Gastric cancer is the second most common cancer worldwide and peritoneal dissemination is one of the most important prognostic factors in patients with gastric cancer (Dicken et al., 2005; Hohenberger and Gretschel, 2004). However, the molecular mechanism causing peritoneal dissemination of cancer cells is still poorly understood and there have been no valid markers indicating the development of peritoneal dissemination.
The anti-inflammatory protein, Annexin A1 (ANXA1) is the characterized member of the annexin family of proteins able to bind to cellular membranes in a calcium-dependent manner. Previous studies have shown that ANXA1 inhibits cytosolic phospholipase 2 activity and arachidonic acid generation, which is related to its anti-inflammatory and antiproliferative functions. Furthermore, ANXA1 expression in tumor cells has been investigated in vitro and in vivo in the cancer of head and neck (Garcia et al., 2004), liver (de Coupade et al., 2000), esophagus (Xia et al., 2002), lung (Lu et al., 2008), breast (Cao et al., 2008), stomach (Yu et al., 2008), and prostate (Siang et al., 2006). Correlation of ANXA1 expression with tumorigenesis and advancement of some types of cancer and patients' survival have also been reported (Paweletz et al., 2000, Wang et al., 2006). ANXA1 is a soluble protein and localized in the cytoplasm of cells, moving to membrane when calcium level is elevated and also can translocate into nucleus upon stimulation. Treatment of A549 and HeLa cells with heat, hydrogen peroxide, or sodium arsenite resulted in an increase in ANXA1 and translocation from the cytoplasm to the nucleus (Rhee et al., 2000). Recently, it is reported that nuclear staining of ANXA1 is a significant predictor of poor overall survival in oral squamous cell carcinoma (OSCC) (Lin et al., 2008). Therefore, different subcellular localizations of ANXA1 may contribute to different functions in a condition-specific manner.
The objective of the current study was to investigate the ANXA1 protein expression level and the subcellular localization in gastric adenocarcinoma specimens and to evaluate the clinical significance of ANXA1 expression by its correlation with clinicopathological parameters. To our knowledge, no studies have been performed to characterize the expression of ANXA1 in subcellular compartments in clinical gastric adenocarcinoma specimens previously.
PATIENTS AND METHODS
Paired specimens of gastric adenocarcinoma and corresponding normal tissues were obtained from 104 patients who underwent surgical resection of gastric adenocarcinoma in the Department of Surgical Oncology, Sir Run Run Shaw Hospital, College of Medicine, Zhejiang University between July 1995 and March 2007, with informed consent under the guideline of Hospital Ethics Committee. The patients comprised 68 males and 36 females aged from 38 to 88 years old (mean = 67.5 years old), including 91 R0 (radical operation) patients and 13 peritoneal carcinomatosis patients. None of the patients received any anti-cancer treatment before surgery. Correlation between expression of ANXA1 and clinicopathologic parameters including age, sex, pathological Tumor-Node-Metastasis (pTNM) (Hermanek et al., 1987), peritoneal dissemination and lymph node metastasis was evaluated.
Immunohistochemical Analysis of ANXA1
An Immunohistochemical analysis of ANXA1 was performed on formalin-fixed, paraffin-embedded sections of surgically resected specimens. The slides were deparaffinized in xylene and dehydrated in a graded ethanol series. Endogenous peroxidase was blocked with 0.3% H2O2 in methanol for 10 min. The slides were immersed in 10 mM citric buffer solution (pH 6.0) for antigen retrieval. Nonspecific binding was blocked by preincubation with 10% fetal calf serum in PBS with 0.01% sodium azide. Blocking serum was drained off, and the slides were separately incubated with antibody against ANXA1 (mouse monoclonal, Santa Cruz; 1:200) in a humidified chamber for 1 hr. After washing three times with PBS, the slides were incubated with the EnVision-HRP complex (undiluted, DAKO) for 30 min. The slides were visualized with diaminobenzidine (DAKO Corp.) and counterstained with hematoxylin. For negative controls, the primary antibody was replaced with phosphate buffered saline. Positive controls included gastric or colonic carcinomas tissue known to exhibit high levels of ANXA1 expression.
All the slides were examined and scored independently by two investigators. Expression of ANXA1 was assessed as overall expression and subcellular distribution in the nucleus and cytoplasm. Immunoreactivity was evaluated by determining the percentage of positive cells in each core and then taking the average of three cores. High expression of ANXA1 was defined as staining of ≥25% of cells, low expression was defined as staining of <25% of cells, and negative staining was defined as no staining of any cells.
