Up-regulation of neogenin-1 increases cell proliferation and motility in gastric cancer

Although elevated expression of neogenin-1 has been detected in human gastric cancer tissue, its role in gastric tumorigenesis remains unclear due to the lack of neogenin-1 studies in cancer. Therefore, we demonstrated here the function and regulatory mechanism of neogenin-1 in gastric cancer. Neogenin-1 ablation decreased proliferation and migration of gastric cancer cells, whereas its over-expression reversed these effects. Xenografted analyses using gastric cancer cells displayed statistically significant inhibition of tumor growth by neogenin-1 depletion. Interestingly, galectin-3 interacted with HSF-1 directly, which facilitated nuclear-localization and binding on neogenin-1 promoter to drive its transcription and gastric cancer cell motility. The galectin-3-increased gastric cancer cell motility was down-regulated by HSF-1 depletion. Moreover, the parallel expression patterns of galectin-3 and neogenin-1, as well as those of HSF-1 and neogenin-1, were detected in the malignant tissues of gastric cancer patients. Taken together, high-expression of neogenin-1 promotes gastric cancer proliferation and motility and its expression is regulated by HSF-1 and galectin-3 interaction. In addition, we propose further studies for neogenin-1 and its associated pathways to provide them as a proper target for gastric cancer therapy.


INTRODUCTION
Gastric cancer is the second most common cause of cancer-related death in the world [1]. Despite recent advances in treatments involving surgery, chemotherapy, and radiation therapy, gastric cancer is difficult to cure unless it is found at an early stage [2]. Even patients who present with the most favorable characteristics and who undergo curative surgical resection often die of recurrent disease due to metastasis and drug resistance [3]. Therefore, to cure gastric cancer, we need to understand the molecular mechanisms behind gastric cancer metastasis and drug resistance.
Through previous studies (GSE 29630), we reported microarray data on galectin-3 silencing in gastric cancer cell lines [4] which indicated that several cell motilityrelated genes changed [5][6][7][8]. Among these changes, in the present study we focus on neogenin-1 expression. Neogenin-1 is a transmembrane receptor belonging to the immunoglobulin superfamily. It shares 50% aminoacid identity with the human tumor suppressor molecule deleted in colon cancer (DCC) [9], and is a netrin receptorlike DCC that binds repulsive guidance molecules (RGM); its functions as a RGM receptor have been studied [10][11]. While DCC expression is restricted to the nervous system, neogenin-1, a cell adhesion molecule, is widely expressed in a variety of developing tissues, such as in the gastrointestinal tract of chickens, as well as on the surface of developing neurons in the heart, lungs, and pancreas [12]. Moreover, due to its similarity to DCC, neogenin-1 was thought to mediate axon guidance in a netrin-dependent manner. However, the knockout of the neogenin-1 gene in mice did not affect axon guidance phenotypes [13], and its DCC-independent functions have been determined, including the regulation of bone formation, mammary gland morphogenesis, and skeletal myofiber development [14][15][16].
Neogenin-1 seems to be involved in tissue growth regulation, cell to cell recognition, cellular transition from proliferation to terminal differentiation, and cell migration. Although neogenin-1 expression has been detected in many adult tissues, there is no information available regarding the status of neogenin-1 expression in the human ovary. Neogenin-1 is present in tissues where active growth takes place, and overexpression of neogenin-1 has been observed in a wide variety of human cancers including those of the breast, pancreas, cervix, colon, medulloblastoma and rectum [17][18][19]. Previous studies have also shown that neogenin-1 may promote cell migration and cell-cell adhesion in human vascular smooth muscle in addition to zebrafish neurectoderm [9,20]. Also, neogenin-1 is involved in various aspects of cancer progression and metastasis [9]. It is expressed and has a role in colon cancer [21], breast cancer [22], pancreatic cancer [23], glioma [24], and esophageal squamous cell carcinoma (ESCC) [25]. Interestingly, its role and expression levels vary significantly among cancers affecting different organs. For example, low expression of neogenin-1 is seen in breast cancers, colon cancers and glioma, moderate expression in pancreatic cancer and high expression in ESCC compared to the same healthy organs. However, the role of neogenin-1 remains unclear because the opposite functions on cancer have been reported in several cancers [9,22,24].
In this study, we first determined the expression of neogenin-1 in gastric cancer patients and gastric cancer cells. The function of neogenin-1 was also determined by assessment of gain and loss of function of neogenin-1 in gastric cancer cell proliferation and migration. Moreover, we also determined that neogenin-1 expression is regulated by transcriptional factor HSF-1, mediated stress response and the activity of specific heat shock proteins (HSPs) [26][27][28]. HSF-1 expression is elevated in various cancers, including human prostate carcinoma cell lines, and has been shown to support proliferation of several cancer cell lines [27][28][29][30][31]. Finally, we assessed how galectin-3 regulates the expression of neogenin-1, specifically through interaction with HSF-1. Therefore, we demonstrate herein the effects of neogenin-1 on gastric cancer cell proliferation and migration, as well as the regulation of its expression by HSF-1 and galectin-3 interactions.

