CENPK Interacts with SOX6 to Promote Cervical Cancer Stemness and Chemoresistance Through Wnt and P53 Signaling

Stemness and chemoresistance contribute to cervical cancer recurrence and metastasis. It is meaningful to develop alternative targets for eliminating cancer stem cells (CSCs) properties, suppressing metastasis, and alleviating chemoresistance in cervical cancer. This study claries the role of CENPK in cervical cancer stemness and chemoresistance and its clinical signicance in cancer progression. Human cervical cancer cell lines, xenografts, and clinical samples were used for expression and functional analysis. CENPK expression in cervical cancer was analyzed by bioinformatics based on a microarray and TCGA database and immunohistochemistry based on 119 paran-embedded cervical cancer specimens and 35 paran-embedded adjacent normal tissues in a tissue chip.


Tumorsphere formation assay
The protocol of tumorsphere formation assay was in line with a previous method [16] . Brie y, cervical cancer cells (5,000/well) were seeded on 6-well ultra-low-attachment plates (Corning, Inc., NY, USA) and cultured in serum-free DMEM/F12 with 2% B27, 20 ng/ml EGF, and 20 ng/ml FGF. Fresh medium was replaced every 2 days. The tumorspheres were recorded and counted on day 14 post-seeding, and the number and size of tumorspheres were analyzed after passaging for three generations.

MTT assay
The MTT assay was used to measure cell proliferation. Cervical cancer cells (1,000 cells/well) were seeded in 96-well plates. At the indicated time, MTT (5 mg/ml; Sigma-Aldrich Corporation, MO, USA) was added to each well. After 4 h of incubation, dimethyl sulfoxide (Sigma-Aldrich Corporation, MO, USA) was added to each well. The absorbance value (OD) of each well was detected at 490 nm.

Colony-formation assay
For the measurement of cell proliferation, cervical cancer cells were plated at a density of 100 cells per well. After 10 days of culture, colonies were xed with methanol and stained with a hematoxylin solution. The number of colonies (≥ 50 cells) was counted under a microscope. For measurement of cell chemoresistance, transfected cervical cancer cells were treated with cisplatin or carboplatin for 6 h at the indicated concentrations before seeding at a density of 500 cells per well. After 10 days of incubation, colonies were xed, stained, and photographed for analysis.

Transwell assays
Transwell assays were applied to evaluate the migration and invasion capacity of cervical cancer cells. Cells suspension were added to the upper chamber of Transwell coated with (for invasion assay) or without (for migration) Matrigel purchased from BD corporation, and the lower chambers of Transwell were lled with Dulbecco's modi ed Eagle's medium containing 10% FBS. The migrated and invaded cells were xed, stained, and photographed under microscopy.

EdU incorporation assays
Following the manufacturer's protocol, EdU incorporation was conducted using the Apollo567 In Vitro Imaging Kit (RiboBio Corporation, Guangzhou, China). After 2 h treatment with EdU (10 μM), cervical cancer cells were xed with 4% paraformaldehyde, permeabilized with 0.3% Triton X-100, and co-stained with Apollo uorescent dyes and DAPI (5μg/ml). The EdU-positive cells were photographed and counted under a microscope.

Immuno uorescence
Cervical cancer cells were seeded and grew on coverslips. After incubation, the cells were subjected to xation with 4% paraformaldehyde, permeabilization with 0.2% Triton X-100, and incubation with antibodies (shown in Additional le 7 Table S2). The cells were then co-stained with DAPI (0.2 mg/ml) and observed under a uorescence confocal microscope.

RT-PCR and QPCR
Total RNA was extracted from cervical cancer cells or xenografts using TRIzol (Invitrogen Corporation, Shanghai, China), and cDNA was generated using a reverse transcription reagent kit (TaKaRa Corporation, Dalian, China). Then, the synthesized cDNA was used as a template for RT-PCR and QPCR with speci c primers (shown in Additional le 8 Table S3) on Bio-Rad T100 and Bio-Rad CFX 96, respectively. β-actin was used as a control. For RT-PCR, the images were visualized by Bio-Rad GelDoc XR+. For QPCR, the relative mRNA expression was quanti ed using the 2 -ΔΔCt method.

