• annexin A1;
  • gastric cancer;
  • cancer invasion;
  • survival;
  • formyl peptide receptor


  1. Top of page
  2. Abstract


Annexin A1 (AnxA1) has been well-known as a glucocorticoid-regulated anti-inflammatory protein, and it is implicated in tumorigenesis in a tumor type–specific pattern. However, the role of AnxA1 in gastric cancer (GC) is indeterminate, and the underlying mechanism is not clear. The purpose of this study was to evaluate the prognostic significance and associated mechanism of AnxA1 in GC.


Immunohistochemical staining was employed to analyze 118 GC patients. Both AnxA1 gain-of-function and loss-of-function approaches were performed in GC cells. Western blotting and reverse-transcription polymerase chain reaction were used for assessment of the AnxA1 regulation mechanism in GC cells. An intraperitoneal inoculation model in severe combined immunodeficient mice was used for an in vivo assay.


High AnxA1 expression was significantly associated with peritoneal metastasis (P = .009) and serosal invasion (P = .044). Cox multivariate analysis showed that high AnxA1 expression was an independent risk factor for poor overall survival in GC patients (P = .037). AnxA1 expression positively correlated with invasiveness of human GC cells both in vitro and in vivo. AnxA1 could regulate the GC cell invasion through the formyl peptide receptor (FPR)/extracellular signal-regulated kinase/integrin beta-1-binding protein pathway, and all 3 FPRs (FPR1 through FPR3) were involved in the regulation process.


High AnxA1 expression was associated with more serosal invasion, more peritoneal metastasis, and poorer overall survival in GC patients. The current study demonstrated a novel mechanism involving FPRs, extracellular signal-regulated kinases 1 and 2, and integrin beta-1-binding protein 1 by which AnxA1 regulated GC cell invasion. Cancer 2012. © 2012 American Cancer Society.

Gastric cancer (GC) remains the second leading cause of cancer-related death in the world.1 Although the 5-year survival rate of GC is 40% to 60% in Japan and some Asian countries, it is approximately 20% in developing countries and Western developed countries where only few GCs are diagnosed at an early stage.2, 3 Surgical resection of the primary tumor and regional lymph nodes is still the treatment of choice for GC. Yet the current surgical staging system, based on clinical and pathological status only, has limited accuracy in predicting clinical outcomes.4 Development of new molecular markers may improve the postoperative prognostic predictability and identify a subgroup of GC patients who should benefit from further adjuvant therapies.

The work of 3 research groups, including ours, has identified the annexin A1 (ANXA1) gene to be a potential marker in gastric carcinogenesis by genome-wide complementary DNA microarray studies.5-7 AnxA1 is a member of a family of calcium and membrane-binding proteins, which include 12 members (AnxA1 through AnxA11, and AnxA13) of mammalian annexins.8 Although AnxA1 initially characterized as a glucocorticoid-regulated anti-inflammatory protein,9 N-terminally truncated AnxA1 plays a proinflammatory role by promoting neutrophil transendothelial migration.10 Recent evidence suggests that AnxA1 has a wide variety of cellular functions, such as membrane aggregation, phagocytosis, proliferation, apoptosis and tumorigenesis.11 AnxA1 expression is up-regulated in patients with hepatocellular carcinoma and adenocarcinomas of the esophagus and pancreas, but down-regulated in adenocarcinoma of the breast and prostate, as well as in squamous cell carcinoma of the esophagus and head and neck.12 However, the role of AnxA1 in GC is still indeterminate,13, 14 and the biological consequence of these expression patterns is unclear.

AnxA1 can function in an autocrine, paracrine, or juxtacrine manner to activate the formyl peptide receptor (FPR) signal on the cell surface.9 The FPR family, including 3 human subtypes (FPR1, FPR2/FPRL1, FPR3/FPRL2), is involved in the host defense response to bacteria-derived formyl peptides.15 AnxA1 may inhibit leukocyte extravasation by interacting with FPRs,16 specifically FPR2, the only family member for which there is existing direct evidence.17 Besides, overexpression of AnxA1 causes constitutive activation of the mitogen-activated protein kinase (MAPK)/extracellular signal-regulated kinase (ERK) pathway in monomyelocytic cells,18 vascular smooth muscle cells, and kidney cells.19 Based on these findings, we hypothesized that AnxA1 promoted invasion ability of GC via activating FPRs and ERK pathway leading to subsequent interference effects.

