Spondin 2 promotes the proliferation, migration and invasion of gastric cancer cells

Abstract Spondin 2 (SPON2), a member of the Mindin F‐Spondin family, identifies pathogens, activates congenital immunity and promotes the growth and adhesion of neurons as well as binding to their receptors, but its role in promoting or inhibiting tumour metastasis is controversial. Here, we investigated its expression levels and mechanism of action in gastric cancer (GC). Western blotting and GC tissue arrays were used to determine the expression levels of SPON2. ELISAs were performed to measure the serum levels of SPON2 in patients with GC. Two GC cell lines expressing low levels of SPON2 were used to analyse the effects of regulating SPON2 expression on proliferation, migration, invasion, the cell cycle and apoptosis. The results revealed that SPON2 was highly expressed in GC tissues from patients with relapse or metastasis. The levels of SPON2 in sera of patients with GC were significantly higher compared with those of healthy individuals and patients with atrophic gastritis. Knockdown of SPON2 expression significantly inhibited the proliferation, migration and invasion of GC cells in vitro and in vivo. Down‐regulation of SPON2 arrested the cell cycle in G1/S, accelerated apoptosis through the mitochondrial pathway and inhibited the epithelial‐mesenchymal transition by blocking activation of the ERK1/2 pathway. In summary, this study suggests that SPON2 acts as an oncogene in the development of GC and may serve as a marker for the diagnosing GC as well as a new therapeutic target for GC.

331 amino acid residues (36 kD). 6,7 SPON2 expression identifies pathogens, activates congenital immunity and promotes the growth and adhesion of neurons as well as binding to their receptors. 8 SPON2 inhibits myocardial hypertrophy through the AKT-GSK3β and TGF-β1-SMAD signal transduction pathways. 9,10 SPON2 is highly expressed in numerous tumours, and elevated serum levels of SPON2 serve as a marker for prostate and ovarian cancers. 11,12 However, the physiological function of SPON2 in tumours and its associated molecular mechanisms are controversial.
For example, SPON2 promotes the infiltration of M1-like macrophages by affecting the activity of the RhoA-Rho kinase signalling pathway, further inhibiting hepatocellular carcinoma (HCC) cells from invading adjacent tissues and migrating to distant sites. Further, this process is positively regulated by thyroid hormones, which is exploited to improve the prognosis of patients with HCC. 13 However, other studies show that high levels of SPON2 are associated with poor prognosis of patients with prostate, hepatocellular and lung cancers. [14][15][16] In colon cancer, SPON2 acts as a downstream effector of the product of the metastasis-associated gene 1. Moreover, overexpression of SPON2 enhances the proliferation, migration, invasion and colony formation by colorectal cancer cells and induces metastasis to the liver. 17 Studies of small numbers of patients with GC show that overexpression of SPON2 is associated with poor prognosis, although the underlying mechanisms and their effects on GC cells are unknown. 18 Therefore, the current study employed a larger number of patients to answer these questions. Here we show that high expression of SPON2 was associated with poor prognosis of GC and may therefore serve as an auxiliary serological marker for early diagnosis and to evaluate the efficacy of treatments for GC. Moreover, F I G U R E 1 SPON2 is overexpressed in GC tissues. A, SPON2 expression is higher in GC tissues than para-tumorous tissues according to ONCOMINE analysis (Cho Gastric and DErrico Gastric, all P < .001). B-D, qRT-PCR and Western blotting analyses of SPON2 mRNA and protein expression in 20 matched pairs of T3N0M0 GC tissues (C) and paratumorous tissues (N). Data represent the mean ± SD of three independent experiments, **P < .01. E, Western blot analysis of SPON2 expression in eight GC tissue samples acquired from patients with relapse or metastasis 3 y after surgery for GC and in 12 samples acquired from patients without relapse or metastasis, *P < .05. F, Representative images of IHC analysis of SPON2 expression in 207 matched pairs of GC tissues and para-tumorous tissues (upper: magnification ×40; lower: magnification ×400). G, IHC analysis of SPON2 expression in 207 matched pairs of GC tissue samples and para-tumorous tissues, **P < .01. H, I, Kaplan-Meier survival analysis of 207 GC patients with low or high SPON2 expression. Patients with high SPON2 expression experienced longer overall survival (OS) and disease-free survival (DFS) after surgery down-regulation of SPON2 expression promoted apoptosis of GC cells and inhibited their abilities to invade and migrate by blocking activation of the ERK1/2 pathway.

| Patients, tissue samples and blood samples
Patients with GC, gastric stromal tumours and atrophic gastritis were definitively diagnosed according to pathological data.