Western blot analysis
Cervical cancer cells were lysed in lysis buffer, and proteins were quanti ed by the BCA method. After loading, proteins were separated, transferred, and immunoprobed with speci c antibodies (shown in Additional le 7 Table S2). Chemiluminescence was used for protein detection, and images were captured using a Bio-Rad ChemiDocTM CRS+ Molecular Imager.

Animal studies
The Institutional Animal Ethical Committee, Experimental Animal Center of Fujian Medical University approved the protocols for animal studies. All experiments conform to all relevant regulatory standards.
BALB/c-nu mice (4-5-week-old, female) (n = 5 per group) were subcutaneously injected with 5 × 10 6 cells/100 μl in the ank. The longest diameter (A) and the shortest diameter (B) of the growing tumors were measured every 3 days with a caliper, and the tumor volume (V) was calculated by the equation V = (AB 2 )/2. Twenty days after cell injection, the mice were euthanized and tissue was taken for further experimentation. The subcutaneous xenograft mouse model was constructed and used to evaluate tumor formation ability and tumor growth. A series of 5×10 5 , 2×10 5 , and 5×10 4 cells inoculated subcutaneously into BALB/c-nu mice (4-5-week-old, female) (n = 6 per group), and the tumor-initiating frequency was evaluated. A pulmonary metastasis model was applied to detect cervical cancer metastatic ability. BALB/c-nu mice (4-5-week-old, female) (n = 5 per group) were injected with 1 × 10 6 cells/100 μl through the tail vein. Four weeks after injection, the formation of lung metastasis was evaluated under light and uorescent microscopy.
The subcutaneous xenograft mouse model was also used to determine the chemoresistance of cervical cancer tumors. Cervical cancer cells (5 × 10 6 cells/100 μl) were inoculated in the ank of the BALB/c-nu mice (4-5-week-old, female) (n = 10 per group). The mice were treated with cisplatin (3 mg/kg, i.p.) once a week for six weeks [17,18] or carboplatin (30 mg/kg, i.p.) once a week for four weeks [19] . Survival curves were plotted using Kaplan-Meier analyses.

Luciferase reporter assay
The assays of Wnt signaling activity, p53 signaling activity, and the transcriptional activity of RAD21/SMC3 on CENPK were carried out using the luciferase assay system. Cervical cancer cells were co-transfected with TOP ash or FOP ash with pRL (Millipore Corporation, MA, USA) to detect Wnt signaling. Cervical cancer cells were transfected with pGL4 luciferase reporter vector (Promega Corporation, Madison, USA) harboring p53 response element to detect p53 signaling. Cervical cancer cells were transfected with pGL4 luciferase reporter vector (Promega Corporation, Madison, USA) harboring RAD21/SMC3-binding sites in the CENPK promoter region to detect the transcriptional activity of RAD21/SMC3 on CENPK. Forty-eight hours after transfection, cells were collected and subjected to luciferase activity measurement using the Dual-Luciferase Reporter Assay System (Promega Corporation, Madison, USA) on a BioTek luminometer. The relative activity of each reaction was calculated as the ratio of re y to Renilla luciferase.

Co-immunoprecipitation (Co-IP)
According to the manufacturer's instructions, Co-IP was conducted with a Pierce Co-Immunoprecipitation kit (Thermo Scienti c, Shanghai, China). In brief, total proteins were lysed from cervical cancer cells and subjected to concentration quanti cation. Proteins (5 mg) were then incubated with speci c antibodies or IgG, which was applied as a negative control. The enriched proteins were eluted and used for western blot analysis.

Chromatin immunoprecipitation (ChIP)
The ChIP assay was conducted using a ChIP assay kit (Thermo Scienti c, Shanghai, China) according to the manufacturer's instructions. Brie y, chromatin from cervical cancer cells was crosslinked, extracted, and clipped with Micrococcal Nuclease to generate DNA fragments. Immunoprecipitation was performed with speci c antibodies (shown in Additional le 7 Table S2) or IgG, which served as a negative control. The DNA fragments were eluted, puri ed, and subjected to PCR and qPCR.
Cycloheximide (CHX) chase assay Cervical cancer cells were treated with cycloheximide (50 µg/ml) at the indicated time. Subsequently, proteins were collected and quanti ed using a BCA method. The proteins were further subjected to western blot analysis, and signals were recorded for evaluating the half-life of the proteins.

