ARHGAP4 regulates the cell migration and invasion of pancreatic cancer by the HDAC2/ -catenin signaling pathway

Yehua Shen1,2△, Litao Xu1,2, Zhouyu Ning1,2, Luming Liu1,2, Junhua Lin1,2, Hao Chen1,2, Zhiqiang Meng1,2△
1Department of Integrative Oncology, Fudan University Shanghai Cancer Center, Shanghai, 200032, China
2Department of Oncology, Shanghai Medical College, Fudan University, Shanghai, 200032, China

Corresponding author

Zhiqiang Meng and Yehua Shen, Department of Integrative Oncology, Fudan University Shanghai Cancer Center, 270 Dong’An Road, Shanghai, 200032, China; Department of Oncology, Shanghai Medical College, Fudan University, 138 Xueyuan Road, Shanghai, 200032, China. Tel:
+86-021-64175590; E-mail: [email protected] (Zhiqiang Meng) and [email protected]

(Yehua Shen)

The Author(s) 2019. Published by Oxford University Press. All rights reserved. For Permissions, please email: [email protected].
Abstract

-catenin is a subunit of the cadherin protein complex and acts as an intracellular signal transducer in the Wnt signaling pathway that mediates multiple cellular processes, such as cell migration and invasion. HDAC2 (histone deacetylase 2), a deacetylase that maintains histone H3 in a deacetylated state in the promoter region of Wnt-targeted genes where β-catenin is bound, negatively regulating -catenin activation. However, the regulation of HDAC2/ -catenin pathway remains unclear. Here we report ARHGAP4 as a new regulator of the -catenin pathway that regulates cell invasion and migration of pancreatic cancer as well as the downstream effector MMP2 and MMP9 expression in vitro. Mechanistically, ARHGAP4 interacts with and ubiquitinates HDAC2, which in turn inhibits -catenin activation. Furthermore, treatment of CAY10683, a HDAC2 inhibitor, and XAV939, a Wnt/β-catenin pathway inhibitor, attenuated the effects of ARHGAP4 silencing on pancreatic cancer cells. Overall, our findings establish ARHGAP4 as a novel regulator of HDAC2/ -catenin pathway with a critical role in tumorigenesis.
Summary

ARHGAP4 as a new regulator of the -catenin pathway regulates cell invasion and migration of pancreatic cancer as well as the downstream effectors MMP2 and MMP9 expression. Mechanistically, ARHGAP4 ubiquitinates HDAC2, contributing to inhibit -catenin activation.
Key words: pancreatic cancer, ARHGAP4, HDAC2, -catenin, ubiquitination

Introduction

Pancreatic cancer (PC) is a highly malignant tumor of the digestive system and mainly consists of ductal adenocarcinoma, accounted for 80%-90% of PC. It is the fourth leading cause of cancer-related deaths only to that of lung, colon and breast cancer(1), but compared with the first three tumors, the prognosis of PC is poor, and the disease at diagnosis is often advanced and metastatic(2,3). The overall 5-year survival rate for PC patients was only 6%(4), and the median survival time was only 6-10 months(5). The traditional treatment of PC mainly includes surgical resection, chemotherapy, and radiotherapy(6,7), but less than 20% of patients are candidates for surgical resection(8). According to the physiological microenvironment of PC, new therapeutic strategies are also actively combining small molecule targeted drugs with immunotherapy and metabolic therapy(6,9).
Epigenetic chromatin remodeling events such as histone modification mediated by histone acetylases (HATs) and histone deacetylases (HDACs) are essential for many biological processes(10). So far, abnormal expression of HDACs has been found in glioblastoma, breast, lymphoma, pancreatic cancer and other malignant tumors(11-13). Overexpression of HDACs can lead to imbalance of histone acetylation and recruitment of transcription factors, leading to abnormal inhibition of specific genes, and thereby participating in the development and progression of cancers(14,15). However, the role of HDACs in tumorigenesis and the key targets for HDAC action are not yet fully understood.

-catenin is an important player in Wnt signaling that has been associated with tumor formation and activation is required for stabilization of -catenin in the cytoplasm, which then translocates and accumulates in the nucleus(16), activating downstream target genes such as MMP2 and MMP9(17). Silence of both HDAC2 and HDAC1 results in stabilization and nuclear translocation of
-catenin in oligodendrocyte lineage cells(10). HDAC4 contributes to glucocorticoid-induced loss of differentiation capacity and β-catenin acetylation in osteoblasts(18). Inactivation of HDAC6 by siRNA inhibits EGF-induced -catenin nuclear localization, leading to inhibition of colon cancer cell growth(15). Therefore, we hypothesize that HDAC function may involve in the regulation of Wnt/ -catenin signaling through deacetylation and nuclear localization of -catenin.
ARHGAPs are negative regulators of Rho family proteins, which can recognize GTP and induce GTP hydrolysis to produce GDP, and have been associated with the tumorigenesis such as head and neck squamous cell carcinoma, glioblastoma, breast, lung, colon and prostate cancer(19-21). ARHGAP4 is a novel RhoGAP inhibiting axon outgrowth and cell motility(22), overexpressed in colorectal cancer, and correlated with T stage, N stage, clinical stage, and metastasis(21). Our bioinformatics demonstrate the significant correlation of ARHGAP4 expression with PC, metastasis, Wnt signaling pathway, and HDACs, but the molecular mechanism by which ARHGAP4 regulates PC tumorigenesis has not been confirmed yet.
The present study found ARHGAP4 down-regulated while HDAC2 up-regulated in PC tissues. ARHGAP4 inhibited PC cell migration and invasion as well as the expression of MMP2 and MMP9 through acetylation and nuclear localization of -catenin via the ubiquitination of HDAC2. Thus,

ARHGAP4/HDAC2/ -catenin/MMP2 and MMP9 may be of great importance in regulating cell migration and invasion during PC development suggesting the clinical potential of ARHGAP4 in the prognosis of PC.

