Recovery from DNA Damage-induced G2 Arrest Requires Actin-binding Protein Filamin-A/Actin-binding Protein 280*

Filamin-A (filamin-1) is an actin-binding protein involved in the organization of actin networks. Our previous study shows that filamin-A interacts with BRCA2, and lack of filamin-A expression results in increased cellular sensitivity to several DNA damaging agents in melanoma cells (Yuan, Y., and Shen, Z. (2001) J. Biol. Chem. 276, 48318-48324), suggesting a role of filamin-A in DNA damage response. In this report, we demonstrated that deficiency of filamin-A results in an 8-h delay in the recovery from G2 arrest in response to ionizing radiation. However, filamin-A deficiency does not affect the initial activation of the G2/M checkpoint. We also found that filamin-A deficiency results in sustained activation of Chk1 and Chk2 after irradiation. This in turn causes a delay in the dephosphorylation of phospho-Cdc2, which is inhibitory to the G2/M transition. In addition, filamin-A-deficient M2 cells undergo mitotic catastrophe-related nuclear fragmentation after they are released from the G2 arrest. Together, these data suggest a functional role of filamin-A in the recovery from G2 arrest and subsequent mitotic cell death after DNA damage.

In response to DNA damage, cell cycle checkpoints are activated to delay or block the progression of the cell cycle. This may prevent damaged cells from progressing into the next phase of the cell cycle, thus facilitating the maintenance of genomic stability (1)(2)(3)(4). When DNA damage is repaired, cells recover from cell cycle arrest and resume their normal cell cycle progression. Studies in yeast suggest that the transition from cell cycle arrest to cell cycle progression is not simply a passive response to the removal of DNA damage. It appears to be an active process in which cells can adapt to prolonged cell cycle arrest even when the DNA damage has not been completely repaired (5)(6)(7)(8). However, the mechanism that regulates the recovery of cell cycle arrest is poorly understood, especially in mammalian cells.
Our previous study has shown that an actin-binding protein filamin-A (filamin-1, FLNa, 1 or ABP-280) interacts with BRCA2, and deficiency of filamin-A renders melanoma cells more sensitive to ␥-rays, bleomycin, and UV irradiation (9). In this study, we further investigated the role of filamin-A in cell cycle regulation using a filamin-A-deficient melanoma cell line (M2) and a C8161 melanoma cell line in which filamin-A expression was knocked down by RNA interference. We found that filamin-A deficiency correlated with an approximate 8-h delay in the recovery from G 2 arrest after irradiation. However, filamin-A deficiency had little effect on the initiation of G 2 arrest, indicating intact activation of the G 2 /M checkpoint. The delayed G 2 recovery correlated with delayed dephosphorylation of the phospho-Cdc2 protein and sustained activation of the Chk1 and Chk2 kinases. Furthermore, we found that deficiency of filamin-A renders cells more prone to mitotic cell death. These data provide the first evidence that filamin-A plays a significant role in the recovery from radiation-induced G 2 arrest and mitotic cell death in mammalian cells.

EXPERIMENTAL PROCEDURES
Cell Culture-Human A7 and M2 melanoma cells were kindly provided by Drs. T. P. Stossel and Y. Ohta (Brigham and Women's Hospital, Harvard Medical School). The M2 is a human melanoma cell line lacking filamin-A expression. The A7 cell line was derived from M2 through the stable transfection of a plasmid expressing full-length filamin-A (10). A7 and M2 cells were subcultured twice a week in minimum essential medium with Earle's salt (EMEM) (M2) or EMEM with 0.5 mg/ml G418 (A7). The C8161 melanoma cells were kindly provided by Dr. D. R. Welch (Pennsylvania State University College of Medicine) and cultured in RPMI 1640 medium. All media contained 10% fetal bovine serum and 1% penicillin/streptomycin. To induce DNA damage, cells were irradiated with ␥-rays ( 137 Cs, 0.938 Gy/min) at room temperature and immediately returned to a 37°C incubator. To inhibit Chk1 and Chk2 activities, 100 nM UCN-01 (7-hydroxystaurosporine, National Cancer Institute) was added to the cell culture 30 min before the cells were exposed to radiation and was maintained in the media until the cells were collected.
