Mucin Expression in Gastric Cancer: Reappraisal of Its Clinicopathologic and Prognostic Significance

We studied a cohort of 412 gastric cancer patients who underwent gastrectomy with lymph node dissection at Pusan National University Hospital between 2005 and 2007. The group comprised 286 men and 126 women with a mean age of 58.5 years (range, 42–75 years). Standard formalin-fixed, paraffin-embedded sections were obtained from the Department of Pathology, Pusan National University Hospital, and the National Biobank of Korea, Pusan National University Hospital. The study was approved by the institutional review board. None of the patients received preoperative radiotherapy and/or chemotherapy. We assessed the several clinicopathologic factors (tumor site, gross appearance and size, histologic classification [ie, intestinal or diffuse], and lymphovascular invasion) according to the Korean standardized pathology report for gastric cancer¹⁴  and the Japanese classification of gastric carcinoma, 3rd English edition,¹⁵  and made staging (depth of invasion, lymph node status) based on the American Joint Committee on Cancer Staging Manual, 7th edition.¹⁶  The clinical outcome of each patient was followed from the date of surgery to the date of death or March 1, 2011. The follow-up period ranged from approximately 1 to 74 months (average, 46.0 months). Cases lost to follow-up or with death from any cause other than gastric cancer were censored for the analysis of survival rates.

Sections were dewaxed and rehydrated according to the standard procedure and washed with phosphate-buffered saline. For immunohistochemical staining, sections were heated in a 600-W microwave oven twice for 5 minutes in 0.01 M citrate buffer (pH 6.0). Sections were immersed in 3% H₂O₂ to quench endogenous peroxidase activity, and unspecified binding was blocked in 5% normal goat serum (0.1% bovine serum albumin in phosphate-buffered saline). Immunohistochemical staining was performed using the avidin-biotin peroxidase complex method with aminoethylcarbazole as a chromogen, using the Vectastain ABC elite kit (Vector Laboratories, Burlingame, California) according to the manufacturer's instructions. Sections were counterstained with Mayer hematoxylin solution. Immunohistochemical staining was carried out with monoclonal antibodies against the mucin antigen (Table 1).

MUC5AC and MUC6 are markers of gastric foveolar cells and antral/cardiac mucous glandular cells, respectively, and they reflect gastric phenotypes (Figure 1, A and B). MUC2 and CD10 exhibit the typical intestinal epithelial cell phenotype, decorating goblet cells and the brush border of intestinal absorptive epithelial cells, respectively (Figure 1, C and D). Adenocarcinomas with at least 10% reactivity for each mucin gene antibody were identified as positive as previously reported.^7,8,13  Mucin phenotypes were further subdivided into gastric (only gastric mucin positive) and gastric-predominant gastrointestinal phenotypes and intestinal (only intestinal mucin positive) and intestinal-predominant gastrointestinal phenotypes and null type based on the combination of predominant patterns of MUC5AC, MUC2, MUC6, and CD10 staining according to Tsukashita et al.¹⁷  We defined gastric cancer with gastric mucin predominant type (GC-GP) as gastric type plus gastric-predominant gastrointestinal phenotypes (gastric type + gastric-predominant gastrointestinal type) and gastric cancer with intestinal mucin predominant type (GC-IP) as intestinal type plus intestinal-predominant gastrointestinal phenotypes (intestinal type + intestinal-predominant gastrointestinal type) for comparison based on the predominant mucin expression.

Clinicopathologic features were analyzed by Student t test, the χ² test, or Fisher exact test to test for differences among the mucin phenotypes. Cumulative survival plots were obtained using the Kaplan-Meier method, and significance was compared using the log-rank test. Prognostic factors were identified using the Cox regression stepwise method (proportional hazard model) adjusted for the patients' age, gender, tumor site, and morphologic type (intestinal versus diffuse). Statistical significance was set at P < .05. Statistical calculations were performed with SPSS version 10.0 for Windows software (SPSS Inc., Chicago, Illinois).

Of the 412 specimens, 67.5% (278 of 412), 44.9% (185 of 412), 35.4% (146 of 412), and 20.6% (85 of 412) were positive for MUC5AC, MUC6, MUC2, and CD10, respectively (Figure 2, A through D). As shown in Table 2, 63.8% (262 of 412), 24.5% (102 of 412), and 11.7% (48 of 412) of gastric cancers were classified as having GC-GP, GC-IP, and null phenotypes, respectively, on the basis of the predominant pattern of mucin expression.

Mucin phenotype was significantly correlated with sex, size, depth of invasion, gross type, histologic type, and lymphovascular invasion. The men had more null phenotype than GC-GPs and GC-IPs (P = .004). The tumor size of GC-GPs was smaller (P = .04) than that of GC-IP and null phenotypes. Early gastric cancer (EGC; limited to the mucosa and submucosa) showed more GC-GPs than advanced gastric cancer (invasion into muscularis propria and beyond) (P = .01). Diffuse histologic type was more strongly associated with GC-GP than with GC-IP or null phenotypes (P = .02). In addition, the presence of lymphovascular emboli was more related to GC-IP and null phenotypes than to GC-GP phenotype (P = .001). Also, there was a tendency of more lymph node metastasis in the GC-IP and null phenotypes than the GC-GP phenotype (P = .09) (Table 2).