Materials and methods

Study subjects

The human PC tissue microarrays used in this study were prepared by Shanghai Outdo Biotech Co., Ltd. (Shanghai, China), including 46 men and 30 women, ages 43 to 85 years (median, 62 years), which stage I (n=4), II (n=59), and III (n=13) diseases according to the criteria of the 8th edition American joint committee on cancer (AJCC) TNM staging system. Clinicopathologic characteristics including tumor size, tumor site, pathological classification, pathological stage, vascular invasion, and survival time were also collected.
Bioinformatics

RNA-sequencing dataset of PC cohort was downloaded from The Cancer Genome Atlas (TCGA, https://tcga-data.nci.nih.gov/tcga/) and gene set enrichment analysis (GSEA) algorithm was used
to identify the pathways that were significantly enriched between ARHGAP4 high and low(23).

Cell culture

Capan-1, Capan-2, CFPAC-1, BxPC-3, SW1990, AsPC-1, PANC-1, and HPAC human PC cell lines were purchased from the American Type Culture Collection (ATCC, Manassas, VA, USA). Capan-1,

Capan-2, CFPAC-1, AsPC-1 and BxPC-3 cells were cultured in 6-well plates (Cellstar, Greiner Bio-one, Germany) with RPMI-1640. SW1990, HPAC and PANC-1 were maintained in DMEM (high glucose). All of the PC cell lines were grown in the medium supplemented with 10% fetal bovine serum (FBS), 100 U/ml penicillin, and 100 μg/ml streptomycin at 37°C with 5% CO2.

Cell line authentication

All cell lines were characterized in the past 6 months by using short tandem repeat (STR) based assays of genomic DNA from the cell lines. Cell lines were periodically confirmed negative for mycoplasma contamination using PCR assays.

Lentivirus production and transduction

ARHGAP4 coding sequence designed and synthesized by Sangon Biotech Co., Ltd. (China) was cloned into the pLVX-Puro lentivirus (pLVX-Puro-ARHGAP4) and blank pLVX-Puro lentivirus (blank vector) was used as negative control. To generate high-titer lentivirus, vectors encoding the target gene and the packaging plasmids were cotransfected into HEK 293T cells using lipofectamine 2000 (Invitrogen Life Technologies, USA) according to the manufacture’s instruction. The packaging, purification and titration of lentivirus were performed as described previously(24). 48 h after transfection, viral particles in cell culture medium were collected and infected BxPC-3 and SW1990 cells.

siRNA transfection

PANC-1 cells were transfected with three siRNAs (siRNA-1, position 42-60,

5’-GGCTGAGTATGAGACGCAA-3’; siRNA-2, position 48-66, 5’-GTATGAGACGCAAGTCAAA-3’;

siRNA-3, position 1697-1715, 5’-GCTGCATTCGCTTCATCAA-3’) targeting human ARHGAP4 or scramble siRNA (siNC) by using Lipofectamine 2000 following the instruction of the manufacturer. 48 h after transfection, PANC-1 cells were collected for further experiments.
Cell proliferation

Cells were seeded onto 96-well plates and treated as follows: BxPC-3 and SW1990 cells were infected with pLVX-Puro-ARHGAP4 or blank vector; PANC-1 cells were transfected with ARHGAP4 siRNA or siNC. At 0 h, 24 h, 48 h and 72 h after treatment, the culture medium were replaced with 10% CCK-8 solution (SAB Biotech., College Park, MD, USA) and the cells were cultured at 37°C for 1 h. Optical density at wavelength 450 nm (OD450) was measured using a microplate reader.
Transwell assay

BxPC-3, SW1990 and PANC-1 cells with the density of 3×105 cell/well were grown in 6-well plates and maintained at 37°C for 24 h. SW1990 and BxPC-3 cells were infected with pLVX-Puro-ARHGAP4 or blank vector. PANC-1 cells were transfected with ARHGAP4 siRNA or siNC in the presence of either -catenin inhibitor (XAV939, 10 M) or HDAC2 inhibitor (CAY10683, 100 nM). 24 h after treatment, cells were serum-starved in free-serum RPMI-1640 or DMEM for 24 h, and the 300 μL of cell suspension adjusted to 6 104 cells was filled in the upper Transwell chamber (8.0-µm with Size 24 Cluster Platel; Costar, USA) with (invasion assay) or without Matrigel coated (migration assay). 700 μL of complete medium containing 10% FBS were added

into the lower chamber, and then cultured for 24 h (migration assay) or 48 h (invasion assay) at 37°C. After incubation, cells were fixed with 1 mL of 4% methanol for 10 min, stained with 1 mL of 0.5% crystal violet for 30 min, photographed and counted in six view fields which were randomly selected under a microscope (Olympus Corporation, Japan) under a 200 × light microscope.
Gelatin zymography gel assay

For the analysis of MMP2 and MMP9 activities in PC cell lines, 25 g of proteins were separated on a 10 % sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) containing 1 % gelatin (Sigma-Aldrich, München, Germany). The gels were incubated at 37°C overnight, stained with Coomassie Blue, de-stained, and then scanned. Signal intensities were quantified by densitometry using Bio-1D software version 15.01 (Vilber Lourmat, Eberhardzell, Germany).
Real-time PCR

