SCRIB Is Involved in the Progression of Ovarian Carcinomas in Association with the Factors Linked to Epithelial-to-Mesenchymal Transition and Predicts Shorter Survival of Diagnosed Patients

SCRIB is a polarity protein important in maintaining cell junctions. However, recent reports have raised the possibility that SCRIB might have a role in human cancers. Thus, this study evaluated the roles of SCRIB in ovarian cancers. In 102 human ovarian carcinomas, nuclear expression of SCRIB predicted shorter survival of ovarian carcinoma patients, especially in the patients who received post-operative chemotherapy. In SKOV3 and SNU119 ovarian cancer cells, overexpression of SCRIB stimulated the proliferation and invasion of cells. Knockout of SCRIB inhibited in vivo tumor growth of SKOV3 cells and overexpression of SCRIB promoted tumor growth. Overexpression of SCRIB stimulated epithelial-to-mesenchymal transition by increasing the expression of N-cadherin, snail, TGF-β1, and smad2/3, and decreasing the expression of E-cadherin; the converse was observed with inhibition of SCRIB. In conclusion, this study presents the nuclear expression of SCRIB as a prognostic marker of ovarian carcinomas and suggests that SCRIB is involved in the progression of ovarian carcinomas by stimulating proliferation and epithelial-to-mesenchymal transition-related invasiveness.


Introduction
SCRIB (scribble) is a protein important in maintaining cell polarity and tight junctions of epithelial cells [1]. The role of SCRIB as a component of cell junctions suggests that SCRIB acts as a tumor suppressor because structural and functional alteration of cell polarity induces tumorigenesis [2,3]. Consequently, loss of SCRIB can induce mammary tumorigenesis and promote prostate neoplasia [4,5]. The tumor-suppressive role of SCRIB is also supported by a report that the loss of SCRIB causes loosening of the cell to cell contact and leading epithelial-to-mesenchymal transition (EMT) [1]. However, subsequent reports presented conflicting evidence. Mislocalization of SCRIB from the cytoplasmic membrane to the cytoplasm or nuclei was presented as a tumorigenic phenotype of SCRIB [4,6].

Immunohistochemical Staining in Human Ovarian Carcinoma Tissue
The expression of SCRIB in human ovarian carcinoma tissue was evaluated via immunohistochemical staining of tissue microarray sections. The tissue microarray cores were 5 mm in diameter, and one core per case was arrayed from areas of the highest histologic grade without degenerative or necrotic changes. An antigen retrieval procedure was performed with a microwave oven. The tissue sections were boiled in pH 6.0 antigen retrieval solution (DAKO, Glostrup, Denmark) for 20 min. The primary antibody for SCRIB (1:50, Santa Cruz Biotechnology, Santa Cruz, CA, USA) was used for staining. Immunostained slides were scored with consensus by two pathologists (K.M.K. and K.Y.J.) without clinical information. Based on the subcellular expression patterns of SCRIB in nuclei and cytoplasm in human cancer tissue, immunohistochemical expression of SCRIB was separately analyzed according to its cytoplasmic and nuclear expression patterns. The scoring of immunostained slides was conducted according to the Allred scoring system [20]. The Allred scoring system utilizes a staining intensity score (0; no staining, 1; weak, 2; intermediate, 3; strong) and a staining area score (0; no staining, 1; 1%, 2; 2-10%, 3: 11-33%, 4; 34-66%, 5; 67-100%) [20][21][22]. The final scores were obtained by adding the staining intensity score and staining area score.

