Role of protein kinase CK2 in antitumor drug resistance

An increased drug efflux is a common event in the multidrug resistance (MDR), a phenotype displayed by cells which become insensitive to a broad range of cytotoxic agents. In these cases, the (over) expression of a pump mediating extrusion of chemotherapeutic drugs from the cell is observed. These pumps are proteins belonging to the ABC family. In humans, three of them are known to mediate MDR: the P-glycoprotein (P-gp, also known as MDR1/ABCB1), the MDR-associated protein (MRP1, ABCC1), and the breast cancer resistance protein (BCRP, ABCG2) [12].

Several evidences support a role of CK2 in the regulation of MDR pumps activity. In 2007, we demonstrated that inhibition of CK2 allowed an increased accumulation of doxorubicin in P-gp expressing cells [58]. Although a direct evidence of a cause/effect relationship of phosphorylation on P-gp activity is still lacking, P-gp is indeed known as a substrate of CK2 [40]; since our results were obtained in cell treated with CK2 inhibitors for a very short time (30 min), they strongly suggest a direct regulation due to phosphorylation. Additionally, more recently CK2 was found to increase drug-induced P-gp amount, through a mechanism involving the phosphorylation of heat shock protein 90β (HSP90β) and subsequent stabilization of Pregnane X receptor (PXR), which dissociates and translocates into the nucleus to interact with RXR (retinoid X receptor) and induce the transcription of ABCB1 gene [28].

A regulatory role of CK2 was later found also on MRP1: a model was proposed in which CK2 potentiates MRP1 function through direct phosphorylation of Thr249; in fact, CK2α knock-down, or MRP1 Thr249Ala mutation, decreased the efflux of doxorubicin and increased doxorubicin cytotoxicity in MRP1-expressing cells [33]. The expression level of ABCG2 was found dramatically decreased in CK2α-silenced lung cancer cells, due to down-regulation of the hedgehog signaling (see below for the CK2 impact on this signaling) [15].

Summarizing, CK2 not only phosphorylates P-gp, with possible consequences on its extrusion activity, but also induces its overexpression, thus amplifying the MDR phenotype; it phosphorylates and activates MRP1, the other major extruding pump mediating MDR, and controls the expression of the third ABC family extrusion pump (BCRP) (Fig. 1).

Our initial studies on the CK2 roles in MDR cells disclosed a possible specific function of the monomeric CK2 α isoform, since the MDR cells analyzed (CEM-R) expressed a higher level of CK2 α, but not CK2 β, compared to their parental line [58]. However, this does not seem to be a general feature of MDR cells, since a subsequent analysis revealed a variable scenery of CK2 expression in resistant compared to sensitive cells [59]. Regardless of the overexpression of CK2 (or of a specific CK2 isoform), the general observation is that MDR cells are sensitive to different kind of CK2 inhibitors, as demonstrated by our and other groups [22, 59, 60] (see also the paragraph on CK2 inhibitors), suggesting that MDR pumps are not active towards chemical compounds targeting CK2.

On this regard, it is worth to mention that CK2 inhibitors have been derivatized for targeting both CK2 and the breast cancer resistance protein ABCG2 [61]. The same group found that the structure-activity relationships for CK2 and ABCG2 are totally different, and they developed compounds blocking the extrusion pump without significantly inhibiting CK2 [62]; however, a dual CK2/ABCG2 inhibitor is particularly interesting, considering that co-administration of pump-inhibitors and cytotoxic agents is one of the strategies proposed to fight MDR [13].

