Casein Kinase 2 Regulates Both Apoptosis and the Cell Cycle Following DNA Damage Induced by 6-Thioguanine

CK2 (formerly, casein kinase II) is a protein Ser/Thr kinase complex that forms a heterotetramer mainly consisting of two catalytic (α and/or α′) subunits and two noncatalytic (β) subunits (for recent reviews, see refs. 1, 2). The highly conserved amino acid sequences of CK2 from yeast to humans suggests the importance of CK2 in cellular functions, although its major function(s) is not clearly understood. CK2 is considered to be essential for basic cell viability because a double knock-out of the two catalytic subunits is lethal in yeast (3). In male mice, a knock-out of CK2 α′ is reported to be sterile and related to a defect in spermatogenesis (4). A knock-out of CK2 β shows embryonic lethality in mice (5). CK2 can interact with Chk1 to increase Cdc25C phosphorylation activity in vitro (6). CK2 also plays an antiapoptotic role by protection of Bid from cleavage by caspase 8 (7).

The purine antimetabolite, 6-thioguanine (6-TG), is an effective chemotherapy drug in the management of acute leukemias (8, 9). However, in spite of 50 years of research, it is still difficult to precisely identify the major mechanism of cytotoxicity. A considerable amount of evidence points to the incorporation of 6-TG into DNA as an initial step in determining cytotoxicity (10). In various mammalian cell lines, 6-TG treatment results in a dose-dependent increase in DNA single-strand breaks, which occur during the second or later cell cycles following initial 6-TG incorporation into DNA (11, 12). Additionally, DNA protein cross-links, DNA double-strand breaks, and a decrease in mRNA synthesis have been reported as other intermediate damage end-points associated with 6-TG cytotoxicity (13, 14).

It is also known that DNA mismatch repair (MMR) plays a role in 6-TG cytotoxicity and that both 6-methyl TG-cytosine and 6-methyl TG-thymidine mismatches are recognized (15). MMR-proficient cells are highly sensitive to 6-TG, showing a prolonged G₂-M arrest followed by delayed cell death (16). We recently reported that a 6-TG-induced G₂-M arrest involves activation of the ataxia telangiectasia–related-Chk1 pathway and that Chk2 can also be activated and linked to a subsequent tetraploid G₁ arrest which blocks cells that escape from the G₂-M arrest (17, 18). Thus, these two signaling kinases seem to work cooperatively in MMR-proficient cells to ensure that 6-TG-damaged cells arrest at these cell cycle checkpoints. However, MMR-deficient cells, which do not show a clear G₂-M arrest with 6-TG treatment, are also sensitive to higher doses of 6-TG (19). Additionally, although MSH2 knock-out mice were relatively resistant to 6-mercaptopurine treatment compared with wild-type mice, lethality was seen at higher 6-mercaptopurine doses (20), suggesting that these purine antimetabolites (6-TG and 6-mercaptopurine) have dose-related cytotoxicity independent of MMR.

CK2 phosphorylates many substrates including p53 (21), Cdc2 (22) and BRCA1 (23), which can regulate the cell cycle. We recently reported that both BRCA1 and a BRCA1-like protein, TopBP1, could regulate the G₂-M checkpoint, partially compensating each function (24). Furthermore, cell cycle progression can be altered by inhibition of CK2 activity (25–27). Therefore, in this study, we examine the role of CK2 in modifying apoptosis and cell cycle regulation following treatment with 6-TG. We find that a reduction of CK2 protein levels by small interfering RNA (siRNA) results in a marked induction of apoptosis and that CK2 also partially regulates cell cycle progression following 6-TG-induced damage.

