CD2AP inhibits metastasis in gastric cancer by promoting cellular adhesion and cytoskeleton assembly

Abstract Diffuse gastric cancer (DGC) is a lethal malignancy lacking effective systemic therapy. Among the most provocative recent results in DGC has been that the alter of the cellular cytoskeleton and intercellular adhesion. CD2‐associated protein (CD2AP) is one of the critical proteins regulating cytoskeleton assembly and intercellular adhesion. However, no study has investigated the expression and biological significance of CD2AP in gastric cancer (GC) to date. Therefore, the aim of our study was to explore if the expression of CD2AP is associated with any clinical features of GC and to elucidate the underlying mechanism. Immunohistochemistry of 620 patient tissue samples indicated that the expression of CD2AP is downregulated in DGC. Moreover, a low CD2AP level was indicative of poor patient prognosis. In vitro, forced expression of CD2AP caused a significant decrease in the migration and invasion of GC cells, whereas depletion of CD2AP had the opposite effect. Immunofluorescence analysis indicated that CD2AP promoted cellular adhesion and influenced cell cytoskeleton assembly via interaction with the F‐actin capping protein CAPZA1. Overall, the upregulation of CD2AP could attenuate GC metastasis, suggesting CD2AP as a novel biomarker for the prognosis and treatment of patients with GC.


| INTRODUCTION
Gastric cancer (GC) is one of the most prevalent cancers worldwide, especially in Asia, and is associated with a high mortality rate, with more than 700 000 people dying of the disease each year. 1 Most cases of GC are diagnosed at a late stage characterized by metastasis. Even with the currently best treatment, the survival of GC patients with distant metastases is often poor. [2][3][4] Despite great progress in identifying novel therapeutic targets for advanced or metastatic GC, the underlying molecular mechanisms are poorly understood, in part because of the biological heterogeneity of GC. 5 However, accumulating evidence indicates that a reduction of intercellular adhesion and alterations in the cytoskeleton structure are the major causes of the development of GC. 6 Moreover, several studies have shown that the intercellular adhesion-related molecules such as E-cadherin, ZO-1, CAPZA1, and SCIN can impact the prognosis of cancer patients. [7][8][9] Thus, identifying novel molecules or mutations involved in GC-related intercellular adhesion or cytoskeleton assembly is an important step for understanding the pathogenesis of GC metastasis.
CD2-associated protein (CD2AP) is an adhesion-related adapter protein that plays an important role in the formation of epithelial cell junctions, [10][11][12][13] and has been reported as a key factor in the pathogenesis of foot-related nephropathy. [14][15][16][17][18] The ablation of CD2AP led to a disordered tissue structure in the lamina propria cells of the gastric mucosa, suggesting that CD2AP plays an important role in maintaining the normal morphological structure of the gastric mucosa. 19 However, no study has investigated the expression and biological significance of CD2AP in GC to date. Therefore, in this study, we examined the expression of CD2AP in 564 GC clinical specimens and explored the role of CD2AP in the metastasis and proliferation of GC in vitro, with the aim of determining its potential application in the diagnosis and treatment of GC. assessed. The patients comprised 402 males and 162 females ranging in age from 20 to 86 years (median, 59 years). The study was approved by the Review Board of the Second Affiliated Hospital of Wenzhou Medical University. The tissue microarray (TMA) was constructed as described previously. 20 All the patients were aware of the research and signed the informed consent form.

| Immunohistochemistry
Immunohistochemistry on a tissue microarray, constructed as described previously, 20 was performed for the detection of CD2AP expression. In brief, the sections were dewaxed by incubation with dimethyl benzene at 45°C for 60 minutes and then immersed in distilled water. Endogenous peroxidase activity was inhibited by incubation in a 0.5% hydrogen peroxide bath for 10 minutes. After washing three times with 0.01 M phosphate-buffered saline (PBS, pH 7.4), the slides were immersed in citrate antigen retrieval buffer (Zhongshan Golden Bridge Biotechnology, Beijing, China). After blocking with sheep serum for 30 minutes, the sections were incubated with anti-CD2AP antibody (sc-25272; Santa Cruz Biotechnology, Dallas, TX; 1:50 dilution; Figure S1B) in a humidified chamber at 24°C for 2 hours. After washing three times with PBS, a Dako EnVision FLEX detection system (Dako, Carpinteria, CA) was used for visualization, according to the manufacturer's instructions.
The sections were counterstained with hematoxylin, dehydrated, and sealed with neutral gum.

| Quantification of CD2AP expression
The TMA was obtained using a digital slice scanner for the Whole Slide Image (Easyscan6; MOTIC Medical Diagnostic Systems, Fuzhou, China) and separated into single spots. Microarray spots with no tumor tissue, missing spots, and spots with the minimal valid area were then eliminated. The tumor-positive area was selected and the integrated optical density (IOD) of CD2AP was determined using Image Pro Plus 6 (IPP; Media Cybernetics, Rockville, MD). Finally, the CD2AP score (IOD/tumor-positive area) was calculated.