In this study, we evaluated AnxA1 expression in GC tissues and determined its prognostic significance by clinical correlation. Furthermore, we attempted to elucidate the mechanistic relationship among AnxA1, FPRs, ERK pathway, and downstream molecules in terms of GC cell invasion.


  1. Top of page
  2. Abstract

Patients, Specimens, and Immunohistochemical Staining

This study included 118 patients with newly diagnosed GC who were undergoing gastrectomy in the inpatient unit of the National Taiwan University Hospital, Taipei, Taiwan, between November 1999 and February 2003. Patients receiving preoperative anticancer therapies were excluded. The corresponding clinical and pathologic data were retrieved for analysis. The histological diagnosis of gastric adenocarcinoma was made according to the recommendations of the World Health Organization. Tumor–node–metastasis status was based on the 7th edition of the American Joint Commission on Cancer Staging System.20 All patients had been followed for at least 7 years. The surgical specimens were obtained from patients after written informed consent was obtained in the National Taiwan University Hospital.


After rehydration, sections (5 μm) of paraffin-embedded tissue blocks were incubated in 3% hydrogen peroxide to block endogenous peroxidase activity. Following citrate buffer antigen retrieval, the sections were blocked by incubation in 3% bovine serum albumin in phosphate-buffered saline. The primary antibody was applied to the slides and incubated at 4°C overnight. After washes in phosphate-buffered saline, the samples were treated with a SuperPicture Polymer Detection kit (Invitrogen, Carlsbad, Calif). The slides were stained with diaminobenzidine, washed, counterstained with hematoxylin, dehydrated, treated with xylene, and mounted. The pathologists assessing immunostaining intensity were blinded to the patients' information. The expression level of AnxA1 in GC was determined by immunohistochemistry using an AnxA1-specific antibody (Fig. 1A). AnxA1 expression in lymphoplasma cells served as an internal positive control. The results of immunohistochemical staining were classified using extent of cells stained: level 0 (negative staining), level 1 (<25% of tumor cells stained), level 2 (25%-50% of tumor cells stained), and level 3 (>50% of tumor cells stained). Institutional review board approval was obtained to procure and analyze the tissues used in this study.

thumbnail image

Figure 1. Immunohistochemical analysis of AnxA1 expression in GC specimens is shown. (A) Specificity of AnxA1 antibody for immunochemical staining. Upper panel: Western blot analysis using AnxA1 antibody showed reduced AnxA1 expression after treatment with AnxA1 small interfering RNA (siRNA). Lower panel: Immunochemical staining with anti-AnxA1 antibody was less detectable in cells treated with AnxA1 siRNA. Scale bar, 50 μm. ICC, immunocytochemistry. (B) AnxA1 levels are shown in representative GC tissue. Panel I: Lack of immunoreactivity for AnxA1 in adjacent clinically normal gastric mucosa. Panel II: Negative AnxA1 expression in GC. Panel III: Low AnxA1 expression in GC. Panel IV: High AnxA1 expression in GC. Panel V: High AnxA1 expression in liver metastases. Panel VI: High AnxA1 expression in peritoneal metastases. Black arrowheads indicate the tumor epithelial cells. White arrows indicate the lymphoplasma cells with AnxA1 immunoreactivity. Scale bars, 50 μm. (C) Kaplan-Meier estimates and log-rank tests of overall survival are shown for GC patients with different AnxA1 expression (P = .0002, n = 118).

Download figure to PowerPoint

Antibodies and Reagents

Antibodies and reagents used were as follows: AnxA1 (A2295-16; US Biological, Swampscott, Mass), α-tubulin, p-ERK1/2, ERK1/2, p-Akt1/2/3 (Ser-473)-R, anti-Akt-1, p-p38, p38, phosphorylated c-Jun N-terminal kinase (p-JNK), JNK (Santa Cruz Biotechnology, Santa Cruz, Calif), and integrin beta-1-binding protein 1 (ITGB1BP1) (Orbigen, San Diego, Calif). Boc2 was bought from MP Biomedicals (Solon, Ohio). U0126 was purchased from Promega (Madison, Wis). AnxA1 and ITGB1BP1 ON-TARGETplus SMARTpool small interfering RNAs (siRNAs) were acquired from Dharmacon (Lafayette, Colo).

