Ninety patients with gastric carcinoma who underwent surgery between 2007 and 2008 were enrolled in this study. None of the patients had received any treatment before surgery. We obtained complete clinicopathological information for all patients, including age, gender, tumor size, histological grade, AJCC tumor stage, depth of invasion, lymph node metastasis and distant metastasis. All of the patients included in this study had adenocarcinoma. The median age of the patients at the time of diagnosis was 65.5 years (range 34–83 years). The histological grade of the tumor was evaluated based on the degree of tumor differentiation, tumor necrosis and mitotic count, according to the criteria of Enzinger and Weiss (23). Follow-up time was calculated as the time from initial surgery to the death of the patient due to the primary tumor or the date of last contact. Tumor tissue and paired adjacent normal tissue samples were obtained at surgery. All tissues were dissected in the operating room, immediately frozen, and stored at −80°C. Informed consent for use of tissue samples in future molecular studies was obtained from each patient. This study was approved by the Ethics Committee of the Third Xiangya Hospital, Central South University (Hunan, China). Clinical and treatment information was extracted by chart review carried out by the surgeon with approval from our institutional review board.

Core needle biopsies (1.5 mm diameter) were extracted from paraffin-embedded tissue samples, and mounted into a recipient paraffin block using a dedicated tissue array instrument (Beecher Instruments, Sun Prairie, WI, USA). Then, 4-μm-thick sections of the TMA were cut, transferred to glass slides and stained with hematoxylin and eosin.

TMA sections (1.5-mm diameter; 4-μm thick) from archival, formalin-fixed, paraffin-embedded tissue specimens were mounted on poly-L-lysine (Muto Chemicals, Tokyo, Japan)-coated slides. The TMA sections were deparaffinized in xylene for 15 min, rehydrated in an ethanol gradient and heated at 95°C for 5 min in 10 mM sodium citrate buffer (pH 6.0) in a microwave oven for antigen retrieval. Endogenous peroxidase was inactivated by incubating the sections in 3% H₂O₂ for 15 min at room temperature. The sections were blocked in 3% normal donkey serum and incubated at 4°C overnight with monoclonal anti-FAS antibody (dilution, 1:50; no. 3180S, Cell Signaling Technology, Danvers, MA, USA), anti-pERK1/2 antibody (dilution, 1:1000; no. 4370, Cell Signaling Technology) and anti-PTEN antibody (dilution, 1:50; no. 9559C, Cell Signaling Technology). Finally, the sections were stained with horseradish peroxidase-conjugated donkey anti-rabbit IgG (H+L) secondary antibody (711-035-152, Jackson ImmunoResearch Europe, Newmarket, UK). Signal detection was carried out using a Dako signaling amplification system (K346811; Dako, Glostrup, Denmark). The TMA sections were counterstained with hematoxylin, dehydrated, and mounted.

Immunohistochemistry was scored based on staining intensity and the percentage of positive cells. The staining intensity was scored as follows: 0, negative; 1, weak; 2, moderate; and 3, high intensity. The immunoreactive score was calculated as staining intensity score x percentage of FAS-positive cells. We also calculated S (T/A), immunore-active score of tumor tissue/immunoreactive score of paired adjacent normal tissue, as an index for the difference in expression between tumor and normal tissue. FAS immunostaining was analyzed under a microscope (Nikon Eclipse E600) and estimated independently by two pathologists.

Statistical analyses were performed using GraphPad Prism 5 (GraphPad Software Inc., San Diego, CA, USA) for Windows. The Student’s t-test was used to compare FAS expression between cancer tissue and adjacent normal tissues. Contingency table analysis and χ² tests were used to investigate the relationship between FAS expression and clinical variables. For outcomes with a small number of cases, Fisher’s exact test was used. Survival was estimated using the Kaplan-Meier method, and differences in survival curves were determined using the log-rank test. Values of P<0.05 were regarded as statistically significant.

Table I summarizes the clinicopathological characteristics data of the patients with previously untreated gastric carcinoma. The study cohort included 67 males and 23 females, ranging in age from 34 to 83 years (median age, 65 years). The histological type of all patients was adenocarcinoma. Tumor size ranged from 0 to 20 cm with a mean size of 6.17 cm and a median size of 5.75 cm. The tumors were classified as grade I in 1 patient, grade II in 23 patients and grade III in 66 patients. The median and mean duration of follow-up was 35 and 31.75 months, respectively, ranging from 1 to 51 months. Most of the patients (77.8%, 70/90) developed distant or regional lymph node metastasis.

We determined FAS expression in all 90 tumor tissue and paired adjacent normal tissue samples by TMA and immunohistochemistry. A representative TMA stained for FAS is shown in Fig. 1A. No signal was detected in the nuclei or on the cell membrane, indicating that FAS was mainly localized to the cytoplasm. The mean immunohistochemical score of FAS expression was significantly higher in tumor tissue than in adjacent normal tissue (1.065±0.099 vs. 0.798±0.074, P<0.05, Fig. 2A). These data indicate that FAS was overexpressed in the cytoplasm in tumor tissue compared to the expression in the adjacent normal tissue. To investigate the difference in expression between tumor tissues and normal tissues, we calculated S (T/A) as described in Materials and methods. Representative tissue sections corresponding to S (T/A) ≥1 to <2 (low), S (T/A) ≥2 to <3 (medium) and S (T/A) ≥3 (high expression) are shown in Fig. 1B. Overall, 38.9% (35/90), 22.2% (20/90), 10% (9/90) and 28.9% (26/90) of the carcinoma tissue specimens were classified as S (T/A) ≥0 to <1, S (T/A) ≥1 to <2, S (T/A) ≥2 to <3 and S (T/A) ≥3, respectively (Fig. 2B).

To investigate the association between FAS overexpression and clinicopathological parameters, three grades of FAS overexpression (S (T/A) ≥1, ≥2 and ≥3) were established. FAS overexpression was not significantly associated with any of the clinicopathological variables recorded, including age, gender, grade, tumor size and lymph node metastasis (Table II). As several reports have shown that FAS expression is modulated by the PTEN/PI3K/AKT and RAS/RAF/MEK/ERK pathways, we determined the expression of several components in these two pathways (4,24). Surprisingly, FAS overexpression was not significantly correlated with total, cytoplasmic or nuclear loss of PTEN (Table II) or pERK expression (data not shown) in gastric carcinomas.

Finally, we conducted survival analysis to determine whether FAS overexpression in tumors was associated with survival by plotting Kaplan-Meier survival curves and calculating the 3-year survival rate for all three grades of FAS overexpression. For cases with S (T/A) ≥1 or ≥2, FAS overexpression was not associated with overall survival or 3-year survival rate (Table III; Fig. 3A and B). However, among cases with S (T/A) ≥3, FAS overexpression was significantly associated with poor survival (Fig. 3C, log-rank test, P=0.0078) and with a decreased 3-year survival rate (Table III; χ² test, P=0.0023).