| Bioinformatics analysis
The Cancer Genome Atlas (TCGA) GC data were downloaded from https://tcga-data.nci.nih.gov/docs/publications/tcga. The data set contains survival data, clinical information, and messenger RNA (mRNA) expression levels. For validation, we obtained four independent microarray data sets (GSE26942, GSE2669, GSE15460, and GSE62254) containing cancer and paracancerous samples, and another four independent microarray data sets (GSE14208, GSE15459, GSE57303, and GSE62254) with prognostic information from GEO. The mRNA levels of the validation set were measured by the Affymetrix Human Genome U133 Plus 2.0 Array (Thermo Fisher, Massachusetts, CA) on a log2 scale. Gene expression in the discovery set was determined using the Illumina HiSeq platform and transformed to the log2 scale. According to the publication guidelines, the data sets may be used for publication without restriction or limitation (https://cancergenome.nih.gov/publications/publicationguidelines, https:// www.ncbi.nlm.nih.gov/geo/info/disclaimer.html).    Subsequently, the membranes were exposed to horseradish peroxidaselabeled IgG for 2 hours and the bands were visualized using a Bio-Rad imaging system.

| Immunofluorescence
Cell focal adhesion was detected by immunofluorescence. In brief, the expression of CD2AP was induced in MGC-803-CD2AP cells (2 × 10 5 cells/well in six-well plates containing glass coverslips) by treatment of DOX (1 µg/mL) for 48 hours. The cells were then fixed in 4% paraformaldehyde at 37°C for 10 minutes and permeabilized with 0.1% Triton X-100 at 37°C for 10 minutes.
After blocking with 10% goat serum (Beyotime) in PBS for XIE ET AL.

| 341
60 minutes at 37°C, the cells were incubated with a rabbit antibody specific to zyxin (Abcam, Cambridge, UK) and DyLight488-conjugated goat antirabbit antibody in PBS supplemented with 10% goat serum for 1 hour at 37°C. The cells were subsequently incubated with Hoechst 33258 (Beyotime) for 5 minutes at room temperature. The coverslips were mounted and the signals were visualized under a fluorescence microscope (Carl Zeiss, Jena, Germany).
The expression of CD2AP is significantly lower in diffuse gastric cancer. A, Immunohistochemical expression of CD2AP in IGC, DGC, paired normal tissue, normal gastric lamina propria, and normal gastric epithelium. B, Select the gland region by IPP, analyze the IOD of the CD2AP positive region and the total area of the gland, and calculate the CD2AP score. C, In the paired specimens, the expression of CD2AP in DGC tissues (n = 15) was significantly lower than that in the matched normal gastric mucosa (***P < .001), but not in IGC (n = 37) and mixed gastric cancer (n = 4) (P > .05). D, The expression of CD2AP in DGC is significantly lower than that in IGC, mixed type and normal gastric mucosa (***P < .001). E, Violin picture show CD2AP expression in DGC, IGC, and normal gastric tissue. For each Violin picture, median and ranges are indicated (*P < .05, ***P < .001). CD2AP, CD2-associated protein; DGC, Diffuse gastric cancer; IGC, intestinal gastric cancer; IOD, integrated optical density, IPP, Image Pro Plus 6 [Color figure can be viewed at wileyonlinelibrary.com] 2.12 | Co-immunoprecipitation

| Statistical analysis
All statistical analyses were performed using the SPSS 17.0 software package (IBM, Chicago, IL). Data are reported as means ± SEM. Analysis of variance and independent-sample t tests were performed to assess the differences between groups. The cutoff value of CD2AP expression in GC was determined by the Youden index based on the overall survivalspecific receiver operating curve. The CD2AP immunohistochemical scores and the CD2AP mRNA levels in GEO databases and TCGA database were then divided into high-expression and low-expression groups. Univariate survival analysis and Kaplan-Meier's analysis with a log-rank test were performed to construct survival curves. A value of P < .05 was considered statistically significant.

| The expression level of CD2AP is significantly lower in diffuse GC
We first used immunohistochemistry to examine the expression of CD2AP in the tissues from 564 patients with GC, including 56 adjacent noncancerous tissues. CD2AP was mainly expressed in the cytoplasm of the cancer cells, with some expression detected in the nucleus. Only a small amount of CD2AP was expressed in the stromal cells. In the normal gastric mucosa, CD2AP was predominantly expressed in the mucosal epithelial layer ( Figure 1A). The scores of CD2AP expression are shown in Figure 1B. Based on tissue structures and biological behaviors, GC can be categorized into intestinal gastric cancer (IGC) and diffuse gastric cancer (DGC). 22 The expression of CD2AP in these two tissue types was significantly distinct. In the paired tissue samples, the expression of CD2AP in DGC tissue (n = 15) was significantly lower than that in the paired normal tissues (P < .001); however, there was no significant difference in CD2AP expression scores between IGC (n = 37) and mixed GC (n = 4) tissues (P > .05) ( Figure 1C). Moreover, the 620 tissue samples were divided into IGC (n = 337), DGC (n = 197), mixed GC (n = 30), and normal gastric mucosa (n = 56). The expression of CD2AP in DGC was significantly lower than that in IGC, mixed type, or normal gastric mucosa (P < .001), but there was no significant difference between that in IGC, mixed GC, and normal gastric mucosa (P > .05) ( Figure 1D). Similarly, in the GEO data sets (GSE26942, GSE2669, GSE15460, and GSE62254), the expression of CD2AP in DGC was lower than that in IGC or normal gastric mucosa, but there was no significant difference in CD2AP expression between IGC and the normal gastric mucosa in these four data sets ( Figure 1E).  Figure 2A) and worse disease-free survival (DFS) rate ( Figure 2B). Considering that the expression of CD2AP in DGC was significantly lower than that in the paired normal tissues and that the expression of CD2AP in IGC was not significantly different from that in the paired normal tissues, we further analyzed CD2AP expression in DGC and found that  Table 2). In addition, a low level of CD2AP expression also indicated a poor prognosis in the GSE14208 (P = .010), GSE57303
This trend was also present in the GSE15459 data set, but the difference was not statistically significant (P = .064) ( Figure 2E-I).
These results suggest that CD2AP may act as a tumor suppressor in GC.