Proliferation Assays
A Cell Counting Kit-8 (CCK8, Dojindo, Kumamoto, Japan) assay and a colony-forming assay were used to evaluate the proliferation of cells. The CCK8 assay was performed by growing 3 × 10 3 SKOV3 and 3 × 10 3 SNU119 cells in 96-well plates for 24, 48, and 72 h. At the indicated time, CCK8 was added for two hours, and the absorbance was measured at 560 nm in a microtiter plate reader (Bio-Rad, Richmond, CA, USA). The colony-forming assay was performed by culturing 1 × 10 3 cells per well in 6-well culture plates. The cultures were grown in triplicate and allowed to grow and form colonies for two weeks. The colonies in the culture plates were fixed with methanol and stained with 0.01% crystal violet. Quantification of the number of colonies was performed using Clono-Counter software (software was downloaded from the supplementary electronic material) [23].

In Vitro Wound Healing and Trans-Chamber Migration and Invasion Assays
The migration and invasion activity of ovarian cancer cells were evaluated with a wound-healing assay and trans-chamber migration and invasion assays. The woundhealing assay was performed by making linear scratches with the tip of a 200 µL pipette in 60 mm culture plates that had reached 100% confluency with cells. The microscopic images for wound healing assay were taken immediately following the application of the linear scratches, and again 24 h after making the scratches. For the migration assay, SKOV3 (5 × 10 4 ) and SNU119 (1 × 10 5 ) cells were grown for 48 h in the 24-transwell migration chamber (Corning Life Sciences, Acton, MA, USA). Invasion assay was performed by growing SKOV3 (1 × 10 5 ) and SNU119 (2 × 10 5 ) cells for 48 h in the bioCoat Matrigel Invasion Chamber (Corning 24-Well Plate 8.0 Micron, REF. 354480, ready-to-use, BD Biosciences, San Jose, CA, USA). The migrated and invaded cells on the underside of the insert were counted in five microscopic fields (magnification ×100) per well after being stained with DIFF-Quik staining solutions (Sysmex, Kobe, Japan).

Western Blotting Assay
The cultured cells were washed twice with phosphate buffered saline and lysed using ice-cold PRO-PREP Protein Extraction Solution (iNtRON Biotechnology, Seong- , and anti-actin (Catalog #; sc-376421, Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA) antibodies were used as the primary antibodies. The membranes were developed with an ECL detection system (Amersham Biosciences, Buckinghamshire, UK). The images were obtained by using a luminescent image analyzer (LAS-3000, Fuji Film, Tokyo, Japan) and quantified using ImageJ software (ImageJ, version 1.38e, NIH, Bethesda, MD, USA).

Quantitative Reverse-Transcription Polymerase Chain Reaction (qRT-PCR)
Total RNA was obtained using an RNeasy Mini Kit (Qiagen Sciences, Valencia, CA, USA). After normalization of RNA, 1.5 µg RNA was reverse transcribed to cDNA using an RT-qPCR kit (Takara Biotechnology Co., Ltd., Dalian, China). A quantitative reversetranscription polymerase chain reaction was performed with the Applied Biosystems Prism 7900HT sequence Detection System and SYBR Green PCR Master Mix (Applied Biosystems, Foster City, CA, USA). All experiments were performed in triplicate, and the values were normalized to the expression of the GAPDH reference gene. To calculate qRT-PCR results, the delta-delta-Ct method was used. In this method, we calculate the differences between gene being Tested Experimental (TE) and Housekeeping gene Experimental (HE) [TE-HE], and the differences between gene keeping Tested Control (TC) and Housekeeping gene Control (HC) [TC-HC]. These are the delta Ct values for the Experimental (delta-CTE or ∆CTE) and Control (delta-CTC or ∆CTC) conditions. Finally, we calculate the difference between ∆CTE and ∆CTC (∆CTE−∆CTC) to get the delta-delta-Ct or what we knew double delta-Ct (∆∆Ct) values. All experiments were performed in triplicate. The primer sequences used in the qRT-PCR are listed in Table 1.

