Casein Kinase 1 Delta Regulates Cell Proliferation, Response to Chemotherapy and Migration in Human Ovarian Cancer Cells

Casein kinase 1 delta (CK1δ) has a tumor-promoting role in different cancers and it is genetically amplified in a portion of human epithelial ovarian cancer (EOC). CK1δ is involved in pleiotropic cellular functions such as cell proliferation, DNA damage, and migration. We specifically knocked down CK1δ by short hairpin RNA (shRNA) in human ovarian cancer cells and we performed proliferation, chemosensitivity, as well as in vitro and in vivo migration assays. CK1δ knocked-down cells displayed reduced proliferation capability both in vitro and in vivo. Nonetheless, these cells were sensitized to the first line chemotherapeutic agent carboplatin (CPT), and this observation could be associated to reduced expression levels of p21(Cip1/Waf1), involved in DNA damage response, and the anti-apoptotic X-linked inhibitor of apoptosis protein (XIAP). Moreover, CK1δ knocked-down cells were affected in their migratory and lung homing capability, even if in opposite ways, i.e., IGROV1, SKOV3 and MES-OV lost, while OVCAR3 gained motility potential. The results suggest CK1δ as a potential exploitable target for pharmacological EOC treatment, but they also advise further investigation of its role in cell migration.


INTRODUCTION
Epithelial ovarian cancer (EOC) is the first cause of death if considering only the gynecological malignancies, and the fifth leading cause of cancer-related death among women (1). This mortality is mainly due to the paucity of symptoms and to the lack of an effective screening program; as a consequence, EOC is usually diagnosed at advanced stages (2). EOC is often considered chemosensitive, with a response rate higher than 80% after first-line treatments. However, about 70% of patients relapse within 18 months (3). Recently, PARP [poly (ADP-ribose) polymerase] inhibitors such as Olaparib and Niraparib have been approved for treatment of BRCA-deficient EOC patients who have completely or partially responded to chemotherapy treatments (4). Nevertheless, the treatment of remaining EOC patients is still challenging and the search for possible molecular targets could help in the identification of promising therapeutic options.
For instance, the highly related CK1δ and ε are 98% identical in their kinase domain, but they differ in the regulatory N-terminal and C-terminal domains. CK1 are monomeric, constitutively active, co-factor independent kinases (6). CK1 recognize a consensus sequence S/T(P)-(X) 1−2 -S/T for phosphorylation. The substrates targeted by the different CK1 are involved in a plethora of cellular functions, including cell cycle progression, chromosome segregation, apoptosis, DNA repair, circadian rhythm, ribosome biogenesis, vesicle trafficking, p53, Wnt, Shh, and Hippo pathways (7)(8)(9)(10).
CK1 activity leads to anti-apoptotic effects, in many different ways. Indeed, CK1α mediates the resistance to TRAIL (tumor necrosis factor related apoptosis inducing ligand) by phosphorylating members of the death-inducing signaling complex (DISC). CK1α, δ, and ε phosphorylate the pro-apoptotic protein BID, thus reducing its caspase 8-mediated cleavage and activation (8).
CK1δ has been shown to be implicated in mitotic spindle dynamics and cell cycle progression (6). Indeed, CK1δ expression increases as cell cycle progresses from G1/S to G2/M phase, in order to efficiently phosphorylate and lead to proteasomal degradation the tyrosine kinase Wee1, which is a negative regulator of CDK1-cyclin B1, the checkpoint that controls the mitotic entry (11). Moreover, CK1δ regulates the total level and phosphorylation of the checkpoint kinase Chk1, which is induced in response to DNA damage. As a consequence, CK1δ silencing disrupts the Chk1-dependent G2/M checkpoint (7,12).
Alteration in the expression of CK1 has been observed in cancer (13,14). However, CK1 mutations seem to be rare in The Cancer Genome Atlas (TCGA) set of cancers (8). Additionally, the detected mutations do not accumulate in hotspots (8). Nonetheless, copy number variations are more frequently found. For example, in ovarian cancer, CK1δ gene (CSNK1D) is amplified in about 4% of the cases reported and is much more rarely deleted, while point mutations are not recorded (15,16). CK1δ is overexpressed in cells of hyperplastic B cell follicles and B cell lymphomas in p53-deficient mice (17), as well as in human choriocarcinoma (18), and in pancreas ductal adenocarcinoma (6,8,19). In breast cancer, CK1δ is strongly expressed in low grade carcinomas and is reduced in less differentiated cancers (6). Moreover, forced expression of a dominant-negative mutant CK1δ impairs SV40-induced transformation of mammary cells (20).
The development of CK1δ isoform-specific inhibitors to be used in the clinical practice has been undertaken. To date, different ATP-competitive inhibitors have been developed, but they have shown no specificity for the CK1δ or ε isoform with further limitations such as off-targets effects, limited antiproliferative activity in cell-based assays, poor solubility and pharmacokinetic properties (20). Therefore, in order to dissect if the possible effects can be ascribed to CK1δ or CK1ε inhibition, it is still necessary to study the specific role of the two kinase isoforms by genetic ablation by means of selective siRNA or shRNA.
Herein, we have investigated the role of CK1δ in ovarian cancer by specific short hairpin RNA approach, and addressed its role in proliferation, chemosensitivity and cell motility. Notably, we have shown that CK1δ knockdown affects ovarian cancer cell growth, and it sensitizes cells to carboplatin treatment. Moreover, CK1δ ablation impacts negatively on cell migration, even if with one exception.

