Predictive value of protease‐activated receptor‐2 (PAR2) in cervical cancer metastasis

Abstract Metastasis is the primary cause of an unfavourable prognosis in patients with malignant cancer. Over the last decade, the role of proteinases in the tumour microenvironment has attracted increasing attention. As a sensor of proteinases, proteinase‐activated receptor 2 (PAR2) plays crucial roles in the metastatic progression of cervical cancer. In the present study, the expression of PAR2 in multiple types of cancer was analysed by Gene Expression Profiling Interactive Analysis (GEPIA). Kaplan‐Meier plotter was used to calculate the correlation between survival and the levels of PAR2, Grb‐associated binding protein 2(Gab2) and miR‐125b. Immunohistochemistry (IHC) was performed to examine PAR2 expression in a tissue microarray (TMA) of CESCs. Empower Stats was used to assess the predictive value of PAR2 in the metastatic potential of CESC. We found that PAR2 up‐regulation was observed in multiple types of cancer. Moreover, PAR2 expression was positively correlated with the clinicopathologic characteristics of CESC. miR‐125b and its target Gab2, which are strongly associated with PAR2‐induced cell migration, are well‐characterized as predictors of the prognostic value of CESC. Most importantly, the Cancer Genome Atlas (TCGA) data set analysis showed that the area under the curve (AUC) of the PAR2 model was significantly greater than that of the traditional model (0.833 vs 0.790, P < .05), demonstrating the predictive value of PAR2 in CESC metastasis. Our results suggest that PAR2 may serve as a prognostic factor for metastasis in CESC patients.


| INTRODUC TI ON
Metastasis is not only the main characteristic of malignancy but also the main factor that affects the therapeutic effect and prognosis of patients. In addition to the genetic background of cancer cells, alterations in the microenvironment have emerged as an important factor in regulating the metastatic progression of cancer. In the past, a great deal of research has focused on the microenvironment that surrounds cells and its role in tumour metastasis. 1 The microenvironment of the tumour invasion front may enable the cancer cells to gain anomalously high motility and penetrate the surrounding stroma. 2 It is worth noting that the invasion front of a tumour is particularly rich in a variety of proteinases, which facilitate cancer invasion and metastasis by remodelling the extracellular matrix and promote cell migration. 3 Protease-activated receptors (PARs) are a subgroup of G protein-coupled receptors (GPCRs). To date, four members of the PAR family have been discovered: PAR 1 , PAR 3 and PAR 4 , which are activated by thrombin, and PAR 2, which is activated by trypsin, tissue factor, neutrophil elastase and other factors. 3,4 Previous evidence has shown that PAR 2 plays an important role in promoting the metastasis of colon cancer cells. [5][6][7] As the sensor of protease, PAR 2 and its activating proteinases are typically observed in the invading frontier cells of cancer, 8 and its expression level is tightly correlated with the switching of a primary tumour from local to metastatic spread. 7,9,10 Dysregulated microRNAs (miRNAs) are highly involved in the initiation and progression of multiple cancers. They function as either proto-oncogenes or tumour suppressors in vivo by repressing their target mRNAs or reducing their transcription. 11,12 The dysregulation of miR-125b is commonly observed in many malignant tumours, such as ovarian, 13 colon 10 and breast 14 tumours. Our previous study revealed that miR-125b 9 not only contributes to cell migration but is also regulated by PAR 2 activation. In view of this evidence, we believe that the molecular characteristics of miR-125b, which is regulated by PAR 2, should be studied in depth to monitor tumour outcomes.
In the present study, we used multiple online tools to analyse the association between PAR 2 levels and tumour prognosis in multiple cancer. Moreover, PAR 2 expression and the clinicopathologic stage of cervical squamous cell carcinoma endocervical adenocarcinoma (CESC) were assessed with a tissue microarray (TMA). miR-125b and its target Grb-associated binding protein 2 (Gab2), which are strongly linked to PAR 2 -induced cell migration, are well-characterized predictors of metastasis in CESC. Most importantly, The Cancer Genome Atlas (TCGA) data sets of CESC analysed by Empower Stats demonstrated the predictive accuracy of PAR 2 in CESC metastasis. Therefore, the PAR 2 expression pattern could serve as a risk factor that indicates a poor prognosis for patients with cervical cancer.

| Cell culture and cell lines
The human colonic epithelial cell line HT-29, and HCT116 as well as the lung adenocarcinoma cell line A549 were obtained from the American Type Culture Collection (Manassas, VA, USA). The cells were grown in Dulbecco's modified Eagle's medium/F12 supplemented with 10% FBS (Gibco, NY, USA). Stably transfected HT29 cells with PAR 2 knockdown were enriched with puromycin according to a previously described protocol. 7

| Gene Expression Profiling Interactive Analysis (GEPIA), Kaplan-Meier plotter and Gene Expression Display Server (GEDS) online database
Multiple tumour vs normal differential PAR 2 expression analysis was performed based on the GEPIA database (http://gepia.cancer-pku. cn), which is a newly developed web-based tool that provides key interactive and customizable functions based on TCGA and genotypetissue expression data. [15][16][17] The prognostic values of PAR 2 , miR-125b and Gab2 in tumour patients were evaluated using Kaplan-Meier plotter (http://kmplot. com/analysis), an open online data set that can be used to assess the effects of 54 675 genes on survival in 21 cancer types. 18,19 The differential expression of miR-125b between tumours and corresponding non-tumour tissues was evaluated using the GEDS data-

| Real-time PCR
The total RNA was isolated from cells using TRIzol reagent (Invitrogen, Carlsbad, CA). After treatment with DNase I, RNA was reverse transcribed into cDNA with a Thermo Scientific Maxima First Strand cDNA Synthesis Kit for mRNA and analysed for miR-125b detection with a TaqMan™ microRNA Transcription Kit. Real-time quantitative PCR was carried out on an Applied Bio-Systems 7500 PCR instrument. PCR data were normalized to those of GAPDH and U6 short hairpin RNA for mRNA and miRNA, respectively.