Total RNA was isolated from PC cell lines, SW1990 and BxPC-3 cells that infected with pLVX-Puro-ARHGAP4 or blank vector, and PANC-1 cells that transfected with ARHGAP4 siRNA or siNC using Trizol reagent (Invitrogen, Carlsbad, CA, USA). The reverse transcription was performed with Superscript III reverse transcriptase kit (Life technologies, USA). PCR amplification was performed using SYBR green PCR master mix (Applied Biosystems, Foster, CA, USA) in an ABI 9700 real-time PCR system (Applied Biosystem). The primers used in the present study were subsequently shown: ARHGAP4-F, 5’-CTACAACCTGGCCGTGTGCTTC-3’ and ARHGAP4-R, 5’-CTTCCAGCTCCGGCTCATTGTC-3’; HDAC1-F, 5’-GCTCCACATCAGTCCTTCC-3’ and HDAC1-R,

5’-GGTCGTCTTCGTCCTCATC-3’; HDAC2-F, 5’-AGGCAAATACTATGCTGTC-3’ and HDAC2-R,
5’-TGAAACAACCCAGTCTATC-3’; HDAC3-F, 5’-CGGGATGGCATTGATGAC-3’ and HDAC3-R,
5’-GGGCAACATTTCGGACAG-3’; GAPDH-F, 5’-AATCCCATCACCATCTTC-3’ and GAPDH-R,

5’-AGGCTGTTGTCATACTTC-3’. The method 2- Ct (25) was used to determine the fold changes of ARHGAP4, HDAC1, HDAC2, and HDAC3 mRNA, meanwhile GAPDH expression was chosen usually as a normalized reference.
Western blotting

Cellular proteins were extracted from PC cell lines, SW1990 and BxPC-3 cells that infected with pLVX-Puro-ARHGAP4 or blank vector, and PANC-1 cells that transfected with ARHGAP4 siRNA-1 or siNC in the presence of either -catenin inhibitor (XAV939, 10 M) or HDAC2 inhibitor (CAY10683, 100 nM) by lysis buffer, and protein concentrations were measured with the BCA Protein Assay Kit (Thermo Scientific, USA). The cytoplasmic and nuclear fractions were prepared as previously described.(26) Proteins that had been adjusted to equal content were fractionated in 10% SDS-PAGE. After transferring protein to a polyvinylidene fluoride (PVDF) membrane (Pall Corporation, NY, USA), the membranes were blocked in 5% non-fat milk and then incubated with 1:1000 dilution primary antibodies against ARHGAP4 (ab96251), MMP2 (ab14311), MMP9 (ab137867), HDAC1 (ab19845), HDAC2 (ab137364), HDAC3 (ab32369; all from Abcam, USA),

acetylated lysine (#9441), -catenin (#8480) and GAPDH (#5174; both from Cell Signaling Technology, USA) at 4°C overnight, washed three times with each time for 10 min, and incubated later with horseradish peroxidase (HRP)-labeled Goat Anti-Rabbit IgG (Beyotime Biotechnology, China) for 2 h at room temperature. The emitted signal of target proteins was detected by the

enhanced chemi-luminescence reagents (Thermo Scientific) and analyzed using Image J 1.44 software (National Institute of Health, Bethesda, MD, USA).

Immunohistochemistry (IHC)

Tissues were fixed in 4% paraformaldehyde, dehydrated and embedded in paraffin. Five-micrometer-thick slices were incubated in 3% H2O2 in methanol and 5% normal horse serum to minimize nonspecific staining. Sections were incubated at 4°C overnight with the primary antibody: rabbit anti-ARHGAP4 (1:100; ab198696; Abcam, USA) or rabbit anti-HDAC2 (1: 500; ab32117; Abcam). Next, the slices were incubated with secondary biotinylated goat anti-rabbit IgG (1:100; ab64256; Abcam) at room temperature for 20-30 min. Subsequently, the sections were stained with diaminobenzidine (DAB), counterstained with hematoxylin for 3 min and washed in water for 10 min. The tumor cells with positive stain more than 25% were defined as higher ARHGAP4 expression and that less than 25% were defined as lower ARHGAP4 expression.
Immunofluorescence (IF) microscopy

PC cell lines or tissues were fixed with 4% formaldehyde and permeabilized with 0.5% Triton X-100 in PBS. After blocking with 1% bovine serum albumin (BSA) in PBS for 30 min, cells were incubated with anti- -catenin antibody (Abcam; ab6302), anti-HDAC2 antibody (Abcam; ab51832), anti-ARHGAP4 antibody (Abcam; ab96251), Alexa Fluor 488-labeled Goat Anti-rabbit IgG (H+L) antibody (Beyotime Biotechnology; A0423), Alexa Fluor 488-labeled Goat Anti-Mouse IgG (H+L) antibody (Beyotime Biotechnology; A0428) and Alexa Fluor 555-labeled Donkey Anti-Rabbit IgG (H+L) antibody (Beyotime Biotechnology; A0453) before staining the nuclei with

DAPI-containing hard-set media (Vectashield, Vector Labs, Burlingame, CA, USA). Laser scanning confocal microscopy (Leica Microsystems Inc., Buffalo Grove, IL, USA) was used to examine stained cells.
Immunoprecipitation and ubiquitination assay

