The present study used formalin-fixed paraffin-embedded sections of tissue samples from 11 patients with gastric adenocarcinoma. These patients were randomly selected at the National Cheng Kung University Hospital (NCKUH) between February, 2001 and May, 2006. Immunohistochemical analysis of ASS expression was performed in accordance with the following methods. Fresh specimens were collected from 35 patients with gastric adenocarcinoma who underwent radical resection at the NCKUH between August, 2003 and August, 2008. A total of 35 pairs of cancerous and matched adjacent normal gastric mucosa tissues were collected. The surgeries were performed by the same surgeon (Professor P.W. Lin). Patients with stage IV cancer who underwent palliative surgery, those who received conservative treatment, and those with gastric cancer arising from other cell types were excluded. Patients who refused to donate samples were also excluded. The demographics of the patients and the histopathological features of the tumors were collected by a retrospective review of each patient’s medical chart. The pathological staging was in accordance with the tumor, node, metastasis (TNM) staging system of the American Joint Committee on Cancer (AJCC) staging manual (16). The specimens were preserved in the Human Biobank within the Research Center of Clinical Medicine of the NCKUH.

All the patients provided written informed consent, and the study was approved by the Institutional Review Board of NCKUH (IRB number: ER-97–148).

Formalin-fixed paraffin-embedded sections of gastric cancer and normal stomach were used. Sections (4 μM) were cut from each tissue array block. The sections were deparaffinized and dehydrated. After the sections were subjected to antigen retrieval in an autoclave, immunohistochemical staining was performed by incubating the sections overnight with a mouse anti-human argininosuccinate synthase antibody (BD Transduction Laboratories, San Jose, CA, USA). The sections were then incubated with an avidin-biotin complex reagent (Dako, Carpinteria, CA, USA) and the final color was developed with 3-amino-9-ethyl carbazole (AEC) (Zymed Laboratories, Inc., San Francisco, CA, USA). The sections were counterstained with hematoxylin. The immunoreactivity of the ASS protein was assessed by using a semiquantitative method and scaled according to the immunoreactive score (IRS) of Remmele and Schicketanz (17). The sections were divided into four categories on the basis of the IRS scores, which ranged from 0 to 12: negative, weak, mild, and strong. Dr H.P. Hsu assessed the lesions.

Total cell lysates were prepared and analyzed by SDS-polyacrylamide gel electrophoresis as previously described (18). Immunodetection was performed by using the HRP-based SuperSignal chemiluminescent Substrate (Pierce, Rockford, IL, USA). For quantification, the bands were measured with the AlphaImager 2200 system (Alpha Innotech, San Leandro, CA, USA) and the densities of the ASS bands were normalized to those of the β-actin bands. ASS expression was quantified and described as a ratio to β-actin expression (the ASS/β-actin ratio).

In the present study, we conducted a search of the Oncomine database (http://www.oncomine.com) (19) to systematically assess the expression level of the ASS gene in gastric cancer. In addition, we used the online biomarker validation tool SurvExpress (http://bioinformatica.mty.itesm.mx:8080/Biomatec/SurvivaX.jsp) (20) to identify information on the gene expression from mRNA datasets of gastric cancer. The P-value was calculated by using Pearson’s linear correlation. Data were presented as mean ± SD. All statistical analyses were carried out by using SPSS version 12.0 (SPSS Institute, Chicago, IL, USA). Univariate analysis between categorical variables was performed by using the Chi-square test. Continuous variables that did not follow the normal distribution were compared by using the non-parametric Mann-Whitney or the Kruskal-Wallis test. The association between ASS expression and recurrence-free survival of patients with gastric cancer was assessed by using the Kaplan-Meier method, and the significance was tested by using the log-rank test. P<0.05 was considered to indicate a statistically significant result.