Cell Culture

The human cell lines AGS, N87, and human embryonic kidney 293T (HEK293T) were purchased from the American Type Culture Collection (Manassas, Va). TSGH9201 was purchased from the Bioresource Collection and Research Center (Hsinchu, Taiwan). MKN45 was purchased from the Health Science Research Resources Bank (Osaka, Japan). These cell lines were cultured in Roswell Park Memorial Institute 1640 (RPMI-1640) medium (GIBCO) containing 10% fetal bovine serum (Bioserum, Victoria, Australia) in a humidified atmosphere containing 5% CO2 at 37°C.

Western Blot Analysis

Protein lysates were prepared as described.21 Western blot analysis was performed with primary antibodies for AnxA1, α-tubulin, p-ERK1/2, ERK1/2, p-Akt1/2/3 (Ser-473)-R, anti-Akt-1, p-p38, p38, p-JNK, JNK, and ITGB1BP1.

Reverse Transcription–Polymerase Chain Reaction

Messenger RNAs were isolated and amplified as described.21 Reverse transcription–polymerase chain reaction (RT-PCR) was performed with human glyceraldehyde 3-phosphate dehydrogenase as an endogenous control.

Construction of AnxA1 Expression Vectors and Transient Transfection

ANXA1 cloned into pcDNA3.1 V5/His-TOPO plasmid (Invitrogen, Carlsbad, Calif) was amplified using specific AnxA1 primer sets, in either the sense or antisense orientation. The cloned DNA fragments were verified by direct sequencing (ABI-Prism 377 DNA sequencer; PE Biosystems, Foster City, Calif). The AnxA1 (sense, antisense) expression vectors were transiently transfected into test cells with Lipofectamine 2000 transfection reagents (Invitrogen).

Lentivirus Production and Infection

The pGIPZ short hairpin RNA (shRNA) constructs were obtained from Open Biosystems (Open Biosystems, Huntsville, Ala). The lentiviral pLKO.1 shRNA constructs were acquired from the National RNAi Core Facility in Academic Sinica, Taipei, Taiwan. Lentiviruses were produced by cotransfecting the shRNA-expressing vector, packing vector pCMVΔR8.91, and envelope vector pMD.G into HEK293T cells, by using calcium phosphate. Briefly, 1 × 106 HEK293T cells per 100-mm dish were seeded 1 day before transfection. The cells were transfected with 10 μg shRNA-expressing vector together with 10 μg of pCMVΔR8.91 and 1 μg of pMD.G. After 5 hours of incubation, the transfection media was replaced with fresh culture media. After 48 hours, viral supernatants were harvested, titered, and used to infect MKN45 or AGS cells with 8 μg/mL of polybrene. Cells were selected using 2 μg/mL puromycin.

siRNA Transfection

GC cells were transfected with ON-TARGETplus SMARTpool siRNA or Risc-free control siRNA, as indicated, for 18 to 24 hours, using Lipofectamine 2000. Treatment was followed by a change of media. Cells were subjected to further assays 2 to 3 days after siRNA transfection.

Two-Chamber Migration/Invasion Assay

Cell migration/invasion ability was determined by the modified 2-chamber migration assay (8 μm pore size; BD Biosciences, Bedford, Mass) according to manufacturer's instructions. Approximately 1 × 106 cells were seeded into the upper chamber and allowed to migrate into the lower chamber for 24 hours. For invasion assays, we used the upper chambers coated with Matrigel (40 μg; Collaborative Biomedical, Becton Dickinson Labware, San Jose, Calif) and observed cells 48 hours later. Cells in the upper chamber were carefully removed with cotton-tipped swabs, and cells at the bottom of the membrane were fixed and stained with crystal violet 0.2%/methanol 20%. Quantification analysis was performed by counting the stained cells. Each data point is representative of 3 independent experiments and is shown as mean ± standard deviation.

