Cyclooxoygenase (COX)-2 overexpression is involved in gastric carcinogenesis. While high-salt intake is a known risk factor for gastric cancer development, we determined the effects of high salt on gastric chemical carcinogenesis in COX-2 transgenic (TG) mice. COX-2 TG mice were developed in C57/BL6 strain using the full-length human cox-2 complementary DNA construct. Six-week-old COX-2 TG and wild-type (WT) littermates were randomly allocated to receive alternate week of N -methyl- N -nitrosourea (MNU, 240 p.p.m.) in drinking water or control for 10 weeks. Two groups of mice were further treated with 10% NaCl during the initial 10 weeks. All mice were killed at the end of week 50. Both forced COX-2 overexpression and high-salt intake significantly increased the frequency of gastric cancer development in mice as compared with WT littermates treated with MNU alone. However, no additive effect was observed on the combination of high salt and COX-2 expression. We further showed that MNU and high-salt treatment increased chronic inflammatory infiltrates and induced prostaglandin E 2 (PGE 2 ) production in the non-cancerous stomach. Whereas high-salt treatment markedly increased the expression of inflammatory cytokines (tumor necrosis factor-α, interferon-γ, interleukin (IL)-1β and IL-6) in the gastric mucosa, COX-2 overexpression significantly altered the cell kinetics in the MNU-induced gastric cancer model. In conclusion, both high salt and COX-2 overexpression promote chemical-induced gastric carcinogenesis, possibly related to chronic inflammation, induction of PGE 2 , disruption of cell kinetics and induction of inflammatory cytokines.
Gastric cancer is the second commonest cancer in the world. It is particularly prevalent in East Asian countries where it remains the leading cancer killer ( 1 ). While advanced gastric cancer is associated with high morbidity and mortality, early detection and prevention is the best approach in reducing gastric cancer incidence and its related mortality. Despite the strong link between Helicobacter pylori infection and gastric cancer development, recent intervention trials fail to show any significant effects of H.pylori eradication alone in preventing gastric cancer development in human subjects ( 2–5 ). On the other hand, there are compelling epidemiological data to suggest that long-term use of non-steroidal anti-inflammatory drugs is associated with a significant reduction in gastric cancer risk ( 6 ). This protection is largely attributed to the inhibition of cyclooxygenase (COX) enzyme, particularly the COX-2 isoform ( 7 ). In this regard, we have shown previously that treatment with celecoxib, a specific COX-2 inhibitor, prevents gastric cancer development in a rodent model of gastric carcinogenesis ( 8 ). Nonetheless, we found that the chemopreventive effect of celecoxib appears to be unrelated to the degree of COX and prostaglandin inhibition in that study ( 9 ). Hence, it is suggested that non-COX-mediated pathway may be involved in celecoxib-mediated chemoprevention. In addition to established cancer, COX-2 expression was demonstrated in other stages of H.pylori -associated gastric carcinogenesis including chronic gastritis, intestinal metaplasia and dysplasia ( 10 , 11 ). Although enhanced COX-2–prostaglandin E synthase pathway was shown to induce hyperplastic tumor development ( 12 ), the role of COX-2 per se on gastric cancer development remains elusive. In particular, it remains unknown whether COX-2 initiates or promotes gastric tumor development.
Apart from H.pylori infection, high dietary salt intake is frequently linked to gastric cancer development. Epidemiological studies showed that high dietary salt intake is a significant risk factor for human gastric cancer development even after adjustment for H.pylori infection and atrophic gastritis ( 13 ). In mouse model, high salt has been shown to induce gastric epithelial hyperplasia, parietal cell loss as well as H.pylori colonization ( 14 ). Experimental studies also showed that salt promotes gastric carcinogenesis in animals after chemical carcinogen treatment ( 8 , 15 ). Recently, Kato et al. ( 16 ) further showed that high salt dose dependently promotes gastric chemical carcinogenesis in H.pylori -infected Mongolian gerbils.
In this study, we sought to determine the effects of forced COX-2 expression alone and in combination with high-salt intake on gastric chemical carcinogenesis in a COX-2 transgenic (TG) mouse model.