Cell Culture, Reagents, and Transfection of Small Interfering RNA. HeLa human cervical carcinoma cells, which are known to be MMR-proficient (28), were grown in RPMI 1640 medium supplemented with 10% FCS in a humidified atmosphere of 5% CO₂ and 95% air at 37°C. HCT116 human colorectal cancer cells are known to be MMR-deficient because of a hemizygous nonsense mutation in the hMLH1 gene on chromosome 3 (29). HCT116 cells were stably transfected with hMLH1 cDNA or an empty vector to generate an isogenic pair of MMR-proficient (MMR⁺) and MMR-deficient (MMR⁻) cells, respectively (30). These HCT116 cells were generously provided by Dr. F. Praz (Centre National de la Recherche, Villejuif, France), and cultured in DMEM supplemented with 10% FCS in a humidified atmosphere of 5% CO₂ and 95% air at 37°C. The CK2 chemical inhibitors, apigenin and emodin, and 6-TG were obtained from Sigma (St. Louis, MO) and dissolved in DMSO. The CK2 chemical inhibitor, 5,6-dichloro-1-β d-ribofuranosyl benzimidazole was obtained from Calbiochem (San Diego, CA). The caspase inhibitor, z-VAD, was obtained from Biomol (Plymouth Meeting, PA).

The anti-CK2 siRNAs used were 5′-GAUGACUACCAGCUGGUUCdTdT (α), 5′-UCAAGAUGACUACCAGCUGdTdT (α10, another siRNA against the α mRNA), and 5′-CAGUCUGAGGAGCCGCGAGdTdT (α′). The control siRNA was 5′-GCUCAGAUCAAUACGGAGAdTdT. The anti-MSH2 siRNA used was 5′-UCUGCAGAGUGUUGUGCUUdTdT. All siRNAs were obtained from Dharmacon (Lafayette, CO). All were annealed with the complementary strand with dTdT overhangs. Unless stated otherwise, α (but not α10) was used for reduction of the α subunit. Transfections were done with OligofectAMINE, as recommended by the manufacturer (Invitrogen Life Technologies, Carlsbad, CA). Transfection was done for 24 to 30 hours, and HeLa cells were then washed and trypsinized. Next, the cells from one well per 24-well plate were divided into two to three wells and RPMI 1640 medium (and 6-TG when indicated) was added to the culture. The efficiencies of transfection were ∼70% to 90% according to the maximum protein level reduction.

Transfection with plasmid DNA was done using Fugene 6, as recommended by the manufacturer (Roche, Indianapolis, IN). This transfection reagent mixed with DNA was incubated with cells during 6-TG treatment. The efficiency of transfection was ∼95% according to green fluorescent protein transfection when a tandem or sequential transfection procedure was used. For a tandem transfection, the genes were initially transfected on day 0 and then cells were treated with 6-TG (0 or 3 μmol/L) for 48 hours. A second transfection was done on day 3 and cells were again treated with 6-TG (0 or 3 μmol/L) for another 24 hours.

Western Blotting. Total cell extracts were prepared by trichloroacetic acid precipitation to detect CK2 α and α′ using specific antibodies (N-18 and C-20, respectively, Santa Cruz Biotechnology, Santa Cruz, CA). An anti-Myc antibody was 9E10 (Babco, Berkeley, CA), anti-β-actin antibody was from Sigma. After extraction of trichloroacetic acid with ether, the DNA was sheared by sonication before loading onto gels. MLH1 (554072, BD Biosciences, Palo Alto, CA) and MSH2 antibodies (Ab-2, Oncogene Research, San Diego, CA) were also used as controls for the siRNA transfections. The γ-H2AX antibody was obtained from Upstate (Chicago, IL).

Flow Cytometry. Cells were fixed with 90% ethanol at −20°C from 60 minutes to a few days, incubated with RNase, stained with propidium iodide, and then subjected to flow cytometry (Coulter, Epics XL-MCL, Miami, FL). For terminal dUTP nick end labeling (TUNEL) staining, the Apo-Direct kit (eBioscience, San Diego, CA) was used. Typically, at least 10,000 events were counted.

