Grainyhead‐like 2 (GRHL2) regulates epithelial plasticity in pancreatic cancer progression

Abstract The epithelial‐mesenchymal transition (EMT) and mesenchymal‐epithelial transition (MET) contribute to cancer metastasis of pancreatic ductal adenocarcinoma (PDAC). We explored the role of grainyhead‐like 2 (GRHL2), a suppressor of EMT, in the progression of PDAC. Expressions of GRHL2 were assessed using surgically resected PDAC tissues by immunohistochemistry analysis, and in vitro using human and mouse PDAC cells. Effects on epithelial plasticity and stemness of GRHL2 were examined in vitro using liver metastatic PDAC cells (CFPAC‐1) with GRHL2 knockdown by specific siRNAs. GRHL2 has a significantly positive correlation with E‐cadherin and CD133 in 155 resected human primary PDAC tissues. GRHL2 is highly expressed in liver metastatic cells than in primary invasive cells of both human and mouse PDAC, accompanied by a positive correlation with E‐cadherin expression. GRHL2 knockdown CFPAC‐1 cells demonstrated morphological changes into mesenchymal appearances and reduced proliferation through EMT. Notably, knockdown studies followed by flow cytometry analysis for a subpopulation of CD133+ showed that GRHL2 facilitates CFPAC‐1 cells to maintain stem‐like characters including self‐renewal capacity and anoikis resistance. GRHL2 regulates epithelial plasticity along with stemness in PDAC, both of which are crucial for metastasis, implicating the possibility of GRHL2 as a therapeutic target for PDAC liver metastasis.


Introduction
Pancreatic ductal adenocarcinoma (PDAC) is the fourth leading cause of cancer-related deaths [1]. Despite the advances in surgery and chemotherapy for PDAC, 5-year overall survival even for patients with surgically resected PDAC remains about 20-25%, accompanied by high rate of recurrent systemic disease [2]. This trait of easily becoming an extensive metastatic disease can be partially explained with the recent finding that hematogenous spread of circulating tumor cells from the pancreas appears to be an early event in PDAC progression [3]. Specifically, cells within histological "preinvasive" pancreatic intraepithelial neoplasia (PanIN) can circulate into the bloodstream, and seed the liver in genetically engineered mouse models.
Epithelial-mesenchymal transition (EMT) plays an important role in tumor progression and metastasis in diverse cancers including PDAC [4]. A variety of intrinsic and extrinsic factors initiate and propagate the signaling during EMT or the reverse process, the

ORIGINAL RESEARCH
Grainyhead-like2(GRHL2)regulatesepithelialplasticityin pancreaticcancerprogression mesenchymal-epithelial transition (MET), leading to alterations in the expression of several transcription factors [5]. However, the molecular mechanisms that regulate EMT and MET in cancer progression remain unclear. Grainyhead-like 2 (GRHL2) is a member of grainyheadlike transcription family, which plays a fundamental role in epidermal integrity, embryonic neural tube closure, and wound healing processes [6]. Emerging evidence has revealed that GRHL2 is a novel proto-oncogene that regulates epithelial plasticity by suppressing EMT in several tumor types [7,8]. Recent studies focused on demonstrating its crucial role in regulating EMT and MET in cancer cells [9,10]. GRHL2 inhibits TGF-β mediated activation of Smad2/3, which leads to the loss of ZEB1 expression and accompanying upregulation of miR-200 expression [11]. It is also revealed that GRHL2 inhibits transactivation of the ZEB1 promoter mediated by the homeodomain proteins Six1, LBX1, and HoxA5. ZEB1 reciprocally repressed GRHL2 expression through a direct interaction with the GRHL2 promoter [7]. Furthermore, Yang et al. have shown the expressional association and clinical relevance of six genes (CDH2, FN1, CITED2, CTNNB1, and CTNNA3) identified as GRHL2-related genes together with GRHL2 for breast cancer metastasis [12]. In contrast, GRHL2 is considered as a tumor suppressor in gastric cancer, cervical cancer, clear cell renal cell carcinoma, and sarcoma [13,14]. Thus, the functional roles and clinical impact of GRHL2 vary with cancer type, and its expression and effect in pancreatic carcinogenesis has not yet been investigated.
Herein, we explored and defined novel functional roles for GRHL2 in the regulation of EMT and MET during cancer progression in PDAC cells. GRHL2 suppresses EMT and drives invasive PDAC cells to alter epithelial phenotype with self-renewal capacity to induce metastatic colonization. This study contributes to a better understanding of the functional roles of GRHL2 in PDAC, which might prove to be a novel therapeutic target for PDAC.