Animal Study of Intraperitoneal Metastasis

Seven-week-old female severe combined immunodeficient mice (NOD.CB17-Prkdcscid mice, supplied by the animal center of the National Taiwan University Hospital, Taiwan) were acclimatized over a period of 1 week while being caged in groups of 5. MKN45/AnxA1 shRNA and MKN45/control shRNA (pGIPZ NS shRNA) cells (1 × 107 cells) were suspended in RPMI-1640 medium in a final volume of 1.0 mL. They were then inoculated into peritoneal cavities of test mice. The mice were sacrificed 4 weeks later, and peritoneal nodules and metastases were evaluated. All mouse studies were performed using protocols approved by the Institutional Animal Care and Use Committee of the College of Medicine, National Taiwan University.

Statistical Analysis

For comparison of clinicopathologic parameters, the Mann-Whitney test was employed for scale variables and the Pearson chi-square test for nominal variables. Logistic regression analyses adjusted for age and sex were conducted to evaluate the associations between AnxA1 expression levels and clinicopathologic features of GC patients. To determine whether AnxA1 expression is an independent prognostic factor for survival, hazard ratios were calculated using the Cox proportional hazards model. Data were analyzed with the SPSS program for Windows, version 11.0 (SPSS Inc., Chicago, Ill). Analysis of variance was used to evaluate the statistical significance of the mean values. All statistical tests included 2-way analysis of variance. Statistical significance was assumed at P < .05.


  1. Top of page
  2. Abstract

AnxA1 Associates With Serosal Invasion and Peritoneal Metastasis in GC Patients

The majority (78.0%, 92/118) of the adjacent clinically normal mucosa showed no AnxA1 immunoreactivity (Fig. 1B, panel I). Among GC specimens, the majority (64.4%, 76/118) showed negative or low expression (level 0 and 1), whereas a subset (35.6%, 42/118) of GC showed high expression (level 2 and 3) of AnxA1 with ≥25% of immunoreactive cancer cells (Fig. 1B, panels II, III, IV). We also found high AnxA1 expression on all available specimens of liver and peritoneal metastases (liver metastases: 3 of 3 specimens; peritoneal metastases: 7 of 7 specimens; Fig. 1B, panels V, VI). The AnxA1 protein was predominantly localized in the cytoplasm of the epithelial cells. Only sporadic cells occasionally showed weak AnxA1 nuclear staining in a small subset (14.4%, 17 of 118) of GC specimens. We then compared the AnxA1 expression levels among GC patients with different clinicopathological characteristics (Table 1). High AnxA1 expression was associated with stage IV disease (P = .005), peritoneal metastasis (P = .009), and serosal invasion (P = .044).

Table 1. Association of AnxA1 Expression and Clinicopathological Features in Gastric Cancer Patients (n = 118)
CharacteristicsNegative/Low AnxA1 (n = 76)High AnxA1 (n = 42)Odds Ratio (95% CI)aP
  • AJCC indicates American Joint Commission on Cancer; AnxA1, annexin A1; CI, confidence interval; NS, nonsignificant.

  • a

    Odds ratio adjusted for age and sex.

  • b

    Compared with stage I.

Age, y    
 >6042221.12 (0.52-2.43)NS
 Male51281.00 (0.44-2.25)NS
 Cardia651.59 (0.45-5.61)NS
Histological type    
 Diffuse + mixed38251.50 (0.68-3.32)NS
 Moderate + well36181 
 Poor40241.19 (0.54-2.63)NS
Pathological stage (AJCC 7th)    
 II1361.26 (0.26-5.98)NSb
 III41181.67 (0.48-5.85)NSb
 IV7148.88 (1.93-40.90).005b
Submucosal confinement    
 Advanced66403.03 (0.63-14.63)NS
Serosal invasion    
 Present50352.64 (1.02-6.81).044
Lymphovascular invasion    
 Present40251.31 (0.61-2.83)NS
Lymph node involvement    
 Present50311.47 (0.63-3.41)NS
Peritoneal metastasis    
 Present6114.32 (1.45-12.89).009

AnxA1 Expression Associates With Survival in GC Patients

On Kaplan-Meier analysis, overall survival rates for GC patients with high AnxA1 expression were significantly lower than those with negative/low expression (P = .0002) (Fig. 1C). On multivariate analysis, only old age, stage III or IV disease, lymphovascular invasion, peritoneal seeding, and high AnxA1 expression were independent risk factors for poor overall survival in GC patients (Table 2).

