CX-4945 Induces Methuosis in Cholangiocarcinoma Cell Lines by a CK2-Independent Mechanism

Cholangiocarcinoma is a disease with a poor prognosis and increasing incidence and hence there is a pressing unmet clinical need for new adjuvant treatments. Protein kinase CK2 (previously casein kinase II) is a ubiquitously expressed protein kinase that is up-regulated in multiple cancer cell types. The inhibition of CK2 activity using CX-4945 (Silmitasertib) has been proposed as a novel treatment in multiple disease settings including cholangiocarcinoma. Here, we show that CX-4945 inhibited the proliferation of cholangiocarcinoma cell lines in vitro. Moreover, CX-4945 treatment induced the formation of cytosolic vacuoles in cholangiocarcinoma cell lines and other cancer cell lines. The vacuoles contained extracellular fluid and had neutral pH, features characteristic of methuosis. In contrast, simultaneous knockdown of both the α and α′ catalytic subunits of protein kinase CK2 using small interfering RNA (siRNA) had little or no effect on the proliferation of cholangiocarcinoma cell lines and failed to induce the vacuole formation. Surprisingly, low doses of CX-4945 increased the invasive properties of cholangiocarcinoma cells due to an upregulation of matrix metallopeptidase 7 (MMP-7), while the knockdown of CK2 inhibited cell invasion. Our data suggest that CX-4945 inhibits cell proliferation and induces cell death via CK2-independent pathways. Moreover, the increase in cell invasion brought about by CX-4945 treatment suggests that this drug might increase tumor invasion in clinical settings.


Introduction
Cholangiocarcinoma (CCA) is a cancer of bile duct epithelia thought to arise as a result of chronic inflammation of the liver and bile ducts. The incidence of CCA has increased gradually worldwide the cytoplasmic vacuoles in different non-canonical cell death mechanisms is different yet not unique. One relatively well known non-apoptotic cell death is autophagy. Autophagy forms autophagosomes that are double-membrane vacuoles [32][33][34]. In contrast, methuosis, a recently described cell death mechanism, utilizes single-membrane vacuoles from macropinocytosis [35][36][37]. A few other types of non-apoptotic cell death mechanisms share a similar origin for the formation of cytoplasmic vacuoles. Oncosis, paraptosis, and necroptosis form vacuoles from endoplasmic reticulum (ER) and mitochondria [35,[38][39][40][41][42].
Methuosis is derived from the Greek methuo (to drink to intoxication) because the hallmark of this form of cell death is the formation of large fluid-filled vacuoles derived from macropinosomes. The demise of the cell resembles many forms of necrosis, with loss of metabolic capacity and plasma membrane integrity. Moreover, cell shrinkage and nuclear fragmentation associated with apoptosis are not observed [35]. Methuosis was initially defined in glioblastoma cells after ectopic expression of activated Ras [36,37], but recent reports have described small molecules, including indole-based chalcone 3-(2-methyl-1H indol-3-yl)-1-(4-pyridinyl)-2-propen-1-one (MIPP) and MOMIPP, which has a 5-methoxy group added to the indole ring of MIPP, that can induce the features of methuosis [37,43].
Herein, we report for the first time that CX-4945 at low doses stimulated CCA cell invasion due to an upregulation of matrix metallopeptidase 7 (MMP-7) and matrix metallopeptidase 9 (MMP-9). Also, CX-4945 induced cytosolic vacuolization, which met the criteria of methuosis, via a CK2-independent mechanism in a dose-and time-dependent manner.

CX-4945 Inhibits CK2 Activity at High Dose
To investigate the effects of CX-4945 on CCA cells, we treated three CCA cell lines with increasing amounts of this drug. CX-4945 reduced CK2 activity in a dose-dependent manner 4 h post-treatment as determined using Western blotting with an antibody that recognizes phosphorylated CK2 sites in multiple proteins (Figure 1a). The CCA cell lines all showed a reduction in multiple phospho-CK2 substrates with apparent molecular masses of approximately 200, 180, 100, 60, 37, 35, and 33 kDa as indicated by arrows in the Figure (Figure 1a).