Immunostaining for Confocal Microscopy. HeLa cells were fixed with methanol and acetone (1:1), and then stained with the specific anti-α or α′ antibodies (N-18 and C-20, respectively; Santa Cruz). Samples were washed and stained with secondary antibodies (Alexa Fluor 633 anti-goat IgG, A21082; Molecular Probes, Eugene, OR) and 4′,6-diamidino-2-phenylindole, and then subjected to confocal microscopy (Zeiss Model LSM510; Zeiss, Oberkochen, Germany). Mitotracker (M7514; Molecular Probes) and transfection of endoplasmic reticulum-targeted green fluorescent protein (31) were used for staining of each organelle. The confocal images were obtained with an inverted confocal microscope system (Zeiss) equipped with a tunable T-sapphire laser (Mira-F-V5-XW-220) with a diode pump (Verdi 5 W) to obtain the images with the different dyes. Image analysis was carried out with Adobe Photoshop Software (Adobe Systems, San Jose, CA).

Site-Specific Mutagenesis. Genes encoding apoptosis repressor with caspase recruitment domain (ARC) and Cdc2 were also amplified as described above, and cloned into a pcDNA3-based vector (Invitrogen Life Technologies). The mutations were introduced using Quikchange (Stratagene, La Jolla, CA). All amplified and mutated sequences were confirmed by DNA sequencing in all coding regions.

Casein Kinase 2 is Essential for Inhibition of Apoptosis Following 6-Thioguanine Damage. We employed a RNA interference technique (32) for reduction of the protein levels of the CK2 catalytic subunits, mainly using transfection of two siRNAs against both catalytic subunits α and α′, because the two subunits are reported to have overlapping functions (1). Transfection of the two siRNAs markedly reduced the protein levels of the CK2 α and α′ subunits, whereas a control siRNA transfection did not (Fig. 1A). Additionally, transfection of a single siRNA directed against either α or α′ reduced the protein levels of each corresponding subunit (Fig. 1A). The anti-α′ siRNA also slightly reduced the protein level of the α subunit. The time course experiments indicate that inhibition of CK2 continued at least until day 3 (Fig. 1B). No changes in MSH2, MLH1, and β-actin protein levels were found following siRNA transfection against α and/or α′ in HeLa cells.

After transfection for 1 day, the different transfectants were then treated with 6-TG (15). After continuous exposure to 6-TG (0 or 3 μmol/L) for up to 4 days, the control transfectant showed a predominant G₂-M population, whereas the transfectants with siRNAs against both α and α′ showed a predominant sub-G₁ population and a markedly reduced G₂-M population (Fig. 2A). We also examined the response to the same 6-TG treatment schedule in HeLa cells following a single siRNA transfection against either the CK2 α or α′ subunits. For CK2 α, two different siRNAs (α and α10) against α mRNA were used. Western blotting showed that anti-α10 siRNA specifically inhibited the protein level of the α subunit (data not shown). After 4 days of treatment with 6-TG, we found a similar induction of a sub-G₁ population following these single siRNA transfections compared with the control (Fig. 2A).

Figure 2B shows that 6-TG treatment (0 or 3 μmol/L) for 3 days in CK2-reduced HeLa cells resulted in a large TUNEL-positive apoptotic population (67%), demonstrating that CK2 is important for inhibition of apoptosis after 6-TG-induced DNA damage. The TUNEL-positive population was significantly lower (29%) following 6-TG treatment in the control cells. Even in the absence of 6-TG damage, a reduction of protein levels of CK2 generated a moderate apoptotic population (21%). It should be noted that the TUNEL-positive populations may not directly correspond to the sub-G₁ populations in Fig. 2A, because TUNEL-positive cells are also distributed throughout the cell cycle.