All statistical analyses were performed using the statistical program SPSS 15.0 for windows (SPSS, Chicago, IL, USA). Two-tailed chi-square test and the two-tailed t-test were used to evaluate association between ANXA1 expression and clinicopathologic factors. The association of ANXA1 nuclear staining and the peritoneal dissemination was assessed by both correlation test and curvilinear regression. Logistic regression was used to determine the independent risk factors for peritoneal dissemination. The level of significance was set for P < 0.05.
ANXA1 Expression and Clinicopathologic Parameters
ANXA1 protein expression was positive in 72 of 104 (69.2%) normal tissues and 47 of 104 (45.2%) gastric adenocarcinoma tissues. ANXA1 staining was predominantly localized in the cytoplasm in all 72 ANXA1-positive normal specimens. In gastric adenocarcinoma tissues, 45 (43.3%) specimens showed positive ANXA1 cytoplasmic staining, and in addition, 10 specimens also showed positive ANXA1 nuclear staining. There were only two gastric adenocarcinoma tissues that showed positive ANXA1 nuclear staining, which made a total of 12 (11.5%) gastric adenocarcinoma specimens in the group of positive ANXA1 nuclear staining (Fig. 1). In the tumor specimens, the levels of cytoplasmic staining of ANXA1 were decreased, whereas an increase in nuclear staining of ANXA1 was observed (Table 1). The correlation between ANXA1 nuclear and cytoplasmic staining in the adenocarcinoma specimens with clinicopathological parameters of 104 gastric cancer patients are shown in Table 1. ANXA1 nuclear staining correlates with tumor size, peritoneal dissemination, and TNM stage. However, no difference was found between positive and negative nuclear staining groups regarding sex, age, depth of invasion, node involvement, differentiation, and location. The decrease in ANXA1 cytoplasmic staining in gastric cancer was not associated with any clinicopathologic parameters either.
Table 1. Correlation between adenocarcinoma tissue and corresponding normal tissue in gastric adenocarcinoma patients
Cytoplasmic ANXA1 staining of ≥25% of cells was scored as 2, staining of <25% of cells was scored as 1, and absence of staining was scored as 0 (P < 0.001; chi-square test).
ANXA1 Protein Nuclear Staining and Peritoneal Dissemination
There were 13 gastric adenocarcinoma patients with peritoneal dissemination enrolled in this study, with five in positive ANXA1 nuclear staining group (totally 12) and eight in negative ANXA1 nuclear staining group (including negative ANXA1 staining, totally 92). The proportion of patients with peritoneal dissemination in positive nuclear ANXA1 staining group (5 of 12, 41.7%) is much higher than those in negative nuclear ANXA1 staining group (8 of 92, 8.7%) (P = 0.007; Table 2). In univariate analysis, ANXA1 protein nuclear staining, poor differentiation, large primary tumor size, serosal invasion, and lymph node metastasis were associated with peritoneal dissemination. When a logistic regression analysis was performed, ANXA1 nuclear staining, poor differentiation, and lymph node metastasis were associated with peritoneal dissemination (P = 0.039, 0.007, 0.013, respectively, Table 3).
Table 2. Comparison of clinicopathologic features in gastric cancer patients
Negative (n = 92)
Positive (n = 12)
56.0 ± 12.0
55.6 ± 11.4
Upper body or whole
Lower or middle body
Tumor size data lost
Table 3. Univariate and multivariate analyses of peritoneal dissemination by logistic regression
Univariate analysis P-value
Age (>55 vs. ≤55)
Gender (male vs. female)
Tumor size (≥ 4 cm vs. <4 cm)
Lymph node metastasis
Annexins are a family of structurally related, calcium-dependent phospholipid binding proteins, which have been proposed to be involved in diverse cellular processes, including phospholipase A2 inhibition, adhesion, exocytosis, and interaction with cytoskeletal proteins (Raynal and Pollard et al., 1994). ANXA1 is a substrate for the epidermal growth factor (EGF)-stimulated tyrosine kinase and is postulated to be involved in mitogenic signal transduction. It is also the substrate for platelet derived growth factor, insulin and hepatocyte growth factor/scatter factor receptor tyrosine kinases (Rothhut et al., 1997). Despite of these known functions, however, other biological functions of ANXA1 still need to be investigated.
