Overexpression of RIPK4 Predicts Poor Prognosis and Promotes Metastasis in Ovarian Cancer

This study was conducted to evaluate the prognostic value of receptor-interacting protein kinase 4 (RIPK4) in ovarian cancer (OC) and its role in tumorigenesis. RNA expression and the corresponding clinical data were obtained from The Cancer Genome Atlas (TCGA) and Genotype-Tissue Expression (GTEx) databases. The relationship between clinical-pathological characteristics and RIPK4 expression was analyzed using the Wilcoxon signed-rank test and logistic regression. The Cox regression and the Kaplan-Meier method were used to evaluate the relationship between clinicopathological features and overall survival (OS). Gene set enrichment analysis (GSEA) was performed using Molecular Signatures Database. Scratch assay, transwell assay, and cell transfection were used to verify the function of RIPK4. Overexpression of RIPK4 was associated with the stage of OC and distant metastasis. Survival analysis revealed that patients with OC and higher expression of RIPK4 had a poorer prognosis. Univariate and multivariate analyses indicated that high expression of RIPK4 was associated with poor OS, as well as age and stage of OC. The areas under the curve (AUC) at 1, 4, and 8 years were 0.737, 0.634, and 0.669, respectively, according to the established OS prediction model. GSEA revealed that adherens junction, cadherin binding, and Wnt signaling pathway were enriched in the high RIPK4 expression group. Cell transfection confirmed RIPK4 was involved in the Wnt signaling pathway. RIPK4 can act as a potential prognostic molecular marker for poor survival in OC. Moreover, RIPK4 is associated with tumor metastasis and implicated in the regulation of the Wnt signaling pathway.


Introduction
Ovarian cancer (OC) is the most fatal malignancy of the female reproductive system with multiple histologic subtypes that are characterized by unique metastatic behaviors [1]. The American Cancer Society estimated that there will be approximately 21,750 newly diagnosed ovarian carcinoma cases and more than half of the women will die from ovarian cancer disease in 2020 in the United States [2]. However, the absence of early specific symptoms and effective screening strategies reduces survival above 5 years in a few women after diagnosis [3]. Although the serum level of CA125 is a widely used biomarker, it does not effectively evaluate prognosis and direct treatment [4]. Biomarker research is a major focus of therapeutic studies on OC.
The Wnt/β-catenin signaling pathway, as a family of proteins, participates in several vital cellular functions such as stem cell regeneration and organogenesis [5]. Hypo or hyperactivation of the Wnt/β-catenin signaling cascade has been associated with human diseases such as cancers [6]. Mutated components of the canonical Wnt signaling are drivers of many cancers [7], including OC [8]. The Wnt/β-catenin pathway has been implicated in the pathogenesis of ovarian carcinomas by regulating proliferation, migration, and stemness of ovarian cancer cells [9,10].
Receptor-interacting protein kinase 4 (RIPK4) is a serinethreonine kinase that is involved in NF-κB and JNK signaling pathways and is processed during apoptosis [11]. RIPK4 knockdown in osteosarcoma can inhibit epithelial-mesenchymal transition by inactivating the Wnt/β-catenin signaling pathway [12]. The overexpression of RIPK4 promotes cell migration and invasion in pancreatic cancer [13]. Furthermore, researchers have revealed a simple but effective method of treating bladder cancer using an active small interfering RNA that targets the silencing of RIPK4 expression [14]. Previous studies have revealed that the expression of RIPK4 in ovarian tumor tissues is higher than that of noncancerous ovarian tissues [15]. However, the correlation between RIPK4 and the prognosis of OC is still unclear.
Therefore, the primary aim of this study was to determine the prognostic value of RIPK4 expression in patients with OC based on data acquired from The Cancer Genome Atlas (TCGA) and the Genotype-Tissue Expression (GTEx) databases. Gene set enrichment analysis (GSEA) and in vitro cellular experiments were performed to establish the potential biological pathways involved in ovarian tumorigenesis that are associated with the RIPK4 regulatory network.
The present study demonstrated the adverse impacts of RIPK4 overexpression on the prognosis of human ovarian cancer and the pivotal role of RIPK4 in promoting tumor metastasis. Gene set enrichment analysis and experimental outcomes indicated that the Wnt/β-catenin signaling pathway is associated with the high expression level of RIPK4.