| Cell culture, viral transfection and plasmid transfection
A human cell line derived from normal gastric mucosa (GES-1) and the GC cell lines SGC-7901, BGC-823, AGS, MKN-27, MKN-28 and MGC-803 were purchased from GeneChem. Cell lines were cultured in RPMI-1640 medium supplemented with 10% foetal bovine serum (FBS) and 100 µg/mL penicillin-streptomycin in an atmosphere containing 5% CO 2 at 37°C. To generate GC cell lines stably expressing low levels of SPON2, we transfected GC cells with a lentivirus vector encoding a SPON2-specific short hairpin RNA-expressing (GeneChem) and a lentivirus encoding a random sequence of shRNA (negative control). The activities of the constructs were verified using Western blotting. The shSPON2 sequence was 5′-CCGGCAGGGAC A ATGAGATTGTAGACTCGAGTCTACA ATCTCATTGTCCCTGTT TTTG-3′. SPON2 ectopic expression and negative control lentiviruses were also purchased from GeneChem, and the overexpressing cells were selected and used for next experiments. We

| Tissue microarray (TMA) construction and immunohistochemical analysis
All tested tissues were fixed with 4% polyoxymethylene for 1 hour, dehydrated through an ethanol gradient and embedded in paraffin. (5%-25%), 2 (>25%-50%), 3 (>50%-75%) and 4 (>75). The final expression score of a target protein was calculated by multiplying the positive score (percentage) by the staining intensity score. The staining scores were defined as follows: high expression (score ≥ 3) and low expression (score ≤ 2). SPON2 in serum was determined using the standard protein sample provided with the kit. A receiver operator characteristic (ROC) curve was plotted to calculate the ROC area under the curve (AUC) and to compare the sensitivity and specificity of serum levels of SPON2 in early-stage GC.

| Cell proliferation and colony-forming assays
To measure proliferation, cells (

| Flow cytometry
Gastric cancer cells (5 × 10 5 GC cells per well) were added to a 6well plate. Cells were harvested after 24 hours and fixed overnight with 70% ethyl alcohol. The cells were then incubated in a solution containing propidium iodide (0.5 mg/mL) and RNase A (10 mg/mL; BD Biosciences Pharmingen) for 20 minutes after washing. A flow cytometer (BD Biosciences Pharmingen) was used to determine DNA content. The rate of apoptosis was measured as previously described. 23 Cells (1 × 10 6 ) were stained with annexin V-PE and 7-AAD (BD Biosciences).

| Mitochondrial membrane potential
Gastric cancer cells were treated with JC-1 (5 µg/mL) to measure mitochondrial membrane potential (JC-1; YEASEN) according to the F I G U R E 2 Serum SPON2 levels are up-regulated in patients with GC. A, ELISA analysis of serum levels of SPON2 in patients with GC (n = 83), healthy controls (n = 25), patients with atrophic gastritis (n = 20) and patients with gastric stromal tumours (n = 20). Data represent the mean ± SD, **P < .01. B, Associations between serum levels of SPON2 with sex, tumour site, T stage, N stage, and M stage (M0, no liver or lung metastasis; M1, liver or lung metastasis), *P < .05, #P > .05. C, Changes in the serum levels of SPON2 before and after surgery in 43 patients with GC, **P < .01. D, Serum levels of SPON2 in patients with complete response and partial response (CR + PR, n = 17) and stable disease and progressive disease (SD + PD, n = 23) were evaluated using computed tomography, before and after three cycles of neoadjuvant chemotherapy or conversion therapy (NC + CT), *P < .05, #P > .05. E, ROC curve analysis of serum levels of SPON2 in patients with GC manufacturer's instructions. 24 The red and green fluorescence intensities of GC cells were observed using an inverted fluorescence microscope (TE-2000S, Nikon).

| Wound healing, Transwell and cell adhesion assays
For wound healing assays, cells ( In cell adhesion assays, 10 μg/mL fibronectin (Solarbio) was used to coat a 96-well plate, which was sealed with 1% bovine serum albumin and then seeded with 3 × 10 4 cells per well. Cells were stained with 10% crystal violet, and an inverted fluorescence microscope was used to count the adherent cells in five randomly selected visual fields.