QuantitativeRT-PCR
Quantitative RT-PCR was performed with SYBR Green in a real-time PCR system (Applied Biosystems, Foster City, CA, USA) for total RNA purified with RNeasy Mini Kit (Qiagen Inc., Valencia, CA, USA) according to the manufacturer's instructions. GRHL2 primers are as follows: GRHL2-F, 5′-GGGCATAGGACTCCAGAGTAGGAA-3′; GR HL2-R, 5′-TAGGGCAGGACTGGCAAACA-3′ (TAKARA BIO INC., Kusatsu, Shiga, Japan). The comparative C T method was used to determine relative gene expression levels for each target gene.

RNAinterferenceandreagents
We chose CFPAC-1 cell line for GRHL2 knockdown experiments as it has strong GRHL2 expression compared to other human PDAC cell lines. Double-stranded small interfering RNAs used to knockdown GRHL2 were as follows: H. Nishino et al. GRHL2 Regulates Epithelial Plasticity in Pancreatic Cancer siRNA1:Hs_GRHL2_2 Cat#SI04150090, siRNA2:Hs_GRHL 2_3 Cat#SI04275271 (Qiagen Inc.). Control cells were treated with negative control siRNA (AllStars negative control siRNA; Qiagen Inc.). Cells precultured for 24 h were transfected with these siRNAs (10 nmol/L final concentration), and knockdown efficiency was assessed by western blot 72 h posttransfection.

Three-dimensionalcellculture
Three-dimensional (3D) organotypic pancreatic ductal cell culture was performed as described previously [18]. Briefly, cells suspended in collagen I solution were seeded onto collagen-coated four-well chamber slides (Thermo Fischer Scientific, Rochester, NY, USA). After 10 days, cells were imaged using the Axiovert 25 inverted microscope (Carl Zeiss, Oberkochen, Germany). The structures were classified into three types-spheroid cysts, irregular cysts, and spindleshaped cells-and their numbers were determined.

Cellcytotoxicityassay
Cells (500 cells/well) were seeded onto 96-well plates, precultured for 24 h and treated with the indicated concentration (200 ng/mL) of gemcitabine hydrochloride (Sigma-Aldrich) for 4 days. We quantified the number of viable cells daily with Cell Counting Kit-8 (Dojindo Laboratories).

Pancreatosphereformationassay
Cells (10 cells/well) were seeded onto 96-well ultralow attachment plates (Corning, New York, NY, USA). After 7 days in sphere medium [19], we counted the number of sphere cells that were defined as cell clusters with over 50 μm diameters on day 7. The sphere formation rate was assessed as the percent increase in the number of spheres on day 7 with respect to the number of spheres observed on day 1.

Anoikisassay
Anoikis assay was performed as described previously [20]. In brief, cells at low density (30,000 cells/mL) were continuously rotated on tube rotator (HB-1000 Hybridizer Hybridization Oven; Ultra-Violet Products, Upland, CA, USA) for 24 h. Then, 3000 cells/well suspended in medium with 0.3% agar were seeded on to the 24-well culture plates coated with bottom layer (medium with 1% agar), and were cultured for 14 days. Tumor colonies were visualized after staining with Giemsa Stain Solution (WAKO, Osaka, Japan, dilution 1:20) to reveal the number and size of the colonies formed.

Patientsandhumantissuesamples
PDAC tissues were obtained from 155 consecutive patients who underwent surgical resection in the Department of General Surgery, Chiba University Hospital, Japan, from February 2006 to November 2011. All patients were diagnosed with primary and liver metastatic PDAC histologically. All small metastatic liver tumors were resected from the surface of liver. The Ethics Committees of our institute approved the protocol of this study and written informed consent was obtained from each patient before surgery.

Statisticalanalysis
Accumulative rates were calculated by the Kaplan-Meier method and the significance of difference in survival rate was analyzed by the log-rank test. Data are expressed as mean ± standard deviation (SD) or the standard error of the mean (SEM). Statistical significance of the results were determined by Student's t-test, Welch's t-test, chisquare test, or Fisher's exact test. P < 0.05 was considered significant in all analyses. All statistical calculations were performed using the Statview-J5.0 software package (SAS Institute Inc., Cary, North Carolina, USA).