Cell fractionation assay
The assay was carried out using NE-PER™ Nuclear and Cytoplasmic Extraction Reagents (Thermo Scienti c, Shanghai, China). Cervical cancer cells were sequentially treated with ice-cold CER I and the CER II extraction reagent. After centrifugation, the supernatant containing cytoplasmic extract was collected. The pellet was further incubated with NER extraction reagent. After centrifugation, the supernatant containing nuclear extract was retained. The collected proteins were further subjected to western blot analysis.

Patient tissues
One hundred and nineteen para n-embedded cervical cancer specimens and 35 para n-embedded adjacent normal tissues in a tissue chip (HUteS154Su01) was purchased from Shanghai Outdo Biotech (Shanghai, China). All the patients underwent surgery and had a con rmed pathological diagnosis. Clinical data were extracted from the medical records of the patients. Some clinicopathological characteristics, such as histology grade and Ki67 status, were not available for all patients. Patient consent and ethics approval were obtained from the Ethics Committee of Shanghai Outdo Biotech Corporation.

Immunohistochemistry (IHC)
Para n-xed sections (4 μm) from tissues were depara nized, rehydrated, and subjected to antigen retrieval in citrate buffer. After eradicating endogenous peroxidase activity with 3% H 2 O 2 and blocking non-speci c antigens with goat serum, the sections were incubated with speci c antibodies (shown in Additional le 7 Table S2). The signal of the sections was measured using the DAB substrate (Maixin Biotech. Corporation, Fuzhou, China). Staining intensities were evaluated as previously described [15] .

Bioinformatics analysis
The mRNA-Seq data of cervical squamous cell carcinoma and endocervical adenocarcinoma (CESC) in The Cancer Genome Atlas (TCGA) database were downloaded and applied for performing differential expression analysis, correlation analysis, gene set enrichment analysis (GSEA), and gene set variation analysis (GSVA) with R software. The median of mRNA expression values was assigned as a cut-off value, and samples were classi ed into high-and low-expression groups according to the cut-off value.

Statistical analysis
The data are presented as the mean ± SD from at least three independent experiments and analyzed by SPSS 22.0 or RStudio. Statistical signi cance was detected using the Student's two-tailed t-test for two groups, one-way ANOVA for multiple groups, and a parametric generalized linear model with random effects for the growth curve. The correlations between CENPK expression and clinical parameters were measured using the Chi-square test. The correlations among gene expressions were measured using Spearman's rank correlation test. The Wilcoxon rank-sum test was applied to investigate the difference of CENPK expression between cancers and adjacent normal tissues. Kaplan-Meier survival curves were plotted for survival analysis, and a log-rank test was used to explore the difference. Cox regression models were adopted to identify the relationship between CENPK and the survival time of cervical cancer patients. All statistical tests were two-sided and a P value of < 0.05 was considered statistically signi cant. * P < 0.05, ** P < 0.01 and *** P < 0.001.