Detection of expression level of neogenin-1 in tissues of gastric cancer patients
We investigated the expression of neogenin-1 in 59 cases of gastric cancer tissues and normal tissues of gastric cancer patients (Fig. 1). The expression of neogenin-1 was detected in normal tissues, especially in cytoplasm of glandular cells (Fig. 1A). Although the case number of patients was not enough, we found the expression levels of neogenin-1 was higher in diffuse type than intestinal type of gastric cancer tissues ( Fig. 1B-C). Interestingly, neogenin-1 was highly detected in signet ring cell carcinoma of diffuse type gastric cancer tissues ( Fig. 1D-E), which has a metastatic property on the various cancer

Effect of the ablation of neogenin-1 or netrin-1 on gastric cancer cell proliferation and migration
Whereas neogenin-1 expression was highly detected in 10 of 12 gastric cancer cell lines, both ligand of neogenin-1, netrin-1 and RGMa were low expressed or unexpressed in most gastric cancer cell lines, respectively (Suppl. Fig. 1). We depleted the expression of neogenin-1 and netrin-1 by each siRNA to determine the effect on cell proliferation in six gastric cancer cell lines (Suppl. Fig. 2 and Fig. 2A). The cell proliferation was reduced by the ablation of neogenin-1, and the strongest reduction was detected in AGS cells and YCC-2 cells. However, the ablation of netrin-1 did not show any visible reduction of proliferation in six gastric cancer cell lines. The cell migration by ablation of neogenin-1 distinctively lessened the migration of 5 of the gastric cancer cells except SNU-216, whereas the ablation of netrin-1 had no significant reduction (Fig. 2B). Our results suppose that the effect of neogenin-1 on the proliferation and migration gastric cancer cells and is independent of its ligand netrin-1.

Effect of ablation of neogenin-1 in vivo in gastric cancer cell xenografted mice
We prepared two clones of neogenin-1 shRNA stable expressing AGS cells and inoculated in nude mice ( Fig. 2C-D). Compared to tumors growth made from negative control LacZ shRNA expressing AGS cells, the neogenin-1 ablation by both shRNAs showed reduced Six gastric cancer cell lines were transfected with scRNAs, neogenin-1 siRNA or netrin-1 siRNA, and analyzed by migration and invasion assays, with the results presented as histograms (* p<0.04). All experiments were performed in triplicate. (C) The effect of neogenin-1 on tumor growth of xenografts in nude mice. 50ul of AGS SQ cells (1 x 10 6 ) with Matrigel were implanted into Balb/c-nude mice to form subcutaneous xenografts. Tumor volumes were measured at different time points and are presented in photographs taken at day 28 and in a graph (D). Data are presented as the mean and standard error of the mean (SEM). Unpaired student's t-test was used for the comparison between the two groups (*, ** p<0.002).
tumor growth in xenografted mice. These data strongly suggest that neogenin-1 accelerates in vivo gastric cancer progression.

Transient transfection of neogenin-1 accelerated the proliferation, migration and invasion of SNU-668 gastric cancer cells
Neogenin-1 over-expression plasmids were transfected into SNU-668 cells, which demonstrate low expression of neogenin-1 in twelve gastric cancer cell lines ( Fig. 3A-C). In all cases, cell proliferation, migration and invasion were increased by neogenin-1 over-expression. Previously, it was reported that neogenin-1 regulates Rho/Rac/ROCK1 activity [35]. We reduced neogenin-1 expression, and then detected an associated decrease in both phosphorylated ROCK1 and total ROCK1 protein levels in AGS cells (Suppl. Fig. 3A). The depletion of ROCK1 also reduced AGS cell migration and invasion (Suppl. Fig. 3B). To establish whether neogenin-1 induced gastric cancer cell migration and invasion through ROCK1 activation, we over-expressed neogenin-1 in SNU-668 cells and decreased the expression of ROCK1 by ROCK1 siRNA. Then, cell migration and invasion assays were performed using these cells (Fig. 3D-E). As shown in Figure 3E, over-expression of neogenin-1 increased cell migration and invasion, whereas ROCK1 depletion reversed these effects. Moreover, the reduction in phosphorylation of ROCK1 was also detected in in vivo neogenin-1-depleted tumors of AGS cell-xenografted mice (Suppl. Fig. 4). Taken together, neogenin-1 appears to regulate ROCK1 activation leading to gastric cancer cell migration and invasion. Expression levels of neogenin-1 and ROCK1-1 were detected by RT-PCR and Western blot analysis in SNU-668 cells, which were infected with lentivirus-containing LacZ or neogenin-1 full-length, and then were transfected with scRNA or ROCK1 siRNA. β-actin was used as a loading control. (E) The cells were analyzed with cell migration and invasion assays, and data are presented as histograms (*, ** p<0.001). All experiments were performed in triplicate. www.impactjournals.com/oncotarget
To determine the effect of neogenin-1 on cell migration and invasion, we performed cell migration and invasion assays after treatment with each siRNA of galectin-3 and neogenin-1 in AGS cells (Fig. 4C-D). The galectin-3 or neogenin-1-depleted cells were no longer mobile, and the total number of migrated cells and invasive cells was reduced by almost half.