Table 2. Cox Proportional Analysis for the Predictors of Mortality in Gastric Cancer Patients
Multivariate AnalysisHazard Ratio(95% CI)P
  1. AJCC indicates American Joint Commission on Cancer; AnxA1, annexin A1; CI, confidence interval.

Age, y   
 >60 vs ≤602.21(1.33-3.68).002
Stage (AJCC 7th)   
 III, IV vs I, II2.84(1.06-7.62).038
Lymphovascular invasion   
 Present vs absent2.36(1.28-4.35).006
Peritoneal seeding   
 Present vs absent2.09(1.01-4.32).046
 High vs negative/low1.74(1.04-2.91).037

AnxA1 Promotes Migration and Invasion of Human GC Cells

Because the AnxA1 expression level correlates with tumor invasion in GC patients, we investigated whether AnxA1 affected the invasiveness of human GC cells. Both sense-AnxA1–transfected AGS and N87 cancer cells showed significantly increased migratory and invasive abilities with a dosing effect in comparison with control (Fig. 2A,B). Functional inhibition of AnxA1 with antisense-AnxA1–expressing vectors significantly impaired the migration and invasiveness of TSGH9201 cancer cells (Fig. 2C). Consistently, shRNA-mediated AnxA1 knockdown in MKN45 cancer cells could reduce the migratory and invasive capabilities to 30% and 20%, respectively (Fig. 2D, AnxA1 shRNA1). To rule out proliferation effects, we compared the growth rates of MKN45/AnxA1 shRNA cells to the control, and the growth rate was the same, suggesting that the functional inhibition was not related to different growth rates.

thumbnail image

Figure 2. AnxA1 regulates migratory and invasive ability of GC cells. (A) Upper panel: Western blot analysis showed AnxA1 expression using protein lysate from AGS cancer cells that were transfected with S-AnxA1-His expression vectors. Lower panel: Two-chamber migration/invasion assay is shown. The histogram shows the percentage of cells that penetrated the membrane. (B,C,D) Western blot analysis and migration/invasion assay is depicted in N87, TSGH9201, and MKN45 cancer cells. (E) Severe combined immunodeficient mice were inoculated intraperitoneally with MKN45 cancer cells stably infected with control (n = 5), AnxA1 short hairpin RNA1 (shRNA1) (n = 5), or AnxA1 shRNA2 (n = 5). The group inoculated with MKN45/AnxA1 shRNA1 cancer cells had significantly reduced intraperitoneal nodules compared with the control group. (F) Representative images are shown of the parietal peritoneum and mesentery from both groups. *P < .05 as compared to the control.

Download figure to PowerPoint

AnxA1 Promotes Intraperitoneal Metastases In Vivo

To elucidate whether modulation of AnxA1 could affect the in vivo intraperitoneal metastasis process of GC cells, we performed intraperitoneal inoculation of GC cells in severe combined immunodeficient mice. Intraperitoneal nodules were significantly reduced in the group with MKN45/AnxA1 shRNA1 cells compared with the group with control cells (Fig. 2E,F). The above results suggest that AnxA1 plays an important role in promoting intraperitoneal metastases of GC in mice.

AnxA1 Up-Regulates Invasion Through FPRs in GC Cells

By western blot analysis, AnxA1 secretion increased in the condition medium after overexpression of AnxA1 in AGS cancer cells (Fig. 3A). After addition of an AnxA1 neutralizing antibody, the increased capability of invasion after overexpressing of AnxA1 was diminished (Fig. 3B). The result suggested an outside-in signaling pathway was involved in AnxA1-induced cell invasion. The invasive ability of AGS cancer cells overexpressing AnxA1 significantly decreased after treatment with Boc2, a nonselective FPR antagonist (Fig. 3C). Furthermore, we performed lentiviral shRNA inhibition of FPR1, FPR2, and FPR3 to examine whether specific FPR patterns existed for GC. GC cell invasion significantly attenuated after we individually inhibited the 3 FPRs (Fig. 3D,E). These data suggest that all 3 FPRs were involved in AnxA1-induced GC cell invasion.