CX-4945 Treatment Inhibits CCA Cell Proliferation
To determine the effects of CX-4945 on CCA cell proliferation we treated the three cell lines described above and examined the effect on cell number and 5-bromo-2 -deoxyuridine (BrdU) incorporation. After    . VC; vehicle control, *; p < 0.05, **; p < 0.01, ***; p < 0.001. All experiments were performed in triplicate and with at least at three biological replicates. Graphs were plotted as mean ± SEM.

CX-4945 Treatment Alters Cell Invasion
Protein kinase CK2 is known to be important in cell migration and cancer cell invasion. To determine the effects of CX-4945 on CCA cell invasion we examined the ability of the cells to traverse a layer of Matrigel in vitro. CX-4945 treatment showed biphasic effects on CCA cell invasion though Matrigel. CX-4945 at 10 µM significantly inhibited cell invasion through Matrigel in the three CCA cell lines tested (Figure 1d). In contrast, lower concentrations of CX-4945 stimulated invasion in all CCA cell lines tested (Figure 1d). The increase in cell invasion at low CX-4945 doses was not due to an increase in cell number as the assays were performed at the same time point (24 h post-treatment) that was shown by BrdU assay to have equivalent proliferation rates between the control and CX-4945 treated groups (1 and 5 µM) (Figure 1c). In addition, MTT assay at a later time point (48 h post-treatment) also showed no difference in cell number between these groups (Figure 1b). The increase in cell invasion was at least in part due to an increase in MMP-9, MMP-7, and matrix metallopeptidase 2 (MMP-2) levels in CCLP-1, and an increase in MMP-7 levels in HuCCA-1 and KKU-M213 (Figure 1e,f). The decrease in cell invasion at 10 µM of CX-4945 was at least in part due to a decrease in MMP-9 and MMP-7 levels in HuCCA-1 and to MMP-7 levels in KKU-M213. In addition to a decrease in MMP levels, a smaller invasion in the 10 µM CX-4945-treated group was also likely to be a consequence of the inhibition of cell proliferation at this dose (Figure 1b,c). We conclude that at lower doses, CX-4945 treatment increased the ability of CCA cells to invade Matrigel, while higher doses inhibited this ability.

CX-4945 Induces Caspase-3 Independent Non-Autophagic Cell Death in CCA Cells
To determine whether vacuolization was accompanied by increased cell death, we treated CCA cells with CX-4945 and stained live cells with calcein and dead cells with propidium iodide.  Cell area of cell lines that formed vacuoles did not altered, but PC3 cell area decreased to approximately 50% as compared to the vehicle control group. VC; vehicle control, ***; p < 0.001, NS; not significant. Pictures are representative of the results from three independent experiments. Graphs were plotted as mean ± SEM.

CX-4945 Induces Caspase-3 Independent Non-Autophagic Cell Death in CCA Cells
To determine whether vacuolization was accompanied by increased cell death, we treated CCA cells with CX-4945 and stained live cells with calcein and dead cells with propidium iodide. We conclude that CX-4945 is likely to induce non-canonical cell death in these cell lines. However, the LC3B-II protein might be involved in the initiation of vacuole formation.