DNA damage is a direct or indirect event in response to many antitumor agents, and a major mechanism that cells develop to evade their effects is an increased DNA repair activity. CK2 is a key player in the cellular response to DNA damage. Its role in phosphorylating the scaffold protein XRCC1, thus promoting DNA single-strand break repair, is known for several years [63]. CK2 phosphorylates also XRCC4 [56], a crucial protein for the nonhomologous end-joining (NHEJ), the major DNA double-strand break (DSB) repair pathway. The phosphorylation of XRCC4 at Thr233 by CK2 is required for its association to PNK, which is essential for optimal DSB repair (and indeed a not-phosphorylatable XRCC4 mutant displayed increased sensitivity to radiation-induced DNA-damage). Many other proteins implicated in DNA repair were later found to be regulated by CK2, such as the heterochromatin protein 1 (HP1) β [64], the DNA damage mediator protein MDC1 [65], the DNA-dependent protein kinase [66], the recombinase Rad51 [67], the deubiquitylase OTUB1 [68], the adaptor protein 53BP1 [69], the MLH1 component of the DNA mismatch repair complex MutLα [70], and the yeast proteins Lif1 (regulatory subunit of the NHEJ-specific DNA ligase IV) [71] and Ctp1 (DNA end-processing factor) [72]. Consistently, the specific CK2 inhibitor CX-4945 (also known as silmitasertib) has been shown to suppress the DNA repair response to anticancer drugs [73]. The global role of CK2 in DNA damage response and repair pathways has been recently reviewed [74]. To the purpose of this review, we focus on those publications where the CK2 action on DNA repair has been found specifically responsible of cancer drug resistance (Fig. 2). In particular, a body of evidence indicates a role of XRCC1-CK2 axis in the resistance to Cisplatin and derivatives. Pt-based compounds are DNA-damaging agents; they are not substrate of P-gp, MRP1 and ABCG2 (they can be exported by other transporter such as the copper efflux transporter, ATP7A and ATP7B [12]), and a main mechanism of resistance to these drugs is an increased DNA repair. Consistently, a protective role of CK2 has been frequently reported. For example, it has been shown that lung fibroblasts from idiopathic pulmonary fibrosis (IPF) patients is due to CK2 hyperactivation, which in turn promotes an abnormally high XRCC1 activity. In fact, CK2 blockade sensitizes IPF fibroblasts to Cisplatin [53]. Moreover, the phosphorylation of XRCC1 by CK2 is required for the action of its regulator JWA in gastric cancer cells resistant to DNA repair, following Cisplatin-induced DSBs [54]. Interestingly, conjugated compounds, with a CK2-inhibiting moiety linked to platin-derived drugs, reverse Cisplatin resistance in cancer cells by suppressing DSB repair by CK2 [30, 32]. One of the conjugated compound has been found effective in reversing drug resistance also in vivo, with a molecular mechanism involving the DSB repair MRE11-RAD50-NBS1(MRN) complex [32]. A similar approach of chimeric agent has been exploited to produce a Cx-platin drug, a CK2-targeting Pt (IV) prodrug, containing the CK2 inhibitor CX-4945 [31]. It is effective in suppressing CK2-mediated DNA damage repair and reversing Cisplatin resistance. The identified mechanism was the prevention of MDC1 phosphorylation by CK2 and its association to the FHA domain of aprataxin at the DSBs. In vivo studies showed a higher antitumor efficacy of Cx-platin compared to Cisplatin [31].

Another in vivo study showed a synergistic effect of Cisplatin used in combination with the CK2-targeting compound CIGB-300, in nude mice xenografted with human cervical tumor cells, and an increased mice survival compared to single-agent treatment was observed [75].

The rationale of targeting CK2 in combination with Cisplatin-based compounds is also supported by the observation of an increased CK2 expression in response to these drugs: this was reported by Yang et al. [76], who observed down-regulation of the tumor suppressor PML in cisplatin-treated lung cancer cells, as a consequence of CK2α overexpression.

The multitude of CK2 targets implies a network where the kinase simultaneously intervenes at different levels in mediating the response to DNA damaging agents. An example is provided by Kang and coworkers showed that XRCC4, PTEN and p53 (that are all CK2 substrates) merge on a concerted signal, which produce resistance to the DNA damage-inducing drug doxorubicin in glioblastoma cells. Sensitivity can be restored by restraining CK2, and the authors suggest that combinatorial therapies based on CK2 targeting might potentially counteract therapeutic resistance in glioblastoma patients with aberrancies in p53, PTEN and CK2 [41].