Transcriptional regulation of neogenin-1 by galectin-3 through interaction with HSF-1 transcriptional factor
In order to further explore how galectin-3 increases neogenin-1 expression in gastric cancer cells, we focused on the transcription factor, heat shock factor 1 (HSF-1), as a major regulator of heat shock protein transcription. We analyzed the binding site of HSF-1 in the neogenin-1 promoter (Fig. 5A), and examined the DNA binding activity of HSF-1 with or without galectin-3 by ChIP assay (Fig. 5B). In the presence of galectin-3, HSF-1 bound to the promoter regions of neogenin-1, but not in the absence of galectin-3. A HSF-1 luciferase assay was also performed to detect transcriptional activity of HSF-1 with or without galectin-3 expression (Fig. 5C). Transcriptional activity of HSF-1 was decreased in the absence of galectin-3. Therefore, we carried out an immunoprecipitation assay, and observed a strong interaction between galectin-3 and HSF-1 (Fig. 5D). Interestingly, the localization of HSF-1 was changed by galectin-3 (Fig. 5E). We separated the cytosol and nuclear fractions of cell lysates treated with galectin-3 siRNA or HSF-1 siRNA. In the absence of galectin-3, most HSF-1 expression was detected in the cytosol, suggesting less transcriptional activity. However, the absence of HSF-1 had no effect on the galectin-3 localization. These results suggest that galectin-3 contributes to the binding of the neogenin-1 promoter by interacting with HSF-1 protein, thereby increasing

HSF-1 promotes motility of human gastric cancer cells through up-regulation of neogenin-1 expression
Because HSF-1 increased neogenin-1 expression, we determined the regulatory effect of HSF-1 on migration of gastric cancer cells (Fig. 6). First, we examined the decreased expression of both neogenin-1 and HSP70 [36], a downstream target of HSF-1, after the depletion of HSF-1 or galectin-3 (Fig. 6A). The ablation of HSF-1 reduced the gastric cancer cell migration significantly (Fig. 6B). Moreover, increased expression of neogenin-1 by galectin-3 over-expression was reduced in the absence of HSF-1. HSP70 expression was also increased by galectin-3 over-expression, and was diminished in the absence of HSF-1 (Fig. 6C). The increased cell migration and invasion by galectin-3 overexpression decreased with HSF-1 depletion (Fig. 6D). Taken together, our results demonstrated that HSF-1 increases gastric cancer cell motility through regulation of neogenin-1 expression, and the transcriptional activation of HSF-1 is regulated by galectin-3.

Correlation between neogenin-1 and galectin-3 expression in malignant and normal tissues from gastric cancer patients
The expression levels of galectin-3 and neogenin-1 were analyzed in 20 gastric cancer patients obtained from the National Cancer Center of Korea. Total RNA of normal and cancer tissues from these patients were prepared and RT-PCR was performed (Suppl. Fig. 6). We quantified galectin-3 and neogenin-1 expression using ImageJ and created a graph (Fig. 7A). The expression levels of (scRNA), or HSF-1 siRNA, the migration assay were performed and quantification of cell migration was presented as a histogram (*, ** p<0.001). (C) Expression levels of galectin-3, HSF-1, HSP70 and neogenin-1 were detected by RT-PCR and Western blotting analysis in SNU-638 cells, which were infected with lentivirus-containing LacZ or galectin-3 full-length, and then followed by transfection with HSF-1 siRNA or scRNA. β-actin was used as a loading control. (D) The cells were analyzed by migration assay and invasion assay, and data are presented as histograms (*, ** p<0.001). All experiments were performed in triplicate. both galectin-3 and neogenin-1 were higher in cancer tissues than in normal tissues. Furthermore, analysis of 20 primary human gastric tumors revealed an overall positive correlation between increased active expression of galectin-3 and elevated neogenin-1 expression, except in 6 tissues (Fig. 7B). We could not determine clinical implications from these samples, maybe due to the small number of samples. However, these results revealed a significant positive correlation between the expression levels of galectin-3 and neogenin-1 in gastric cancer patient tissues.
Moreover, we also established the parallel protein expression and co-localization of HSF-1 and neogenin-1 in malignant tissues of gastric cancer patients (Fig. 7C). Therefore, our results strongly suggest that neogenin-1 was more detected in malignant tissues and its expression was positively correlated with both HSF-1 and galectin-3 in the malignant tissues of gastric cancer patients.