We also used the CK2 chemical inhibitors, apigenin, emodin (33) and 5,6-dichloro-1-β d-ribofuranosyl benzimidazole (34) to assess the response of CK2 to 6-TG damage in nontransfected HeLa cells (Fig. 2C). However, because we recently reported that the DNA damage response at the G₂-M checkpoint requires a long 6-TG treatment period (∼4 days) in HeLa cells (17), the CK2 inhibitors alone had significant toxicity, requiring the use of low concentrations for these experiments. Although the data are less clear than those using CK2 siRNAs (Fig. 2A), the CK2 chemical inhibitors also produced an increased sub-G₁ population on day 4 with concomitant 6-TG treatment (Fig. 2C). These results also suggest that CK2 activity is required for apoptosis inhibition following 6-TG induced damage.

Casein Kinase 2 α Relocalizes to the Endoplasmic Reticulum After 6-Thioguanine Treatment. To examine the localization of the CK2 α or α′ catalytic subunits after 6-TG treatment, we stained nontransfected HeLa cells with the specific anti-CK2 antibodies followed by confocal microscopy. CK2 α was found to localize in both the nucleus and the cytoplasm without 6-TG treatment (Fig. 3A). However, following 6-TG treatment for 3 days, a large subset of CK2 α became localized in extranuclear sites. On the other hand, no significant relocalization was found in CK2 α′, although α′ in the cytoplasm was slightly reduced following 6-TG treatment.

To further investigate the localization change in CK2 α following 6-TG treatment, Mitotracker and endoplasmic reticulum-targeted green fluorescent protein were used. A majority (but not all) of CK2 α was localized to the endoplasmic reticulum with significantly less localization in the mitochondria (Fig. 3B). Because the endoplasmic reticulum is reported to participate in the regulation of apoptosis (31), the significance of this localization change is undergoing further investigation.

Apoptosis Can Be Induced by 6-Thioguanine in Both Mismatch Repair-Deficient Cells and -Proficient Cells. We next examined the possible dependency of apoptosis induction by 6-TG on the MMR system, using MMR-deficient and -proficient human colon cancer cells derived from the HCT116 cell line (30). Unexpectedly, the sub-G₁ fraction produced by 6-TG treatment was similar in both vector and MLH1 cDNA-transfected HCT116 cells, although MLH1 transfection into MMR⁻ HCT116 cells did significantly change the cell cycle progression in G₂-M following 6-TG treatment (Fig. 4A), consistent with our previous data (16–18). The use of hMLH1 siRNA did not result in a significant reduction of MLH1 protein expression in HeLa cells (data not shown). However, although MSH2 siRNA efficiently inhibited the MSH2 protein level in HeLa cells (Fig. 4B), it did not significantly change the extent of apoptosis induction by CK2 siRNA in the presence or absence of 6-TG treatment (Fig. 4B). Similar results of 6-TG-induced apoptosis were also obtained by using three different isogenic pairs of MMR-deficient and -proficient cells, including vector and hMLH1-transfected RKO human tumor cells (16), MSH2 knock-out and MSH2 transfected mouse cells (35), and MSH2-deficient and the MSH2-containing chromosome 2 transferred Hec 59 human tumor cells (ref. 36; and data not shown). However, we did find an enhanced G₂-M arrest in the MMR-proficient cells, similar to our prior data (19). These results may be related to the observation that both MMR-deficient and -proficient cells are sensitive to relatively high concentrations of 6-TG (19), and that MSH2-deficient mice are still sensitive to mercaptopurine, another purine antimetabolite used as a chemotherapeutic drug (20).

We also checked γ-H2AX levels, a marker of DNA double-strand breaks (37), following 6-TG treatment in the HCT116 vector-only and hMLH1-transfected cells. A moderate induction of γ-H2AX was observed at days 3 and 4 in both MMR-deficient and -proficient cells (Fig. 4C). Ionizing radiation (10 Gy) resulted in a more robust increase (5×) in γ-H2AX expression after 90 minutes to similar levels in both MMR-deficient and -proficient cells. These results suggest that DNA damage induced by 6-TG results in a delayed production of DNA double-strand breaks independent of MMR processing of 6-TG mismatches, consistent with our prior data (16).