Whole cell lysates obtained by centrifugation were incubated with 2 g anti-ARHGAP4 (ab226835; Abcam), anti-HDAC2 (ab32117; Abcam), or normal IgG antibody (sc-2027; Santa Cruz, USA) and Protein A/G PLUS-Agarose beads (sc-2003; Santa Cruz) at 4°C overnight. The immunocomplexes were then separated by SDS-PAGE and blotted with anti-ARHGAP4 (ab96251; Abcam), anti-HDAC2 (ab137364; Abcam), and anti-ubiquitin (ab7780; Abcam) antibody.
Animal experiments

A total of 1×106 BxPC-3 cells infected with pLVX-Puro-ARHGAP4 or blank vector were harvested and resuspended in 100 L of PBS, and then intravenously injected through the tail vein into 5-6 week-old female BALB/c nude mice (weight: 16-18 g; 6 per group). Mice were sacrificed at 21 days after injection, and the lungs tissues removed from the xenograft mice were harvested and stained in hematoxylin and eosin (H&E). Animal experiments were approved by the institutional ethical committee of the Fudan University Shanghai Cancer Center (Shanghai, China) and performed according to the legal requirements.
Statistical analysis

Data were represented as mean ± standard error of the mean and analyzed with GraphPad Prism

5 (GraphPad Software, La Jolla, CA, USA). Cells were seeded in triplicates for each group and the experiment was independently repeated thrice. The results were analyzed by analysis of variance (ANOVA) or Student’s t-test followed by unpaired comparison. P < 0.05 was considered as statistical significance.
Results

Higher ARHGAP4 expression correlates with favorable prognosis in PC

To examine ARHGAP4 expression patterns in PC, we examined ARHGAP4 protein expressions by immunohistochemical staining on tissue microarrays containing 76 cases of PC and 70 cases of adjacent-normal pancreatic tissues. We observed that ARHGAP4 protein expressions were significantly decreased in PC tissues compared with adjacent-normal pancreatic tissues (Figure 1A and 1B). 29 cases of tumor tissues showed higher ARHGAP4 expression, and other 47 cases of tumor tissues showed lower ARHGAP4 expression (Figure 1A and 1B). As shown in Table 1, ARHGAP4 expression was correlated with pathological stages and vascular invasion, but not with gender, age, tumor size, TNM stages, pathological classification and tumor site. Kaplan-Meier survival analysis revealed that low ARHGAP4 expression levels in tumors were significantly associated with reduced survival in PC patients (Figure 1C).
To determine the biological effects of ARHGAP4 on PC tumorigenesis, ARHGAP4 expression in PC cell lines was measured. As shown in Figure 1D and 1E, PANC-1 cells showed a higher ARHGAP4 expression, while BxPC-3 as well as SW1990 cells demonstrated a lower ARHGAP4 expression,

compared with the other PC cell lines. We therefore infected pLVX-Puro-ARHGAP4 lentivirus or blank pLVX-Puro lentivirus (vector) into BxPC-3 and SW1990 cells. ARHGAP4 overexpression significantly induced ARHGAP4 mRNA and protein expression by 27.4 fold and 1.72 fold in BxPC-3 and by 13.6 fold and 1.47 fold in SW1990 cells (Figure S1A and S1B).
ARHGAP4 overexpression inhibited PC cell metastasis in vitro and in vivo

In view of the GSEA data that higher expressions of ARHGAP4 were negatively correlated with Metastasis up pathway (Figure S2A), we then evaluated the effects of ARHGAP4 overexpression or silencing on cell migration and invasion of BxPC-3 and SW1990 cells. Transwell analysis demonstrated that overexpression of ARHGAP4 markedly reduced the migration and invasion of BxPC-3 cells by 59.8% and 52.8% (Figure 2A and 2B) and of SW1990 cells by 50.7% and 42.2% (Figure S3A and S3B) compared with vector groups. However, overexpression of ARHGAP4 inhibited cell proliferation of BxPC-3 and SW1990 by 25.9% and 22.9% at 48 h, and by 32.4% and 26.9% at 72 h, respectively, compared with vector groups (Figure 2C and S3C).
We further investigated whether ARHGAP4 overexpression could suppress PC cell metastasis in vivo. Tumor metastasis was examined after intravenously injected with BxPC-3 cells that infected with pLVX-Puro-ARHGAP4 lentivirus or blank pLVX-Puro lentivirus (vector) into the tail vein of nude mice. Compared with the mice injected with BxPC-3 cells upon vector infection, less tumor cell metastasis was found in the lung tissues of the nude mice 21 days after they were injected with BxPC-3 cells upon pLVX-Puro-ARHGAP4 lentivirus infection (Figure 2D).

In view of the GSEA data that higher expressions of ARHGAP4 were also negatively correlated with KEGG Wnt/ -catenin Signaling Pathway, Senese HDAC1 Targets Up, Senese HDAC2 Targets Up, and Senese HDAC3 Targets Up pathways (Figure S2B-E), we then evaluated the effects of ARHGAP4 overxpression on the expression of HDAC1, HDAC2, HDAC3, -catenin, MMP2, and MMP9 in BxPC-3 or SW1990 cells. ARHGAP4 overxpression in BxPC-3 cells had no effect on the mRNA levels of HDAC1, HDAC2 and HDAC3, but significantly inhibited HDAC2 and HDAC3 protein levels (Figure S1D-F). It also decreased the localization of -catenin in nucleus of BxPC-3 cells (Figure 2E), induced the acetyl- -catenin level in cytoplasm, and inhibited the levels of -catenin in nucleus (Figure 2F). Moreover, ARHGAP4 overexpression also significantly inhibited the expression level and activity of MMP2 and MMP9 (Figure 2G and 2H). Similar results were also found in SW1990 cells with ARHGAP4 overxpression (Figure S4A-D).
ARHGAP4 regulated PC cell metastasis through HDAC2/ -catenin pathway