Bioinformatic Analysis of CK1δ Expression in the Ovarian Tissue
The expression data of CK1δ gene (CSNK1D) and protein were obtained from Oncomine TM1 (22) and The Human Protein Atlas 2 (23) platforms, respectively. Fold changes of CSNK1D mRNA expression in EOC over normal ovary tissues and the respective p-values were retrieved from the TCGA, Bonome, Lu, Hendrix, Adib, and Toshihara ovarian datasets within Oncomine TM database. Immunohistochemistry images of CK1δ in normal ovarian tissue 3 and in ovarian cancer 4 were downloaded from the Human Protein Atlas, as well as the description of the protein expression levels.

Lentiviral Vector Production and Cell Transduction
For CSNK1D gene knockdown, MISSION R TRC shRNA bacterial glycerol stocks transformed with plasmids encoding short hairpin RNA (shRNA) specifically targeting human CSNK1D (sh599, sh1552) or a scramble control sequence (shCTRL) were purchased from Sigma-Aldrich (St. Louis, MO).

Patient-Derived Xenograft Generation and in vivo Experiments
Non-Obese Diabetic/Severe combined immunodeficiency (NOD/SCID) and NOD/SCID gamma (NSG) mice were obtained from internal breeding.
Patient-derived xenografts (PDX) were generated by injecting NOD/SCID mice intraperitoneally (i.p.) with 10 6 tumor cells derived from ascitic effusions of EOC-bearing patients (PDOVCA), collected after obtaining written informed consent. Briefly, patients' cancer cells were obtained by centrifugation of the ascitic fluid and subsequent red blood cell lysis, if needed (24). Cells were injected into NOD/SCID mice and ascitic fluid from mice was collected after its accumulation and processed in the same way as patients' clinical samples.
For in vivo tumor growth assay, 1 × 10 6 shCTRL, sh599, and sh1552 OVCAR3 and IGROV1 cells were injected subcutaneously (s.c.) in 200 µl of Matrigel R (Corning, New York, NY) in the dorso-lateral flank of NSG mice, and the growth rate was monitored by caliper measurements. Mice were sacrificed when the tumors of the shCTRL group reached 600-900 mm 3 volume. For protein extraction, tumors were snapfrozen in liquid nitrogen and homogenized with a T18 basic Ultra-Turrax R disperser (Ika, Staufen im Breisgau, Germany) in RIPA buffer.
For lung colonization assay, 1 × 10 6 shCTRL, sh599, and sh1552 Fluc-OVCAR3 and IGROV1 cells were injected into the tail vein of NOD/SCID mice. At 2 and 24 h after cell injection, mice received 200 µL of D-luciferin (15 mg/mL) i.p. for 8 min. Then, mice were sacrificed and lungs harvested and subjected to bioluminescence analysis with IVIS Imaging System, as previously described (25).