In Vivo Tumorigenic Assay
Five-week-old female FoxnN.Cg/c nude mice (Orient Bio, Seongnam, Korea) were used in vivo tumorigenic assay with the approval of the institutional animal care and use committee of Jeonbuk National University (approval number: CBNU 2019-067). The mice were randomly assigned to three groups, with four mice in each group. According to the experimental groups, SKOV3 cells were transfected with empty vector, SCRIB-overexpressing vector, or vector for knock-down of SCRIB. Next, 2 × 10 6 SKOV3 cells were transfected with the indicated vectors via subcutaneous injection into the flank after mixing with Corning matrigel Basement Membrane Matrix (1:1) (Catalog #; 356234, Bedford, MA, USA). Tumor size was measured every week and calculated by the following equation: tumor volume = length × width × height × 0.52 [10,22,24]. The mice were euthanized according to humane end points at six weeks after tumor cell inoculation. The mice were euthanized after anesthetizing with sodium pentobarbital. The size and weight of the tumors were measured. Hematoxylin and eosin staining was performed on the resected tumor tissue, lung, liver, and kidney.

Statistical Analysis
The immunohistochemical positivity of the nuclear expression of SCRIB (nSCRIB) and cytoplasmic expression of SCRIB (cSCRIB) was determined by receiver operating characteristic (ROC) curve analysis [25]. The cut-off points in ROC curve analysis for nSCRIB and cSCRIB were determined to predict the cancer-related death of patients [10,25,26]. The prognosis of ovarian carcinoma patients was evaluated for overall survival (OS) and relapse-free survival (RFS) through June 2015. An event in OS analysis was the death of a patient from ovarian cancer. An event in RFS analysis was the relapse of cancer and death of a patient from ovarian cancer. Univariate and multivariate Cox proportional hazards regression analysis were used to present a hazard ratio (HR) and 95% confidence interval (95% CI). Survival curves were derived from Kaplan-Meier survival analysis. The relationship and difference between factors were analyzed with the Pearson's chi-square test and the Student's t-test, and the values are presented as the mean ± standard deviation. All experiments were done in triplicate and performed three times, with representative data presented. SPSS software (IBM, version 20.0, Armonk, NY, USA) was used in statistical analysis, and a p-value less than 0.05 was considered statistically significant.

The Expression of SCRIB Is Associated with the Progression of Ovarian Carcinomas
Immunohistochemically, SCRIB was expressed in both the cytoplasm and nuclei of ovarian carcinoma cells ( Figure 1A). As shown in Figure 1A, there were cases that did not express SCRIB, and cases that expressed SCRIB predominantly in the cytoplasm, predominantly in the nuclei, or both the cytoplasmic and the nuclei. Therefore, the expression of SCRIB was separately analyzed according to its cytoplasmic or nuclear expression. The cutoff points of immunohistochemical staining scores for the nuclear expression of SCRIB (nSCRIB) and cytoplasmic expression of SCRIB (cSCRIB) were determined with ROC anal-ysis to predict the death of cancer patients ( Figure 1B). The cut-off points were six in both nSCRIB and cSCRIB ( Figure 1B). The cases with a score equal or greater than six for both nSCRIB and cSCRIB were considered positive. With these cut-off points, the factors significantly associated with both nSCRIB and cSCRIB were CA125 level, cancer stage, ascites, bilaterality of cancer, histologic grade, histologic type, and platinum resistance (Table 2).
Immunohistochemically, SCRIB was expressed in both the cytoplasm and nuclei of ovarian carcinoma cells ( Figure 1A). As shown in Figure 1A, there were cases that did not express SCRIB, and cases that expressed SCRIB predominantly in the cytoplasm, predominantly in the nuclei, or both the cytoplasmic and the nuclei. Therefore, the expression of SCRIB was separately analyzed according to its cytoplasmic or nuclear expression. The cut-off points of immunohistochemical staining scores for the nuclear expression of SCRIB (nSCRIB) and cytoplasmic expression of SCRIB (cSCRIB) were determined with ROC analysis to predict the death of cancer patients ( Figure 1B). The cut-off points were six in both nSCRIB and cSCRIB ( Figure 1B). The cases with a score equal or greater than six for both nSCRIB and cSCRIB were considered positive. With these cut-off points, the factors significantly associated with both nSCRIB and cSCRIB were CA125 level, cancer stage, ascites, bilaterality of cancer, histologic grade, histologic type, and platinum resistance ( Table 2).