| Statistical analysis
Data are expressed as the mean ± standard deviation (SD), and sta-

| SPON2 overexpression is associated with poor prognosis of patients with GC
We first analysed SPON2 expression using the ONCOMINE cancer microarray database, which revealed that SPON2 levels in GC tissues are higher compared with those in para-tumorous tissues ( Figure 1A).
Western blotting and qRT-PCR analyses revealed that the levels of SPON2 in 20 T3N0M0 GC tissues were higher compared with those of the corresponding para-tumorous tissues ( Figure 1B,C). Further, SPON2 levels in eight cancer tissue samples of patients with relapse or metastasis were significantly higher compared with those without relapse or metastasis ( Figure 1D).
To analyse the associations between SPON2 levels and clinicopathological features of patients with GC, we used IHC to analyse 207 pairs of GC tissue samples on a tissue chip ( Figure 1E).
Although we detected the expression of SPON2 in both GC cells and stroma, we found that SPON2 was mainly expressed in GC cells. Moreover, the purpose of this study is to observe the effect of SPON2 expression of GC cells on their own biology. Therefore, we only observed the expression of SPON2 in GC cells during IHC detection. The levels of SPON2 in GC tissue were significantly higher compared with those in para-tumorous tissue ( Figure 1F).
Univariate analysis revealed that high levels of SPON2 were significantly associated with differentiation, T-and N-staging but not significantly associated with age, sex, tumour site or tumour size ( Kaplan-Meier survival analysis revealed that postoperative OS and DFS of patients GC with high levels of SPON2 were significantly lower compared with those with low levels ( Figure 1G). Multivariate analysis revealed that a high level of SPON2 was an independent risk factor for postoperative OS and DFS ( Table 2).

| Serum SPON2 levels are increased in patients with GC
To analyse whether serum SPON2 levels can be used as a marker of GC, we analysed sera acquired from patients with different gastric lesions. Serum levels of SPON2 in patients with GC (n = 83) were significantly higher compared with those of healthy controls (n = 25), patients with atrophic gastritis (n = 20) and patients with gastric stromal tumours (n = 20; Figure 2A), but there was no difference in serum SPON2 levels among healthy controls, patients with atrophic gastritis and patients with gastric stromal tumours (P > .05).
Pearson's correlation analysis indicated that there was a positive correlation between the SPON2 expression levels of the serum and that of the GC tissue (r = .750, P < .01). Serum levels of SPON2 were not significantly associated with sex, tumour site, and T-and N-staging. However, serum levels of SPON2 in patients with advanced GC (n = 20) with liver or lung metastasis were significantly higher compared with those without such metastases (n = 63; Figure 2B).
Spearman's rank correlation analysis revealed that serum levels of SPON2 were significantly increased (R = .621, P < .001) in association with disease progression (healthy → non-metastatic GC → metastatic GC). Moreover, we detected changes in serum levels of SPON2 before and after surgery in 43 patients with GC who underwent radical surgery. Furthermore, there was a significant decrease in SPON2 levels 1 week after surgery compared with those before surgery ( Figure 2C). When we compared the serum levels of SPON2 before and after three cycles of chemotherapy administered to patients with GC who received neoadjuvant chemotherapy (n = 20) and conversion therapy (n = 20), we found that the levels in responders (CR + PR) to neoadjuvant chemotherapy and conversion therapy were significantly lower, while those in non-responders (SD + PD) to neoadjuvant chemotherapy and conversion therapy were not significantly increased ( Figure 2D).

| Down-regulation of SPON2 expression inhibits the proliferation of GC cells
To identify the mechanism through which SPON2 contributes to the progression of GC, we determined the levels of SPON2 in GC cells Consistent with these results, we found that the growth rates of subcutaneous tumours generated by the SPON2 knockdown cell lines were lower compared with those of the controls ( Figure 3F). The volumes and weights of the subcutaneous tumours after 5 weeks were significantly reduced compared with those of the control groups ( Figure 3G,H).

| Inhibition of SPON2 expression inhibits the migration and invasiveness of GC cells in vivo and in vitro
To analyse the mechanism through which SPON2 contributes to metastasis and the relapse of GC, we analysed the influence Wound healing (Figure 4A,B) and invasion assays ( Figure 4C,D) showed that the ability of the SPON2-knockdown cells to migrate and penetrate Matrigel was significantly reduced compared with those of the controls. Moreover, the adhesiveness of the SPON2knockdown cells was significantly lower compared with that of the controls ( Figure 4E,F). Overexpression of SPON2 in GES-1 cells and AGS cells enhances their migration and invasion ability ( Figure E-H).
Using a nude mouse model of abdominal metastasis, we found that the number of GC cell deposits (BGC-823 and SGC-7901) in the abdominal cavities of nude mice engrafted with SPON2 knockdown cells after 4 weeks was significantly lower compared with that of the controls ( Figure 4G,H).

| Down-regulation of SPON2 promotes G 1 arrest
To identify the mechanism through which SPON2 influences the pro-  Figure 5H).

