proteosomal degradation and thus inhibition of cyclin A activity. The G2/M checkpoint is controlled by the mitosispromoting factor (MPF), cyclin B1/CDK1. Cyclin B1/CDK1 induces mitosis by phosphorylating and activating enzymes regulating chromatin condensation, nuclear membrane breakdown and microtubule reorganization (12). In normal cells, cyclin B1 accumulates in the cytoplasm during S and G2, before translocating to the nucleus prior to prophase (13). The G2/M damage checkpoint is initiated by DNA damage activation of ATM/ATR cascades, with dephosphorylation of Cdc25B/C initiating the replicative block. Cdc25C phosphatase is itself regulated by phosphorylations at 2 major sites, serine 216 and the uncharacterized amino-terminus which is phosphorylated by Chk1 and Chk2 after DNA damage (14). While the causal relationship between radiation-induced DSB and cell cycle delay are well established, evidence supporting the correlation between cyclin expression and radiation-induced division delay in normal cells is, at best, circumstantial, since much of the data reported to date was
3311
ANTICANCER RESEARCH 26: 3311-3316 (2006) generated from malignant cell lines or cells exposed to doses ≥8 Gy (9, 10). Studies have shown that low doses (≤1 Gy) are able to initiate division delay (15). This raises the question of whether a quasi – threshold level of DNA damage exists for initiating cyclin/CDK-CKI-mediated division delay. Until recently, methods to establish such a correlation between threshold levels of DNA damage and the initiation of cell cycle delay based on cyclin/CDK-CKI expression were not available. However, the identification of the phosphorylated histone, Á-H2AX makes the establishment of these correlates possible. Presently, however, it is unknown to what extent the expression of the Á-H2AX foci is related to division delay and, vis-à-vis, to the expression of the cyclin/CDK-CKI complexes that regulate cell cycle progression. For this reason, correlations between the magnitude of molecular damage (Á-H2AX foci), the cellular (cell cycle) response and the expression of specific cyclin/CDK – CKI complexes were examined in the NIH3T3 fibroblast cell line. The cyclin proteins chosen are involved in specific checkpoint activation and include cyclin E (G1/S), cyclin A (Intra – S), and cyclin B1 (G2/M) and their corresponding inhibitor, p21WAF-1.
Materials and Methods Exponentially growing NIH/3T3 cells were plated into 35 mm dishes containing a 22 cm2 coverslips at densities from 5x103 - 6.0x104 per coverslip and incubated for 24 h prior to washing in PBS and fixation in –20ÆC methanol. For irradiation studies, 5x103 were grown as described above, irradiated (2 or 4 Gy) with a Siemen’s Stabilipan at a dose rate of 0.3 Gy/min, and harvested from 0.25 – 24 h, respectively. Radiation induced expression of the cyclins, p21 and Á-H2AX. Exponentially growing cultures of NIH/3T3 cells were plated into 35 mm dishes containing a sterile 22 cm2 coverslip at densities of 1.0x104 cells per coverslip and incubated in growth media for 12 h. Following a 12-h incubation, samples were irradiated to a total dose of 2 or 4 Gy and were then incubated for times ranging from 0.25-24 h as previously described (16). Following incubation, the cells were washed with PBS and fixed in –20ÆC methanol for 10 min, air dried and mounted on slides. Immunocytochemistry for the cyclins, p21 and Á-H2AX. Immunocytochemistry was incorporated as previously described (16). Briefly, cells were permeabilized in (TBS)/0.1% Triton X-100, blocked with 1.5% goat serum in PBS and 100 ml of primary antibody were added, either rabbit polyclonal p21 (Santa Cruz Biotech., sc-756, 2 Ìg/ml), cyclin B1 (Santa Cruz Biotech., sc-752, 4 Ìg/ml), cyclin E (Santa Cruz Biotech., sc-481, 2 Ìg/ml), cyclin A (Santa Cruz Biotech., sc-751, 2 Ìg/ml) or rabbit polyclonal Á-H2AX (Upstate Biotech., 07164, 0.4 Ìg/ml) and the cells were then incubated overnight at 4ÆC. The cells were subsequently washed in TBS and treated with biotinylated goat anti-rabbit IgG (Santa Cruz, sc-2040, 5 Ìg/ml). They were then rinsed twice in TBS, incubated in ABC reagent (Vector, PK-4000) and resolved with 3,4-diaminobenzidine (DAB, Vector, SK4100) and counterstained with eosin. Four samples from three
3312
Figure 1. The average number of Á-H2AX foci formation per nucleus following 2 or 4 Gy irradiation. The error bars represent the standard deviation. The means of 3 independent experiments are provided.
independent experiments for a total of 12 counts per data point were then scored. Unless otherwise noted, the number of positive cells (cells staining for gH2AX, p21, or cyclins) was determined from a total of 500 cells. The mean number of positive cells from 3 independent experiments was used to determine data points and statistical significance was set at p≤0.001.
Results The initial response of the cell to radiation-induced DSB involves the rapid phosphorylation of the histone homolog H2AX. When phosphorylated (Á-H2AX), this homolog forms microscopic foci at the site of the DSB. To better assess the relationship between the formation of these foci and thus the extent of DNA damage, the expressions of cyclin E, A, B1 and p21 were correlated with the temporal formation of DSB in NIH3T3 cells treated with 2 or 4 Gy irradiation (Figure 1). As shown, foci formation occured within 15 min of 2 Gy treatment (13.7 DSB) and reached a maximum of 22 DSB per nucleus within 30 min. Within an hour, the number of DSB decreased to 13.8 and continued to decline over the next 2 h to 8.8 and 7.2 DSB, respectively. By 6 h following 2 Gy, all values have returned to pretreatment levels, i.e., there were no visible foci present. A dose increase to 4 Gy led to a greater number of DSB within the cell (Figure 1). Within
Cariveau et al: Á-H2AX Mediates Cell Cycle Checkpoint Activation in X-irradiated Fibroblasts
15 min after irradiation, 25 DSB are present and by 1 h, 33 DSB were evident. The level of damage remained elevated (29.7 and 22.6) at 2 and 4 h, respectively, before declining to 11 DSB by 6 h. Within 12 h of irradiation, the distribution of DSB has returned to pre-irradiated control levels. While the trends at 2 and 4 Gy were similar, there were noticeable and significant differences between DSB formation in the 2 and 4 Gy groups. Specifically, both the magnitude and temporal expression of foci increased at 4 Gy. Using Á-H2AX as a marker for DNA damage, correlations were established between the formation of Á-H2AX foci and the expressions of the cyclin E, A, B1 and p21 proteins following 2 or 4 Gy irradiation. To establish the statistical significance of cyclin/CKI response over time and dose, a univariate general linear model was incorporated and the lack of significance (p