To further investigate the molecular mechanism by which ARHGAP4 regulates PC cell migration and invasion, PANC-1 cells that expressed higher levels of ARHGAP4 were also introduced and transfected with ARHGAP4 siRNA or scramble siRNA (siNC) in the absence or presence of either
-catenin inhibitor (XAV939) or HDAC2 inhibitor (CAY10683). ARHGAP4 silencing with siRNA-1, siRNA-2, or siRNA-3 significantly inhibited ARHGAP4 mRNA expression by 96.1%, 96.1% and 71.5% and ARHGAP4 protein expression by 84.2%, 74.9% and 44.1% in PANC-1 cells (Figure S1C). ARHGAP4 siRNA-1 was therefore chosen for further experiments. As shown in Figure 3A-C, ARHGAP4 silencing induced PANC-1 cell migration and invasion, while XAV939 or CAY10683 treatment demonstrated an inverse effect. Meanwhile, ARHGAP4 silencing significantly induced

PANC-1 cell proliferation at 48 and 72 h, respectively, compared with siNC group (Figure 3D and 3E). However, XAV939 or CAY10683 treatment attenuated the effects of ARHGAP4 silencing on PANC-1 cells. These data suggest that ARHGAP4 silencing promoted PC cell migration and invasion in vitro.
Subsequently, we found that ARHGAP4 silencing also increased the localization of -catenin in nucleus of PANC-1 cells (Figure 4A), inhibited the acetyl- -catenin level in cytoplasm, and induced the nuclear level of -catenin (Figure 4B). Moreover, ARHGAP4 silencing also significantly increased the expression level and activity of MMP2 and MMP9 (Figure 4C and 4D). However, CAY10683 treatment demonstrated an inverse effect and attenuated the effects of ARHGAP4 silencing on PANC-1 cells (Figure 4A-D). These results indicate that HDAC2/ -catenin signaling is a possible downstream effector of ARHGAP4 in PC.

ARHGAP4 ubiquitinated HDAC2 and correlation between ARHGAP4 and HDAC2 in PC

Given the role of ARHGAP4 in regulating the HDAC2 expression and the genes of Ubiquitin Mediated Proteolysis pathway (Figure S1D-S1F and S2F), we suggested that ARHGAP4 may regulate HDAC2 expression through ubiquitination. As shown in Figure 5A, ARHGAP4 coimmunoprecipitated with HDAC2 and, reciprocal immunoprecipitation with HDAC2 antibodies also brought down ARHGAP4, suggesting an interaction between ARHGAP4 and HDAC2 in cells. Moreover, ARHGAP4 overexpression also promoted ubiquitination of HDAC2 (Figure 5B).
To examine HDAC2 expression patterns in PC, we examined HDAC2 protein expressions by

immunohistochemical staining on tissue microarrays in 76 cases of PC and 70 cases of adjacent-normal pancreatic tissues. We observed that HDAC2 protein expressions were significantly increased in PC tissues compared with adjacent-normal pancreatic tissues (Figure 5C and 5D). Moreover, immunofluorescence followed by Pearson correlation analysis in PC tissues showed that ARHGAP4 protein was negatively correlated with HDAC2 protein in PC tissues (Figure 5E and 5F). These data further supported the findings in PC cell lines.

Discussion

ARHGAP4 is a member of the RhoGAP family, and has a SH3 structure at the C-terminal that interacts with proline rich domains of many proteins, involving in cell scaffold formation, intracellular vesicle transport and endocytosis(27). ARHGAP4 mutation was found in patients with systemic lupus erythematosus(28). The role of RhoGAP proteins has been confirmed in development and progression of cancers(19,20), while the involvements of ARHGAP4 have not been fully concerned. Here, ARHGAP4 was down-regulated in PC tissues compared with adjacent normal pancreatic tissues, while HDAC2 was up-regulated, and associated with the prognosis of patients with PC. ARHGAP4 inhibited PC cell invasion and migration, and also interacted with and ubiquitinated HDAC2. The molecular mechanism by which ARHGAP4 regulated PC tumorigenesis is through the HDAC2/ -catenin/MMP2 and MMP9 signaling pathway.
ARHGAP4 was up-regulated in human colorectal cancer tissues compared with non-tumor tissues and had been characterized as a novel prognosis marker in colorectal cancer, correlated with T

stage, N stage, clinical stage, and metastasis(21). However, results of this study demonstrated that ARHGAP4 was down-regulated in PC tissues, and its lower expression was associated with advanced pathological stages, vascular invasion and poor survival of patients with PC, but not with gender, age, tumor size, TNM stages, pathological classification, and tumor site. ARHGAP4 seemed to be of great importance in the suppression of PC cell invasion and migration as supported by the overexpression of full-length ARHGAP4 in SW1990 and BxPC-3 cells. This suppressed effect is further confirmed by the induction of migration and invasion following a knockdown of endogenous ARHGAP4 expression in PANC-1 cells using siRNA. Additionally, ARHGAP4 overxpression in BxPC-3 cells also inhibited the lung metastasis in nude mice. These results indicate that ARHGAP4 activation contributes to the decrease of cell motility. This was further supported by similar results from GSEA analysis that ARHGAP4 high expression was negatively correlated with the Metastasis up pathway and from the correlation between ARHGAP4 expression and clinicopathological parameters of PC patients such as pathological stages and vascular invasion, which are considered as independent prognosticators for metastasis(29). Similarly, ARHGAP4 inhibited the functions of RhoGTPase such as Rac1 and Cdc42 that are typically associated with inducing cell motility(22).
ARHGAP4 overexpression significantly decreased -catenin nuclear localization and the expression and activity of its downstream effector MMP2 and MMP-9 in BxPC-3 and SW1990 cells, suggesting the inhibitory effect of ARHGAP4 on activating -catenin signaling pathway(16,17). Our results also revealed the cytoplasmic accumulation of acetylated -catenin after ARHGAP4 overexpression, suggesting a role of acetylation in -catenin-dependent MMP2

and MMP-9 expression. However, increase of nuclear accumulation and nuclear acetylation of