thumbnail image

Figure 3. AnxA1 regulates GC cell invasion with signaling through FPRs. (A) Western blot analysis shows AnxA1 expression using protein lysate from AGS cancer cells that were treated with concentrated condition medium (CM). (B) Two-chamber invasion assay results are shown. AGS cancer cells were treated with concentrated condition medium only or followed by treatment with AnxA1 neutralizing antibody. (C) Two-chamber invasion assay results are shown. AGS cancer cells were transfected with S-AnxA1-His expression vectors followed by treatment with Boc2 at the indicated dose. (D) RT-PCR assay showed FPR1 through FPR3 expression in AGS cancer cells that were infected with FPR1 through FPR3 shRNAs. GAPDH, human glyceraldehyde 3-phosphate dehydrogenase. (E) Two-chamber invasion assay is shown. AGS cancer cells overexpressing AnxA1 were treated with FPR1 through FPR3 shRNAs individually. *P < .05 as compared to the control.

Download figure to PowerPoint

AnxA1 Up-Regulates Invasion Through Phosphorylation of ERK in GC Cells

Western blot analysis for downstream signaling molecules of both AGS and N87 cells transfected with AnxA1 showed increased expression of phosphorylated ERK1/2 (Fig. 4A). No significant change of phosphorylated p38 or JNK was noted after AnxA1 overexpression. Increased invasive ability of AGS cancer cells after overexpressing AnxA1 was reversed after treatment with the MAPK/ERK kinase (MEK)/ERK inhibitor, U0126 (Fig. 4B).

thumbnail image

Figure 4. AnxA1 regulates GC cell invasion by activating the FPR/ERK/ITGB1BP1 pathway in GC cells. (A) Western blot analysis for signaling molecules is shown in AGS (left panel) and N87 (right panel) cancer cells with AnxA1 overexpression. (B) Two-chamber invasion assay is shown. AGS cancer cells overexpressing AnxA1 were treated with U0126 at the indicated dose. (C) RT-PCR assay shows AnxA1-associated expression levels of the integrin family members and integrin regulators. Both AnxA1 gain-of-function and loss-of-function approaches were performed. GAPDH, human glyceraldehyde 3-phosphate dehydrogenase. (D) RT-PCR and western blot assays show AnxA1 and ITGB1BP1 expression in AGS cancer cells with AnxA1 overexpression and TSGH9201 cancer cells with AnxA1 down-regulation. (E) Two-chamber invasion assay is shown. AGS cancer cells overexpressing AnxA1 were treated with ITGB1BP1 siRNA at the indicated dose. (F) Western blot analysis shows phosphorylated ERK1/2 (p-ERK1/2), total ERK1/2 (t-ERK1/2), and ITGB1BP1 expression in AGS cancer cells that were treated with concentrated condition medium followed by treatment with FPR1 through FPR3 shRNAs individually. (G) Western blot analysis showed AnxA1 and ITGB1BP1 expression in AGS cancer cells overexpressing AnxA1 that were treated with or without U0126. *P < .05 as compared with the control.

Download figure to PowerPoint

AnxA1 Activates ERK/ITGB1BP1 Pathway Through FPRs in GC Cells

Because previous reports had emphasized the importance of integrins in tumor invasion and peritoneal involvement in GC,22-24 we investigated the possible role of integrins in AnxA1-related GC invasion. We screened the integrin family members and associated integrin regulators by RT-PCR. Both β4 integrin and ITGB1BP1 were found to be up-regulated with AnxA1 overexpression and down-regulated with AnxA1 silencing in GC cells (Fig. 4C). On western blot analysis, β4 integrin failed to show significant association with AnxA1 expression. Yet, AGS cells overexpressing AnxA1 showed increased levels of ITGB1BP1, whereas TSGH9201 cells transfected with AS-AnxA1 expression vectors showed decreased expression of ITGB1BP1 (Fig. 4D). Increased invasive ability of AnxA1-overexpressing AGS cells was attenuated after treatment with ITGB1BP1 siRNA (Fig. 4E). It suggested that ITGB1BP1, at least in part, was responsible for AnxA1-induced cell invasion. We found that shRNAs of all 3 FPRs inhibited AnxA1-activated phosphorylation of ERK and subsequent ITGB1BP1 expression (Fig. 4F). Furthermore, the MEK/ERK inhibitor, U0126, decreased activated ITGB1BP1 expression after AnxA1 overexpression on AGS cancer cells (Fig. 4G). Taken together, we demonstrated that AnxA1 induced GC cell invasion through activating FPRs and subsequent ERK phosphorylation and ITGB1BP1 induction.