GRHL2expressionpositivelycorrelateswith E-cadherinandCD133inprimaryPDACand ishighlyexpressedinlivermetastaticPDAC humantissues
At first, we analyzed GRHL2 expression in human primary PDAC tissues by IHC staining. GRHL2, predominantly localized in the nucleus of cells, is strongly expressed in normal pancreatic duct compared to invasive pancreatic ductal cells (Fig. 1A). The expression pattern of GRHL2 is similar to that of E-cadherin localized in the membrane and cytoplasm of cells (Fig. 1B). Among 155 cases, 100 cases (64.5%) showed high GRHL2 expression, while 55  (Fig. 1C). The patient characteristics of GRHL2 high/low groups are summarized in Table 1. There were no significant differences between the backgrounds of these two groups in primary PDAC tissues. We next assessed GRHL2 expression in primary tumors and liver metastases of PDAC patients. GRHL2 is highly expressed in liver metastases of PDAC (Fig. 1D) (7/7: 100%, high group), and notably, GRHL2 expression in liver metastasis is significantly higher than in paired primary tumor obtained from the same PDAC patient (n = 7) (P = 0.017; chi-square test; Fig. 1E). Considering the biological functional roles for GRHL2 in the EMT/MET plasticity, we performed IHC staining for epithelial/mesenchymal markers, E-cadherin and Vimentin to investigate the correlation between GRHL2 and, these markers in primary PDAC tissues. Interestingly, GRHL2 and E-cadherin expression had a strong positive correlation (P = 0.0002; Fig. 1F), whereas there was no correlation between GRHL2 and Vimentin expression in primary PDAC tissues (Fig. 1G). These observations implicated that GRHL2 is associated with epithelial features in human PDAC.
We also analyzed the correlation of GRHL2 and CD133 expression in primary PDAC tissues. CD133 showed predominant membrane localization in primary PDAC cells (Fig. 1H). High CD133 expression significantly correlated with high GRHL2 expression in primary PDAC tissues (P = 0.02). These results also implicated that GRHL2 expression might be linked to the cancer stem cell (CSC)like property in human PDAC tissues.

GRHL2ishighlyexpressedinmetastatic pancreaticcellscharacterizingepithelial property
Next, we investigated GRHL2 expression in human pancreatic ductal cells. Increased GRHL2 mRNA and protein expression was found in human pancreatic ductal epithelial (HPDE) cells as well as CFPAC-1, liver metastatic PDAC cells, whereas primary invasive PDAC cells displayed decreased expression ( Fig. 2A and B). We also assessed the correlation of protein expression between GRHL2 and representative epithelial/mesenchymal markers, E-cadherin and Vimentin, in these cell lines. E-cadherin is highly expressed in both HPDE and CFPAC-1 cells with high GRHL2 expression, whereas, Vimentin expression is high in primary PDAC cells (Fig. 2B). Next, we examined GRHL2 protein expression in murine pancreatic cells isolated from pancreas of KC mice or pancreas/liver metastasis of KPC mice, genetically engineered mouse models. Consistent with the results from human cell lines, GRHL2 showed high expression in KC precancerous cells and liver metastatic KPC cells characterizing epithelial phenotype compared to primary KPC cells exhibiting mesenchymal phenotype (Fig. 2C). Interestingly, E-cadherin is synchronously expressed with GRHL2 in these murine cells. Additionally, liver metastatic cells compared to pairmatched primary tumor cells show a significant increase in relative GRHL2 expression (Fig. 2D). Taken together, these data supported our findings of human PDAC tissues to confirm the association between GRHL2 and cells characterizing epithelial phenotype in pancreatic cells.