Some studies have shown ANXA1 cytoplasm staining correlation with esophageal and esophagogastric junction adenocarcinoma patients' survival (Wang et al., 2006), and loss of ANXA1 expression correlated with gastric cancer patients' poor survival rates (Yu et al., 2008). A more recent study evaluating the prognostic significance of nuclear staining of ANXA1 in OSCC patients showed a lower overall survival, whereas changes in ANXA1 expression levels and decrease of ANXA1 membranous staining in OSCC may not be involved in oral carcinogenesis (Lin et al., 2008). Upregulation of ANXA1 was also observed in multidrug resistant K562 cells (Zhu et al., 2009). In addition, it has been suggested that ANXA1 nuclear translocation participates in the regulation of cellular proliferation (Alldridge and Bryant, 2003; Alves et al., 2008; Gerke and Moss, 2003), and treatment of cells with heat, hydrogen peroxide, sodium arsenite, and EGF resulting in translocation of ANXA1 from the cytoplasm to the nucleus (Rhee et al., 2000). Therefore, we postulated that the different subcellular distribution of ANXA1 may also contribute to tumorigenesis and progression in gastric cancer.
In the current study, no correlation between overall ANXA1 expression and patients' survival was observed. We analyzed the correlation of both cytoplasmic and nuclear expression of ANXA1 with different clinicopathological parameters. Nuclear ANXA1 expression was associated significantly with tumor size, peritoneal dissemination, and TNM stage. However, cytoplasmic ANXA1 expression did not correlate with any of these pathologic parameters.
In this study, peritoneal dissemination was noted in 13 (12.5%) of 104 patients with gastric cancer. More than one-third of these patients had nuclear staining of ANXA1 (5 of 13 patients). Because patients with nuclear staining of ANXA1 had a significantly higher rate for peritoneal dissemination than patients without nuclear (41.7% vs. 8.7%, respectively; P = 0.007), the nuclear staining of ANXA1 may likely be associated with poor prognosis. Based on these data, we propose that ANXA1 nuclear staining correlates with advanced disease and peritoneal dissemination. Although it is not yet clear why ANXA1 nuclear localization correlates with the advanced disease stage and peritoneal dissemination in patients with gastric cancer, there exist some underlying clues. Recently, it is hypothesized that ANXA1 nuclear translocation reflects the activity of oncogenic tyrosine kinases, such as EGF receptor family (EGFR) and c-Met (Lieto et al., 2008, Rhee et al., 2000, Zeng et al., 2008). Overexpressions of the EGFR, Her-2 and Her-3, have been frequently detected in gastric cancer and are thought to be a recognized indicator of poor prognosis (Moutinho et al., 2008, Poller et al., 1992). These finding may provide useful information for further investigation on the role of nuclear localization of ANXA1 in gastric adenocarcinoma.
In this study, correlation between positive ANXA1 nuclear staining and peritoneal dissemination was established; however, there were also seven positive ANXA1 nuclear staining specimens in patients without peritoneal dissemination. Because of the multiple alterations of upstream and downstream pathways affecting ANXA1 expression and subcellular localization, ANXA1 should be recognized as one of the multiple factors, which correlate with gastric adenocarcinoma peritoneal dissemination but exceptions may exist.
Although the exact roles ANXA1 plays and the underlying mechanisms remain to be elucidated, recent works by other researchers provide helpful clues. One study showed that reduced ANXA1 protein level by siRNA led to a significant decrease of melanoma cell invasion, and ANXA1 might enhance melanoma dissemination through formyl peptide receptors (Rondepierre et al., 2009). In another investigation on global screening of protein profiles, ANXA1 is also proposed to play an important role in genesis, progression, recurrence, and metastasis of lung squamous cell carcinoma (Nan, et al. 2009). A more recent work utilizing ANXA1-knockout mice model showed that ANXA1 plays important roles in tumor growth, metastasis, angionesis, and wound healing (Yi, et al., 2009).
In summary, our results showed that ANXA1 is extensively expressed in normal gastric tissues and decreased expression were observed in gastric adenocarcinoma tissues. Compared with negative nuclear staining of ANXA1 in corresponding normal tissue, positive nuclear staining of ANXA1 presented in gastric cancer tissue correlates with advanced disease stage and peritoneal dissemination. Alteration of ANXA1 subcellular distribution may be explored further for its value in monitoring the progress of gastric cancer and predicting patients' outcome and prognosis.