CX-4945 Induced Methuosis in KKU-M213
All cell lines used in these studies, except PC3 cells which did not form vacuoles (Figure 3a), showed no change in cell area ( Figure 3b) and had no membrane blebbing (Figure 3a). Although their cytoplasm was filled with vacuoles, the cell membrane was well-spread and cell shape was normal as compared to the vehicle control group (Figure 3a). To further identify the cell death mechanism induced by CX-4945 in CCA cell lines, the nucleus and ER of cells treated with CX-4945 were examined. First, nuclear staining with DAPI showed that the nucleus of cells treated with 25 µM CX-4945 for 24 h were still intact. No chromatin condensation or nuclear fragmentation was observed ( Figure 6a). Second, KKU-M213, which are cells that are most sensitive to CX-4945 in terms of vacuole formation, were incubated for 72 h with CellLight ® ER-GFP BacMam 2.0 (C10590, Fisher Scientific, Hampton, NH, USA). This reagent transfected and expressed GFP-fused ER signal sequence, calreticulin, and KDEL, in live cells, thereby labeled ER with GFP. The ER of KKU-M213 treated with CX-4945 covered a large part of cytoplasm but did not overlap with the vacuoles (Figure 6b). Some large vacuoles reside near the nucleus and, therefore, share a similar space as the ER in the cytoplasm, however, many other medium and small vacuoles as well as some other large vacuoles were observed in the area not overlapping with the ER (Figure 6b). In addition, dextran blue/yellow was added in the culture media during CX-4945 treatment to track the endocytosis process. KKU-M213 engulfed dextran blue/yellow into the vacuoles, which showed as blue dots of various sizes in the cytoplasm (Figure 6c). Dextran blue/yellow remained blue indicating that pH of the vacuoles was neutral since it changes to green or yellow when the environment is acidic. We conclude that the vacuoles formed from CX-4945 treatment did not alter nuclear and membrane integrity. Furthermore, vacuoles in KKU-M213 are not formed from the ER, instead they are formed from the endocytosis process as their pH is neutral. These finding meet the criteria for methuosis.

CX-4945 Induced Vacuolization Is CK2-Independent
To determine whether the effects of CX-4945 on vacuole formation were mediated by the inhibition of CK2, we next knocked down the expression of the α and α′ CK2 catalytic subunits using specific siRNAs (Figure 7a). Knockdown of both CK2 α and α′ in the HuCCA-1, CCLP-1, and KKU-M213 CCA cell lines was verified using Western blotting (Figure 6a). In all three cell lines, knockdown of CK2 α and CK2 α′ decreased the levels of multiple phospho-CK2 substrates (Figure 7b). A reduction in phospho-CK2 substrate levels was observed for proteins with apparent molecular masses of 180, 110, 60, and 37 kDa as indicated by arrows in the Figure (Figure 7b). However, CK2 α and CK2 α′ knockdown failed to induce vacuolization in any of the CCA cell lines (Figure 7c). To confirm these results, we next treated CCA cell lines with 4,5,6,7-Tetrabromobenzotriazole (TBB), a well-characterized inhibitor of CK2 activity. Treatment with TBB did not induce the formation of vacuoles in any of the three CCA cells lines (Figure 7e-g) although it did block the phosphorylation of multiple CK2 substrates in HuCCA-1 and KKU-M213 cell lines (Figure 7d). We conclude that the induction of vacuoles by CX-4945, at least in HuCCA-1 and KKU-M213 cells, is not mediated by the inhibition of CK2 activity.

CX-4945 Induced Vacuolization Is CK2-Independent
To determine whether the effects of CX-4945 on vacuole formation were mediated by the inhibition of CK2, we next knocked down the expression of the α and α CK2 catalytic subunits using specific siRNAs (Figure 7a). Knockdown of both CK2 α and α in the HuCCA-1, CCLP-1, and KKU-M213 CCA cell lines was verified using Western blotting (Figure 6a). In all three cell lines, knockdown of CK2 α and CK2 α decreased the levels of multiple phospho-CK2 substrates (Figure 7b). A reduction in phospho-CK2 substrate levels was observed for proteins with apparent molecular masses of 180, 110, 60, and 37 kDa as indicated by arrows in the Figure (Figure 7b). However, CK2 α and CK2 α knockdown failed to induce vacuolization in any of the CCA cell lines (Figure 7c). To confirm these results, we next treated CCA cell lines with 4,5,6,7-Tetrabromobenzotriazole (TBB), a well-characterized inhibitor of CK2 activity. Treatment with TBB did not induce the formation of vacuoles in any of the three CCA cells lines (Figure 7e-g) although it did block the phosphorylation of multiple CK2 substrates in HuCCA-1 and KKU-M213 cell lines (Figure 7d). We conclude that the induction of vacuoles by CX-4945, at least in HuCCA-1 and KKU-M213 cells, is not mediated by the inhibition of CK2 activity.   The authors apologize for any inconvenience caused and state that the scientific conclusions are unaffected. The original publication has also been updated.