DISCUSSION
In this study, we confirmed for the first time that neogenin-1 is highly expressed in gastric cancer and promotes gastric cancer proliferation and migration. It is unclear why neogenin-1 is differentially expressed across organs. For example, the loss of expression of neogenin-1 in breast cancer and colorectal cancer has already been studied, and it appears that the RGMa/neogenin-1 interaction inhibits cancer growth and migration [21][22]. In our study, the expression of neogenin-1 was relatively high and easily detectable in twelve gastric cancer cell lines. In human tissues of gastric cancer patients, the expression level was highly detected in malignant tissues than normal counterparts, especially in diffuse type of signet ring cell carcinomas. It may provide the cue that neogenin-1 is involved in motility of gastric cancer cells, because signet ring cell carcinoma has a metastatic property on the various cancer types [32][33][34]. Neogenin-1 ligands netrin-1 and RGMa were expressed at low levels or not at all in gastric cancer cells, respectively. Moreover, only ablation of neogenin-1 reduced gastric cancer cell proliferation and migration, whereas ablation of netrin-1 had little effect. Neogenin-1 is thought to be up-regulated during gastric cancer development and is involved in gastric cancer growth and metastasis in a ligandindependent manner. In contrast to DCC, neogenin-1 appears to have varying expression and function depending on the organ from which the cancer originated [9]. We suggest that alternatively spliced isoforms of neogenin-1 are expressed in various developmental tissues and development stages, and that these isoforms may function differently. However, the regulatory mechanisms for these spliced isoforms remain to be determined. Moreover, detailed analyses of the expression and function of DCC and neogenin-1 are also needed at the single-cell level as well as in different organs.
To understand how neogenin-1 expression is upregulated in gastric cancer, we searched for transcription factor binding sites in the neogenin-1 promoter region using a transcription factor binding site prediction program. Next, we verified that the HSF-1 transcriptional factor interacts with the neogenin-1 promoter. Many cancers histologically exhibit activated HSF-1 and increased HSP levels, and activate oncogenes to regulate cell proliferation and cell motility [37][38]. Another study also demonstrated that HSF-1 plays an essential role in the development of lymphoma in p53-deficient mice, and of carcinomas in the Ras tumor model [28,39]. Whether the multifaceted HSF-1-mediated stress response plays a causal, supportive, or inhibitory role in mammalian oncogenesis remains unclear. Interestingly, we first demonstrated that HSF-1 activation is induced by galectin-3 interaction. Usually HSF-1 is activated by extracellular stresses, e.g., heat shock, cytokines or ROS [40]. Over-expression of galectin-3 facilitated the nuclear-localization of HSF-1 and binding to the promoter of neogenin-1. Related to this, HSF-1-mediated cell motility was caused by an increase in neogenin-1. Moreover, parallel and high expression of neogenin-1 and galectin-3, as well as neogenin-1 and HSF-1, was detected in human gastric cancer tissue. These data strongly suggest that over-expression of HSF-1 and galectin-3 jointly increases neogenin-1 expression both in vivo and in vitro. Neogenin-1 also appears to regulate gastric cancer cell motility through the activation of ROCK [35]. The function of ROCK1 is cell type-specific and remains controversial. However, there are several studies demonstrating that up-regulation of ROCK signaling contribute to cancer progression and metastatic behavior [41][42]. Moreover, ROCK1 increases gastric cancer cell migration and invasion [43], and the inhibition of ROCK1 induces apoptosis in gastric cancer cells [44], suggesting that neogenin-1 increases the activation of ROCK1 and, in turn, promotes gastric cancer cell survival and motility. Taken together, our results suggest that neogenin-1 increases gastric cancer cell motility, and its expression is highly regulated in gastric cancers through interactions with galectin-3 and HSF-1. Further study will be critical in determining the role and regulation of neogenin-1 in various cancer types, as well as the roles of neogenin-1, galectin-3 and HSF-1 regulation in cancer metastasis.

