Regulation of matriptase and HAI-1 system, a novel therapeutic target in human endometrial cancer cells

The effects of specific and non-specific regulation of matriptase on endometrial cancer cells in vitro were investigated. Messenger ribonucleic acid (mRNA) and protein expression of matriptase and hepatocyte growth factor activator inhibitor-1 (HAI-1) in RL-952, HEC-1A, and HEC-1B endometrial cancer cells were detected by real-time quantitative PCR (RT-qPCR) and western blot. The cells were infected with lentivirus-mediated small-interfering RNA (siRNA) targeted on matriptase (MA-siRNA) or treated with different cisplatin (DDP) concentrations. After treatment, invasion, migration, and cellular apoptosis were analyzed. Matriptase mRNA and protein expression significantly decreased to 80% after infection with MA-siRNA (P < 0.01), and scratch and trans-well chamber assays showed significant inhibition of invasiveness and metastasis. Upon incubation with cisplatin at concentrations higher than the therapeutic dose for 24 h, the expressions of matriptase and HAI-1 significantly decreased (P < 0.001). Moreover, the invasiveness, metastasis, and survival rate of HEC-1A and RL-952 endometrial cancer cells were significantly decreased (P < 0.001) due to the down-regulation of matriptase and HAI-1 upon increasing cisplatin concentration. However, a slight increase in matriptase and HAI-1 expression was observed in cells treated with low cisplatin concentration (P = 0.01). Moreover, matriptase expression was associated with metastasis and invasiveness. Down-regulation of matriptase by specific Ma-SiRNA or non-specific cisplatin in matriptase/HAI-1–positive endometrial cancer cells showed promising therapeutic features.


INTRODUCTION
Endometrial cancer is one of the most common malignancies of the female reproductive tract, and its incidence is currently increasing. The American Cancer Society estimated there would be 60,050 new cases and 10,470 deaths from endometrial cancer in 2016 [1]. Asian nations such as China, Japan, and Korean have lower incidence than do western industrialized countries. However, the incidence of endometrial cancer in China has increased over the past 30 years, and it is currently the second most common gynecologic malignancy [2]. Although most of endometrial cancer patients are diagnosed early due to irregular uterine bleeding and abnormal vaginal discharge or other symptoms, there are still 15-25% of patients at advanced stage [3]. Moreover, patients with either advanced stage endometrial cancer at diagnosis or recurrent disease present a considerable therapeutic challenge [4]. Advanced-stage endometrial cancer seriously threatens the patient's health and is responsible for most deaths [4]. Optimal treatment approaches yield response rates of 40-70% in patients

Research Paper
Oncotarget 12683 www.impactjournals.com/oncotarget with primary advanced cancer and 15-30% in patients with recurrent disease. Furthermore, among these patients, median progression-free survival is only 6 months and median overall survival is 12 months [5]. For the advanced endometrial cancer, metastases are the major cause of treatment failures and mortality. Thus, studies on the invasive and metastatic mechanism are essential to improve advanced endometrial cancer-related survival and cure rate.
Tumor invasion and metastasis formation are complex biological processes depending on the matrixdegrading proteolytic system, which allows tumor cells to detach from their primary site and migrate to distant sites. Matriptase is a type-II transmembrane serine protease (TTSP) of about 855 amino acids and belongs to the family of S1 trypsin-like protease [6,7]. It combines an amino-terminal hydrophobic transmembrane region with an extracellular section of several domains including trypsin-like catalytic and low-density lipoprotein regions. Moreover, matriptase is expressed not only in epithelial cells, but also in mast cells, B-cells, and blood monocytes [8,9,10]. The expression levels of matriptase reflect the degree of tumor progression in several types of cancerous cells, which indicates the crucial role of this protein in malignant cells metastasis [11,12,13]. Moreover, matriptase is implicated in a number of other diseases and induces cancer itself [13]. Therefore, matriptase has become a promising target for anti-cancer treatments. Interestingly, matriptase is controlled by its endogenous inhibitor HAI-1 (Hepatocyte growth factor activator inhibitors-1). It has been reported that matriptase and HAI-1 are closely related to the development and progression of many malignant tumors such as esophageal cancer, breast cancer, and prostate cancer [14,15,16] and have strong potential oncogenic effect, which may influence tumor invasion and metastasis [17]. However, only two studies about the expression of matriptase and HAI in endometrial cancer have been published on the PUBMED until now [18,19]. The role of matriptase in endometrial cancer remains unclear. In this study, we analyzed the expression of matriptase and HAI-1 in endometrial cancer cell lines and analyzed their relationship with the invasion and migration of endometrial cancer cells.