-catenin could enhance Wnt signaling in cancer(30). These differences may due to the different cell fractions such as cytoplasm and nucleus used to detect the levels of acetylated -catenin. ARHGAP4 knockdown in PANC-1 cells showed an inverse effect. Interestingly, this ARHGAP4 siRNA-enhanced cell migration and invasion was inhibited by treatment with the -catenin inhibitor XAV939, suggesting that the anti-motility effects of ARHGAP4 are mediated by its ability to suppress the transcriptional activity of endogenous -catenin and that this may require deacetylase activity. This was further supported by similar results from GSEA analysis that ARHGAP4 high expression was negatively correlated with the Wnt/ -catenin signaling pathway.
Given the correlation of ARHGAP4 with the Senese HDAC1 Targets Up, Senese HDAC2 Targets Up, and Senese HDAC3 Targets Up pathways, the expression of HDAC1, HDAC2 and HDAC3 in BxPC-3 cells with ARHGAP4 overexpression was also examined. ARHGAP4 overexpression significantly inhibited HDAC2 and HDAC3 protein levels, but did not influence their mRNA levels and HDAC1 mRNA and protein levels. The regulation of HDAC2 in SW1990 cells with ARHGAP4 overexpression was also observed. In the canonical Wnt/ -catenin pathway, recruitment of multiple chromatin remodeling components such as HDACs is required for TCF factor transduces signals to activate downstream target genes(10,15,18). Moreover, HDAC2 inhibitor CAY10683 treatment in PANC-1 cells showed decreased -catenin nuclear localization and its downstream effector MMP2 and MMP-9 expression as well as increased cytoplasmic accumulation of acetylated -catenin, suggesting that HDAC2 promotes the cytoplasmic deacetylation and nuclear localization of β-catenin, thereby potentiating -catenin transcriptional activity. In consistent with

our findings HDAC6 has been found to deacetylate -catenin and induce -catenin nuclear localization in colon cancer cells(15). SIRT1 inhibits β-catenin by promoting the deacetylation of β-catenin and inhibiting nuclear localization of β-catenin(31). Interestingly, this ARHGAP4 siRNA-enhanced cell migration and invasion as well as HDAC2/ -catenin activation was inhibited by treatment with the HDAC2 inhibitor CAY10683, suggesting that ARHGAP4 regulates PC cell migration and invasion through HDAC2/ -catenin/MMP2 and MMP-9.
ARHGAP4 expression significantly regulated HDAC2 protein level but not mRNA level, and our GSEA analysis showed the correlation between ARHGAP4 and Ubiquitin Mediated Proteolysis pathway. We hypothesize that ARHGAP4 may regulate HDAC2 through ubiquitin-dependent manner, which was confirmed in our immunoprecipitation and ubiquitination assay. Recruitment of HDAC2 down-regulated the expression of E-cadherin(32), which regulated metastasis of pancreatic cancer(33), and attenuated TRAIL-induced apoptosis of pancreatic cancer cells(34), suggesting an important role of HDAC2 in PC carcinogenesis. Up-regulated expression of HDAC2 was observed in PC tissues compared to adjacent-normal pancreatic tissues, while ARHGAP4 expression was negatively correlated with HDAC2 expression, which further supported the findings in PC cell lines.
In summary, this study preliminarily provides insights into the anti-cancer effect of ARHGAP4 in PC cell lines through HDAC2/ -catenin signaling pathway. More importantly, our study suggests the clinical value of ARHGAP4 as a prognostic marker in PC.

Acknowledgements

This study was supported by the National Natural Science Foundation of China (81573757).

Conflict of interest

The authors declare that they have no conflict of interest.

References

1. Siegel, R.L., et al. (2016) Cancer statistics, 2016. CA Cancer J. Clin., 66, 7-30.

2. Vincent, A., et al. (2011) Pancreatic cancer. Lancet, 378, 607-620.

3. Wolfgang, C.L., et al. (2013) Recent progress in pancreatic cancer. CA Cancer J. Clin., 63, 318-348.
4. Lin, Q.J., et al. (2015) Current status and progress of pancreatic cancer in China. World J. Gastroenterol., 21, 7988-8003.
5. Muniraj, T., et al. (2013) Pancreatic cancer: a comprehensive review and update. Dis. Mon., 59, 368-402.
6. Mohammed, S., et al. (2014) Pancreatic cancer: advances in treatment. World J. Gastroenterol., 20, 9354-9360.
7. Delitto, D., et al. (2017) Human Pancreatic Cancer Cells Induce a MyD88-Dependent Stromal Response to Promote a Tumor-Tolerant Immune Microenvironment. Cancer Res., 77, 672-683.
8. Rucki, A.A., et al. (2017) Heterogeneous Stromal Signaling within the Tumor

Microenvironment Controls the Metastasis of Pancreatic Cancer. Cancer Res., 77, 41-52.