Results
CENPK is elevated in cervical cancer and is associated with stemness, chemoresistance, DNA damage repair signaling, and DNA replication Our previous study demonstrated an oncogenic role for iRhom2 in cervical cancer [15] , and its microarray data indicated that iRhom2 knockdown induced the downregulation of 124 genes and the upregulation of 33 genes (fold change ≥ 2.0, P 0.05). Differential and correlation analyses were then conducted in the TCGA CESC dataset and revealed 19 of the 157 genes that were differentially expressed between tumor and normal tissues and were correlated with iRhom2 expression. Since iRhom2 promotes cell cycle progression and DNA replication in our previous study, GSEA was performed to investigate the correlations between the expression of those 19 genes and the cell cycle and DNA replication. Therein, CENPK, TMPO, UNG, and RCC2 were identi ed as the genes that potentially mediated the pro-proliferative effect of iRhom2 in cervical cancer ( Fig. 1a-c and Additional le 1 Fig. S1a). Moreover, CENPK was con rmed to be the most prominent downstream effector of iRhom2 among these four genes (Additional le 1 Fig. S1b).
CENPK was upregulated in pan-cancer tissues compared to the corresponding normal tissues in TCGA (Additional le 1 Fig. S1c). GSEA and GSVA showed that CENPK was involved in regulating the Wnt, DNA damage repair signaling, cell cycle, and DNA replication ( Fig. 1d-f). Since DNA damage repair is closely related to platinum drug resistance and radioresistance, we further collected the platinum drug resistanceassociated gene set from KEGG websites and constructed the radioresistance-associated gene set based on a previous study [20] . Intriguingly, GSEA showed that CENPK participated in the modulation of platinum drug resistance, which was more prominent than its regulation of radioresistance ( Fig. 1g and Additional le 1 Fig. S1d). In addition, the correlation analyses showed a positive association between CENPK and stemness markers, including EPCAM, CD133, SOX2, and OCT4 (Fig. 1h).
Silencing CENPK suppresses cell CSCs properties, chemoresistance, metastasis, and proliferation in cervical cancer Considering the bioinformatics results, we then focused on the function of CENPK in cell CSCs properties, chemoresistance, and proliferation in cervical cancer. siRNA and lentivirus were used to knock down CNEPK expression in cervical cancer cells (Additional le 2 Fig. S2). Tumorsphere formation assays and immuno uorescence indicated that CENPK knockdown reduced CSCs properties of HeLa and SiHa cells ( Fig. 2a and b). Clonogenic assays revealed that inhibition of CENPK impaired the resistance to cisplatin and carboplatin in HeLa and SiHa cells (Fig. 2c). Transwell assays indicated that silencing of CENPK impaired the migration and invasion capacity in HeLa and SiHa cells (Fig. 2d). MTT and colony formation assays also showed that CENPK knockdown inhibited cell growth in HeLa and SiHa cells ( Fig. 2e and f). Interestingly, immuno uorescence showed that downregulation of CENPK enhanced DNA damage as indicated by γH2AX expression in cervical cancer cells treated with platinum drugs (Fig. 2g).
Consistent with the bioinformatics results, EdU assays suggested that CENPK knockdown suppressed the percentage of S phase and DNA replication in HeLa and SiHa cells (Fig. 2h). In addition, downregulation of CENPK facilitated the expression of DNA damage repair-associated genes p53 and p21, and suppressed the expression of cell stemness, epithelial-mesenchymal transition (EMT), and DNA replication-associated genes c-Myc, Vimentin, CCND1, and c-Jun (Fig. 2i).
These ndings were then tested in animal models using stable CENPK-silenced cervical cancer cells. Subcutaneous tumor xenograft mouse models were applied to evaluate the role of CENPK in cervical cancer. Decreased CENPK expression was con rmed in the tumors treated with CENPK shRNA (Fig. 3a).
In the animal model, tumors with CENPK downregulation presented a lower tumor formation ability, a less lung metastasis formation, and a slower growth rate than controls (Fig. 3b-f). IHC also showed a lower expression of Ki67 expression in the CENPK-silenced tumor compared to the controls (Fig. 3g). Interestingly, CENPK silencing prolonged the survival time of the mice, and CENPK downregulation synergized with chemotherapeutic treatment to further extend the survival time of the mice (Fig. 3h).
CENPK interacted with SOX6 to activate Wnt signaling and inactivate p53 signaling in cervical cancer CENPK was shown to modulate Wnt signaling (Fig. 1d). TOP/FOP-ash assays con rmed the facilitation of Wnt signaling by CENPK (Fig. 4a). To explore the speci c mechanism by which CENPK exerted its oncogenic role, we used the BioGRID database (https://downloads.thebiogrid.org/BioGRID) and identi ed a potential interaction between CENPK and SOX6. In addition, previous studies have demonstrated that SOX6 can activate p53 signaling and inactivate Wnt signaling [21] , and can suppress cervical cancer progression [22] . Therefore, SOX6 was selected as a candidate protein interacting with CENPK. Co-IP analyses veri ed the interaction between CENPK and SOX6 (Fig. 4b). Moreover, colocalization of CENPK and SOX6 was detected in HeLa and SiHa cells (Fig. 4c).
SOX6 inhibits the transcription of target genes regulated by β-catenin through interacting with β-catenin and thereby inactivating Wnt signaling [21] . Moreover, TOP/FOP-ash assays veri ed the inhibition of Wnt signaling by SOX6 (Fig. 4a), implying that the interaction between CENPK and SOX6 might disrupt the interaction between SOX6 and β-catenin. Co-IP analyses indeed demonstrated the interaction between SOX6 and β-catenin and revealed that CENPK could competitively bind to SOX6 to prevent interaction between SOX6 and β-catenin (Fig. 4d), thereby activating Wnt signaling. Moreover, immuno uorescence analyses suggested an inhibitory impact of CENPK knockdown on β-catenin expression and nuclear translocation (Fig. 4e). However, we detected no effect of CENPK on SOX6 location or expression (Fig. 4e).
SOX6 could increase the stability of p53 protein by transcriptionally suppressing c-Myc transcription, thus activating p53 signaling in HeLa cells [22] . Accordingly, the effect of CENPK on p53 signaling was further investigated. Dual-luciferase assays showed that CENPK knockdown facilitated p53 signaling and that SOX6 silencing abolished that effect (Fig. 4f). Intriguingly, ChIP assays showed that CENPK knockdown facilitated SOX6 binding to the c-Myc transcription regulatory region (Fig. 4g). Furthermore, CHX chase assays, Co-IP analyses, and cell fraction assays all supported a stimulatory role of CENPK knockdown on p53 stability and an inhibitory role of CENPK knockdown on p53 ubiquitination and nuclear export, these effects all being terminated by SOX6 knockdown (Fig. 4h-j).
Collectively, these data suggested SOX6 mediation of CENPK's effect on Wnt and p53 signaling.
CENPK promotes cervical cancer cell CSCs properties, chemoresistance, metastasis, and proliferation through Wnt and p53 signaling Next, CENPK function mediated by Wnt and p53 signaling in cervical cancer was further investigated.
Tumorsphere formation assays, immuno uorescence, clonogenic assays, MTT, and Edu assays showed that β-catenin overexpression partly reversed the inhibitory effect of CENPK knockdown on cell CSCs properties, resistance to cisplatin and carboplatin, migration, invasion, and growth in HeLa and SiHa cells (Fig. 5a-f). Furthermore, p53 knockdown partially eliminated the suppressive impact of CENPK inhibition on stemness, cisplatin and carboplatin resistance, migration, invasion, and proliferation of HeLa and SiHa cells (Fig. 5a-f). In addition, immuno uorescence and western blot showed that the multiple effects of CENPK knockdown on the expression of DNA damage repair, cell stemness, EMT, and DNA replicationassociated genes (γH2AX, p53, p21, c-Myc, Vimentin, CCND1, and c-Jun) could be reversed by β-catenin overexpression or p53 knockdown ( Fig. 5g and h).