| Statistical analysis
All statistical analyses were performed using GraphPad Prism 5.0, Empower Stats software (www.empow ersta ts.com, X&Y solutions, Inc Boston MA) and R (http://www.R-proje ct.org). 22 The data are presented as the mean ± SD, and a P value less than .05 was considered statistically significant.

| PAR 2 is up-regulated in multiple types of tumours
We initially found that PAR 2 was expressed in nearly all human tissues after searching Gemini online tools ( Figure S1). Moreover, we used the GEPIA online tool to further evaluate whether PAR 2 expression was different between non-tumour and tumour tissues in multiple human cancers. Notably, PAR 2 expression was markedly up-regulated in tumour tissue relative to control adjacent tissue. As shown in Figure  and uterine carcinosarcoma (UCS). In contrast, PAR 2 expression was down-regulated in kidney chromophobe (KICH). Taken together, these results suggest that PAR 2 up-regulation is highly related to the progression of multiple tumours.

| PAR 2 correlates positively with poor survival in CESC, LUAD and PAAD
To These results confirmed that PAR 2 expression had an impact on the prognosis (both OS and RFS) of the CESC, LUAD, PAAD and READ cohorts. In the present study, we performed an in-depth investigation on whether the activation of PAR 2 was associated with a poor prognosis of CESC.

| PAR 2 expression is associated with tumour metastasis
Metastasis is a major cause of death for patients with malignant tumours. To better understand the relevance and fundamental mechanisms of PAR 2 in tumours, the correlation between PAR 2 expression and the metastasis characteristics of clinical CESC tumours was assessed.
PAR 2 expression in 119 CESC patient samples and matched adjacent cervical mucosa samples was assessed with a TMA and IHC staining. The IHC results revealed that PAR 2 was strongly expressed in CESC tissues, but weakly expressed in normal epithelial tissues of the cervical mucosa and in cervicitis tissues ( Figure 2A). Moreover, PAR 2 expression was more easily observed in poorly differentiated cervical tumours than in moderately or well-differentiated cervical tumours ( Figure 2B). We experienced difficultly in detecting significant differences between tumour stages I and II, but PAR 2 was up-regulated in the lymphatic metastasis relative to the local tissue in stage III tumours ( Figure 2C). Notably, the PAR 2 -positive cells were typically stacked in advance of the invasive margin of the tumour tissue ( Figure 2D). These results confirm that abnormal PAR 2 expression is closely associated with the metastasis of CESC.

| The levels of miR-125b and its target Gab2 are closely correlated with a poor prognosis in CESC
Our previous study suggests that miR-125b mediates PAR 2induced cancer cell migration by regulating Gab2 expression. 9 To better understand the roles of miR-125b and Gab2 in clinical CESC progression, the impacts of factors on cell migration were tested. We found that altering the miR-125b or Gab2 expression changed the cell migration abilities. Notably, the overexpression of miR-125b by the mimic-miR-125b or the knockdown of Gab2 by the siRNA significantly blocked the PAR 2 -induced cell migration ( Figure S3).

TA B L E 1
The prognostic value of PAR 2 in patients with multiple cancers

| The prognostic value of PAR 2 in CESC metastasis
To determine whether PAR 2 could serve as a prognostic factor for CESC metastasis, the TCGA data set of CESC was downloaded from cBioPortal (https://www.cbiop ortal.org/datasets, TCGA Pan-Cancer Atlas) and analysed by Empower Stats software.
The data set contains detailed information on 255 patients (Table 3).
Based on the 7th edition of the American Joint Committee on Cancer staging system, we collected recorded items (N = 127) and the risk factors (P < .05) that are closely related to tumour metastasis. Then, we formed two predictive models. Data in the CEA model included age at diagnosis, body mass index (BMI), total number of pregnancies, patient smoking history category, tumour type, primary lymph node presentation assessment, neoplasm cancer status and the expression level of CEACAM5, which represents a typical oncofetal antigen. [23][24][25][26] In the PAR 2 model, the expression level of PAR 2 was taken into consideration instead of CEACAM5 (used in the CEA model). Figure 4A  These data demonstrated that PAR 2 can serve for an important indicator to predict the potential for metastasis in CESC patients.
Machine learning methods were used to validate the importance of risk factors in the PAR 2 model. Figure 4C shows that the PAR 2 expression level, following the BMI, was the second most important predictor in the random forest model. The decision curve analysis evaluating the benefit and risks of the two models is presented in Figure 4D.    In summary, we demonstrated that PAR 2 expression was higher in cervical cancer tissues than in normal tissues and correlated with advanced cancer metastasis and short survival. PAR 2 activation regulates miR-125b repression, which may result in its migration-promoting effect on cancer cells. In addition, our multivariable analysis indicated that PAR 2 could increase the predictive accuracy of the metastatic prognosis of CESC. We believe that PAR 2 is an important factor for predicting CESC metastasis, and the change in its expression level should be emphasized in the treatment process of CESC.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no competing interests.

AUTH O R S' CO NTR I B UTI O N S
HS: Acquisition of data; analysis and interpretation of data; drafting of the paper; and statistical analysis. XM and WL: Acquisition of data; study supervision; and critical revision of the paper. XZ, HY, HQ, LJ and LY: Technical and material support. YL: Study concept and design; obtained funding; drafting of the paper; and study supervision.