| Modulation of CD2AP expression alters cell migration and invasion
To examine how CD2AP influences the function of GC cells in vitro, we depleted CD2AP in BGC-823 and MGC-803 GC cells by siRNA ( Figure S2A), respectively. The results from Transwell assays showed  Figures 3D and 4E).

| Overexpression of CD2AP promoted cellular adhesion and cytoskeleton assembly in GC cells
CD2AP is an adhesion-related adapter protein that plays an important role in the formation of epithelial cell junctions. 23 The process of GC cell migration includes the extension of flaky pseudopods, formation of focal adhesions, and deagglomeration of tail adhesion spots. 24  expression of CD2AP did not impact the expression of EMT-related markers such as E-cadherin and vimentin ( Figure S3). Next, we induced the expression of CD2AP in MGC-803-CD2AP cells by DOX.
Immunofluorescence analyses with phalloidin labeling showed that overexpression of CD2AP enhanced the assembly of cytoskeleton Factins compared with that of the cells without treatment of DOX ( Figure 5A, P < .01). Moreover, as zyxin is a LIM domain protein localized mainly at focal adhesion plaques, 27 we used an antibody against zyxin as a marker to detect cell adhesion spots. The number of focal adhesion spots was significantly increased in MGC-803-CD2AP cells treated with DOX when compared to that of the cells without DOX treatment ( Figure 5B, P < .01). Moreover, the cell adhesion ability was significantly increased in MGC-803-CD2AP cells treated with DOX compared with that of the cells without treatment of DOX ( Figure 5A, P < .01).

| The interaction of CD2AP with CAPZA1 mediates cellular adhesion and cytoskeleton assembly
The F-actin network is required for presynaptic assembly and for specificity-determining adhesion spots. 28 CAPZA1 can regulate the structural stability of F-actin. 29 Hence, we transfected the plasmid, but not in the cells transfected with pcDNA3.1 empty plasmid ( Figure 6A). We also extracted MGC-803 cell proteins, which were incubated with anti-CAPZA1 antibody to form an antigen-antibody mixture, and immunoprecipitation with Protein A/G magnetic beads was carried out. Western blot analysis analyses confirmed that CAPZA1 could form a complex with CD2AP in the GC cells ( Figure 6B). These results suggested that CAPZA1 might act as a key mediator for CD2AP protein to promote cytoskeleton assembly and focal adhesion formation.
To examine how CAPZA1 influences the function of CD2AP of GC cells, we depleted CAPZA1 in MGC-803-CD2AP GC cells by siRNA ( Figure 6C). Transwell assays showed that the depletion of CAPZA1 RhoA, a member of the Rho family, is a small GTPase that plays a fundamental role in regulating diverse cellular processes, including cell junction assembly, cell-matrix adhesion, and cell migration.
RhoGAP plays an important role in regulating the activation of RhoA. [39][40][41][42] In 2014, a high RhoA mutation rate and RhoGAP fusion were found in GC samples in the TCGA, and these mutations were present almost exclusively in DGC. 43 The alteration in intercellular adhesion is an important feature of DGC. 44 CD2AP is an important adhesion-related adapter protein that plays a role in the formation of epithelial cell junctions. 10   shown that enhanced focal adhesion contributed to tumor progression by affecting the cytoarchitecture and motility, 27,50 which is consistent with our results. This could provide a mechanism for the inhibition of GC cell migration with a high expression of CD2AP.

| CONCLUSIONS
In summary, our study illustrates an important function of CD2AP in the malignant behavior of GC. We demonstrated that CD2AP was expressed at a significantly low level in DGC and that a low level of CD2AP expression in GC tissue correlated with a poor prognosis.
Furthermore, the downregulation of CD2AP expression enhanced GC metastasis by interacting with CAPZA1 to promote intercellular adhesion and to influence cell cytoskeleton. Therefore, CD2AP may represent a novel biomarker associated with a good prognosis for patients with GC.