RAD21/SMC3 transcriptionally promotes CENPK expression and forms a positive regulatory circuit through Wnt signaling
To further illustrate how CENPK expression was regulated, we focused on the CENPK transcription process. Firstly, 1558 identi ed transcription factors were extracted from the Cistrome Data Browser (http://cistrome.org/db/) database (gene set A). Secondly, the top 1000 genes correlated with CENPK expression were identi ed using the TCGA CESC dataset (gene set B). From the intersection of these two gene sets, a total of 91 transcription factors were identi ed. Thirdly, the GTRD database (http://gtrd.biouml.org/) was applied to predict the factors that bound to the promoter region of CENPK based on ChIP-seq experiments performed in cervical cancer cells, and 37 transcription factors were identi ed with this strategy (gene set C). From the intersection of these three gene sets, we identi ed RAD21, SMC3, and AFF4 as potential factors regulating CENPK transcription (Fig. 6a).
Interestingly, RAD21 and SMC3 were shown to be target genes of β-catenin [23,24] . Since CENPK could activate Wnt signaling, we proposed that there was a positive feedback between RAD21/SMC3 and CENPK through Wnt signaling. TOP/FOP-ash assays showed that RAD21 or SMC3 knockdown impaired Wnt signaling activity, but this effect was obliterated by overexpressing CENPK (Fig. 6g). ChIP assays also indicated that inhibition of Wnt signaling suppressed the binding of RAD21/SMC3 to the CENPK promoter region (Fig. 6h). Furthermore, dual-luciferase assay assays showed that inhibition of Wnt signaling facilitated p53 signaling (Fig. 6i). Collectively, these results indicate that RAD21/SMC3 and CENPK formed a feedback loop to drive Wnt/β-catenin signaling, and that there was a link between Wnt and p53 signaling.
iRhom2 modulates the RAD21/SMC3-CENPK/SOX6-mediated Wnt and p53 signaling to potentiate cervical cancer stemness, chemoresistance, metastasis, and proliferation We previously found that iRhom2 downregulated Wnt signaling to inhibit cell cycle progression and DNA replication in cervical cancer [15] . We thereby investigated the relationship between iRhom2 and the CENPK/SOX6-β-catenin-RAD21/SMC3 feedback loop. RT-qPCR and western blot showed that silencing of iRhom2 decreased CENPK expression (Additional le 4 Fig. S4a), and TOP/FOP luciferase reporter assays and dual-luciferase assays suggested that silencing of iRhom2 decreased Wnt signaling activity and elevated p53 signaling activity. However, these effects were reversed by CENPK overexpression (Additional le 4 Fig. S4b and c). Moreover, iRhom2 knockdown inhibited RAD21/SMC3-mediated CENPK transcription, an effect that was abolished by activating Wnt signaling (Additional le 4 Fig. S4d). Functional and mechanistic analyses suggested that iRhom2 knockdown had a suppressive role in cervical cancer stemness, chemoresistance, metastasis, and proliferation that was terminated by overexpressing CENPK (Additional le 4 Fig. S4e-l).