The Expression of SCRIB Is Associated with Shorter Survival of Ovarian Carcinoma Patients in Univariate Analysis
In univariate survival analysis, age, cancer stage, cancer size, presence of ascites, bilaterality of cancer, CA125 level, histologic grade, nSCRIB, and cSCRIB were significantly associated with OS or RFS (Table 3) ( Figure 2).  In addition, we performed survival analysis according to histologic subtype of ovarian carcinomas. In high-grade serous carcinoma (HGSC), age, cancer stage, and nSCRIB were significantly associated with OS or RFS in univariate analysis (Table 3) ( Figure 3A). In addition, despite the low number of cases, nSCRIB and cSCRIB were significantly associated with survival of mucinous carcinoma patients ( Figure 3B).  In addition, we performed survival analysis according to histologic subtype of ovarian carcinomas. In high-grade serous carcinoma (HGSC), age, cancer stage, and nSCRIB were significantly associated with OS or RFS in univariate analysis (Table 3) ( Figure 3A). In addition, despite the low number of cases, nSCRIB and cSCRIB were significantly associated with survival of mucinous carcinoma patients ( Figure 3B).  In addition, we performed survival analysis according to histologic subtype of ovarian carcinomas. In high-grade serous carcinoma (HGSC), age, cancer stage, and nSCRIB were significantly associated with OS or RFS in univariate analysis (Table 3) ( Figure 3A). In addition, despite the low number of cases, nSCRIB and cSCRIB were significantly associated with survival of mucinous carcinoma patients ( Figure 3B).

Nuclear Expression of SCRIB Predicts Shorter Survival of Ovarian Carcinoma Patients in Multivariate Analysis
The factors significantly associated with OS or RFS in univariate analysis except for CA125 level were included in multivariate analysis. In overall 104 ovarian carcinomas, age, cancer stage, histologic grade, and nSCRIB were associated with OS or RFS (Table 4). The nSCRIB positivity predicted a 2.252-fold (95% CI; 1.205-4.207, p = 0.011) greater risk of death of patients. In HGSC, age and nSCRIB were independent indicators of OS, and tumor stage was an independent indicator of RFS (Table 4). nSCRIB-positive patients had a 2.236-fold (95% CI; 1.097-4.558, p = 0.027) greater risk of death compared with nSCRIBnegative patients with HGSC (Table 4).

The Expression of SCRIB Predicts Shorter Survival of Ovarian Carcinoma Patients Who Received Adjuvant Chemotherapy
In our results, nSCRIB and cSCRIB were significantly associated with platinumresistance and shorter survival of patients. These results suggest that SCRIB might be related to chemoresistance. Therefore, we further analyzed for survival in the sub-population of ovarian carcinoma patients who received adjuvant chemotherapy. In eighty patients who received adjuvant chemotherapy, both nSCRIB and cSCRIB were significantly associated with OS and RFS in univariate analysis (Table 5) (Figure 4). In multivariate analysis, age, cancer stage, bilaterality, and nSCRIB expression were independent indicators of OS (Table 5). Higher tumor stage and nSCRIB positivity predicted shorter RFS ( Table 5). Positivity of nSCRIB predicted a 2.363-fold (95% CI; 1.149-4.859, p = 0.019) greater risk of death and a 2.530-fold (95% CI; 1.313-4.875, p = 0.006) greater risk of relapse or death of patients (Table 5).