RNA Extraction, Reverse Transcription, and Quantitative RT-PCR
Total RNA was extracted following the TRIzol method (Ambion, Thermo Fisher Scientific) as per manufacturer's instruction, as previously described (26). cDNA was retrotranscribed from 1 µg of total RNA using the High capacity RNA-to-cDNA kit (Applied Biosystems, Thermo Fisher Scientific), then it was mixed with Platinum TM SYBR TM Green qPCR SuperMix-UDG (Invitrogen, Thermo Fisher Scientific) and the gene-specific primers; samples were run in duplicate. The PCR reaction was performed on ABI PRISM R 7900HT Sequence Detection System (Applied Biosystems, Thermo Fisher Scientific).
Ct values were utilized to calculate the fold change = 2 − Ct . Data were expressed as the difference in gene expression (normalized to the housekeeping β 2 -microglobulin gene) relative to the shCTRL sample. Primer sequences are:

Proliferation Assay
ShCTRL, sh599, and sh1552 OVCAR3 and IGROV1 cells were plated in triplicate in four 24-well plates at 5,000 cells/well. Cells were fixed with 4% PFA after overnight culture (day 1) and after 2, 3, and 4 days, and crystal violet-stained (Sigma Aldrich), as reported elsewhere (27). Crystal violet was solubilized in 1% sodium dodecyl sulfate (SDS), and the absorbance was measured at 595 nm using the plate reader Victor TM X4 (Perkin Elmer). The absorbance values were normalized to the corresponding day 1.
Flow cytometry analysis was performed using a FACS LSRII (BD Bioscience, Franklin Lakes, NJ); data were collected using a low flow rate and elaborated with FlowJo software (TreeStar, Ashland, OR).

In vitro Migration Assays
ShCTRL, sh599, and sh1552 OVCAR3, IGROV1, SKOV3 and MES-OV cell migratory capacity was determined in vitro by both wound healing and transwell migration assays, as previously described (28).
For wound healing assay, confluent cells were scratched with a p200 pipet tip. Wells were washed to remove detached cells, and medium was replaced with serum-free RPMI-1640 or DMEM (for MES-OV cells). At time 0 and after 24 and 48 h, pictures of the wounded area were taken with Leica DM IL LED microscope (Wetzlar, Germany). The distance between scratch edges was quantified using ImageJ software.
For transwell migration assay, 5 × 10 4 cells resuspended in 200 µL of RPMI-1640 supplemented with 0.2% FBS were seeded into 8 µm pore cell culture insert (migration chambers, Falcon, Corning) in 24-well plates. Wells were filled with 800 µL of RPMI-1640 medium containing 20% FBS, and cells were incubated at 37 • C. After 18 h, cells that had not crossed the membrane were removed with a cotton swab, and inserts were fixed with 4% PFA. Cells on the bottom of the membrane were stained with crystal violet. Images of five fields per insert were taken with a Leica DM IL LED microscope and the area covered by migrated cells was quantified using ImageJ software.

Statistical Analysis
Data from replicate experiments were shown as mean values ± Standard Deviation (S.D.) or Standard Error of the Mean (S.E.M.), as indicated. Comparisons between groups were done by the two-tail Student's t-test and Mann-Whitney test, as appropriate. The association between CK1δ staining score, reported in The Human Protein Atlas, and the status of either normal ovary or cancer tissue was analyzed using χ 2 -test. Statistical analyses were performed by using the Sigmaplot software (Systat Software, San Jose, CA).