Construction of pPUR/U6 Vector for Short Hairpin RNA (shRNA) Expression and to Confer Puromycin Resistance-The BamHI/EcoRI restriction fragment of pBS/U6 (a gift from Y. Shi, Harvard Medical School of Medicine, Boston, MA) that contains the U6 promoter for short hairpin RNA expression was inserted into the BamHI/EcoRI sites of the pPUR vector (Clontech, Palo Alto, CA). This resulted in a new vector pPUR/U6 to express the short hairpin RNA and to confer puromycin resistance in mammalian cells. The strategy to construct an shRNA expression vector was described by Sui et al. (11). Briefly, two inverted sequences of 21 bases (5ЈGGG CTG ACA ACA GTG TGG TGC3Ј) of the filamin-A cDNA were linked by a 9-base linker sequence and cloned downstream of the U6 promoter, resulting in the RNA interference vector pPUR/U6/filamin-A. This vector directs the synthesis of a filamin-A-specific shRNA in vivo. Vacant vector was used as the negative control.
Analysis of DNA Content by Flow Cytometry-Cells were trypsinized, washed with phosphate-buffered saline, and fixed with 70% ethanol. Fixed cells were then pelleted, washed with phosphate-buffered saline, and suspended in 200 l of citrate buffer (250 mM sucrose, 0.05% Me 2 SO, 40 mM trisodium citrate, pH 7.6). Nine hundred l of Solution A (0.003% trypsin in stock buffer (3.4 mM trisodium citrate, 0.1% Nonidet P-40, 1.5 mM spermine tetrahydrochloride, 0.5 mM Trizma, pH 7.6)) were added to the cells. Cells were then incubated at room temperature for 10 min with 5-6 vortexes during the incubation. Afterward, 750 l of Solution B (0.025% trypsin inhibitor, 0.01% ribonuclease A in stock buffer) was added for another incubation of 10 min at room temperature with two vortexes during the incubation. Then, 750 l of Solution C (0.0416% propidium iodide, 3.3 mM spermine tetrahydrochloride in stock buffer) was added to the cells on ice. Afterward, the cells were covered with aluminum foil and were ready for flow cytometry analysis. In each flow cytometry assay, 20,000 cells were collected using FACScan (BD Biosciences) (12).
Mitotic Index Assay-The cells were continuously treated with 1 g/ml nocodazole for various time periods. Then the cells were harvested, swollen hypotonically for 5 min in 75 mM KCl, gently pelleted, and fixed in 0.5 ml of Carnoy's fixative (3:1 methanol:glacial acetic acid, v/v) for 5 min. Subsequently, the cells were dropped on pre-wetted slides and allowed to air dry. Fixed cells were stained with DAPI, and mitotic cells were counted with fluorescent microscopy (Zeiss Axioskop 2) as described previously (13). For each sample, the percentage of mitotic cells was determined from 200 -300 cells counted, and the experiment was repeated three to four times.
Visualization of Nuclear Fragmentation-To demonstrate fragmen-tation of the nucleus, cells were fixed with methanol for 15 min at Ϫ20°C. The cells were then stained with rabbit anti-␣-tubulin antibody (Sigma) and visualized by Texas Red-conjugated secondary antibody (Jackson ImmunoResearch Laboratories Inc., West Grove, PA). DAPI was used to visualize the nucleus.

RESULTS
Delayed Recovery from G 2 Arrest in Filamin-A-deficient M2 Melanoma Cells-The M2 melanoma cell line does not express filamin-A, whereas the A7 cell line is an isogenic cell line derived from M2 in which filamin-A expression is restored to its normal level (10). To investigate whether absence of filamin-A in M2 cells affects DNA damage-induced checkpoint activation, we studied the cell cycle profiles of A7 and M2 cells after irradiation. Specifically, asynchronized populations of A7 and M2 cells in the exponential growth phase were irradiated with 8 Gy of ␥-rays, and their cell cycle profiles were analyzed at various time points from 4 -24 h after irradiation. As shown in Fig. 1, nonirradiated control A7 and M2 cells displayed a similar pattern of heterogeneous log phase distribution throughout the 24-h period. When the cells were irradiated with ␥-rays, both the A7 and M2 cells started to accumulate into the G 2 phase at 4 h and peaked at 16 h. At 24 h, the percentage of A7 cells in the G 2 phase decreased significantly, suggesting a release from G 2 arrest. However, the percentage of M2 cells in G 2 was still high at 24 h, suggesting a slower release from the G 2 phase.