CK2 Knockdown Inhibited CCA Cell Proliferation and Cell Invasion
To investigate the effects of CK2 α and CK2 α knockdown on CCA cell proliferation, we next examined cell number and assayed BrdU incorporation. Combined CK2 α and CK2 α knockdown had no effect on HUCCA-1 cell number over 5 days in culture and only a modest effect on the number of CCLP-1 and KKU-M213 cells (Figure 8a). CCLP-1 and KKU-M213 cell number was decreased by 25% at 5 days after gene knockdown. In contrast to the biphasic effects of CX-4945 treatment on CCA cell invasion, combined CK2 α and CK2 α knockdown inhibited CCA cell invasion in HuCCA-1 and KKU-M213 cells. HuCCA-1 cell invasion was inhibited to approximately 60% of the control, while KKU-M213 cell invasion was inhibited to approximately 45% of the control (Figure 8b). Moreover, CK2 α and CK2 α knockdown had no effect on invasion of CCLP-1 cells. We conclude that the effects of CX-4945 treatment on CCA cell proliferation and cell invasion are unlikely to be mediated by the inhibition of CK2 in these cells.

CK2 Knockdown Inhibited CCA Cell Proliferation and Cell Invasion
To investigate the effects of CK2 α and CK2 α′ knockdown on CCA cell proliferation, we next examined cell number and assayed BrdU incorporation. Combined CK2 α and CK2 α′ knockdown had no effect on HUCCA-1 cell number over 5 days in culture and only a modest effect on the number of CCLP-1 and KKU-M213 cells (Figure 8a). CCLP-1 and KKU-M213 cell number was decreased by 25% at 5 days after gene knockdown. In contrast to the biphasic effects of CX-4945 treatment on CCA cell invasion, combined CK2 α and CK2 α′ knockdown inhibited CCA cell invasion in HuCCA-1 and KKU-M213 cells. HuCCA-1 cell invasion was inhibited to approximately 60% of the control, while KKU-M213 cell invasion was inhibited to approximately 45% of the control (Figure 8b). Moreover, CK2 α and CK2 α′ knockdown had no effect on invasion of CCLP-1 cells. We conclude that the effects of CX-4945 treatment on CCA cell proliferation and cell invasion are unlikely to be mediated by the inhibition of CK2 in these cells.