Materials and Methods
2.1. Data Acquisition. RNA sequencing data for normal ovarian samples (88 cases) were downloaded from the GTEx database [16]. RNA-Seq data and the corresponding clinical information for patients with ovarian cancer were obtained from TCGA data portal (https://portal.gdc.cancer.gov/) in April 2020. Clinical and demographic data for 587 patients with OC were extracted for further analysis. The inclusion criteria of the study were patients underwent hysterectomy and bilateral adnexectomy and the histology type was serous adenocarcinoma. The patients who underwent other types of surgery or without RNA sequencing dataset were excluded from this study.

Screening for Differentially Expressed
Genes. The differentially expressed mRNAs were selected for subsequent analysis using the |log 2 fold change | >2 with an adjusted false discovery rate of P < 0:05 and >10 -50 as the criteria. The volcano and heat map plots were visualized separately using the ggplot2 and pheatmap R packages. The Limma and Beeswarm R packages were used to visualize differential expression of RIPK4 between normal ovarian tissues and ovarian carcinoma tissues.

Enrichment
Analysis. GSEA is a statistical algorithm used to determine whether members of a gene set are correlated with the phenotypic class distinction [17]. Each gene set analysis was performed 1000 times, and gene set enrichment analysis was conducted to generate significant prognostic genes with phenotype labels "high-RIPK4" vs. "low-RIPK4" using the Molecular Signatures Database (MSigDB, https://data.broadinstitute.org/gsea/msigdb/collections.jsp). The enriched results considered to be statistically significant in each group were arranged according to nominal P value of <0.05 and |normalized enrichment score | >1.

Cell
Culture. The human OC cell line SKOV3 was obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA). The human OC cell line HO8910 was bought from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China). Both cell lines were maintained in DMEM (Gibco, Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 10% (v/v) fetal bovine serum (FBS) (Gibco) and 1% (w/v) penicillin-streptomycin (Gibco), in 5% CO 2 at 37°C.
2.6. Wound Healing Assay. The cells seeded into 6-well plates. When cell confluency reached above 90%, a 200 μl pipette tip was used to make parallel linear scratches. Then, the plates were washed using phosphate-buffered saline (PBS) gently. After that, cells were incubated in serum-free DMEM medium at 37°C for 24 hours. The wound areas were measured under an optical microscope at 0 h and 24 h, respectively.  BioMed Research International with secondary antibodies (1 : 5000; Proteintech) for an hour at room temperature. Finally, the bands were detected by the enhanced chemiluminescence (ECL) Plus kit (Beyotime) and imaged on X-ray film.
2.9. Statistical Analysis. R platform (v.3.6.1) was used to perform all statistical analyses. The median value of RIPK4 expression was considered as the cut-off value. The Wilcoxon signed-rank test and logistic regression were used to analyze the correlation between clinical-pathological characteristics and the expression level of RIPK4. The relationship between clinicopathological features and overall survival (OS) was analyzed using Cox regression and the Kaplan-Meier method. The areas under the curve (AUC) at 1, 4, and 8 years were generated from the time-dependent receiver operating characteristic (ROC) curve using the "survivalROC" package.

Differentially Expressed
Genes. Analysis of differentially expressed genes using R packages revealed that the mRNA expression levels of 1192 genes were downregulated and 1017 genes were upregulated in patients with OC from TCGA database in comparison with normal ovarian samples from the GTEx database (Figure 1(a)). We selected top 50 relevant genes from these differentially expressed genes for further analysis (Figure 1(b)). Excluding the genes that have been reported and studied, we found RIPK4 that is not fully investigated in ovarian carcinoma according to the existing researches. The expression of RIPK4 in ovarian tumor tissues was considerably higher than that in normal ovarian tissues (Figure 1(c)).  b)). The patients with higher RIPK4 expression had more aggressive clinical stage and exited distant metastasis lesions more commonly. The logistic regression was used to perform univariate analysis, and results indicated that upregulation of RIPK4 expression was associated with the stage (OR = 3:08 for I+II vs. IV (P = 0:026) and OR = 2:08 for III vs. IV (P = 0:015)) and distant metastasis (OR = 2:06 for no vs. yes (P = 0:016)) in OC (Table 1), while there was no significant relation between the expression of RIPK4 and age and race (P > 0:05). The wound healing and transwell migration and invasion assay were performed to detect the role of RIPK4 in OC metastasis. The wound healing assay indicated that the ability of OC cells' migration was weakened after inhibiting RIPK4 expression (Figures 2(c) and 2(d)). The transwell assay results showed that the downregulation of RIPK4 decreased the migration and invasion abilities of SKOV3 and HO8910 cells, compared with the control groups, respectively (Figures 2(e) and 2(f)). These results indicate that increased expression of RIPK4 is a risk factor for the development and metastasis of OC.