Our previous study showed that M2 cells are ϳ10 times more sensitive to irradiation than A7 cells as measured by a clonogenic survival assay (9) (see also Fig. 2A). It is possible that the delayed release from G 2 arrest in the M2 cell line is because of their higher sensitivity to radiation. To verify this hypothesis, we compared the cell cycle distribution of A7 and M2 cells at the same survival rate, namely 8 Gy for A7 and 6 Gy for M2 (pair I), as well as 12 Gy for A7 and 8 Gy for M2 (pair II) ( Fig.  2A). As shown in Fig. 2B, M2 cells again displayed a much higher percentage in the G 2 phase at 24 h for both pairs of iso-survival doses. This further supports the notion that the lack of filamin-A expression in M2 cells results in a slower recovery from G 2 arrest, which is not necessarily because of the reduced survival in M2 cells.
To determine the difference in G 2 arrest recovery between A7 and M2 cells, we conducted a nocodazole-trapping assay. Briefly, the microtubule-disrupting agent nocodazole was added to cells immediately after irradiation to block cells at the anaphase of mitosis. The accumulation of mitotic cells was determined by counting the cells that had condensed chromosomes. Under normal conditions, the mitotic index increases immediately after nocodazole treatment. However, a delayed accumulation of mitotic cells is expected if the G 2 /M checkpoint is activated to arrest the cells in the G 2 phase. The mitotic index will increase if the G 2 -arrested cells recover from this arrest. As shown in Fig. 3, the mitotic indices of both A7 and M2 cells accumulated in a similar pattern as time progressed in the control groups (0 Gy), indicating that filamin-A is not necessary for normal progression of G 2 cells into the M phase. However, after a prolonged mitotic blockage by nocodazole, the percentage of mitotic cells tended to decline. This might reflect an escape of mitotic cells from the nocodazole blockage, which is not unusual for many tumor cell lines. There was no mitotic accumulation in the A7 cells during the first 8 h. The mitotic index started to increase between 8 -16 h post-irradiation, indicating that the G 2 /M checkpoint was activated upon irradiation in the A7 cells and lasted for ϳ8 h. However, the mitotic index in the M2 cells remained low during the first 16 h post-irradiation. It started to increase at 16 -24 h, suggesting that there was normal activation of the G 2 /M checkpoint in the M2 cells, but the G 2 arrest lasted for at least 16 h, ϳ8 h longer than that of the A7 cells. Therefore, data here are consistent with the data in Figs. 1 and 2. Together, these data demonstrated that a lack of filamin-A expression in M2 cells results in an ϳ8-h delay in the recovery from G 2 arrest after irradiation.
Delayed Recovery from G 2 Arrest in C8161 Melanoma Cells with Filamin-A Knockdown Using RNA Interference-To confirm the results obtained with the M2/A7 cell lines (Figs. 1-3), the filamin-A expression in the melanoma cell line C8161 was knocked down using RNA interference, resulting in the cell line C8161/FLNa-KD (Fig. 4A). The mitotic index assay was carried  Cells were irradiated at 8 Gy, and 1 g/ml nocodazole was added immediately. At various time points (4 -48 h) after the nocodazole treatment, the percentage of mitotic cells was counted (see "Experimental Procedures" for details). The distance between the curves of irradiated A7 and M2 cells reflects the delayed recovery from G 2 arrest. out using this cell line, and its parental cell line C8161 transfected with an empty vector as the control. As shown in Fig. 4B, in the absence of irradiation, a similar accumulation of mitotic cells was observed in C8161 and C8161/FLNa-KD after nocodazole block. When the cells were irradiated, both cell lines displayed a delayed accumulation of mitotic cells when compared with nonirradiated cells, indicating an activation of the G 2 /M checkpoint and subsequent G 2 arrest. Recovery from G 2 arrest in C8161/FLNa-KD was about 8 h longer than that observed in the C8161 parental cells. These data confirmed that the delayed recovery from G 2 arrest in the M2 cells was because of the lack of filamin-A expression (Fig. 3).
