CK2 (alpha') was expressed in both cell lines with a marked peri-nuclear accumulation in the tumor cell line. This finding is in accordance with the previously reported higher ratio of nuclear to cytosolic activity of CK2 in cancer cells than in normal cells . Despite the observed difference in basal CK2 localization, the cytotoxic effect of CK2 inhibition was very similar with both cell types indicating that the subcellular distribution of CK2 was not the prevailing determinant of its pro-survival function. But this particular aspect was not further investigated.
A critical issue in the use of a chemical kinase inhibitor is its specificity for a given target enzyme. TBB was previously shown to exhibit a remarkable in vitro selectivity for CK2 among a panel of about 80 kinases, where only three other groups of kinases were inhibited by TBB with comparable efficacy . They are not known to be involved in DSB processing but two of the respective kinases (HIPK2 and DYRK2) regulate p53-dependent differential transactivation of growth arrest genes versus pro-apoptotic genes in response to DNA damage severity . The reported in vitro IC50 for TBB is 0.15 μM  which is a factor of about 100 lower than respective numbers obtained when cells were exposed to TBB before protein extraction and assessment of in vitro phosphorylation with synthetic CK2 target peptide [32, 33]. The 2 hours exposure of cells at 20 μM TBB prior to irradiation used in the present study is therefore considered as an effective treatment for the reduction of CK2 activity while still being only moderately toxic in the clonogenic assay. The efficacy of this TBB exposure with respect to the inhibition of CK2 was further confirmed by a functional assay (Figure 3) that measured intracellular phosphorylation of a protein (Xrcc1) for which CK2 is known to be the major kinase [27, 28].
The major result of the present investigation, however, is the discrepancy between DSB rejoining (PFGE) and the disappearence of γH2AX foci, where only the latter was delayed upon CK2 inhibition. A comparison of the results obtained with these two methods needs to consider the widely different doses applied. Analysis of DNA fragmentation requires sufficiently small fragments that may be resolved during electrophoresis and which are only produced at high doses. On the contrary, the focus assay is applicable at doses of only a few Gy. A comparison of absolute repair/rejoining rates derived with these different assay may thus be problematic. But particularly with the PFGE assay, it is not known that a repair modification, by whatever reason, would become differentially detectable depending on the dose level at which it is investigated.
Therefore, CK2-phosphorylated factors known to be involved in the response to radiation-induced DSB (Xrcc4, MDC1 or HP1-β; referred to in the Introduction section) need to be discussed.
Xrcc4 is constitutively phosphorylated at Thr233 by CK2 and this was shown to mediate the recruitment of DSB end-processing factors  which aid NHEJ . Utilizing a Thr233 mutant system, it was also shown that lack of the CK2-targeted site resulted in a delayed removal of γH2Ax foci, but an actual rejoining defect, which may well have existed, was not assessed by these authors . One has to keep in mind that the absence of constitutive CK2-dependent Xrcc4 phosphorylation represents a quite different situation compared to the addition of the CK2 inhibitor shortly (2 hours) prior to irradiation and assessment of DSB rejoining. Therefore, the phosphorylation status of Xrcc4 (at the CK2 residue) is sufficiently long-lived, or does not impact on NHEJ efficiency to an extent that could be resolved by our PFGE assay.
MDC1 is constitutively phosphorylated at multiple residues by CK2 . But unlike Xrcc4, MDC1 functions - via MRN-complex recruitment - in the propagation and retention of the chromatin changes that spread over large regions surrounding a DSB, particularly the ATM-dependent H2AX phosphorylation (see: Introduction section), rather than being involved in the DSB rejoining reaction [11, 15]. The present experiments could not detect an inhibitory effect of TBB on the number of initially formed γH2AX foci. This is in accordance with an earlier observation where TBB was unable to abrogate MRN recruitment (NBS1 foci) whereas downregulation of CK2 by siRNA was effective, leading to the conclusion that chemical CK2 inhibition was not potent enough to sufficiently reduce CK2 activity towards MDC1 .
Recently, ATM was shown to act as a repair factor for DSB in heterochromatic regions of the genome by phosphorylating heterochromatin protein KAP-1 to allow for localized and transient changes of chromatin organization that would otherwise inhibit repair [34–36]. Notably, knockdown of another heterochromatin protein, HP1-β, relieved the requirement for ATM in heterochromatic DSB repair [34, 35]. HP1-β is phosphorylated by CK2 in response to DNA damage leading to its mobilization from chromatin and allowance for H2AX phosphorylation . Accordingly, CK2 inhibition by 20 μM TBB resulted in a decreased fluorescence intensity of the individual γH2AX foci formed shortly after irradiation, which is at least qualitatively confirmed by our observation. Whether a TBB treatment affected DSB rejoining was not measured by these authors , but our results strongly argue against this possibility. Because foci did in fact form with initial numbers being independent from TBB preexposure, we conclude that the role of CK2 in the DNA damage response is to aid the propagation of chromatin changes, i.e. by the suggested mobilization of HP1, distal to the DSB site rather than being involved in initial damage recognition and rejoing. This appears to be different to the role of ATM in heterochromatin repair, where cells with defective or downregulated ATM fail to rejoin that particular portion (about 15%) of all induced DSB .