A Caspase Inhibitor and Apoptosis Repressor with Caspase Recruitment Domain Inhibit Apoptosis Following 6-Thioguanine Damage. The enhanced apoptotic response in CK2-reduced cells to 6-TG treatment (Fig. 2) suggests that CK2 may regulate apoptosis in the presence of specific chemically induced damage. We next examined the effect of a general caspase inhibitor, z-VAD on this 6-TG effect. When z-VAD was added to CK2-reduced cells, apoptosis was strongly inhibited after 6-TG treatment, suggesting that caspases are involved in the induction of apoptosis (Fig. 5A).

Additionally, the G₂-M population decreased in CK2-reduced cells after 6-TG and z-VAD treatment (Fig. 5A, compare z-VAD+ and 6-TG+), suggesting that CK2 can affect both apoptosis and the cell cycle. Recently, it was reported that ARC can be phosphorylated by CK2 at Thr¹⁴⁹, regulating the apoptosis inhibitory function of ARC (38). Therefore, we expressed ARC and its mutants, 149A (Thr to Ala, an unphosphorylatable form) and 149E (Thr to Glu, a phosphorylation-mimicking form) in HeLa cells (Fig. 5B). After treatment with both anti-CK2 α and α′ siRNAs, the wild-type ARC strongly inhibited apoptosis when the gene was transfected twice (tandem transfection; see protocol in Fig. 5C) in both 6-TG-treated and -nontreated cells (vector and wild-type ARC; Fig. 5D and E), indicating that ARC can also inhibit apoptosis in CK2-reduced cells. It should be noted that the tandem transfection procedures of the vector plasmid did increase the sub-G₁ populations (vector, tandem, and single transfection; Fig. 5D and E). However, the mutant ARC transfectants did not show significant differences in the extent of apoptosis compared with the wild-type ARC transfectants. These results suggest that Thr¹⁴⁹ is not critical for the inhibition of apoptosis following 6-TG DNA damage, which differs from the apoptosis triggers used in the previous report (38).

Phosphorylation of Cdc2 Ser³⁹ Can Regulate the Cell Cycle after 6-Thioguanine Treatment Cdc2 can be phosphorylated at Ser³⁹ by CK2 in vivo (22). Because Cdc2 is an essential protein involved in cell cycle regulation, we examined whether the cell cycle response to 6-TG-induced DNA damage was changed by a Cdc2 39A mutation (Ser³⁹ was mutated to Ala). When HeLa cells were transfected with the 39A mutant, the G₂-M peak induction was partially inhibited after 6-TG treatment (3 μmol/L × 4 days; 62% versus 27%; Fig. 6A). These cell cycle data are also consistent with the results of the effect of z-VAD in CK2-reduced cells (Fig. 5A). It should be noted that Ser³⁹ is close in the three-dimensional structure of Cdc2 to Tyr¹⁵ (39), which is critical for Cdc2 regulation. These data suggest that CK2 may participate in a proposed CK2-Cdc2 cell cycle regulatory pathway. CK2 was not involved in a G₂-M checkpoint control (data not shown) when we examined this checkpoint using nocodazole as we recently described (17).

In this study, we show that apoptosis is strongly enhanced following 6-TG treatment by a reduction of the CK2 catalytic subunits through siRNA transfection (Fig. 2). Transfection of three different siRNAs (α, α10, and α′) against CK2 catalytic subunits resulted in a similar induction of apoptosis following 6-TG treatment. The use of three chemical inhibitors of CK2, apigenin, emodin, and 5,6-dichloro-1-β d-ribofuranosyl benzimidazole (33, 34) also enhanced apoptosis after 6-TG damage in nontransfected cells (Fig. 2). A pan-caspase inhibitor, z-VAD, or cotransfection with ARC, blocked this apoptotic response to 6-TG treatment in CK2-reduced cells, suggesting that CK2 regulates caspase activity (Fig. 5). These results also suggest that a 6-TG-induced DNA damage response involves a CK2-apoptosis inhibitory pathways (Fig. 6B).