SCRIB Is Associated with the Proliferation and Invasion of Ovarian Cancer Cells
In ovarian cancer cells, we evaluated the effect of SCRIB expression on the proliferation and invasiveness of cells. The proliferation of SKOV3 and SNU119 ovarian cancer cells were significantly inhibited by knockout of SCRIB and significantly stimulated by overexpression of SCRIB as evidenced by CCK8 proliferation assay ( Figure 5A) and colony forming assay ( Figure 5B). In addition, the migration and invasion ability of SKOV3 and SNU119 ovarian cancer cells were significantly inhibited by knockout of SCRIB and

SCRIB Is Associated with the Proliferation and Invasion of Ovarian Cancer Cells
In ovarian cancer cells, we evaluated the effect of SCRIB expression on the proliferation and invasiveness of cells. The proliferation of SKOV3 and SNU119 ovarian cancer cells were significantly inhibited by knockout of SCRIB and significantly stimulated by overexpression of SCRIB as evidenced by CCK8 proliferation assay ( Figure 5A) and colony forming assay ( Figure 5B). In addition, the migration and invasion ability of SKOV3 and SNU119 ovarian cancer cells were significantly inhibited by knockout of SCRIB and were significantly stimulated by overexpression of SCRIB as evidence by wound-healing, migration, and invasion assay ( Figure 6). Furthermore, overexpression of SCRIB significantly increased in vivo growth of SKOV3 cells, and knockout of SCRIB significantly decreased in vivo tumor growth compared with controls ( Figure 7). However, histologically, there were no identifiable metastatic lesions in the lung, liver, or kidney of the mice in the three experimental groups. In addition, there were no significant differences in the shape of implanted cancer cells between experimental groups ( Figure 7C). were significantly stimulated by overexpression of SCRIB as evidence by wound-healing, migration, and invasion assay ( Figure 6). Furthermore, overexpression of SCRIB significantly increased in vivo growth of SKOV3 cells, and knockout of SCRIB significantly decreased in vivo tumor growth compared with controls ( Figure 7). However, histologically, there were no identifiable metastatic lesions in the lung, liver, or kidney of the mice in the three experimental groups. In addition, there were no significant differences in the shape of implanted cancer cells between experimental groups ( Figure 7C).

SCRIB Is Associated with a Change in Expression of Factors Linked to the EMT of Cancer Cells
In malignant epithelial tumors, the molecules associated with EMT are closely associated with cancer progression and resistance to therapy. Therefore, we evaluated the ex-

SCRIB Is Associated with a Change in Expression of Factors Linked to the EMT of Cancer Cells
In malignant epithelial tumors, the molecules associated with EMT are closely associated with cancer progression and resistance to therapy. Therefore, we evaluated the expression of proteins and mRNA transcripts associated with EMT. The protein and mRNA expression of cyclin D1, N-cadherin, snail, TGF-β1, and smad2/3 decreased, and the expression of E-cadherin increased with knockout of SCRIB in both SKOV3 and SNU119 cells (Figure 8). Overexpression of SCRIB increased the protein and mRNA expression of cyclin D1, N-cadherin, snail, TGF-β1, and smad2/3 and decreased expression of E-cadherin ( Figure 8). However, the expression of β-catenin mRNA was unchanged with knockout or overexpression of SCRIB despite the change in β-catenin protein expression (Figure 8).

Discussion
In this study, the expression of SCRIB was associated with advanced clinicopathological factors of ovarian carcinomas, such as elevation of CA125 level, high cancer stage, and higher histologic grade. In addition, both nSCRIB and cSCRIB positivity were inde- Figure 8. Western blot and quantitative reverse-transcription polymerase chain reaction after knockout or overexpression of SCRIB in ovarian carcinoma cells. (A) Western blot was performed for SCRIB, β-catenin, active β-catenin, cyclin D1, E-cadherin, N-cadherin, snail, TGF-β1, smad2/3, and actin after knockout or overexpression of SCRIB in SKOV3 and SNU119 ovarian carcinoma cells. The western bands were quantified using ImageJ software and the values are indicated below the bands. (B) Quantitative reverse-transcription polymerase chain reaction was performed for SCRIB, β-catenin, cyclin D1, E-cadherin, N-cadherin, snail, TGF-β1, smad2, and smad3 after knockout or overexpression of SCRIB in SKOV3 and SNU119 ovarian carcinoma cells. EVs; the cells transfected with both empty vector for knockout and empty vector for overexpression of SCRIB. ** p < 0.001.