  1. Top of page
  2. Abstract

GC invasion with subsequent peritoneal metastasis is a major cause of death in patients with advanced GC.25 This study demonstrated that high AnxA1 expression was associated with more serosal invasion and peritoneal metastasis. GC patients with high AnxA1 expression in tumors had worse overall survival in comparison with patients who had negative/low AnxA1 expression in tumors. Furthermore, high AnxA1 expression was an independent prognostic factor for predicting poorer overall survival on Cox multivariate analysis. Higher AnxA1 expression was shown in primary tumors and metastatic lymph nodes of colorectal cancer,26 and was associated with higher pathologic T stage and distant metastasis in esophageal adenocarcinoma.27 However, loss of AnxA1 expression was associated with more advanced stage in prostate adenocarcinoma28 and poorer tumor differentiation in esophageal squamous cell carcinoma.29 The conflicting expression of AnxA1 in different tumors might suggest a tumor type–specific pattern.

Using tissue microarray analysis, Xue et al showed overexpression of AnxA1 in advanced GC,13 but Yu et al reported that loss of AnxA1 expression was associated with advanced T stage and nodal metastasis in GC.14 In tissue microarray analysis for cancer translational studies, loss of immunoreactivity in stored sections of specimens has been reported to be an important problem, with a false negative rate up to 32.3%.30 In this study, we used whole tissue sections of GC for immunohistochemistry analysis to avoid such bias. In addition, the antibody preparations without appropriate controls may provide variable and even erroneous results.31 For antibody specificity confirmation, we performed immunostaining in an in vitro knockout test. We also applied both in vitro and in vivo experiments to evaluate AnxA1 effects and found a positive correlation between AnxA1 expression and increased invasion ability of cancer cells in GC. Our study is the first complete study to investigate the AnxA1 effects from clinical observation to laboratory validation in GC.

With respect to the mechanisms by which AnxA1 regulates the invasive ability of GC cells, our data suggest a novel mechanism: expression of AnxA1 within GC cells could transcriptionally up-regulate membrane FPRs and, in turn, promote ITGB1BP1 via ERK phosphorylation activation. Although the cell-specific mechanism is only partially unraveled in neutrophils and macrophages, AnxA1 appears to undergo externalization and/or secretion before further interaction with membrane FPRs.9 We found that AnxA1 secreted in the condition medium could promote GC cell invasion in an outside-in signal transduction pattern, whereas an AnxA1-neutralizing antibody could counteract this effect. These results suggest an autocrine/paracrine regulatory mechanism for AnxA1 in GC cells. Recently, extracellular AnxA1 has been shown to regulate SKCO-15 colon cancer cell migration and/or invasion via FPRs, but the chemical pan-antagonist Boc2 could not further clarify the specific FPR(s) involved.32 We took advantage of specific shRNAs for each human FPR (FPR1, FPR2, or FPR3) and found that functional inhibition of any 1 of the 3 FPRs attenuated GC cell invasion by down-regulating ERK phosphorylation and subsequent ITGB1BP1 expression. Our results suggested, for the first time, that all 3 FPR members took part in the signal regulation process of AnxA1 in GC. Recently, Yi et al showed lack of AnxA1 expression in the tumor stroma may reduce tumor progression and metastasis in an AnxA1-knockout mouse model.33 These data supported the outside-in signal transduction mechanism of AnxA1, as proposed in this study.