It is of interest to note that in the absence of irradiation, the percentage of mitotic cells decreased after 16 h of blockage with nocodazole. This is consistent with the observation that in the absence of irradiation, the percentage of mitotic A7 and M2 cells decreased 24 -32 h after nocodazole blockage (Fig. 3). The mechanism behind this phenomenon is not clear, but it is not unusual for some tumor cells to escape nocodazole-induced mitotic blockage, and it may reflect a defect of the spindle checkpoint in these cells.

Delayed Recovery from G 2 Arrest in Filamin-A-deficient Cells Correlates with Slower Dephosphorylation of Phospho-Cdc2-It
is well established that the onset of mitosis is triggered by the activation of cyclin B/Cdc2 and that the activation of cyclin B/Cdc2 is absolutely required for the transition of G 2 cells into the M phase (4, 14 -19). Generally, cyclin B/Cdc2 is inactivated by phosphorylations of the Thr-14 and Tyr-15 residues during the G 2 phase (4,(15)(16)(17). When the cells are ready to enter mitosis, the mitosis-promoting phosphatase Cdc25C dephosphorylates Cdc2 at Thr-14 and Tyr-15 (4,(15)(16)(17). To determine whether filamin-A-mediated recovery from G 2 arrest is regulated by Cdc2 dephosphorylation, we measured the status of Cdc2 phosphorylation in the A7 and M2 cells from 4 to 48 h after irradiation. As shown in Fig. 5, the Cdc2 triplet band (from top to bottom, panels A and C) represents double-phosphorylated, single-phosphorylated, and nonphosphorylated Cdc2, respectively. The double-phosphorylated Cdc2 is the form that sustains G 2 arrest. We confirmed that the top band in the triplet contained Tyr-15-phosphorylated Cdc2 using an an-tiphospho-Tyr-15-Cdc2 antibody (see Fig. 6C). In irradiated A7 cells (Fig. 5A), the relative abundance of double-phosphorylated Cdc2 (as compared with the other two forms) remained high from 8 to 16 h and started to decline after 16 h. These data are in agreement with those in Fig. 3, where the recovery from G 2 -arrested A7 cells started at 8 -16 h after irradiation. However, the relative abundance of the double-phosphorylated Cdc2 in M2 cells remained high during the first 24 h and declined slowly afterward. This delayed dephosphorylation of Cdc2 in M2 cells is consistent with the data in Fig. 3, where the recovery from G 2 arrest started 16 -24 h after irradiation. Thus, the dephosphorylation of the double-phosphorylated Cdc2 (Fig. 5) coincides with the recovery from G 2 -arrested cells (Fig. 3), suggesting that the filamin-A-dependent recovery from G 2 arrest is most likely related to the Cdc2-mediated G 2 /M control mechanism.
The Slower Dephosphorylation of Cdc2 in M2 Cells Is Attributed to the Sustained Activation of Chk1 and Chk2 Kinases-Activation of Chk1 and Chk2 is the upstream event that leads to the increased phosphorylation of Tyr-15 on Cdc2. Chk1 and Chk2 (also named hCds1) are the human homologues of the fission yeast checkpoint kinases Chk1 and Cds1, respectively. Chk2 is activated mainly in response to DNA damage caused by ionizing radiation and requires phosphorylation on its Thr-68 residue by ATM kinase (20,21). In contrast, Chk1 activation requires phosphorylation at Ser-345 by ATR kinase and is primarily elicited by DNA replication blocking agents such as hydroxyurea or nonionizing radiation such as UV light (22)(23)(24). It was recently reported that the ATR/Chk1 pathway can also be activated by ionizing radiation to elicit G 2 arrest in the absence of ATM (25). Therefore, the ATM/Chk2 and ATR/ Chk1 pathways probably overlap and cooperate after DNA damage. Both activated Chk1 and Chk2 phosphorylate Cdc25C at Ser-216, preventing Cdc25C from dephosphorylating Cdc2 at Thr-14 and Tyr-15 (26). Therefore, we measured the phosphorylation of Chk1 and Chk2 in the A7 and M2 cells following irradiation to determine whether the delayed G 2 recovery in the M2 cells is mediated by these upstream kinases.