Topoisomerase inhibitors are a class of antitumor agents with final effects on DNA integrity and functions. CK2 is deeply implicated in the functionality of both topoisomerase I and II; however, since this is not a pure matter of CK2 control on DNA repair, and given the complexity of the issue, it will be dealt below, in a specific paragraph.

Chaperone proteins are often expressed at high levels in tumors and closely associated with a poor prognosis and resistance to therapy [77]. CK2 is considered a master regulator of chaperones [27, 78, 79], by which it exerts its protective function on onco-kinases and other survival proteins. In this sense, the CK2-dependent protection from apoptosis mediated by the chaperone machinery is obvious. More specifically focusing on drug resistance, it has been found that acquired MDR in response to rifampin treatment is correlated to the phosphorylation of HSP90β at Ser225 and Ser254 by CK2: phospho-HSP90β forms a more stable complex with the Pregnane X receptor (PXR), the transcription factor of the P-gp gene (MDR1), and this explains the final induction of ABCB1 expression due to CK2 [28].

Moreover, it has been observed that CK2 inhibition reduces the association between HSP90 and the co-chaperone CDC37 in cancer cells resistant to conventional therapies, with a consequent down-regulation of HSP90-client proteins (EGFR, PTEN, mTOR, Raptor and Tuberin/TSC2) [22].

Targeting chaperones belonging to the family of heat shock proteins (HSP) is a promising antitumor strategy, since many of their clients are involved in tumor development and progression [77]. In particular, several HSP90 inhibitors are in clinical trials [80], but different factors may influence cellular susceptibility to them, and resistance can occur due to redundant pathways or increased levels of other pro-survival chaperones in response to prolonged treatment. Therefore, combination therapies are considered the more promising approach to prevent compensatory mechanisms [80]. Consistently, the co-treatment of glioblastoma cells with a CK2 inhibitor (D11) was found effective in preventing the increase in HSP70 amount in response to the HSP90 inhibitor 17-AAG, and concomitantly also a reduction of the co-chaperone HSP27 was observed [26].

A different story has been found by a yeast based assay: it has been shown that the phosphorylation of Thr22 in yeast HSP90 by CK2 regulates the chaperone function, but increases the sensitivity to HSP90 inhibitors in vivo [29]. Results have been produced in a yeast strain devoid of the drug resistance pump PDR5, the major mediator of HSP90 inhibitor efflux. Although they need to be confirmed in mammals, they suggest caution against too cursory generalization of CK2 targeting as a sensitization strategy towards HSP90 inhibitors.

Also HSP27 inhibitors are considered valuable tools against drug resistance [81], and we have recently found that CK2 strictly controls the turnover of this chaperone in cervical and liver cancer cells [27].

The co-chaperone CDC37, which has a prominent role on the onco-kinome, is also controlled by CK2 [79]. The suppression of CDC37 phosphorylation and stabilization by CK2 was recently identified in the mechanism exploited by the microtubule-targeting pyrrolo-1,5-benzoxazepine compound for overcoming resistance to imatinib of gastrointestinal stromal tumor cells [21].

Figure 3a summarizes the major chaperone proteins controlled by CK2.

CK2 has been widely implicated in survival signaling (as reviewed in [7]), and it represents a target of general validity to downregulate different hyperactive pathways that can cause abnormal survival of cancer cells. Here below, we separately describe the pathways where the intervention of CK2 has been found crucial for apoptosis resistance.