GRHL2regulatesepithelialmorphologyand cellproliferationinPDACcells
To explore the functional roles of GRHL2 in PDAC cells, endogenous GRHL2 expression was suppressed by two specific siRNAs for GRHL2 (GRHL2 siRNA1, GRHL2 siRNA2). Knockdown of GRHL2 resulted in significant downregulation of E-cadherin in CFPAC-1 cells (Fig. 3A  and B). Using an established three-dimensional (3D) organotypic culture system, we investigated whether GRHL2 influences the morphology of PDAC cells. CFPAC-1 control cells expressing GRHL2 formed well-organized spheroid cysts, whereas GRHL2 knockdown cells demonstrated altered morphology and larger population of mesenchymal spindle shaped cells compared to control cells ( Fig. 3C and D). These results suggest that GRHL2 plays a functional role in maintaining the epithelial phenotype in PDAC cells.
We also examined whether GRHL2 affects PDAC cell proliferation. GRHL2 knockdown resulted in significantly reduced proliferation as compared to CFPAC-1 control cells (P < 0.05) (Fig. 3E). Recent experimental evidences have proposed that EMT is associated with chemo resistance [4]. Toward understanding, if GRHL2 is crucially involved in the maintenance of epithelial phenotype in PDAC cells, we hypothesized that GRHL2 knockdown induces mesenchymal phenotype and renders them more resistant to chemotherapy-induced cytotoxicity through EMT process. To analyze this hypothesis, we performed cytotoxicity assays with gemcitabine, which is a widely used chemotherapeutic reagent for PDAC. Upon gemcitabine treatment, GRHL2 knockdown CFPAC-1 cells exhibited enhanced resistance against cytotoxicity as compared to CFPAC-1 cells (Fig. 3F). This result demonstrated that GRHL2 knockdown facilitates EMT and strengthens their chemo resistance. Taken together, these results indicate that GRHL2 induces cell proliferation and sensitizes PDAC cells to chemotherapy-induced cytotoxicity along with epithelial phenotype.

GRHL2maintainsstem-likecharacteristicsin livermetastaticPDACcells
In order to colonize at a distant organ, self-renewal properties are required for the disseminating cancer cells. We performed pancreatosphere formation assay to investigate whether GRHL2 confers the property of putative cancer stem cell (CSC) in CFPAC-1 cells. In pancreatosphere formation assays, GRHL2 knockdown CFPAC-1 cells demonstrated significantly reduced number of spheres compared to control CFPAC-1 cells (Fig. 4A and B). These results indicate that GRHL2 maintains the self-renewal capacity in liver metastatic PDAC cells.
We next examined whether GRHL2 expression correlates with that of CD133, a representative CSC marker for PDAC [21]. Flow cytometry analyses showed increased CD133+ cells in CFPAC-1, liver metastatic PDAC cells compared to PANC-1, primary PDAC cells (data not shown). Notably, the percentage of CD133+ cells dramatically reduced in GRHL2 knockdown CFPAC-1 cells compared to control cells (P < 0.05; Fig. 4C and D). Additionally, the percentage of EpCAM High , another CSC marker for PDAC significantly decreased in GRHL2 knockdown CFPAC-1 cells compared to control cells (P < 0.05; Fig. 4E). These results suggest that GRHL2 might play a crucial role of stem-like property of liver metastatic PDAC cells.

GRHL2isnecessaryforanoikisresistancein metastaticPDACcells
Lastly, we sought to evaluate whether GRHL2 affects anoikis resistance of PDAC cells. In order to form metastatic colonization, disseminating tumor cells that enter into the bloodstream need to survive from apoptosis termed anoikis that results from losing the attachment of those cells to the extracellular matrix [22]. The anoikis resistance ability is a major characteristic of CSC that promotes metastasis [20,23]. GRHL2 positive control cells showed increased colony formation and the colonies were larger compared to GRHL2 knockdown cells (Fig. 4F and G), suggesting that GRHL2 contributes to the maintenance of anoikis resistance in CFPAC-1 cells in vitro. Taken together, these results imply that GRHL2 enhances the CSC-like characters in liver metastatic PDAC cells with the ability to form spheres and anoikis resistance, which facilitates these cells to survive and colonize metastasis at distant organs.