In overall and advanced gastric cancers, the depth of invasion and nodal status were significantly associated with cumulative survival (P < .001) (Figures 3, A through C, and 4, A through C). Alternatively, there was no association between mucin phenotype and survival in the cases of overall gastric cancer (P = .41) or advanced cancers (P = .29). In contrast, there was a significant correlation between mucin phenotype and survival in EGC (P = .02) (Figure 5, A through C). The GC-GPs showed longer survival rate than GC-IP (P = .01) and null phenotypes (P = .01) in EGC. Furthermore, MUC5AC-positive EGCs had a longer survival rate than MUC5AC-negative EGCs (P = .009) (Figure 6, A through C). The expression of MUC5AC was identified as an independent prognostic factor in EGC adjusted for age, sex, histologic classification, tumor location, depth of invasion, and lymph node metastasis in the Cox regression proportional hazard model (P = .045) (Table 3). Notably, there was no significant association between survival rate and the expression of other mucins (ie, MUC2, MUC6, CD10) or mucin phenotypes (data not shown).

The present study demonstrates that the expression of a gastric mucin phenotype (especially MUC5AC expression) is significantly associated with favorable outcome, and should be considered as a possible prognostic predictor in the gastric cancer, especially in the EGCs.

From the perspective of carcinogenesis, the mucin plays an important role in the development of neoplasm. Alterations in mucin expression or glycosylation of tumor cells control the local microenvironment, playing a role in invasion and metastasis of tumor cells.⁴  In addition, mucins regulate the differentiation, proliferation, and tumor suppression of tumor cells.⁴ 

It has been reported that EGC exhibits a gastric mucin phenotype independently regarding the histologic type, whereas advanced gastric cancers have a higher tendency to express an intestinal phenotype.^18–20  In the present study, the gastric mucin phenotype was more commonly expressed in EGC whereas the intestinal phenotype was more common in advanced gastric cancer, which is concordant with previous reports.^18–20  In addition, Saito et al²¹  revealed that small intestinal-type EGCs (<1 cm) commonly express gastric mucin but transform into more undifferentiated and intestinal phenotypes with progression. Taking into account previous reports and our study, we confirm that during the progression of gastric cancer, there is a frequent phenotypic switch, with a mucin intestinal phenotype being acquired in most gastric cancers.

The clinical and, particularly, the prognostic significance of the mucin immunophenotype of gastric cancers are debated (Table 4). Lee et al⁶  reported that the mucin intestinal phenotype was associated with a significantly better prognosis. In contrast, Wakatsuki et al⁷  found that the same immunophenotype was associated with a significantly worse prognosis and a higher rate of postoperative liver metastasis. Notably, many different antibodies have been used to evaluate mucin expression, as well as different cutoff values for the definition of mucin positivity, leading us to speculate that these differences are responsible for the inconsistent results.

Compared with previous findings,^5–13  no relationship was found in our study between survival and mucin phenotypes in overall gastric cancer or in the advanced cancer subgroup. Instead, there was a significant relationship solely in EGC; expression of a mucin gastric phenotype was associated with a longer survival rate in EGC. Furthermore, foveolar gastric mucin expression, for example MUC5AC, was also associated with longer survival. Concordant with other reports, depth of invasion and lymph node metastasis were noted as prognostic markers in our entire cohort and the advanced tumor subgroup.¹⁶  However, in the EGCs of the present study, lymph node metastasis and depth of invasion (mucosa versus submucosa) were determined to be not predictive of outcome in EGC. Alternatively, retained MUC5AC expression was associated with extended survival in EGC, and this independently of the depth of invasion. It is noteworthy that in contrast to our results, Koseki et al²²  reported that in their experience, the expression of a gastric phenotype was associated with increased risk of lymph node metastasis among intestinal-type EGC. Yet, in contrast to these 2 reports and in agreement with our data, Wang et al¹⁰  noted that in a series of 76 gastric cancers, MUC5AC expression was associated with a favorable prognosis. In addition, others have detected an increased number of genetic alterations in adenocarcinomas with an intestinal mucin phenotype compared with those with a gastric mucin phenotype.^23–24  For example, EGCs with an intestinal mucin phenotype have been shown to display more p53 and nuclear β-catenin expression and undergo more frequent chromosomal alteration compared with those with a gastric mucin phenotype, suggesting a more aggressive behavior. And our findings suggest that MUC5AC expression can be used to predict outcomes in EGC. However, out study is not sufficient for ending conflicts about the relationship between the mucin phenotype and the gastric cancer prognosis. Because the reasons underlying these discrepancies are still undetermined, additional molecular and clinical studies are needed to sort out these results.

This study was supported by grant 0920050 from the National R&D Program for Cancer Control, Ministry for Health, Welfare and Family Affairs, Republic of Korea, and a clinical research grant from Pusan National University Hospital 2011.