Ethics Approval
Human samples of ascitic effusions from EOC-bearing patients were obtained after receiving written informed consent. The study was approved by IOV Institutional Review Board and Ethics Committee, and was performed in accordance with the declaration of Helsinki.
The animal studies presented in this work were approved by the Italian Ministry of Health (authorization n • 250/2017-PR).
Procedures involving animals and their care were performed according to institutional guidelines that comply with national and international laws and policies (EEC Council Directive 86/609, OJ L358, 12 December 1987).

CK1δ knockdown Causes a Growth Braking in Human Ovarian Cancer Cells
CK1δ has been demonstrated to have a pro-tumorigenic role in a variety of cancers, including B cell lymphoma (17), choriocarcinoma (18), pancreas ductal adenocarcinoma (19), colorectal cancer (29) and breast cancer (30). In the latter, CK1δ inhibitors efficiently slowed in vivo tumor growth (30). Retrieving expression data from Oncomine TM5 and The Human Protein Atlas platforms 6 , CK1δ showed a trend toward higher expression levels in ovarian cancer tissue compared to normal ovary at mRNA level (Figure 1A). At the protein level (Figure 1B), the analysis of the CK1δ expression score (low, medium, high) from the Human Protein Atlas showed a statistically significant difference between normal and cancer tissue (p < 0.001, χ 2 test), i.e., a higher expression in the latter.
Thus, we confirmed these observations checking the expression of CK1δ protein in a panel of human primary samples of ovarian cancer (PDOVCA), ovarian cancer cell lines, PDX samples and immortalized ovarian epithelial cell lines (31). As shown in Figure 1C, all the PDOVCA, cell lines and the PDX tested resulted strongly positive for CK1δ compared to the non-tumoral counterpart.
In order to verify whether CK1δ perturbation could have any effect on some typical features of cancer cells, such as proliferation, resistance to apoptosis, migratory ability, and in vivo tumor growth, we decided to knockdown CK1δ in two human EOC cell lines, OVCAR3 and IGROV1 cells. To this end, we transduced cells with lentiviral vectors bearing shRNA directed against CSNK1D, named sh599 and sh1552; control cells were generated by transduction with scramble shRNA (shCTRL). CSNK1D specific shRNA significantly downregulated CK1δ at both mRNA (Supplementary Figures 1A,B) and protein levels (Supplementary Figures 1C,D) in both cell lines.
Afterwards, cell growth was assessed by proliferation assay. Growth curves showed that CK1δ knockdown significantly affected both cell lines in their proliferative potential (Figures 2A,B). However, the observed differences in cell proliferation rate were not due to alterations in cell viability linked to CK1δ knockdown, since no significant differences were detected in the apoptotic cell rate (Annexin V-positive cells, Supplementary Figure 2). Furthermore, EdU-labeled cells were let proceed through the cell cycle for a few hours (7-8 h) and then analyzed with propidium iodide staining to detect their distribution in the different phases. According to the proliferation assay, cell cycle analysis highlighted that CK1δ knocked-down OVCAR3 ( Figure 2C) and IGROV1 ( Figure 2D) cells progressed more slowly through cell cycle phases than control cells. Indeed, after EdU incorporation, CK1δ knocked-down OVCAR3 and IGROV1 cells displayed a lower percentage of cells in G1 and a higher percentage in S/G2-M compared to shCTRL, meaning a braking of the progression through the cell cycle.
Eventually, we assessed the in vivo tumor growth potential of knocked-down OVCAR3 and IGROV1 cells injected s.c. into NSG mice. All cell lines were able to generate tumors, but in agreement with in vitro experiments, and with previously published data (30), sh599 and sh1552 tumors grew slower compared to shCTRL ones (Figures 2E,F). CK1δ silencing efficiency was checked at the end of the experiment, confirming that CK1δ was still knocked-down (Supplementary Figure 3).