Discussion
In this study, we demonstrated for the first time that GRHL2 has a functional role in the regulation of epithelial plasticity of PDAC cells (Fig. 5). IHC analyses revealed that GRHL2 expression showed strong positive correlation with both E-cadherin and CD133 in clinical samples. Our in vitro data also suggested that GRHL2 facilitates invasive PDAC cells to undergo MET in order to prevent anoikis.
EMT and MET are thought to be vital cellular characteristics in the invasion-metastasis cascade of cancers [24]. EMT and MET enable cancer cells to survive and induce metastatic colonization post dissemination and extravasation into the parenchyma of distant organs. The metastatic tumors almost invariably display epithelial features, such as well-organized adherens junctions, suggesting that tumor cells disseminating from the primary site revert to an epithelial phenotype through MET as they form macroscopic metastases in secondary sites [25]. Recent studies revealed the significance of EMT in PDAC carcinogenesis [3]; however, the involvement of MET in PDAC metastasis is poorly understood. In this study, we hypothesized that GRHL2 could be a potential therapeutic target regulating epithelial plasticity and CSCs-like property of PDAC. We examined the correlation between GRHL2 and Vimentin, a representative mesenchymal marker for human PDAC tissues. Unexpectedly, there was no inverse correlation between GRHL2 and Vimentin expression; however, we confirmed that GRHL2 expression strongly correlates with E-cadherin or CD133 expression in PDAC tissues. The fact that CD133 is expressed specifically in pancreatic ductal epithelium supports our results [26].
In this study, we demonstrated that GRHL2 is upregulated in human pancreatic ductal cells and mouse PanIN cells compared to primary invasive PDAC cells. Notably, both GRHL2 and E-cadherin in human and mouse PDAC cell lines with liver metastases appeared to be significantly upregulated, which suggests the observation that reestablishing epithelial integrity is required for disseminating cancer cells to form overt metastasis [10,27]. We confirmed that GRHL2 knockdown promotes the morphological change from epithelial to mesenchymal phenotype in three-dimensional culture. It is also speculated that GRHL2 knockdown accelerates cells to acquire mesenchymal traits in precancerous PanIN cells. This finding suggests that GRHL2 might be required to determine epithelial phenotype of pancreatic ductal cells during PDAC progression.
Cell proliferation is an epithelial characteristic crucial for macrometastasis. Previous studies demonstrated that GRHL2 knockdown resulted in a significant reduction in cell proliferation in various cancer cells [28]. Consistent with this, we demonstrated that GRHL2 downregulation decreased the PDAC cell proliferation. Collectively, GRHL2 appears to be the essential transcription factor that determines the epithelial plasticity of PDAC and contributes to maintain epithelial functional traits. The role of EMT in metastasis is a longstanding source of debate. Recent studies support a novel proposal that EMT is not always essential for metastasis [29]. However, EMT cells significantly contribute to recurrent lung metastasis formation after chemotherapy [30], and suppression of EMT leads to an enhanced sensitivity to gemcitabine treatment [31]. Our study demonstrated that knockdown of GRHL2 induces cytotoxicity resistance as well as EMT in PDAC cells. The drug resistance appears to be more influenced by mesenchymal phenotypic change induced by GRHL2 knockdown in liver metastatic PDAC cells.
In this study, we also demonstrated that GRHL2 knockdown decreases the subpopulation of CD133+ cells or EpCAM High cells in metastatic PDAC cells, supporting the observation that GRHL2 regulates liver metastatic PDAC cells to maintain stem cell-like properties. Increasing evidence indicates that the tumor cells that initiate metastatic outgrowth are CSCs or, at least possess several attributes of these cells [32]. Some studies indicate the effects of CSC on metastasis; a distinct subpopulation of CD133+ CXCR4+ CSCs was identified that determines the metastatic phenotype of the individual tumor [21]. It has been demonstrated that EMT-inducing transcription factors confer mesenchymal as well as stem cell properties [26,33], whereas results of the counterpart that EMT can suppress CSC-like characteristics including self-renewal capacity were also shown in previous studies [34]. Understanding this discrepancy, we investigated selfrenewal capacity and anoikis resistance, which are two stem cell-like properties required to survive and outgrow in the hostile environment of a distant organ in vitro. Notably, both self-renewal capacity and anoikis resistance were GRHL2-dependent in liver metastatic PDAC cells with epithelial phenotype. Collectively, our results suggest that GRHL2 plays a crucial role in regulating epithelial plasticity with CSC-like properties to form overt liver metastatic colonies.
In conclusion, our study demonstrated that MET and CSC-like properties might be coupled by GRHL2, which is associated with epithelial characteristics and enhanced stemness of PDAC. These findings provide an insight into plasticity regulation during PDAC progression. Further studies are warranted to examine the functional relationship between GRHL2 and epithelial phenotype in in vivo assays such as orthotopic transplantation and metastatic models, and determine whether GRHL2 could be a promising therapeutic target for PDAC.