These considerations, however, do not answer the question why the disappearance of foci was delayed due to CK2 inhibition. γH2AX focus decay may proceed through in-situ dephosphorylation or by histone exchange (followed by dephosphorylation of displaced γH2AX). In mammalian cells, protein phosphatase PP2A seems to be crucially involved in γH2AX dephosphorylation . It was shown that PP2A directly interacts with CK2 catalytic subunit α kinase thus getting phosporylated and activated  implicating a role of CK2 in γH2AX turnover. The respective scheduling, however, is still controversial. The early decrease in the number of foci was not associated with a significant change in the global γH2AX level as measured by flow cytometry or western blotting  favouring the idea of a histone exchange mechanism being responsible for the dissolution of foci which would then not a priori require phosphatase activity. A similar conclusion was reached when the inhibition of PP2A by calyculin A, an agent known to suppress γH2AX dephosporylation , had only a small effect on foci elimination . In contrast, other authors did in fact observe a strong inhibition of foci decay by PP2A inhibitor calyculin A following radiation exposure (10, 42). Notably, this finding was not accompanied by a respective rejoining defect (using PFGE analysis) similar to what is found in the present investigation. A slower removal of γH2AX foci was also noted when a more specific inhibitor of PP2A (fostriecin) was used, or when employing PP2A catalytic subunit knockdown . In the latter study, foci formation was induced by stalling replication forks due to treatment with topoisomerase inhibitor camptothecin and, contrary to the radiation studies [10, 43], a concurrently expressed repair defect was measured by means of neutral comet assay. Whether this discrepancy could relate to the different mechanisms of damage induction or the distinct experimental approaches to assess residual breakage remains unclear.
The present data together with the above considerations argue for an involvement of CK2 in DNA damage response relaxation through its ability to stimulate in-situ γH2AX dephosphorylation, most likely via PP2A recruitment/activation. Because foci decay is thought to reflect a timely response to finalized damage repair, CK2 appears to exert a respective coordinating function which can be uncoupled upon its chemical inhibition whereby the end joining reaction is not affected (at least under the experimental conditions, used here). The apparent increase and persistance of the DNA damage checkpoint function at the G2/M border when CK2 was inhibited (as demonstrated for the WIDR cells in Figure 7) would be consistent with this idea.
The slight, though definitive, increase of clonogenic radiosensitivity of both cell lines when CK2 was inhibited could thus relate to an imbalanced DNA damage response. Different from an earlier study using HeLa tumor cells with CK2 being depleted by means of RNA interference , this effect was not due to triggering apoptosis which could potentially result from persistant damage signaling. Accordingly, the present observations reflect the modification of other major mechanisms through which clonogenicity upon irradiation is abrogated, i.e. a permanent growth arrest such as expressed in non-transformed fibroblasts or the prevailing mitotic catastrophe when cells bearing residual DNA damage enter mitosis . While a persistant damage signaling could in fact enhance permanent growth arrest (such as in the fibroblasts), it is not immediately clear how a prolongation of the radiation-induced G2/M transition delay (at undisturbed repair efficacy) would potentiate mitotic failures. CK2 has been implicated in the efficacy of either checkpoint by interactions with other key regulatory elements including the tumor suppressor p53 [45, 46] or the phosphatase cdc25B and C isoforms [47, 48]. But given the highly promiscious nature of CK2, other coordinating factors of cell cycle progression after DNA damage (i.e. SMC3 for intra-S phase checkpoint ) may have been affected by the TBB treatment, as well. The phenotypes of increased radiation sensitivities of the two cell systems may thus reflect distinct and differentially CK2-dependent factors of the DNA damage response. This could tentatively explain the intriguing qualitative difference in the dose-dependence of TBB-induced radiosensitization where the fibroblasts but not the tumor cells exhibited a significant threshold-type behaviour. Whether this difference is maintained when high radiation doses are delivered by small consecutive doses in a fractionation schedule is presently under investigation. In the absence of more pronounced differentially expressed phenotypes in tumor cells versus normal cells, however, a potential therapeutic benefit of a TBB-radiation combination can presently not be implied.