Discussion
CX-4945 is in clinical phase I/II trials for cholangiocarcinoma (ClinicalTrials.gov identifier: NCT02128282). Moreover, this drug has shown a therapeutic potential in several other types of cancer [17,22,28,[44][45][46]. Herein, CX-4945, at 10 μM and above, was shown to decrease protein kinase CK2 activity when delivered to CCA cells and to inhibit CCA cell proliferation (Figure 1). Unexpectedly, treatment with CX-4945 had biphasic effects on the ability of CCA cells to invade Matrigel. At 10 µ M, CX-4945 inhibited CCA cell invasion through a reduction of MMP-9 and MMP-7 levels plus an inhibitory effect on cell proliferation. Interestingly, lower doses of CX-4945 (1 and 5 µ M) increased the levels of MMP-9, MMP-7, and MMP-2 (Figure 1e,f). A biphasic effect of CX-4945 on cell proliferation has been reported previously in s22Rv1 prostate cancer cells [25], UM-SCC1 head and neck cancer cells [47], and A549 lung cancer cells [48]. CX-4945 at 0.5, 0.1, and 0.01 µ M slightly stimulated cell proliferation in the aforementioned cell lines, whereas higher doses inhibited cell proliferation [25,47,48]. Knockdown of the α and α′ catalytic subunits of CK2 simultaneously also significantly reduced the phosphorylation of multiple CK2 substrates indicating that CK2 knockdown and CX-4945 treatment both inhibited CK2 activity. We measured CK2 activity indirectly by using an antibody against the phospho-CK2 substrate (motif pS/pTDXE) to measure the amount
It has been reported that CX-4945 induces apoptosis via caspase-3 activity in lung cancer cells [17] and in primary CCA tumor lines [28]. Autophagy was induced by CX-4945 in pancreatic and lung cancer cell lines [45,89]. Surprisingly, we have shown that CX-4945-induced cell death in CCA cell lines was not accompanied by caspase-3 cleavage and did not appear to involve autophagy (Figure 5a,b) despite the involvement of intensive cytoplasmic vacuolization. This phenomenon was observed in a range of cell lines including CCA cell lines, immortalized cholangiocyte cell lines, breast and prostate cancer cell lines, and HEK293T cells. CX-4945-induced vacuoles originated from an engulfment of extracellular fluid, not from the ER. (Figure 6b,c). Interestingly, non-canonical cell death involving intensive cytoplasmic vacuolization is a hallmark of methuosis, oncosis, paratosis, and necroptosis [35]. We propose that the non-apoptotic, non-autophagic cell death mechanism induced by CX-4945 is not oncosis, paraptosis, or necroptosis according to the following criteria. First, no membrane blebbing was observed (Figure 3a). Second, chromatin condensation and nuclear fragmentation were absent (Figure 6a). Third, the vacuoles originated from an engulfment of extracellular fluid, not from the ER (Figure 6b,c) [35]. Furthermore, the vacuoles merged with others into large vacuoles ( Figure 2) and their pH remained neutral (Figure 6c). All of the aforementioned data fit the criteria of methuosis, therefore, we propose that CX-4945 induces methuosis at least in the CCA cell line KKU-M213.
Interestingly, PC3 prostate cancer cells did not form vacuoles. This absence of the vacuoles in PC3 cells suggests a cell line specific response to CX-4945. It seems likely that CX-4945 activity and its effects on cell death are cell-type specific, that is, dependent on, for example, whether tumor suppressors such as TP53 are wild type or mutant in the cell line being tested. Further study on the PC3 cell line might elucidate the molecular mechanism underlying the CX-4945-induced vacuolization and cell death reported in this study. Further investigation of the molecular mechanisms underlying CX-4945-induced vacuolization and cell death might lead to a better understanding of possible combinations of this treatment with other current chemotherapeutic drugs to optimize treatment outcomes in patients.

Cell Culture
Two cholangiocarcinoma cell lines, KKU-M055 and KKU-M213, established as previously described [90], were purchased from the Japanese Collection of Research Bioresources (JCRB) Cell Bank (Osaka, Japan). These two cell lines were from a Thai patient with an Opisthorchis viverrini record. The immortalized cholangiocytes cell line, MMNK-1, was also purchased from the JCRB Cell Bank. Intraheptic cholangiocarcinoma cell lines, HuCCA-1 were established and kindly provided by Prof. Stitaya Sirisinha [91]. HuCCA-1 was derived from a Thai patient whose serum was positive for the O. viverrini antigen [91].

BrdU Incorporation Assay
Cells were plated in tissue culture treated 96-well plates. BrdU incorporation was assessed by using a BrdU Cell Proliferation kit (ab126556, Abcam, Cambridge, MA, USA). Cells were treated with CX-4945 for 24 h. During the last 6 h, cells were incubated in a working stock of BrdU. After incubation with BrdU, media was removed and a fixative was added and the plates were incubated for 30 min at 23 • C. Reaction cocktail was added according to the manufacturer's protocol. The plate was read with a spectrophotometer plate reader at an absorbance at 450 nm. The average no BrdU background measurement was subtracted from all experimental measurements.