As shown in Fig. 6A, the Ser-345-phosphorylated Chk1 (Chk1 (Ser-345-P)) level was low in both the A7 and M2 cells in the absence of irradiation. Irradiation significantly increased Chk1 phosphorylation at Ser-345 in both cell lines. In the A7 cells, the level of Ser-345-phosphorylated Chk1 peaked at 16 h and returned to normal by 24 h. However, in the M2 cells the level of phosphorylated Chk1 peaked at 8 h and returned to normal at 32 h. This suggests a more sustained phosphorylation of Chk1 at Ser-345 in M2 cells than in A7 cells. Chk2 was also activated by irradiation in both cell lines as shown in Fig.  6B. However, the activation was greater in M2 cells than in A7 cells (Fig. 6C). Prolonged phosphorylation of Chk1 and Chk2 in M2 cells agrees with the observed change of Tyr-15 phosphorylation of Cdc2 (Figs. 5 and 6C) and the recovery from G 2 arrest shown in Fig. 3. These data suggest that the delayed recovery from G 2 arrest in filamin-A-deficient M2 cells is because of the sustained activation of the G 2 /M checkpoint pathway involving the Chk1/Chk2-dependent signal transduction cascades.
To further confirm this notion, a Chk1 and Chk2 inhibitor, UCN-01, was used. UCN-01 was originally identified as an inhibitor of Chk1 (27,28), but it also inhibits the Chk2 kinase activity as shown recently (29). Because sustained activation of both Chk1 and Chk2 was observed in the filamin-A-deficient M2 cells after irradiation, we anticipated that inhibition of Chk1 and Chk2 by UCN-01 would abolish the delayed dephosphorylation of Cdc2 at Tyr-15. In other words, the abnormally high level of Tyr-15-phosphorylated Cdc2 in M2 cells would be reduced to a level comparable with that observed in A7 cells. To examine this possibility, 100 nM UCN-01 or Me 2 SO (as a control) was added to the cells 30 min before irradiation and kept in the medium until the cells were harvested. Twenty-four h after irradiation, the cells were harvested to measure the level of Tyr-15-phosphorylated Cdc2. As expected, UCN-01 treatment did not change the level of Tyr-15-phosphorylated Cdc2 in nonirradiated A7 (Fig. 7A, lanes 1 and 2) and M2 (Fig. 7A, lanes  5 and 6) cells. However, the radiation-induced Tyr-15 phosphorylation on Cdc2 in A7 (Fig. 7A, lane 3) and M2 (Fig. 7A, lane  7) cells was reduced to the level of nonirradiated A7 and M2 cells after UCN-01 treatment (lanes 4 and 8). This suggests that Chk1 and Chk2 are responsible for the sustained Tyr-15 phosphorylation of Cdc2 at 24 h after irradiation in the M2 cells (Figs. 5 and 6).
In addition, the nocodazole-trapping assay was carried out to demonstrate that inhibition of Chk1 and Chk2 also abolishes G 2 arrest in M2 cells. Briefly, 100 nM UCN-01 was added to the cell culture and maintained in medium until the cells were collected. Cells were then irradiated, and 1 g/ml nocodazole was added immediately. As seen in Fig. 3, the biggest difference in mitotic indices between A7 and M2 cells occurred at 24 h after irradiation. Therefore, the mitotic cells were scored at 24 h, and the mitotic index of each group was calculated. As expected (Fig. 7B), the addition of UCN-01 did not affect the mitotic indices of the A7 cells and nonirradiated M2 cells. As shown previously, in irradiated M2 cells there was a significant decrease in the percentage of mitotic cells (Fig. 7B), reflecting a delayed release of G 2 -arrested cells into mitosis as shown in Fig. 3. This delayed release was abolished by the addition of UCN-01 (Fig. 7B). These data further demonstrate that the G 2 arrest in M2 cells is controlled by the pathway involving Chk1 and Chk2 kinases, which are upstream effectors of Cdc25C and Cdc2.