Fractionation of cellular extracts and Western blot analysis
Nuclear or cytoplasmic extracts were prepared from AGS cells after galectin-3 siRNA and HSF-1 siRNA treatments. These experiments have been described in a previous study [46]. Western blot analysis was carried out using the methodology from a previous study [47]. Briefly, cells were lysed in RIPA buffer (Biosesang, Korea) containing a protease inhibitor cocktail (Sigma, USA). The primary antibodies were: anti-β-actin, antigalectin-3, anti-neogenin-1, anti-HSF-1, and anti-HSP70 from Santa Cruz, and anti-ROCK-1 from Cell Signaling, also, anti-pROCK-1(T455/S456) antibody from Bioss. Proteins of interest were detected using ECL solutions (Amersham Life Science) with an LAS-3000 (Fujifilm) detector according to the manufacturer's directions.

Immunoprecipitation assays
Cell lysates containing 750 μg proteins was precleared by incubation with 40 μl protein-A/G linked agarose beads (Santa Cruz, CA, USA) for 1 hour at 4 o C. After the beads were spun down, the supernatant was incubated with 1 μg its specific antibody (anti-galectin-3 and anti-HSF-1) overnight at 4 o C, followed by incubation with 40 μl protein-A/G agarose beads for 1 hour. Mouse IgG (Santa Cruz, CA, USA) was used as the negative control. After the incubation, beads were washed three times in a RIPA buffer before being dissolved in a SDS-PAGE loading buffer. Western blot analysis was performed as described above.

Crystal violet staining.
Cells were transfected with 4ug of pcDNA3.0 Eempty vector and pcDNA3.0-Neogenin-1 full length by reverse transfection method using Lipofectamine 2000. For crystal violet staining, cells were plated into 6 well culture dishes and stained with crystal violet 2 days after transfection [49].

Cell proliferation analysis.
Inhibition of cell proliferation by netrin-1, neogenin-1 siRNA treatment was measured by WST assay. The AGS, MKN-28, YCC-2, SNU-216, SNU601 and SNU668 cells were plated in 96-well culture plates (3 × 10 3 per well). After incubation for 24 h, the cells were treated with 20nM of netrin-1, neogenin-1 siRNA treatment for 48 hr. WST solution (Daeil, korea) was subsequently added to each well. After 1-3h of additional incubation, the plate was shaken gently. The absorbance was measured on an ELISA reader at a test wavelength of 450 nm.

Luciferase activity of HSF-1 promoter
For HSF-1 reporter assays, a HSF-1 binding site luciferase construct (pGL3-HSE, containing 3 copies of HSE third pentamer consensus sequence reporter plasmid [50] was used. Gastric cancer cells were transfected with pGL3-HSE and a β-gal control construct, and HSF-1 promoter activities were analyzed by luciferase gene activity as described previously [51].

Preparation of Neogenin-1-depleted AGS gastric cancer cell xenografted mice
All animal experiments were approved by the Institutional Review Board of the National Cancer Center (NCC Korea) and performed in specific pathogen-free facilities and under conditions in accordance with the Guidelines for the Care and Use of Laboratory Animals of NCC (NCC-11-034D). The preparation of xenografted mice has been described in a previous study [52].

Fractionation of cellular extracts
Nuclear or cytoplasmic extracts were prepared from AGS cells after galectin-3 siRNA and HSF-1 siRNA treatments. These experiments have been described in a previous study [5].

Human gastric cancer tissue micro-array analyses and Preparation of frozen tissues from gastric cancer patients
Core tissue biopsy specimens (2-mm diameter) were obtained from individual paraffin-embedded gastric carcinomas (donor blocks) and arranged in new recipient paraffin blocks (tissue array blocks) using a trephine apparatus (Superbiochips Laboratories, Seoul, Korea). Immunohistochemical analysis of HSF-1 and Neogeinin1 was performed as described previously [53]. Two pairs of 2-mm-sized biopsy specimens were obtained from 20 patients with gastric adenocarcinoma during diagnostic endoscopic submucosal dissection. Immediately after biopsy, these tissue samples were frozen in liquid nitrogen in a deep freezer at -70℃ until experimental use. All participants provided written informed consent. All experiments were approved by the Institutional Review Board of the National Cancer Center as approval number NCCNSH 03-024.

Statistics analysis
All data were obtained from at least three independent experiments and are presented as mean ± SD, unless otherwise indicated. Statistical analysis was performed using one-way ANOVA and student T-test methods. Data were considered significant if p<0.05.