Expression of matriptase and HAI-1 in endometrial cancer cells
The mRNA expression of matriptase and HAI-1 was detected by The levels of matriptase and HAI-1 mRNA expression in HEC-1A, HEC-1B, and RL-952 endometrial cancer cell lines were determined using quantitative PCR. The relative mRNA expression of matriptase was 0.212 ± 0.021 in HEC-1A and 0.178 ± 0.013 in RL-952, while it was only 0.00695 ± 0.0012 in HEC-1B (F = 122.629, P < 0.001) ( Figure 1A). The relative mRNA expression of HAI-1 was 0.283 ± 0.049 in HEC-1A and 0.242 ± 0.032 in RL-952. Similar to matriptase mRNA expression, HAI-1 mRNA expression was only 0.0263 ± 0.0043 in HEC-1B (F = 32.875, P < 0.01) ( Figure 1A). The matriptase/ HAI-1 mRNA expression ratio in HEC-1A, RL-952, and HEC-1B was 0.75, 0.73, and 0.02, respectively ( Figure  1B). The western blot showed the same protein expression pattern ( Figure 1C). Both matriptase and HAI could only be detected extreme weak mRNA expression in the HEC-1B cells. Thus, we concluded that, in HEC-1B cells, the matriptase/HAI signal pathway may play a different role with the other two cells.

Inhibition of migration and invasion ability in endometrial cancer cells by down-regulation of matriptase
Migration distances at 24 h were significantly longer in the HEC-1A and RL-952 CON groups than they were in the HEC-1A KD group (CON, 206.67 ± 28.38; KD, 79.2 ± 6.82, P < 0.001) and RL-952 KD group (CON, 184.57 ± 21.97; KD, 76.8 ± 5.48; P < 0.001) ( Figure 3A-3C). The NC group showed results similarly to those of the CON group in both HEC-1A and RL-952 cells (P > 0.05). Person's correlation analysis showed that matriptase mRNA expression level was positively correlated with the migration distance with r = 0.97 and r = 0.982 in HEC-1 and RL-952, respectively. Compared with the CON group, the transmembrane cell number in the KD group was significantly reduced in both HEC-1A cells (CON, 139.25 ± 12.3112; KD, 48.6 ± 4.8496) and in RL-952 cells (CON, 150 ± 7.0710; KD, 53.3 ± 5.6376, P < 0.05). www.impactjournals.com/oncotarget Again, the NC group gave results similarly to those of the CON group. Compared to the control group, the NC gave similar result in both HEC-1A (132 ± 8.3120) and RL-952 cells (145 ± 6.0711, P > 0.05) ( Figure 3B, 3D). Pearson's correlation analysis showed positive correlation between matriptase mRNA expression levels and transmembrane cell number with r = 0.975 and r = 0.994 in HEC-1A and RL-952, respectively.