Mitotic Catastrophe Followed the Delayed Recovery from G 2 Arrest in Filamin-A-deficient M2 Cells after Irradiation-When mammalian cells cannot repair their DNA damage, they can execute two major cell death pathways, apoptosis and mitotic catastrophe (30 -34). If apoptosis is the major cause of cell death, a sub-G 1 population of cells would be observed in a DNA content-based cell cycle analysis. Our data in Figs. 1 and 2 did not exhibit a sub-G 1 population of M2 cells, suggesting that apoptosis was not the cause of increased cell death. In addition, we used terminal deoxynucleotidyltransferase-mediated dUTP nick end labeling (TUNEL) assay to visualize potential apoptotic cells, but no positive TUNEL staining was observed (data not shown). Therefore, apoptotic response to radiation is unlikely to be the cause of enhanced toxicity in M2 cells. This observation is consistent with a recent report that a defect of filamin-A did not enhance the apoptotic response in M2 cells but rather that filamin-A might be a caspase substrate and is cleaved directly by protease granzyme B during apoptosis (35).
Because we did not observe apoptosis in M2 cells, we investigated whether the absence of filamin-A enhances mitotic catastrophe in these cells after irradiation. Mitotic catastrophe is a mode of cell death exhibited by most nonhematopoietic cell lineages in response to ionizing radiation (33,34). It is well known that the onset of nuclear fragmentation is a hallmark of mitotic catastrophe (36). Mitotic catastrophe is distinguishable from apoptosis because it lacks DNA degradation that can be observed by either DNA laddering on agarose gel or TUNEL staining (37). To test whether increased mitotic catastrophe is the cause of increased toxicity in M2 cells, cells displaying nuclear fragmentation were scored at various time points after irradiation. Fig. 8A, panel IV, shows the typical nuclear fragmentation observed in M2 cells after irradiation. As quantified in Fig. 8B, the M2 cells consistently displayed a higher percentage of cells undergoing nuclear fragmentation after irradiation compared with A7 cells. Based on these analyses, we conclude that the deficiency of filamin-A in M2 cells renders them more sensitive to radiation through the mitotic catastrophe pathway. In addition, the application of UCN-01 resulted in identically elevated levels of nuclear fragmentation between A7 and M2 cells. This suggests that abrogation of the G 2 /M checkpoint activation by UCN-01 abolishes the difference of nuclear fragmentation between M2 and A7 cells. DISCUSSION The activation of cell cycle checkpoints plays a significant role in the DNA damage response. It prevents damaged cells from entering the next phase of the cell cycle. Although much is known about the genes that activate the G 2 /M checkpoint, little is known about the genes that control the recovery from G 2 arrest in mammalian cells. The data presented here shed new light on this important aspect. We showed that filamin-A deficiency in the melanoma cell lines M2 and C8161/FLNa-KD caused a delayed recovery from G 2 arrest after irradiation. This correlated with the sustained activation of Chk1 and Chk2 kinases and delayed dephosphorylation of the phospho-Cdc2. However, deficiency of filamin-A in M2 and C8161/FLNa-KD cells did not affect the initiation of the G 2 /M checkpoint. Although the molecular mechanism by which filamin-A may affect the Cdc2-dependent recovery from G 2 arrest is not clear, several possible scenarios are worthy of further discussion. First, a certain level of filamin-A and a fragment of filamin-A normally reside in the nucleus (9,38). The nuclear filamin-A and the actin network may serve as a supporting matrix for the BRCA2-related DNA damage response process via an interaction between filamin-A and BRCA2. Thus, lack of filamin-A may result in delayed DNA repair that in turn leads to delayed recovery from G 2 arrest. Second, it is possible that filamin-A directly regulates the Cdc2-dependent recovery from G 2 arrest. Protein shuffling between the nucleus and the cytoplasm is required for precise regulation of Cdc2 phosphorylation and dephosphorylation (39 -41), and it has been recognized that filamin-A plays a significant role in protein translocation between the cytoplasm and the nucleus (42). Last but not least, disruption of actin organization in yeast cells activates a morphogenesis checkpoint that arrests cells at the G 2 /M border (43)(44)(45)(46). It is possible that filamin-A and its associated actin network contribute to G 2 /M regulation in a way mechanistically resembling the morphogenesis checkpoint observed in yeast cells. Nevertheless, our data present an important link between the DNA damage response and the cellular cytoskeletal components.