CK1δ Knockdown Is Associated With Sensitization to Carboplatin Treatment Mediated by p21 and XIAP Downregulation
Since CK1δ has a role in the regulation of DNA-damage response and apoptosis (14), we tested knocked-down and control cell sensitivity to carboplatin (CPT), an alkylating agent used as first line chemotherapy for ovarian cancer treatment (32). Interestingly, after 72 h CPT treatment, we observed a significant increase in the percentage of apoptotic cells (Annexin V-positive) in CK1δ knocked-down OVCAR3 and IGROV1 cell lines (Figures 3A,B).
P21(Cip1/Waf1) is known to protect against apoptosis induced by DNA damage following radiation and cytotoxic agents, and to be involved in DNA repair (33), hence we wondered whether its expression was affected in CK1δ knockeddown, CPT-sensitive OVCAR3 and IGROV1 cells. Notably, we observed a significant decrease in p21 protein expression following CK1δ knockdown in both cell lines (Figures 3C,D).
Finally, we sought to further characterize the molecular basis underlying CPT cell sensitization. Thus, we checked the expression of X-linked inhibitor of apoptosis protein (XIAP),  a member of the family of anti-apoptotic proteins, whose overexpression in cancer cells is associated to resistance to a variety of apoptotic stimuli (34,35). Indeed, previous literature showed that the CK1 inhibition and CK2 silencing brought to XIAP downregulation, thus sensitizing cells to pro-apoptotic stimuli (19,36,37). Accordingly, we found that XIAP was significantly downregulated in CK1δ knocked-down cells (Figures 3C,D). This result is not in contrast with the absence of difference in cell viability between CK1δ-deficient and proficient cells in basal culture conditions (Supplementary Figure 2). Indeed, previous works already showed that XIAP downregulation does not induce basal cell death by itself, but only in the presence of pro-apoptotic stimuli that lead to caspase activation (38,39).
In order to further prove the role of CK1δ in the response of ovarian cancer cells to CPT, we silenced CK1δ in CPTresistant OVCAR3 cells (OVCAR3 CBP, IC 50 75 µg/mL), obtained from parental OVCAR3 cell line (IC 50 15 µg/mL) by 72 h exposure to increasing doses of CPT (21). Notably, in agreement with previous results shown in Figure 3A, CK1δ-knocked down CPT-resistant cells ( Figure 3F) displayed a significant higher percentage of apoptotic cells after 72 h CPT treatment compared to shCTRL cells ( Figure 3E). Again, both p21 and XIAP resulted to be significantly downregulated in CK1δ-knocked down OVCAR3 CBP cells ( Figure 3F). Similar results were also obtained in another CPT-resistant ovarian cancer cell line, MES-OV CBP (IC 50 75 µg/mL), obtained from parental MES-OV cell line (IC 50 20 µg/mL) by 72 h exposure to increasing doses of CPT (Supplementary Figures 4A,B) (21). Therefore, we may conclude that CK1δ knockdown is functionally associated with p21(Cip1/Waf1) and XIAP downregulation, and this could be a possible explanation for ovarian cancer cell sensitization to CPT treatment, probably due to an impaired DNA-damage response and apoptosis control.