DISCUSSION
To date, there are no effective therapeutic options for patients with advanced-stage or recurrent endometrial cancer. Despite many efforts, this disease inevitably progresses with great morbidity, to eventually cause death. Matriptase and its inhibitor HAI-1, which may play Oncotarget 12687 www.impactjournals.com/oncotarget HEC-1B cells display extreme low expression of the matriptase/HAI. We speculated that these data might relate to the different cell's genetic profile such as the expression of the estrogen receptor. Some studies concluded that matriptase mRNA levels were not significantly increased in breast cancer compared to in normal breast tissue, but other studies reported that high matriptase expression is predictive of poor survival for breast cancer as assessed by immunohistochemistry [20,21]. The process of invasion and metastasis of malignant tumor is complex and it may involve various signaling pathways such as those linked to protease-activated receptors [22], prostasin [23], and micro-RNAs [24]. Several reports suggested that the expression of matriptase is involved in the initiation of malignant progression in epithelial cell carcinogenesis and indicated the potential value of matriptase as prognostic marker in various human cancers [11,25,26]. Moreover, matriptase may play an important role in cell invasiveness and metastasis [27,28]. Nakamura et al. reported that the expression of matriptase in endometrial cancer may be associated with aggressive biological characteristics and may play an important role in prognosis and/or recurrence [18]. Interestingly, strong matriptase expression at both the mRNA and protein level was associated with high metastatic ability and invasiveness [17]. Therefore, we targeted suppressed matriptase expression in HEC-1A and RL-952 cells using lentivirus-mediated siRNA. Real-time qualitative PCR and western blotting revealed that siRNA transfection significantly decreased matriptase expression in both cells (P < 0.01), resulting in significant decreases of the cell invasiveness and migratory activity.
The expression of HAI-1, which has an important role in ECM degradation in endometrial cancer, was obviously lower than that in normal endometrium specimens. Increased expression levels of the HAI-1 gene accompanied with decreased expression of matriptase inhibits endometrial cancer cell proliferation and invasive migration [19]. Some scholars have suggested that maintaining matriptase/HAI-1 equilibrium is crucial in the process of tumor progression. According to previous reports, matriptase/HAI-1 ratio imbalance, which includes increased expression of matriptase or decreased expression of HAI-1, may promote tumor development. In pancreatic cancer cells, increased matriptase/HAI-1 ratio by decreased HAI-1 expression levels promoted tumor invasion [29]. Similarly, increased matriptase/HAI-1 ratio with increased matriptase expression made prostate cancer more aggressive [14,25]. The matriptase/HAI-1 ratio declined in advanced colorectal cancer and infiltrating breast cancer, while it increased in invasive breast cancer [30,31]. To our knowledge, this is the first study reporting on the regulation of the matriptase/HAI-1 ratio by specific (i.e., siRNA) and non-specific (i.e., cisplatin) targeting associated with decreasing matriptase expression in endometrial cancer cells. In our previous study, we concluded that the expression of matriptase is directly and positively correlated with ovarian cellular invasion and metastasis [17]. Usage of siRNA specifically reduced matriptase expression in matriptase/HAI-1-positive RL-952 and HEC-1A cells. Although HAI-1 gene expression did not change, matriptase/HAI-1 ratio decreased from 0.77 to 0.11 after the first knockdown, and cell invasion and migration decreased significantly.
Since platinum-based combination chemotherapy is currently the main therapy for endometrial cancer, in this study, we treated endometrial cancer cell lines with different cisplatin concentrations. The results showed that invasiveness, metastasis, and apoptosis of matriptase/ HAI-1-positive HEC-1A and RL-952 endometrial cancer cells after treatment with normal therapeutic concentration or higher cisplatin doses are positively related to the expression of matriptase and HAI-1. Interestingly, when the cells were treated with low cisplatin concentration, both HEC-1A and RL-952 cells showed increased expression of matriptase and HAI-1, which resulted in enhanced invasion and migration of endometrial cancer cells. Liu et al. [32] reported that short-time low cisplatin concentration treatment leads to elevated invasiveness of prostate cancer cell in vitro, which is possibly due to epithelial mesenchymal transition (EMT). In addition, Chen et al. [33] demonstrated that low-dose cisplatin contributes to the development of drug resistance to promote proliferation of ovarian cancer. The above findings provide novel insights into the mechanisms underlying how low cisplatin concentration treatment leads to elevated tumor progression. Since too low cisplatin doses do not induce tumor cell apoptosis and increased expression of matriptase and HAI-1, these could be clinical indexes for the determination of whether cisplatin doses were administered at effective therapeutic concentrations.
In summary, our findings suggest that the mRNA and protein levels of matriptase and HAI-1 are reliable biomarkers that reflect the aggressive nature of endometrial cancer cells. Matriptase and HAI-1 are potential therapeutic targets for the inhibition of endometrial cancer invasion and metastasis, and could be used as indicators of the curative effect of cisplatin.

Fluorescence Real-time quantitative PCR
Real-time quantitative PCR was performed by using a LightCycler ® 480 SYBR Green I Master Mix (Roche, Germany). The following primer sets were synthesized by Sanggong Biotech (Shanghai, China). For matriptase, sense 5′-TCG TCA CTT GTA CCA AAC ACA CCT A-3′ and anti-sense 5′-GAG CCT GTC TCG TGA ATG ACC-3′ (150 bp); for HAI-1: sense 5′-GGC AAC AAG AAC AAC TTT GAG GA-3′ and anti-sense 5′-CAA TGC AGA TGA CCA GGA ACAC-3′ (154 bp); for GAPDH, sense 5′-GCACCGTCAAGGCTGAGAAC-3′, anti-sense: 5'-TGGTGAAGACGCCAGTGGA-3' (138 bp) [17]. The program for real-time PCR was as follows: 95°C for 15 s, 45 cycles of denaturation at 95°C for 5 s and annealing at 60°C for 20 s, 95°C for 1 min, and cooling to 55°C.  The relative mRNA levels were calculated using the comparative cycle threshold (C t ) method (ΔΔC t ). Briefly, the C t value for GADPH was subtracted from the C t value of the target gene to achieve the ΔC t value. The 2 −ΔCt value was calculated for each sample and each value was then divided by that of the control to determine the relative mRNA levels (ΔΔCt).