Discussion
In this study, the expression of SCRIB was associated with advanced clinicopathological factors of ovarian carcinomas, such as elevation of CA125 level, high cancer stage, and higher histologic grade. In addition, both nSCRIB and cSCRIB positivity were independent indicators of shorter OS and RFS. Furthermore, nSCRIB positivity predicted shorter OS of HGSC, the most common histologic type of ovarian cancer with a relatively unfavorable prognosis. These findings suggest that SCRIB is involved in the progression of ovarian carcinomas and might be used as a prognostic marker of ovarian carcinoma patients. Consistently, nSCRIB was an independent predictor of survival of gastric carcinoma patients [10]. In addition, higher expression of SCRIB mRNA was associated with poor prognosis of breast cancers [4] and shorter survival of hepatocellular carcinoma patients [7]. However, controversially, higher immunohistochemical expression of SCRIB was associated with favorable prognosis of lung cancer patients [12]. Therefore, because of controversial reports on the prognostic impact of SCRIB expression in human cancers, further study is needed in various types of human cancers.
Regarding the subcellular localization of SCRIB, its expression was expected in the basolateral side of the cytoplasmic membrane because SCRIB is involved in maintaining tight junctions [1][2][3]. However, the expression of SCRIB was observed in the cytoplasm and/or nuclei of human cancer cells [4,10,27]. The subcellular localization of SCRIB was different between non-neoplastic cells and cancer cells [7,27]. In hepatocytes of healthy livers, SCRIB is localized in the cytoplasmic membrane [7]. However, cytoplasmic localization of SCRIB was detected in 69% (22/32 cases) of hepatocellular carcinomas [7]. Moreover, enrichment of cytoplasmic SCRIB was associated with the overall increase of intracellular SCRIB by inducing overexpression of SCRIB [7]. In a study with human and mouse liver cancer models, cytoplasmic and nuclear expression of SCRIB was more prevalent in liver cancer tissue, and non-neoplastic liver tissue showed SCRIB expressed solely on the cytoplasmic membrane [27]. The cytoplasmic and nuclear expression of SCRIB in liver cancer tissue was supported by Western blots performed with subcellular fractionated protein lysates [27]. Therefore, higher expression of SCRIB supports its translocation to the cytoplasm and nuclei and might be involved in the tumorigenesis of epithelial cells [7]. Furthermore, in our results, nSCRIB is an independent indicator of shorter survival of ovarian carcinoma patients. These results suggest that nuclear SCRIB is important in cancer progression. Thus, the prognostic significance of nuclear localization of SCRIB might be related to the association between SCRIB and nuclear proteins involved in cancer progression, such as snail and β-catenin [8,10]. In gastric carcinomas, SCRIB stabilizes β-catenin by binding it and consequently activating TCF/LEF transcription [10]. In our results, the expression of proteins of β-catenin were decreased with knockout of SCRIB and increased with overexpression of SCRIB. In addition, as a consequence of the SCRIB-mediated increase of β-catenin protein, the expression of cyclin D1 and snail, the down-stream molecules of Wnt/β-catenin signaling, were increased with overexpression of SCRIB. Supportively, mislocalization of SCRIB from the cytoplasmic membrane to cytoplasm promoted mammary tumorigenesis by activating the Akt/mTOR pathway [4]. Together, these results suggest that subcellular localization of SCRIB, especially in nuclei and cytoplasm, might be important in the tumorigenic role of SCRIB, which warrants additional study.