To determine the clinical relevance of ASS in human gastric cancer, we analyzed the expression profile of ASS in the Oncomine cancer microarray database (Table I). We compiled information on the expression of ASS in normal and cancerous gastric tissues from all of the microarray studies in the database (21–27). Eight studies demonstrated that ASS expression was significantly increased in several types of gastric cancer, such as gastric intestinal-type adenocarcinoma (GITA), diffuse gastric adenocarcinoma (DGA) and gastric mixed adenocarcinoma (GMA) (Table I) (21,22,24–26). In addition, three studies demonstrated that GITAs had a significantly higher expression of ASS than did DGAs (Table I) (22,23,27).

To confirm that ASS protein was expressed in human gastric cancer, we used immunohistochemistry to determine the expression in formalin-fixed paraffin-embedded sections of whole-mount specimens. We detected ASS expression in 11 specimens (100%) from patients with gastric cancer. The protein was localized primarily to the cytoplasm of the normal and cancerous gastric epithelial cells. Examples of the expression in gastric cancer specimens are shown in Fig. 1. In a specimen from stage II cancer, we observed focal ASS expression with heterogeneity (Fig. 1C and D). Strong expression of ASS was also detected in a specimen from stage III cancer (Fig. 1E and F).

We used western blots to quantify the expression of ASS in gastric cancer tissues and paired normal tissues. Quantitative analysis of ASS protein expression was achieved by normalizing the expression to that of β-actin. Fig. 2A shows the relative expression of ASS, described as the ASS/β-actin ratio, for three pairs of tissue (the results for all samples are shown in Fig. 2B). Our analysis determined that ASS expression in 19 (54%) tumor tissue samples was greater than that in the matched normal tissue samples.

We also investigated the clinicopathological characteristics of NCKUH patients with gastric cancer (Table II) and the association between these factors and ASS expression. Although the relative expression of ASS in the tumor samples was not associated with gender, age or poor predictors of histopathological factors, a higher tumor/normal ratio of relative ASS expression was associated with several clinical features (Table III; Fig. 3). First, the tumor/normal ratio of ASS expression was positively correlated with tumor size (R²=0.2032, P=0.007) (Fig. 3A). Second, stage II, but not stage III, cancer was correlated with a higher tumor/normal ratio of ASS expression (P=0.031) (Fig. 3B). However, one extreme value in a stage II cancer may have had substantial influence on our results (Fig. 3B). Third, while not statistically significant, the tumors that exhibited perineural invasion tended to have a higher tumor/normal ratio of ASS expression (P=0.072) (Fig. 3C). Taken together, these results suggested that ASS is involved in tumorigenesis and that it may regulate the invasive potential of gastric cancer.

To determine whether ASS expression is correlated with outcome, we used the Kaplan-Meier method to construct survival curves. The relative expression of ASS in cancer cells and the tumor/normal ratio of ASS expression were compared with the clinical outcomes and cancer recurrence of patients with gastric cancer (Table IV). The increased expression of ASS in cancer cells and the tumor/normal ratio of ASS expression did not correlate with recurrence (Table IV). To analyze the relationship between ASS expression and recurrence, we divided the patients into 2 groups on the basis of the relative expression of ASS (the ASS/β-actin ratio). One group consisted of patients whose tumors had a relative expression less than the median value of 0.39; the other of those whose tumors had a relative expression ≥0.39. The patients in the latter group had a trend toward poor recurrence-free survival (P=0.178) (Fig. 4; Table V).

To analyze the association of ASS expression with survival and risk, we used the online biomarker validation tool SurvExpress (20). This program derives Kaplan-Meier curves and risk groups on the basis of differences in expression. Low- and high-expression groups are shown in green and red, respectively. We determined that the expression of ASS in 57 samples of stomach AD from The Cancer Genome Atlas (TCGA) dataset was not correlated with survival (Fig. 5A). However, the survival risk curves derived for each group demonstrate that the expression of ASS was significantly increased in the high-risk group (P=7.84e-14) (Fig. 5B). In summary, the overexpression of ASS in gastric cancer is predictive of a poor prognosis. Further studies are needed to determine the manner in which ASS contributes to carcinogenesis.