Lentivirus-mediated small interfering RNA construction and infection
Three lentivirus-mediated small interfering RNA constructs were constructed and named siRNA-Ma-1-3. The following siRNA target sequences in the matriptase gene (ST-14, GenBank accession No. NM_021978) were selected: Ma-SiRNA-1, CCGGCTTCTTAGCTGAATA; Ma-SiRNA-2, TGTCCAGAAGGTCTTCAAT; and Ma-SiRNA-3, ACGAGAAAGTGGAATGGCTT [17]. Then, three pairs of complementary oligonucleotides were designed, and stem-loop oligonucleotides were synthesized and cloned into a lentivirus-based vector carrying the green fluorescent protein (GFP) gene (GV115, Genechem, Shanghai, China). A universal sequence (PSC-NC: TTCTCCGAACGTGTCACGT, named NC) was used as the negative control for RNA interference. Lentiviral particles were prepared as previously described [34]. Three Ma-siRNA-carrying and NC-carrying lentiviral vectors were constructed and infected into endometrial cancer cells each at multiplicities of infection (MOIs) of Oncotarget 12692 www.impactjournals.com/oncotarget 20 (low MOI) and 80 (high MOI). After infection for 72 h, GFP expression was detected to calculate the infection efficiency. Five days after infection, cells were harvested. Real-time PCR was performed to determine the efficiency of matriptase depletion and screen for the siRNA with the highest silencing efficiency, which was then used for subsequent experiments. Three groups of cells were set: cells infected with lentivirus-mediated siRNA target on matriptase (group KD), cells infected empty lentivirus (group NC), and cells treat with DMSO as the control (group CON).

In vitro cellular scratch test
Cells, either treated with drug or siRNA, were allowed to grow to confluence in 6-well plates. A 200-μL tip was used to introduce a scratch in the monolayer. The scratch areas in wells were washed with PBS and 1 mmol/L R-flurbiprofen until the cells in those areas were removed thoroughly. The wells were imaged at ×40 magnification with an Olympus IX70 invertedfluorescence microscope (Olympus, Japan) at 0 and 24 h post-scratching. Scratch healing was determined by measuring the shortest distance between scratch edges in each field of view. Three different fields were measured per scratch, and 20 different measurements were taken per field. The distance between scratches was measured using the Image ProExpress C software (Olympus, Japan) and the horizontal migration rate was calculated using the following formula: (width 0 h − width 24 h )/width 0 h × 100% [17,35].

Trans-well chamber migration assay
After thawing overnight at 4°C on ice, 50 μL of Matrigel™ Basement Membrane Matrix (BD, USA) was added to a Millicell Hanging Cell Culture Insert (Millipore, USA) to coat the membrane and incubated at 37°C for 30 min. The membrane was re-hydrated with FBS-free DMEM thrice. Cells in the logarithmic growth phase were suspended in the purpose medium containing 0.5% FBS after conventional digestion. Then, 200-μL cell suspensions (5.0 × 10 5 /mL, untreated with drug or siRNA for 72 h) were added to the Hanging Cell Culture Insert placed in 24-well places containing 1,300 μL of DMEM supplemented with 10% FBS. The plates were incubated for 24 h at 37°C. At the end of incubation, non-migrating cells on the inside of the filter were removed with a cotton swab, and the filters were fixed with methanol and stained with crystalline violet for 15 min. The filters were removed from the inserts and mounted onto slides for imaging and quantification. The number of migrating cells on the underside of the filter was determined by counting cells in 5 random fields from 3 filters for each treatment at ×200 magnification using an inverted microscope (Olympus) [17,36].

Treatment with cisplatin
Cisplatin (Sigma) was prepared in 100% dimethylsulphoxide (DMSO). Before drug treatment, endometrial cancer cells were seeded in 6-well plates at a density of 1 × 10 5 cells per well and cultured in 1 mL of serum-free DMEM for 12 h to achieve adherence. For each cell line, four groups were set. Three groups were treated with cisplatin at different final cisplatin concentration (i.e., 2 mg/L, 10 mg/L, and 50 mg/L for HEC-1A and HEC-1B; and 1 mg/L, 2 mg/L and 5 mg/L for RL-952), and one group was treated with DMSO alone as the control. At different checkpoint times (between 0 and 24 h), Real-time PCR analysis and western blot were performed on 3 wells in RL-952 group per point.

Cellular survival rate analysis via flow cytometry
For flow cytometric analysis, cells were seeded onto 6-well plates and cultured to 80% confluence. After harvesting, the cell pellets were washed twice with precooled PBS and fixed with pre-cooled 70% ethanol. Then, suspended cells were filtered through a 400-mesh sieve and stained with propidium iodide (PI, 100 μg/ mL RNase in PBS) at 37°C for 30 min. Then, the cell cycle distribution was determined using flow cytometry as previously described [37]. Apoptosis was detected using the Annexin-V-FLUOS staining kit (Roche, USA) according to the manufacturer's instructions. Fluorescein and PI fluorescence was measured using a FACSCanto II flow cytometer (BD, USA).

Statistical analysis
All experiments were performed in triplicate. Statistical analysis was performed using the average results of three experiments under identical conditions. Numerical data are presented as the mean ± SD. Differences between two means were compared by Student's t-test, and related parameters were analyzed using Pearson's correlation analysis. Data were analyzed using SPSS 17.0 for Windows (SPSS Inc., Chicago, IL, USA). Differences were considered significant at P < 0.05.

CONFLICTS OF INTEREST
The authors declare no potential conflicts of interest.