The transfectants with wild-type ARC, or its mutants 149A and 149E, show similar levels of apoptosis after 6-TG damage in CK2-reduced cells (Fig. 5D and E). These data suggest that this unique phosphorylation site by CK2 in ARC (38) is not critical for a DNA damage response induced by 6-TG treatment. Therefore, we speculate that CK2 may have other target(s) affecting apoptosis. Interestingly, we found that a subset of CK2 α became localized to the endoplasmic reticulum following 6-TG treatment (Fig. 3). On the other hand, no significant relocalization was found in CK2 α′. Currently, we are further investigating the significance of these localization changes of CK2 in this response to 6-TG-induced damage. It is interesting to note that CK2 is required for adaptation of a checkpoint arrest in yeast (40).

We also find that the G₂-M population is reduced in the presence of z-VAD and 6-TG in CK2-reduced cells (Fig. 5A). These data are supported by a prior study suggesting that CK2 is required for cell cycle progression during G₁ and G₂ phases (3). Furthermore, we show that a 39A mutation of Cdc2 partially inhibits cell cycle progression in G₁ and G₂ following 6-TG treatment (Fig. 6A). Because this site is the phosphorylation site of CK2 in vivo, our results suggest that CK2 partially regulates cell cycle progression during G₁ and G₂. Thus, it seems that a 6-TG-induced DNA damage response may also involve a CK2-Cdc2 pathway (Fig. 6B). We have tried several times, without success, to generate a phospho-specific antibody against the phosphorylated Cdc2 Ser³⁹. It is possible that the epitope around the site containing similar residues (Ser³⁹-Glu-Glu-Glu) might inhibit the specific antibody production.

Apoptosis was similarly induced by 6-TG in isogenic MMR-deficient and -proficient cells derived from HCT116, suggesting that 6-TG-induced apoptosis is independent of the MMR system (Fig. 4). Consistently, MSH2 siRNA transfection in HeLa cells did not significantly change apoptosis in the presence of 6-TG when CK2 protein levels were reduced (Fig. 4). Our results may be explained by published data that both MMR-deficient and -proficient cells are sensitive to relatively high concentrations of 6-TG (19), and that MSH2-deficient mice are still sensitive to mercaptopurine, another purine antimetabolite (20). However, we do not know the kind of DNA damage that is critical for apoptosis induced by 6-TG in CK2-reduced cells. Because 6-TG treatment also caused a modest induction of γ-H2AX (Fig. 4C), it may be possible that CK2 participates in the different types of 6-TG damage, including DNA double-strand breaks, single-strand breaks, and DNA-protein cross-linking (13, 14). In contrast, our recent study indicated that persistent DNA single-strand breaks produced during MMR processing of 6-TG DNA mismatches are likely the initial chemical signals to an initial MMR-mediated G₂-M arrest and later cell death (16).

In conclusion, we have identified two possible pathways involving CK2 in a 6-TG-induced response, a CK2-apoptosis inhibitory pathway and a CK2-Cdc2 pathway. Further elucidation of these proposed pathways may provide insight in our understanding of drug resistance for 6-TG.

We thank Drs. Nancy Oleinick for a critical reading of this manuscript and Song-mao Chiu for discussion; Michael Sramkoski of CWRU for flow cytometry techniques, Drs. Minn Lam and Anna-Liisa Nieminen for confocal microscopy techniques, Clark Distelhorst and Michael Malone for the endoplasmic reticulum targeted-GFP plasmid, Paul Nurse for the cdc2 gene, and Françoise Praz for HCT116-derived cells.