Survival Analysis.
Patients with OC and low expression of RIPK4 had a better prognosis than patients with high expression of RIPK4 using Cox regression and the Kaplan-Meier method (P < 0:001) (Figure 3(a)).

Prediction Model for Overall Survival.
Results of the survival analysis indicated that age, stage, and RIPK4 expression can serve as independent prognostic factors for unfavorable OS in OC. Therefore, a multivariate Cox model was used to establish the OS prediction model. The risk scores for all patients selected from TCGA database were calculated using the following formula: risk score = age × 0:026 + stage × 0:295 + RIPK4 × 0:182. Survival analysis was performed using a median value (1.02) of the risk scores as the cut-off point for dividing patients into two groups (high-risk and low-risk groups). As shown in Figure 3( (Figure 3(c)). In addition, the age, stage, and RIPK4 expression of every patient in the high-and low-risk  ADAMTS10  CNRIP1  IL11RA  MYLK  HERC2P2  N4BP2L2  DDHD2  RNA5-8SN2  ADIRF  ABCA8   PDE1B   GNB3   CHRD  WDR73  CES5AP1  PTPRD  ABCA9  HAND2-AS1  OR2A20P  SLC26A10  CBLC  HSPB1P1  TROAP  TYMS  FOXM1  ECT2  CD52  DTX3L  RAB3D  BMPR1B-DT  KRT7-AS  RIPK4  TTC30B  IQGAP3  HMMR  ADAMTS1   TCF25  ZNF207   WDR13   CENPT  SIDT2  TTC17  NDUFS6  BGN  HIGD2A  ELF3  UCP2  PODXL2 MEST PODXL Type Type Normal Tumor 3.6. Functional Enrichment. GSEA was performed using c2 (c2.cp.kegg.v7.1.symbols.gmt) and c5 (c5.all.v7.1.symbols.gmt) with the Molecular Signatures Database gene sets as the reference gene sets, and we focused on the top 10 relevant gene sets in each collection. Citing c5 collection to perform the gene ontology (GO) analysis (Figure 4(a)), it was established that varied expression of RIPK4 was strongly correlated with adherens junction and cadherin binding. The analysis of c2 collection using the Kyoto Encyclopedia of Genes and Genomes (KEGG) database indicated that the Wnt/β-catenin signaling pathway was enriched under high expression levels of RIPK4 (Figure 4(b)). In order to verify the result of KEGG analysis, we selected β-catenin for further investigation, which is the core molecule of the Wnt/βcatenin signaling pathway. As Figures 4(c) and 4(d) show, we used si-RNA to inhibit the expression of RIPK4 and RIPK4 expression was descended efficiently compared with the control group. The western blot assay revealed that decreasing the expression of RIPK4 could negatively affect β-catenin expression in SKOV3 and HO8910 cells.