Cell Invasion Assay
Eight micrometer polycarbonate membrane transwell inserts (Corning costar, New York, NY, USA) were coated with 50 µL of a 1:9 ratio of Matrigel™ (BD Biosciences, San Jose, CA, USA) to serum free medium, that was allowed to congeal at 37 • C for 30 min. A total of 5 × 10 4 cells was added to the top chamber of each well in serum free medium. The bottom well was filled with complete medium. Treatment was added in both the transwell and the bottom well. Cells were incubated for 24 h at a 37 • C in a humidified incubator with 5% CO 2 . The chambers were swabbed to remove cells and the Matrigel™ remaining on the top of the membrane. Cells on the underside of the membrane were fixed in methanol for 5 min and stained with 0.5% (w/v) crystal violet in 12% glutaraldehyde in water for 5 min. Following a brief distilled H 2 O wash, cells were counted using a Nikon Eclipse T2S phase contrast inverted fluorescent microscope.

Cell Viability Assay
Cells were plated in tissue culture treated 6-well plates and incubated overnight. CX-4945 treatments were added and incubated for either 4 h or 48 h. Calcein-acetoxymethyl (AM) (Fisher Scientific) and propidium iodide (Fisher Scientific) were added 30 min before imaging at a final concentration of 0.125 µM and 1.5 µM, respectively. Images were taken with Nikon Eclipse T2S phase contrast inverted fluorescent microscope. Live cells were stained with calcein and showed a green fluorescent signal while dead cells were stained with Propidium iodine and showed a red fluorescent signal. Green and red fluorescence pixels were counted using Adobe Photoshop software. Average live and dead cells sizes were measured using Adobe Photoshop software.

siRNA Transfection
Transfection was performed by adding 4 × 10 5 cells to 100 nM siRNA mixture in Optimem medium (Fisher Scientific) in a total volume of 500 µL. siRNAs were commercially available from Qaigen (CK2α SI02660504 and SI02660497, CK2α SI00605416 and SI00605409). The mixture of siRNA and cells was plated into 6-well plates and incubated in a 37 • C humidified incubator with 5% CO 2 for 4 h. Complete growth medium was then added and incubated for 2 days before cells were used in further experiments.

Organelle Labeling and Imaging
Cells were plated in tissue culture treated 6-well plates and incubated overnight. CellLight™ ER-GFP and BacMam 2.0 (C10590, Fisher Scientific) were added to cells and incubated for 72 h at 50 particles per cell. LysoSensor™ Yellow/Blue DND-160 (L7545 Fisher Scientific) was added 15 min before the treatment. Treatment (25 µM CX-4945) was added and incubated for 6 h. Fresh culture media was changed prior to imaging. Live cells were imaged with a fluorescent invested microscope with cell culture chamber module (Olympus) at the Olympus Imaging Center, Mahidol University. Cells were plated on circle cover slides in tissue culture 24-well plates and incubated overnight. Twenty-five micromolars of CX-4945 was added and incubated for 24 h. Cells were fixed in 3.7% formaldehyde in Phosphate buffered saline (PBS) for 10 min, and then permeabilized in 0.1% Triton X-100 in PBS for 2 min. After incubation, cells were incubated in 300 nM DAPI in PBS to stain the nuclei. Fixed cells were imaged with a confocal microscope (Olympus) at the Olympus Imaging Center, Mahidol University.

Statistical Analysis
Data were graphed as mean ± standard error of the mean (SEM) using Prism Graphpad software (version 6.01, GraphPad Software, La Jolla, CA, USA). Statistical analyses were performed using ANOVA with Dunnett's test by using JMP software (version 10.0.0, SAS Institute Inc., Cary, NC, USA). All experiments were performed in triplicate.

Conclusions
Our data suggest that CX-4945 inhibits cell proliferation and induces methuosis via CK2-independent pathways. A biphasic effect of CX-4945 on CCA cell invasion was, at least in part, due to an alteration of MMP-2, MMP-7, and MMP-9.