9. Gharibi, A., et al. (2016) Cellular and molecular aspects of pancreatic cancer. Acta Histochem., 118, 305-316.
10. Ye, F., et al. (2009) HDAC1 and HDAC2 regulate oligodendrocyte differentiation by disrupting the beta-catenin-TCF interaction. Nat. Neurosci., 12, 829-838.
11. Li, S., et al. (2018) Histone deacetylase 1 promotes glioblastoma cell proliferation and invasion via activation of PI3K/AKT and MEK/ERK signaling pathways. Brain Res., 1692, 154-162.
12. Huang, R., et al. (2018) Clinicopathological features and prediction values of HDAC1, HDAC2, HDAC3, and HDAC11 in classical Hodgkin lymphoma. Anticancer Drugs, 29, 364-370.
13. Li, Y., et al. (2016) HDACs and HDAC Inhibitors in Cancer Development and Therapy. Cold Spring Harb. Perspect. Med., 6, a026831.
14. Kiweler, N., et al. (2018) The histone deacetylases HDAC1 and HDAC2 are required for the growth and survival of renal carcinoma cells. Arch. Toxicol., 92, 2227-2243.
15. Li, Y., et al. (2008) HDAC6 is required for epidermal growth factor-induced beta-catenin nuclear localization. J. Biol. Chem., 283, 12686-12690.
16. Tortelote, G.G., et al. (2017) Complexity of the Wnt/betacatenin pathway: Searching for an activation model. Cell. Signal., 40, 30-43.
17. Wang, L.H., et al. (2018) The Antihelminthic Niclosamide Inhibits Cancer Stemness, Extracellular Matrix Remodeling, and Metastasis through Dysregulation of the Nuclear beta-catenin/c-Myc axis in OSCC. Sci. Rep. 8, 12776.
18. Ko, J.Y., et al. (2013) MicroRNA-29a ameliorates glucocorticoid-induced suppression of

osteoblast differentiation by regulating beta-catenin acetylation. Bone, 57, 468-475.

19. Wang, J., et al. (2014) Downregulation of miR-486-5p contributes to tumor progression and metastasis by targeting protumorigenic ARHGAP5 in lung cancer. Oncogene, 33, 1181-1189.
20. Lazarini, M., et al. (2013) ARHGAP21 is a RhoGAP for RhoA and RhoC with a role in proliferation and migration of prostate adenocarcinoma cells. Biochim. Biophys. Acta., 1832, 365-374.
21. Xuehu, X.U., et al. (2017) Expression and clinical significance of ARHGAP4 in colorectal cancer.

J. Pract. Med. 33, 705-708.

22. Vogt, D.L., et al. (2007) ARHGAP4 is a novel RhoGAP that mediates inhibition of cell motility and axon outgrowth. Mol. Cell Neurosci., 36, 332-342.
23. Yang, X., et al. (2013) mir-30d Regulates multiple genes in the autophagy pathway and impairs autophagy process in human cancer cells. Biochem. Biophys. Res. Commun., 431, 617-622.
24. Wu, Z., et al. (2018) Serine/Threonine Kinase 35, a Target Gene of STAT3, Regulates the Proliferation and Apoptosis of Osteosarcoma Cells. Cell. Physiol. Biochem., 45, 808-818.
25. Livak, K.J., et al. (2001) Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods, 25, 402-408.
26. Dignam, J.D., et al. (1983) Accurate transcription initiation by RNA polymerase II in a soluble extract from isolated mammalian nuclei. Nucleic Acids Res., 11, 1475-1789.
27. Fujimoto, M., et al. (2008) Immunological profile in a family with nephrogenic diabetes insipidus with a novel 11 kb deletion in AVPR2 and ARHGAP4 genes. BMC Med. Genet, 9, 42.
28. Kaufman, K.M., et al. (2013) Fine mapping of Xq28: both MECP2 and IRAK1 contribute to risk

for systemic lupus erythematosus in multiple ancestral groups. Ann. Rheum. Dis., 72, 437-444.
29. Slaton, J.W., et al. (2001) Tumor stage, vascular invasion and the percentage of poorly differentiated cancer: independent prognosticators for inguinal lymph node metastasis in penile squamous cancer. J. Urol., 165, 1138-1142.
30. Chocarro-Calvo, A., et al. (2013) Glucose-induced beta-catenin acetylation enhances Wnt signaling in cancer. Mol. Cell., 49, 474-486.
31. Firestein, R., et al. (2008) The SIRT1 deacetylase suppresses intestinal tumorigenesis and colon cancer growth. PLoS One, 3, e2020.
32. Aghdassi, A., et al. (2012) Recruitment of histone deacetylases HDAC1 and HDAC2 by the transcriptional repressor ZEB1 downregulates E-cadherin expression in pancreatic cancer. Gut, 61, 439-448.
33. von Burstin, J., et al. (2009) E-cadherin regulates metastasis of pancreatic cancer in vivo and is suppressed by a SNAIL/HDAC1/HDAC2 repressor complex. Gastroenterology, 137, 361-371, 371.e1-5.
34. Schuler, S., et al. (2010) HDAC2 attenuates TRAIL-induced apoptosis of pancreatic cancer cells.

Mol. Cancer, 9, 80.