Discussion
The incidence and mortality of ovarian cancer have reduced over the last few decades, but the deaths caused by OC are substantially higher than the deaths caused by other gynecologic malignancies [3]. Current treatments of OC primarily depend on debulking surgery, platinum-based chemotherapy, antiangiogenic agents, and maintenance therapy of poly (ADP-ribose) polymerase inhibitors. Nonetheless, women suffering from OC still face the challenges of recurrence and drug resistance [18]. Thus, it will be imperative to identify novel biological markers and therapeutic targets that can be used in the treatment of OC.
Studies conducted in the recent past have reported that RIPK4 is involved in the processes of tumor initiation, development, and metastasis. RIPK4 can combine with EZH2, which is a polycomb group of genes that regulates epidermal-mesenchymal transformation and angiogenesis via suppressing tumor suppressor genes, to enhance lymph node metastasis in cervical cancer [19]. Besides, RIPK4 is associated with STAT3 signaling that is activated in tumor metastasis by regulating extracellular matrix remodeling enzymes in lung adenocarcinoma [20]. Moreover, in nasopharyngeal carcinoma, RIPK4 can activate the NF-κB signaling pathway by interacting with the IKK complex to accelerate tumorigenesis and metastasis [21]. It has been reported that RIPK4 plays a prognostic role in a variety of tumors. For example, a study about squamous cell carcinoma of the skin shows that RIPK4 is helpful for the discovery of cancer pathogenesis and for potential treatments [22]. In tongue squamous cell carcinoma, RIPK4 is confirmed to act as a promising biomarker for prognosis and treatment [23]. Furthermore, RIPK4 can also function as a potential diagnostic and independent prognostic biomarker for cervical squamous cell carcinoma [24]. In the case of OC, the expression of RIPK4 was high in tumor tissues and RIPK4 was associated with tumor growth [15]. However, the potential role of RIPK4 in the prognosis of OC still remains unknown. Therefore, the present study was aimed at determining the role of RIPK4 in the prognosis of OC.
The present study indicated that RIPK4 expression in ovarian tumor tissues based on TCGA database RNA sequencing data was remarkably higher than in normal ovarian tissues with reference to data obtained from the GTEx database using bioinformatic analysis. In addition, RIPK4 overexpression was positively correlated with advanced clinical stage, distant metastasis, and poor prognosis of patients with OC. In vitro experiments confirmed that the high expression of RIPK4 was associated with tumor metastasis. The results of GSEA revealed that adherens junction, cadherin    BioMed Research International binding, and Wnt signaling pathway were enriched in patients with phenotypes of high RIPK4 expression. Meanwhile, it was clearly seen that the expression of β-catenin which is known as the core molecule in the Wnt signaling pathway significantly decreased when RIPK4 expression was reduced. This indicates that RIPK4 has a crucial role in the evaluation of development and prognosis of OC. RIPK4 is a novel member of the receptor-interacting protein kinase family, and its overexpression causes activation of the NF-κB signaling pathway [11]. The upregulation of RIPK4 promotes activation of the NF-κB signaling pathway, upregulates VEGF, and ultimately promotes tumor cell aggressiveness [25]. Conversely, silencing of RIPK4 inhibits the Wnt signaling pathway thus disrupting the pathological process [26]. The present study has demonstrated that upregulation of RIPK4 expression is associated with cellular adhesion and involvement in the Wnt signaling pathways in OC. Nevertheless, a limitation of this study is that we did not elucidate the specific mode of regulation; thus, further studies are required to evaluate the role of RIPK4 in OC.
Notably, we developed a valuable OS assessment model. In this study, we just needed to know the mRNA expression    9 BioMed Research International level of RIPK4 and the age and stage of patients with OC and then used the formula, risk score = age × 0:026 + stage × 0:295 + RIPK4 × 0:182, to calculate the risk scores, which were subsequently used to evaluate the prognosis of patients with OC. The prediction model for OS generated by this study will be highly convenient and easily accessible if it is successfully applied in clinical diagnosis and treatment. However, the prediction model has just been verified in this study and the study has some limitations undeniably. More data from multiple cohorts and more evidence from laboratory are needed to verify the prognostic role of RIPK4 in ovarian cancer and the regulatory pathway which RIPK4 is involved in. In addition, it still needs more clinical studies to support the effectiveness of the prediction model for clinical application in the future.

Conclusions
RIPK4 can function as a potential independent risk factor for overall survival and biomarker in evaluating the prognosis of ovarian cancer. RIPK4 can regulate the Wnt signaling pathway thereby having an effect on ovarian carcinoma. However, further investigations are needed to elucidate the specific molecular mechanisms of RIPK4 in ovarian cancer.

Data Availability
Data can be obtained by contacting the corresponding author.

Conflicts of Interest
The authors declare that they have no conflicts of interest.