It has been controversial whether an abnormal G 2 /M checkpoint contributes directly to radiation sensitivity (for a review, see Ref. 4). Some reports suggest that the activation of the G 2 /M checkpoint in response to DNA damage enhances cell survival, because chemicals that abolish the G 2 /M checkpoint reduce clonogenic cell survival (47,48). Others have suggested that prolonged G 2 arrest contributes to the ability of the cells to survive radiation (49 -51), whereas Xu et al. (52) suggest that the G 2 /M checkpoint status has no direct influence on the cellular sensitivity to DNA damage. Previously, we reported an interaction between BRCA2 and filamin-A. We found that deficiency of filamin-A in M2 melanoma cells renders them more sensitive to ␥-radiation, bleomycin, and UV-C radiation (9). In the work reported here, we found that filamin-A deficiency does not affect the activation of the G 2 /M checkpoint, but rather it results in a delayed release from G 2 arrest. Our data suggest that delayed recovery from G 2 arrest may also be associated with sensitivity to DNA damage. It has been recognized that there are two main cell death pathways that regulate clonogenic cell survival after DNA damage. They are the interphase apoptotic cell death pathway and the mitotic cell death pathway (30 -32). It is the balance and individual effectiveness of those pathways that determine the overall clonogenic survival. In M2 cells, enhanced apoptosis was not observed based on DNA content analysis (Figs. 1 and 2) and TUNEL assay (data not shown). However, we did observe a significant increase in mitotic catastrophe after irradiation in M2 cells compared with A7 cells (Fig. 8). Together, these data suggest that deficiency of filamin-A in M2 cells might promote mitotic catastrophe that causes the enhanced cell death in M2 cells after irradiation.
Mitotic catastrophe is a mode of cell death exhibited by most nonhematopoietic cell lineages in response to ionizing radiation (33,34). Although the descriptions of mitotic catastrophe are diverse, its hallmark is nuclear fragmentation (32,34,36,(53)(54)(55). However, these fragmented nuclei generally lack DNA degradation, which is characteristic for apoptotic cells and can be detected by TUNEL assay. In this study, delayed release from G 2 arrest correlated with a reduced clonogenic survival and enhanced mitotic catastrophe in response to ionizing radiation.
Furthermore, it is intriguing to note that, starting 48 h after the irradiation, the percentage of mitotic catastrophe in M2 cells increases more dramatically than in the first 48 h (Fig.  8B). This time point coincides with the release of G 2 -arrested cells (24 -32 h for A7 cells and 32-48 h for M2 cells, as indicated in Fig. 3). These data suggest that release (or adaptation) from a sustained G 2 arrest may be associated with mitotic catastrophe, an idea supported by others (54,56,57). However, it is currently unclear whether the delayed recovery of G 2 arrest in M2 cells is the cause of increased cell death in M2 cells or is simply a correlation between them. Because filamin-A interacts with BRCA2, it is possible that the DNA repair process was much slower in filamin-A-deficient M2 cells than in filamin-A-proficient A7 cells, which is supported by our data 2 that suggest a less efficient RAD51 focus formation in response to radiation in M2 cells than in A7 cells. Therefore, it is plausible to suggest that the delayed recovery from G 2 arrest in M2 cells may be a result of the slowed DNA repair process, which results in a more sustained DNA damage signal to the Chk1/ Chk2 signal transduction pathway. This in turn causes a delayed recovery from G 2 arrest in M2 cells, a scenario that warrants further investigation.