Loss of cell polarity protein might serve to induce tumors by promoting the uncontrolled proliferation of cells [28,29]. Therefore, as a component of polarity protein, SCRIB has been suggested as a tumor suppressor, and loss of SCRIB induced tumorigenesis in MYC-induced transformed epithelial cells [6]. However, in addition to the loss of SCRIB, mislocalization of SCRIB also induced mammary tumorigenesis [6]. Furthermore, cytoplasmic and nuclear localization of SCRIB, induced by overexpression of SCRIB, stimulated the proliferation of liver cancer cells [27]. Consistently, in our results, SCRIB positively regulated the proliferation of ovarian cells. In line with these results in ovarian cancers, SCRIB stimulated in vitro proliferation and in vivo growth of gastric cancer cells by acti-vating the Wnt/β-catenin pathway [10]. Based on the association between SCRIB and the Wnt/β-catenin pathway, recent reports have shown the molecular relationship between MYC, FAM83H, SCRIB, and the Wnt/β-catenin pathway in human cancers [10,11,30]. MYC transcriptionally controls FAM83H expression, and the expression of SCRIB was regulated by FAM83H [10,30]. Thereafter, FAM83H and SCRIB form a complex to stabilize β-catenin [10]. These molecular relationships might be supported by the report that SCRIB is involved in the initiation of MYC-driven lymphoma [11]. Knock-out of SCRIB delayed expansion of B cells in the development of MYC-mediated lymphoma development [11]. Therefore, these results suggest there is a close relationship between MYC, FAM83H, SCRIB, and the Wnt/β-catenin pathway in cancer progression. However, there are conflicting reports that SCRIB acts as a tumor suppressor. Conditional knock-out of SCRIB in the epidermis enhanced epidermal carcinogenesis [31] and SCRIB repressed Wnt signaling in HEK293T cells [32]. Therefore, further study is needed to clarify the role of SCRIB in human cancer.
In our results, SCRIB stimulated invasiveness and the factors linked to EMT of ovarian cancer cells. EMT is an important process in cancer progression and is involved in cancer invasiveness, and the characteristic phenotype of EMT is a loss of E-cadherin in epithelial cells and activation of the snail, TGF-β1, and N-cadherin [9,14]. In our results, SCRIB induced an EMT phenotype; decrease of E-cadherin and increase of N-cadherin, snail, TGF-β1, and smad2/3 with overexpression of SCRIB. Interestingly, the expression of the protein of smad2/3 increased with overexpression of SCRIB. When considering the role of TGF-β1 in EMT as an activator of smad2/3 by inducing phosphorylation of smad2/3 [33], the increase in total protein amount of smad2/3 in SCRIB-overexpressing cells is questionable. However, in addition to the increase in smad2/3 protein, mRNA of smad2/3 also increased with overexpression of SCRIB. Therefore, higher expression of smad2/3 protein in SCRIBoverexpressing cells might be related to increased transcription of smad2/3. Consistently, when we searched the GEPIA public database (http://gepia.cancer-pku.cn. accessed on 21 February 2021) [34], there was a significant correlation between the expression of TGF-β1 mRNA and smad2 mRNA (Pearson's R = 0.33, p < 0.001) and smad3 mRNA (Pearson's R = 0.31, p < 0.001) in ovarian cancers. Furthermore, mRNA expression of SCRIB was also significantly correlated with smad2 mRNA (Pearson's R = 0.12, p = 0.013) and smad3 mRNA (Pearson's R = 0.17, p < 0.001) in ovarian cancers. In addition, SCRIB induced expression of factors linked to EMT phenotype and increased invasion of gastric cancer cells [10] and cytoplasmic SCRIB increased the invasion activity of liver cancer cells [7]. Therefore, these results suggest that SCRIB is involved in the progression of ovarian cancer by inducing the expression of factors linked to EMT. However, SCRIB is a component of the Scribble polarity complex and important in maintaining the apicobasal cell polarity of epithelial cells. Therefore, based on the role of SCRIB as a component of tight junctions, the maintenance of SCRIB is expected as a tumor suppressor or suppressor of EMT [35,36]. In corneal epithelium, conditional deletion of SCRIB induced EMT phenotype; elongation of the shape of epithelial cells and loss of E-cadherin and upregulated expression of snail TGF-β, and smad3/4 [35]. Knock-down of SCRIB in cancer-associated fibroblasts increased invasiveness of both tumor-associated fibroblasts and co-cultured Lewis lung cancer cells [36]. However, in contrast, recent reports have shown that the role of SCRIB in human cancer is closely related to its subcellular localization instead of its expression level. In a study using tissues from normal endometrium, endometriosis, and endometrial adenocarcinoma, membranous expression of SCRIB in normal endometrium changed to be localized in the cytoplasm and nuclei in endometriosis and endometrial carcinomas [37]. In this respect, the loss of SCRIB from the cytoplasmic membrane and its aberrant translocation to the cytoplasm and/or nucleus might serve in the development and progression of cancers. When SCRIB was overexpressed, SCRIB moved to the cytoplasm, and that was associated with cancer development. In addition, transfection of mutant SCRIB caused a cytoplasmic enriched induced EMT phenotype of hepatocellular carcinoma cells and increased invasiveness of cancer cells [7]. Therefore, SCRIB mediated EMT might be related to the subcellular expression pattern of SCRIB, and further study is needed to focus on this point.
One of the interesting findings of our study is that SCRIB might be involved in chemoresistance of ovarian cancer cells. In the comparison of clinicopathologic factors and SCRIB expression in ovarian carcinomas, nSCRIB and cSCRIB positivity were significantly associated with platinum-resistance. Furthermore, in subpopulation survival analysis of patients who received adjuvant chemotherapy, nSCRIB positivity was an independent indicator of shorter survival. In addition, SCRIB expression was closely associated with the expression of EMT-related molecules. The Wnt, TGF-β, Notch, and mitogenic growth factors related pathways are known as the signaling pathways that activate EMT [8,9,14]. Furthermore, EMT mediates chemoresistance by inducing stem-cell-like properties in cancer cells, inducing resistance to drug-induced apoptosis, and modulating tumor microenvironment on cancer-associated fibroblasts and immune cells [14]. In addition, the EMT phenotype was associated with chemoresistance and progression of ovarian cancer cells [15,38]. Therefore, when considering the effect of EMT on cancer drug resistance [8,14,16], SCRIB might be involved in chemoresistance during the treatment of ovarian carcinomas by activating EMT pathways. However, there are controversial reports on the role of SCRIB on EMT and chemoresistance. In corneal epithelium, the EMT phenotype occurred in SCRIB deficient cells [35]. In non-small cell lung cancer cells, SCRIB supported the anti-cancer effects of cisplatin by increasing cisplatin-mediated generation of reactive oxygen species and knock-down of SCRIB induced cisplatin resistance [12]. Therefore, the role of SCRIB in tumorigenesis and cancer progression might be different according to the type of tumor and stage of progression of cancer, and further study is needed.

Conclusions
In this study, we demonstrated that SCRIB stimulates proliferation and invasiveness of ovarian cancer cells in conjunction with activation of the factors linked to the EMT pathway. Moreover, we presented SCRIB expression as a potential prognostic indicator of ovarian carcinoma patients, especially in patients who received chemotherapy. Therefore, our results suggest that SCRIB is involved in cancer progression and chemoresistance by activating some EMT characteristics. Furthermore, our results suggest that the SCRIB-EMT pathway might be a new therapeutic target for the ovarian carcinomas which highly express SCRIB and are refractory to the conventional anti-cancer therapies.

Informed Consent Statement:
The IRB approval (CUH 2019-09-034) contained a waiver for written informed consent based on the retrospective and anonymous character of the study.

Data Availability Statement:
The datasets used in the current study are available from the corresponding author upon reasonable request.