CK1δ Knockdown Affects Migration in Human Ovarian Cancer Cells
Since CK1δ has been shown to regulate cell migration of triple negative breast cancer cells (40), we verified its possible involvement in the modulation of ovarian cancer cell motility. For this purpose, we firstly performed wound healing assay. To exclude biases due to cell proliferation, cells were maintained in serum-free medium. Interestingly, sh599 and sh1552 OVCAR3 cells were able to cover a larger area than shCTRL cells after 24 and 48 h (Figures 4A,B). Therefore, CK1δ knocked-down OVCAR3 cells displayed a higher migratory capacity.
This finding was further confirmed by transwell migration assay. Indeed, sh599 and sh1552 OVCAR3 cells passed through the transwell filter at a higher extent than shCTRL cells (Figure 4C).
On the contrary, wound healing (Figures 5A,B) and transwell migration assays ( Figure 5C) showed that sh599 and sh1552 IGROV1 cells migrated at a lower extent than shCTRL cells.
Moreover, since homing in secondary organs is a common trait of cells endowed with metastatic potential, we also performed an in vivo short-term experimental metastasis assay, in order to exclude any effect due to differences in cell proliferation, as already shown in Figures 2A-F. To this aim, we evaluated lung colonization ability of CK1δ knockeddown OVCAR3 and IGROV1 cells. First, shCTRL, sh599 and sh1552 cells were transduced with firefly luciferase in order to make cells detectable in vivo. Subsequently, cells were injected into the tail vein of immunocompromised NOD/SCID mice. Imaging performed after 2 h from injection showed no significant differences among cell lines, while 24 h after injection a higher bioluminescence signal was observed in the lungs harvested from mice injected with sh599 or sh1552 OVCAR3 cells (Figures 4D,E), further proving that CK1δ ablation increased the homing potential of OVCAR3 cells.
According to the above in vitro data, in vivo lung colonization assay confirmed that sh599 and sh1552 IGROV1 cells have an impaired homing capacity as compared to shCTRL cells (Figures 5D,E).
In view of the contrasting results described above, we decided to investigate more deeply CK1δ role in cell motility. To this end, we knocked down CK1δ in two additional ovarian cancer cell lines, SKOV3 and MES-OV (Supplementary Figures 5A,B).
In agreement with what observed for IGROV1 cells, both cell lines displayed an impaired migratory capability in both wound healing (Figures 6A,B,D,E) and transwell migration (Figures 6C,F) assays.
To sum up, the available data suggest that CK1δ-knockdown influences negatively ovarian cancer cell migratory and homing capability, but some exceptions could be present due to the complex cellular context.