Figure legends

Figure 1. Higher ARHGAP4 protein expression in PC patients and cell lines. (A, B) ARHGAP4 protein expressions were determined by immunohistochemical staining on tissue microarrays

containing tumor tissues (n=76) and adjacent-normal tissues (n=70) from PC patients. Scale bars: 100 μm. (C) Kaplan-Meier survival analysis revealed that low ARHGAP4 expression levels in tumors (n=47) were significantly associated with reduced survival in PC patients compared with high ARHGAP4 levels tumors (n=29). (D, E) ARHGAP4 mRNA and protein expressions were determined by Real-time PCR and Western blotting in PC cell lines. ARHGAP4 expression in BxPC-3, SW1990 and PANC-1 cells was analyzed by Real-time PCR and Western blotting.
***P<0.001 vs. Vector.

Figure 2. ARHGAP4 overexpression inhibits PC cell metastasis in vitro and in vivo. Cell migration and invasion (A, B) as well as cell viability (C) was determined by Transwell and CCK-8 analysis, respectively, in BxPC-3 with ARHGAP4 overexpression. Scale bars: 100 μm. (G) Histology of metastasized lungs from mice intravenously injected with BxPC-3 cells stably infected with pLVX-Puro-ARHGAP4 lentivirus or blank pLVX-Puro lentivirus (vector). Scale bars: 200 μm. BxPC-3 cells were infected with pLVX-Puro-ARHGAP4 lentivirus or blank pLVX-Puro lentivirus (vector), (E) the localization of -catenin was measured by immunofluorescence analysis immunostained with
-catenin (green) antibody and counter-stained with DAPI (blue), (F, G) expression of HDAC2, MMP and MMP9 protein expressions was determined by Western blotting, and (H) activity of MMP2 and MMP9 was measured by Gelatin zymography gel assay. Scale bars: 100 μm.
***P<0.001 vs. Vector.

 
Figure 3. ARHGAP4 silencing promotes PC cell metastasis through HDAC2/ -catenin signaling pathway. PANC-1 cells were transfected with ARHGAP4 siRNA-1 or scramble siRNA (siNC) in the

absence or presence of either -catenin inhibitor (XAV939, 10 M) or HDAC2 inhibitor (CAY10683, 100 nM), and (A-C) cell migration, cell invasion and (D, E) cell proliferation were measured by Transwell and CCK-8 analysis, respectively. Scale bars: 100 μm. ***P<0.001 vs. siNC. ###P<0.001 vs. siRNA-1.
Figure 4. ARHGAP4 silencing increases MMP2 and MMP9 levels trhough HDAC2/ -catenin signaling pathway. PANC-1 cells were transfected with ARHGAP4 siRNA-1 or scramble siRNA (siNC) in the present of either -catenin inhibitor (XAV939, 10 M) or HDAC2 inhibitor (CAY10683, 100 nM), and (A) the localization of -catenin was measured by immunofluorescence analysis immunostained with -catenin (green) antibody and counter-stained with DAPI (blue), (B, C) expression of HDAC2, MMP and MMP9 protein expressions was determined by Western blotting, and (D) activity of MMP2 and MMP9 was measured by Gelatin zymography gel assay. Scale bars: 100 μm. ***P<0.001 vs. siNC. ###P<0.001 vs. siRNA-1.
Figure 5. ARHGAP4 induces ubiquitination of and negatively correlated with HDAC2. (A) BxPC-3 cell lysates were subjected to immunoprecipitation with control IgG, anti-ARHGAP4, or anti-HDAC2 antibody. The immunoprecipitates were then blotted with the indicated antibodies.
(B) Cells were infected with pLVX-Puro-ARHGAP4 lentivirus or blank pLVX-Puro lentivirus (vector).

HDAC2 was immunoprecipitated and immunoblotted with the indicated antibodies. (C, D) HDAC2 protein expressions were determined by immunohistochemical staining on tissue microarrays containing tumor tissues (n=76) and adjacent-normal tissues (n=70) from PC patients. (E, F) PC tissues were immunostained with HDAC2 (green) and ARHGAP4 (red) antibodies, and

Downloaded from https://academic.oup.com/carcin/advance-article-abstract/doi/10.1093/carcin/bgz067/5431283 by Australian Catholic University user on 09 April 2019

Table 1. Correlation of the expression of ARHGAP4 with clinicopathological parameters in patients with PC.
Clinicopathological parameters
Protein expression of ARHGAP4
P -value
Low
(n=47, 61.8%) High
(n=29, 38.2%)
Gender
Male 31 (40.8%) 15 (19.7%) 0.2175
Female 16 (21.1%) 14 (18.4%)
Age (years)
62 21 (27.6%) 17(22.4%) 0.2377
62 26 (34.2%) 12 (15.8%)
Tumor size (cm)
4 29 (38.1%) 16 (21.1%) 0.5737
>4 18 (23.7%) 13 (17.1%)
TNM stages
I 2 (2.6%) 2 (2.6%) 0.7575
II 36 (47.4%) 23 (30.3%)
III 9 (11.8%) 4 (5.3%)
Pathological stages
I 2 (2.6%) 5 (6.7%) 0.0280*
II 21 (27.6%) 18 (23.7%)
III 21 (27.6%) 6 (7.9%)
IV 3 (3.9%) 0 (0.0%)
Pathological classification 0.2263
Ductal adenocarcinoma 42 (55.3%) 23 (30.3%)
Adenosquamous carcinoma 5 (6.5%) 6 (7.9%)
Vascular invasion 0.0251*
Yes 27 (35.6%) 9 (11.8%)
No 20 (26.3%) 20 (26.3%)
Tumor site 0.0614
Pancreatic head 21 (27.6%) 21 (27.6%)
Pancreatic body 13 (17.1%) 4 (5.3%)
Pancreatic tail 13 (17.1%) 4 (5.3%)
Differences between groups CAY10683 were done by the Chi-square test, *P<0.05