MAPK/ERK pathway activation is a common event in tumorigenesis, and plays a key role in cancer progression and invasion by regulating cell migration, proteinase induction, and apoptosis.34 Integrins, cytokine receptors, and G protein–coupled receptors including FPRs could activate the MAPK/ERK pathway.35 Besides, overexpression of AnxA1 has been shown to cause constitutive activation of the MAPK/ERK pathway in monomyelocytic cells,18 vascular smooth muscle cells, and kidney cells.19 In this study, we found that AnxA1 expression had a dose-dependent effect on regulating ERK phosphorylation activation and subsequent invasion in GC cells. We found that only ERK1/2 MAPKs, but not p38 or JNK MAPKs, were phosphorylated in GC cells that were activated by overexpression of AnxA1. The results were different from previous findings obtained with myeloid cells or glioblastoma cells activated by the peptide agonist formyl-methionyl-leucyl-phenylalanine (fMLF),36 but similar to those obtained with kidney cells activated by N-terminal peptide Ac2-26 of AnxA1.37

In this study, we focused on the AnxA1-related cancer cell invasiveness in GC. However, because AnxA1 has diverse roles in cancer development, the ERK pathway may mediate more actions on tumor progression. Although we observed that the in vitro growth rates between AnxA1 knockdown MKN45 cells and the control cells were not different within 3 days of the invasion assay, we could not exclude the possibility of long-term effects of AnxA1 on in vivo tumor proliferation. Recently, impaired tumor growth of lung cancer and melanoma in the AnxA1-knockout mice model has been shown.33 Future studies will clarify whether ERK1/2 or other MAPKs are involved in the regulation process of AnxA1-related tumor growth.

Several studies have shown that expression of β1 integrin was associated with peritoneal metastasis of GC,22-24 and functional blocking of β1 integrin significantly reduced GC cell adhesion in vitro and peritoneal metastasis in vivo in our previous study.21 However, we did not find β1 integrin to be affected by AnxA1 expression in this study. Notably, we observed that the β1-specific integrin regulator ITGB1BP1 acted in accordance with AnxA1 levels. ITGB1BP1 is a small cytoplasmic protein specifically interacting with the carboxy-terminal region of the β1 integrin cytoplasmic domain,38 and is regarded as a member of the specific cytosolic integrin regulators.39 ITGB1BP1 overexpression has negative regulatory effects on β1 integrin with subsequently increased cell motility.40 We observed that enforced expression of AnxA1 in GC cells up-regulated ITGB1BP1 level, whereas reduced AnxA1 expression down-regulated ITGB1BP1, which was evidenced in both RNA and protein levels. Rodrigues-Lisoni et al observed that ITGB1BP1 was down-regulated after either AnxA1 or N-terminal AnxA1 peptide treatment of HEp-2 laryngeal squamous cancer cells.41 Tumor type specificity might contribute to this discrepancy. In our study, stepwise interference with blocking agents in the postulated FPRs/ERK/ITGB1BP1 signaling pathway did affect downstream molecule expression and cell invasion. These results provide a new mechanism that demonstrates how AnxA1 regulates GC invasion.

Recently, Zhu et al reported that a small subset (11.5%) of GC specimens showed nuclear localization of AnxA1, which was associated with peritoneal dissemination and advanced tumor stage.42 In our study, only sporadic cells showed weak nuclear staining of AnxA1 in some GC specimens. The staining pattern is different from that of oral squamous cell carcinoma.43 Our findings suggest that nuclear translocation of AnxA1 might play a minor role in signal transduction in GC cells.

In summary, high AnxA1 expression was significantly associated with more serosal invasion, more peritoneal metastasis, and poorer overall survival in GC patients. These data suggest that AnxA1 expression may be used to predict patient outcome in GC. Furthermore, the in vitro studies illustrate a novel regulatory mechanism involving FPRs, ERK1/2, and ITGB1BP1 by which AnxA1 regulates GC cell invasiveness. These findings contribute important information about the role of AnxA1 in GC cells.


  1. Top of page
  2. Abstract

Grant support is provided by National Science Council, Taiwan grants NSC97-2314-B-002-117 and NSC99-2314-B-002-086.


The authors made no disclosure.


  1. Top of page
  2. Abstract