| Down-regulation of SPON2 promotes apoptosis
Flow cytometry was used to detect apoptosis of GC cells. The rates of apoptosis of SPON2-knockdown BGC-823 and SGC-7901 cells were higher compared with those of the controls ( Figure 6A,B). We applied JC-1 fluorescence labelling to detect changes in the mitochondrial membrane potential of GC cells (SGC-7901). In SPON2-knockdown cells, mitochondrial membrane potential was significantly decreased ( Figure 6C). Western blotting revealed that the levels of BCL2 apoptosis regulator (BCL2) in SPON2-knockdown cells were significantly lower compared with those of the controls, while the levels of BCL2-associated X, apoptosis regulator (BAX), cleaved caspase-3 (CASP3) and cleaved poly (ADP-ribose) polymerase 1 (PARP1) were significantly increased ( Figure 6D).

| SPON2 promotes the epithelial-mesenchymal transition (EMT) of GC cells through the ERK1/2 pathway
The EMT mediates tumour metastasis and relapse, and activation of the mitogen-activated protein kinase (MAPK) pathway contributes to the development and progression of malignant tumours. [26][27][28][29] To identify the molecular mechanism through which SPON2 promotes

| D ISCUSS I ON
SPON2 is highly expressed in cancer tissue, although the association with patients' clinicopathological features varies. 13  significantly associated with T-and N-staging, OS and DFS, which is consistent with the findings of Jin et al. 32 Moreover, SPON2 expression was significantly increased in GC tissues from patients with metastasis and relapse 3 years after surgery, suggesting that SPON2 overexpression is associated with relapse and metastasis.
The high expression of SPON2 in tumour tissues is related to the activation of its promoter region. It has been reported that the lightactivated thyroid hormone receptor (TR) directly transactivated SPON2 transcription via critical TR-binding sites located within the promoter region of the gene. This is the reason why SPON2 has increased transcription levels in some types of HCC tissues. 33 The non-coding RNAs that are abnormally expressed can also first increase the expression level of its target gene SOX13, and then activating the promoter region of SPON2 by SOX13 to increase the expression level of SPON2 in glioma tissues. 34 There may be similar thyroid hormone and non-coding RNA expression abnormalities in GC tissues, which requires further study in the future. The death receptor and mitochondrial pathways represent the major mechanisms controlling apoptosis. We found that down-regulation of SPON2 expression did not cause changes in the level of caspase-8 (CASP8), a component of the death receptor pathway (results not shown). However, down-regulation of SPON2 expression reduced the mitochondrial membrane potential of GC cells and significantly increased the levels of cleaved CASP3.
The activation of signalling pathways such as the MAPK cascade contributes to relapse and metastasis of GC. MAPKs are serine/threonine kinases, among which ERK1/2, MAPK8 and MAPK14 are the most widely studied. 40 In tumour cells, excessive activation of ERK regulates the expression of certain oncogenes, further mediating and amplifying signals during tumour invasion and metastasis. 41 Our present study shows that high levels of SPON2 promoted the migration and invasion of GC cells in vivo and in vitro and that p-ERK1/2 expression in SPON2-knockdown cells, compared with that of the control cells, was significantly decreased, although the levels of ERK1/2, p-MAPK8, MAPK14, MAPK8 and p-MAPK14 were not significantly different.
We detected changes in the expression of EMT-related molecules and found that low levels of SPON2 inhibited the EMT of GC cells. We presumed therefore that SPON2 influenced the EMT of GC cells by regulating the activity of the ERK pathway, further affecting migration and invasion. Consistent with this hypothesis, we found that ERK1 overexpression increased the levels of p-ERK1/2 and activated the ERK pathway, further promoting the EMT. Wound healing and Transwell assays revealed that overexpression of ERK1 reversed the effects of down-regulating SPON2 on invasion and migration of GC cells.
In conclusion, high expression of SPON2 was significantly associated with relapse and metastasis of patients with GC. Serum levels of SPON2 may be used as a auxiliary diagnostic marker to detect early disease and to evaluate the responses to therapy of patients with GC SPON2 promoted the EMT of GC cells by activating the ERK pathway.
Thus, SPON2 may serve as a new serological marker for diagnosing GC and as a potential therapeutic target for GC. and Y. Liu revised the manuscript; and Q. Mao and W. Xue supervised the project.

E TH I C A L A PPROVA L
This study was conducted with the approval of the Institutional Ethical Standards Committee.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.

H U M A N R I G HT S S TATE M E NT A N D I N FO R M E D CO N S E NT
All procedures followed were in accordance with the ethical standards of the responsible committee on human experimentation (institutional and national) and with the Helsinki Declaration of 1964 and later versions. Informed consent or substitute for it was obtained from all patients for being included in the study.

A N I M A L S TU D I E S
All institutional and national guidelines for the care and use of laboratory animals were followed.