Relationships between CENPK expression and clinicopathological features of cervical cancer patients
To further establish the role of CENPK, immunohistochemistry was conducted in 119 cervical cancer and 35 adjacent normal tissues. The results showed that CENPK was elevated in cervical cancer relative to adjacent normal tissues (Fig. 7a). Moreover, CENPK protein levels were positively associated with Ki67 protein levels (Fig. 7b) and positively correlated with cancer recurrence (Table 1). Survival analysis suggested that cervical cancer patients with high CENPK expression exhibited poor overall survival (Fig. 7c). Since CENPK expression was correlated with cancer recurrence, we further investigated the relationship between CENPK expression and the recurrence-free survival of cervical cancer patients. Survival analysis showed that cervical cancer patients with high CENPK expression had a poor recurrence-free survival (Fig. 7d). Univariate and multivariate COX hazard analyses were performed to correlated CENPK, clinicopathological features expression, and overall survival and recurrence-free survival of cervical cancer patients. CENPK expression de ned by IHC was an independent and unfavorable prognostic factor for the overall survival and recurrence-free survival of cervical cancer patients ( Fig. 7e and f). A nomogram showing the role of CENPK, age, and T classi cation in predicting cancer recurrence of cervical cancer patients is shown in Fig. 8g; even strati ed by age or T classi cation, CENPK expression was still correlated with the overall survival and recurrence-free survival of cervical cancer patients (Additional le 5 Fig. S5a and b). Consistently, the correlation analyses suggested a positive association between RAD21, SMC3, β-catenin, Ki67, and CENPK expression (Additional le 5 Fig.  S5c).