TKIs have been widely employed as antitumor drugs, since the deregulated activity of tyrosine kinases is among the most frequent causes of tumors [94]. However, the occurrence of resistance is often observed, due to different mechanisms, frequently represented by gene mutations or amplifications, but also by compensatory signaling [95]. Among the pathways which contribute to TKI failure, CK2 has been reported. We have found that, in imatinib-resistant CML cell lines, without BCR-ABL1 mutations, CK2 differently potentiates the resistant phenotype, either by overexpressing its protein-level [17] or by a BCR-ABL-independent activation of rpS6 [16]. A proteomic study in non-small-cell lung cancer cells identified CK2 as a central element to mediate TKI resistance, and the phosphorylation of its substrate HMGA1 as a critical node to promote resistance to the EGFR inhibitor gefitinib [25]. CDC37 is another CK2 substrate critically important in TKI resistance: in imatinib-resistant gastrointestinal stromal tumors (GIST) cells with constitutively active c-KIT receptor tyrosine kinase, a mechanism has been described where CK2 has a role through the phosphorylation of CDC37, which in turn, in a positive loop, contributes in maintaining high levels of CK2 itself. In fact, treatments with PBOX-15 (pyrrolo-1,5-benzoxazepine, a microtubule-targeting compound), which reduce both CDC37 and CK2 levels, increase sensitivity to imatinib [21]. In CML bone marrow samples derived from patients resistant to imatinib, due to BCR-ABL T315I point mutation, a pro-apoptotic effect of CK2 inhibition was reported, mediated by PTEN reactivation [85].

In EGFR mutant lung cancer cells resistant to TKIs, CX-4945 was reported to sensitize cells and induce autophagy; however, cell recovery was observed upon CX-4945 withdrawal [96]. This is in accordance with our previous finding on CX-4945 short persistence [97], and suggests that this feature should be considered in planning protocols for in vivo treatments with CX-4945. A further note of caution comes from a study where a different CK2 inhibitor, Quinalizarin, was exploited to reduce viability, proliferation and migration of lung adenocarcinoma cells with different EGFR genotypes [98]. The authors found variability in the responses, according to the mutation harbored, and concluded that the effects were especially evident in cells harboring TKI sensitive EGFR mutations; actually, one of the TKI-resistant mutant cell lines was almost insensitive to Quinalizarin in an MTT viability assay. The reasons of this different sensitivity to Quinalizarin were not analyzed, neither other CK2 inhibitors were tested; however, this study suggests speculations on compensative mechanisms possibly occurring, in a background of TKI resistance, which imply resistance also to CK2 inhibition, and will deserve future investigation. Similarly, an only modest effect of CK2 inhibition was recently reported in KRAS-active non–small cell lung cancer cells resistant to EGFR inhibitors [99].

Interestingly, CK2 can regulate also the expression of EGFR itself, as shown by its down-regulation in response to CK2 inhibition [22].

Regardless of sensitive or resistant phenotype, several studies observed synergistic effects of CK2 inhibitors and TKI in tumor cells. Among them, Bliesath and colleagues, by combining the CK2 inhibitor CX-4945, and the EGFR tyrosine kinase inhibitor erlotinib, observed a synergistic antitumor effect in non-small cell lung carcinoma and squamous cell carcinoma, in vitro and in vivo, and demonstrated that it was mediated by an enhanced attenuation of the PI3K/AKT/mTOR pathway [100]. Similarly, an in silico study suggested that the PI3K/PTEN/AKT pathway could be synergistically reduced by simultaneous targeting of the receptor tyrosine kinase HER2 and CK2 [42]. Very recently, a study in KRAS-active non–small cell lung cancer cells resistant to EGFR inhibitors reported that single treatment with a CK2 inhibitor was not sufficient to completely impair cell viability, and an informatic analysis revealed MEK as a possible second target to overcome resistance; indeed, CX-4945 in combination with the MEK inhibitor AZD6244 displayed synergistic effects [99].

In summary, CK2 targeting is widely suggested as a potential therapeutic strategy for improving the response to TKI, although the issue might deserve a deeper investigation.