DISCUSSION
EOC carcinogenesis is supported by numerous genetic and epigenetic alterations and aberrant signaling molecule activation (41)(42)(43)(44)(45)(46)(47). Due to its reported genetic amplification in EOC and its tumor-favoring role described in different cancers (8), we focused on CK1δ, a member of a kinase family characterized by pleiotropic cellular functions, including cell cycle progression, p53 control, mitotic spindle arrangement, and circadian rhythm protein turnover (6). Interestingly, proteins involved in circadian rhythm control, e.g., PER, CRY, BMAL1, CLOCK, are highly expressed in the ovaries, where they regulate ovulation and hormonal cycles in general (43,48), and alterations in their expression levels are associated with an increased risk of ovarian cancer and affect cancer growth, invasiveness, and drug sensitivity (49)(50)(51).
Previous work highlighted a negative impact of CK1δ inhibition on cell cycle progression and proper mitosis, eventually leading to apoptosis (18,20,29,52). Moreover, in vivo breast and pancreatic tumor growth was delayed by CK1δ inhibitor administration (19,30), and recently, impaired cell migration and metastases of triple negative breast cancer were also assessed (40). Since CK1δ resulted to be expressed at a higher extent in ovarian cancer tissue compared to the normal counterpart at both mRNA (Oncomine TM ) and protein levels (The Human Protein Atlas), we investigated its potential role in the regulation of EOC cellular functions.
In agreement with Oncomine TM and The Human Protein Atlas data, EOC cell lines, PDOVCA, and PDX samples displayed higher CK1δ levels in all the specimens compared to the nontumoral counterpart. Genetic ablation of CK1δ in OVCAR3 and IGROV1 ovarian cancer cells impaired cell proliferation in both cell lines, and accordingly, CK1δ-deficient cells gave rise to smaller tumors in immunocompromised mice in agreement with previous literature (19,30). P21(Cip1/Waf1) is a well-known tumor-suppressor that arrests cell cycle progression by disrupting CDK/cyclins complexes and by associating to proliferating cell nuclear antigen (PCNA) (53). Nonetheless, high p21 protein levels have been shown to convey mitogenic signals in vascular smooth muscle cells, augmenting the transit through cell cycle (54).
Besides its role in cell cycle control, p21 (Cip1/Waf1) is involved in DNA repair (33) and it determines an apoptosisresistant phenotype, acting as a tumor-promoting factor (55). For instance, renal cancer is usually difficult to treat because chemo-resistant, due to a highly effective DNA repair response mediated by high p21 levels (56). P21 attenuation by antisense oligonucleotides (57), small molecules (56), or sorafenib (58), is associated to increased chemo-sensitivity, with higher apoptosis rates after doxorubicin, paclitaxel and cisplatin treatment. In prostate cancer, infiltrating mast cells induced p21 expression, thus increasing docetaxel resistance (59). Furthermore, Pavan et al. (60) reported that IRF-1 knockdown sensitized ovarian cancer cells to cisplatin because of impaired p21 induction. Accordingly, CK1δ knocked-down CPT-sensitive and resistant cells showed p21 attenuation and were similarly sensitized to CPT treatment, displaying a higher proportion of apoptotic cells than control ones.
XIAP is a member of the "inhibitor of apoptosis" family, a group of proteins characterized by the presence of baculoviral IAP repeats, responsible for their inhibitory binding to caspases. XIAP in particular can bind both initiator and effector caspases (61). Izeradjene et al. described XIAP downregulation in colorectal cancer cells following the administration of the CK1 inhibitor CKI-7 only if combined with TRAIL (37) and in CK2α knockeddown rhabdomyosarcoma cells (36). Furthermore, Brockschmidt et al. (19) reported that CK1δ/ε inhibitor IC261 exerted pro-apoptotic effects in pancreatic ductal adenocarcinoma by reducing the expression of many anti-apoptotic proteins, including XIAP. Moreover, even though McManus et al. (38) demonstrated that XIAP knockdown sensitized breast cancer cells to taxanes, but not to CPT, recent data by Zhang et al. (62) highlighted that XIAP high-expressing ovarian cancer patients were less likely to respond to CPT treatment and showed that XIAP knockdown in ovarian cancer cell lines slowed their proliferation kinetics, on one hand, and sensitized them to this drug in vitro, on the other. Previously, Asselin et al. (63) demonstrated that XIAP overexpression prevented cisplatin-induced apoptosis in A2780 ovarian cancer cell line in a PI3K/Akt-dependent manner. Accordingly, we observed a reduction in XIAP expression in CK1δ knocked-down CPTsensitive and resistant cells, sensitized to carboplatin treatment.
A recent paper reported that CK1δ knockdown impaired breast cancer cell migration and invasion and upregulated epithelial markers such as claudin1 and occludin (40). Accordingly, specific knockdown of the CK1δ highly related kinase, CK1ε, demonstrated a 60-80% reduction in the migration ability of the EOC cell line SKOV3 (64).
Likewise, different migration assays performed on IGROV1, SKOV3, and MES-OV cells revealed a reduced migratory capacity of CK1δ knocked-down cells. On the contrary, the opposite outcome was observed when the same experiments were performed on OVCAR3 cells. Indeed, sh599 and sh1552 OVCAR3 cells presented a higher motility both in vitro and in vivo. On the whole, our experiments, and previous literature, support a positive role of CK1δ in cancer cell migration. However, an opposite behavior could also be observed. Likely, this kinase interacts directly or indirectly with different partners depending on the complex cellular context, and further work is required to unveil the molecular mechanism underlying this CK1δ-dependent phenomenon.
In summary, we have identified CK1δ as an important player in the regulation of cell proliferation, response to chemotherapeutic drugs, and migration in ovarian cancer cells. Our results would suggest CK1δ as an attractive target for ovarian cancer treatment, but further investigation regarding its role in migration is recommended before the introduction of CK1δ inhibitors into the clinics for EOC management.
The raw data supporting the conclusions of this manuscript will be made available by the authors, without undue reservation, to any qualified researcher.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by IOV Institutional Review Board and Ethics Committee. The patients/participants provided their written informed consent to participate in this study. The animal study was reviewed and approved by Italian Ministry of Health (authorization n • 250/2017-PR).