Discussion
Stemness, chemoresistance, metastasis, and tumor size are responsible for cancer recurrence and/or the predicted poor prognosis of cervical cancer patients [25,26] . In the present study, we demonstrated that CENPK augmented Wnt signaling and attenuated p53 signaling, thus affecting the expression of key regulators of DNA damage repair, EMT and DNA replication, and facilitating the stemness, chemoresistance, metastasis, and proliferation of cervical cancer cells. Downregulation of CENPK delayed the progression of cervical cancer in vitro and in vivo. Importantly, high CENPK expression in cervical cancer was positively correlated with poor overall survival and recurrence-free survival of cervical cancer patients.
Bioinformatics analyses suggested an oncogenic role of CENPK that may be associated with cell proliferation in cancer [6] . In this work, we consistently found that CENPK stimulated cell CSCs properties, chemoresistance, metastasis, and proliferation in cervical cancer. Interestingly, our mechanism research suggested the interaction between CENPK and SOX6. SOX6 plays a role as a tumor suppressor gene in a variety of cancers [27][28][29] , including cervical cancer [30] . Moreover, SOX6 is a Wnt/β-catenin pathway inhibitor and p53 pathway activator [31,32] . Here, we demonstrated that CENPK interacted with SOX6 to impair the interaction between SOX6 and β-catenin and SOX6 suppression of c-Myc transcription, thus activating the Wnt pathway and inactivating the p53 pathway. In addition, CENPK's positive regulation of β-catenin expression and nuclear translocation could contribute to the activation of Wnt signaling, which was consistent with previous reports [33,34] . Nucleotide excision repair is the principal DNA damage repair mechanism in cancer cells with platinum treatment [35] , and p53 is a key regulator in nucleotide excision repair that mediates cancer cell resistance to platinum [36] . Wnt signaling could also facilitate cancer cell resistance to platinum through other mechanisms, including regulation of cancer stemness, EMT, and DNA damage repair [16,37] . In addition, both Wnt and p53 signaling participate in regulating cancer cell stemness, metastasis, and proliferation [38,39] . In our investigation, we further discovered that CSCs properties, DNA damage repair, EMT, and DNA replication contributed to the promotive role of CENPK on cell resistance to cisplatin and carboplatin, metastasis, and proliferation of cervical cancer through Wnt and p53 signaling.
Previous studies demonstrated that RAD21 and SMC3 belong to the cohesin complex that is responsible for DNA repair, DNA replication, transcription, and p53 signaling [40][41][42] . In cervical cancer, RAD21 upregulates cell proliferation and downregulates p53 signaling [43] . In this study, we found that RAD21 and SMC3 served downstream of Wnt/β-catenin signaling and directly enhanced CENPK transcription. This suggested that CENPK interacted with SOX6 and worked with the RAD21/SMC3 to form a positive feedback circuit, thus facilitating the activation of Wnt/β-catenin signaling and augmenting cervical cancer progression. SOX6 mediated the effect of RAD21/SMC3-enhanced CENPK on cervical cancer progression, demonstrating that our results are in line with the conclusion drawn in a previous study [43] .
We previously found that iRhom2 contributed to cervical cancer progression through activation of Wnt signaling [14] . Here, we further demonstrated that iRhom2 potentiated Wnt signaling and thereby promoted cell stemness, chemoresistance, metastasis, and proliferation in cervical cancer. Considering that CENPK was transcriptionally enhanced by the downstream of Wnt/β-catenin, we assumed that iRhom2 might promote CENPK to affect Wnt and p53 signaling. As expected, iRhom2 upregulated CENPK expression and facilitated the activation of the CENPK/SOX6-β-catenin-RAD21/SMC3 feedback loop. Moreover, we also elucidated a link between Wnt and p53 signaling in cervical cancer, further hinting at their synergistic action in promoting cervical cancer progression.
Bioinformatics and immunohistochemical assays con rmed that CENPK functioned as an oncogene enhancing cervical cancer progression and serving as a prognostic indicator for overall survival and recurrence-free survival of cervical cancer patients. GSEA veri ed that CENPK was involved in regulating Wnt, chemoresistance, DNA damage repair, the cell cycle, and DNA replication signaling in the TCGA database. Importantly, the identi cation of CENPK as an independent and unfavorable factor in clinical samples further supported the results obtained from cervical cancer cells.

Conclusions
Overall, our present study provides a new positive regulatory circuit consisting of CENPK, SOX6, β-catenin, RAD21, and SMC3 in the modulation of cervical cancer progression. Furthermore, our study elucidated a novel mechanism by which CENPK could interact with SOX6 to induce a separation between SOX6 and βcatenin, promote β-catenin expression and nuclear translocation, and facilitate p53 ubiquitination and nuclear export, thereby upregulating Wnt signaling and downregulating p53 signaling. Finally, the activation of the CENPK/SOX6-β-catenin-RAD21/SMC3 feedback loop was facilitated by the iRhom2 and synergistically stimulated cervical cancer progression through Wnt signaling (Fig. 8). This work delineates the roles of speci c signaling pathways in network crosstalk and highlights the biological basis for the clinical use of CENPK both as a prognostic indicator and as a potential target for cervical cancer treatments.

Consent for publication
Not applicable Availability of data and materials The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.       Supplementary Files