Materials and methods
COX-2 TG mice were generated by cloning of the full-length human cox-2 complementary DNA downstream of the cytomegalovirus promoter of the pcDNA3.1 plasmid. The sequence of the clone was confirmed by DNA sequencing. The linearized and purified CMV - cox-2 DNA transgene was used for microinjection into pronuclei of fertilized C57BL/6 F1 oocytes. Survived embryos were transferred into the oviducts of pseudopregnant female mice. Integration of transgene was checked by polymerase chain reaction (PCR) of tail genomic DNA. Founders were backcrossed to C57BL/6 mice to establish a heterozygote COX-2 TG line. The heterozygous TG mice line with COX-2 upregulation in the stomach was selected for subsequent study ( Figure 1 ). All TG mouse lines reproduced were phenotypically normal. Wild-type (WT) littermates were analyzed in parallel as controls. All C57BL/6 mice were maintained in metal cages on a 12 h light–dark cycle. All mice were obtained from the Laboratory Animal Services Center of the Chinese University of Hong Kong. The study protocol was approved by the Animal Experimentation Ethics Committee of The Chinese University of Hong Kong and all experiments were performed in accordance with local guidelines.
Gastric carcinoma was induced in WT and TG mice by using N -methyl- N -nitrosourea (MNU) as described previously ( 17 , 18 ). Briefly, 6-week-old WT or TG mice were randomly allocated to six different groups to receive MNU and/or high-salt treatment as shown in Figure 2 . MNU was given in five alternate biweekly cycles during the initial 10 weeks. It was freshly prepared twice a week in distilled water at concentration of 240 p.p.m. and given to mice ad libitum as drinking water in light-shielded bottles on alternate week for 10 weeks. In addition, two groups of mice (Groups C and F) were given 0.1 ml of 10% NaCl once weekly during the initial 10 weeks MNU treatment. All animals were killed at the end of week 50 by cervical dislocation. Stomach was opened along the greater curvature, examined macroscopically for tumor and fixed in 10% buffered formalin or snap frozen for subsequent histological and molecular analysis.
Histopathology and immunohistochemistry
Gastric specimens were formalin fixed and paraffin embedded for histological examination. Sections of 5 μm were stained with hematoxylin and eosin for histological diagnosis by an experienced pathologist who was unaware of the treatment allocation and genetic background of the mice. Gastric cancer was defined as the unequivocal presence of invasive adenocarcinoma in the glandular stomach ( Figure 3 ). Severity of gastric inflammation including acute inflammation, chronic inflammation, atrophy and intestinal metaplasia was determined based on the updated Sydney classification ( 19 ).
COX-2 immunostaining was performed by the ABComplex/HRP method (Dako, Carpinteria, CA) using the anti-COX-2 rabbit monoclonal antibody (NeoMarkers, Fremont, CA). Heat-induced antigen recovery was performed before the immunoreactions in a moist chamber. The reactions were visualized with diaminobenzidine substrate and counterstained with hematoxylin solution.
Quantification of apoptotic and proliferation index
Apoptosis was determined by terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate-digoxigenin nick end labeling and proliferation was quantitated by Ki-67 immunostaining as described previously ( 9 ). Terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate-digoxigenin nick end labeling assay was performed with DeadEndTM Colorimetric TUNEL System (Promega, Madison, WI) as suggested by the manufacturer. Ki-67 immunostaining was performed by the LAB-SA Detection System (Zymed, S. San Francisco, CA) using the anti-Ki-67 monoclonal antibody (Lab Vision, Fremont, CA). Heat-induced antigen recovery was performed and reactions were visualized with diaminobenzidine substrate counterstained with hematoxylin solution. All histological sections were examined in high-power fields (400×). A random starting field was selected and then other field was examined for a total of >1000 gastric epithelial cells. The apoptotic index and Ki-67 labeling (or proliferation) index were presented as the number of positive-stained nuclei among 1000 epithelial cells.
Prostaglandin E 2 assay
Prostaglandin E 2 (PGE 2 ) levels of gastric tissues were determined by enzyme immunoassay kit (Amersham Pharmacia Biotech, Piscataway, NJ). Briefly, 20 mg of snap-frozen gastric tissues were homogenized in 20 volume of ethanol using a ground glass homogenizer. Ice-cold water was added to give a final ethanol concentration of 15% and centrifuged for 10 min at 400 g . A 10 ml volume of glacial acetic acid was added to each sample to pH 3.0. The supernatant was then applied to a preprimed Amprep C18 mini column (Amersham Pharmacia Biotech). The eluted PGE 2 was evaporated to dryness under nitrogen and stored at −80°C. Samples were resuspended in 1 ml of buffer and assayed in 96-well plates by enzyme-linked immunosorbent assay.
The COX-2 protein expression in the normal non-cancerous stomach of different groups of rats was determined by western blot. Mouse gastric tissue was homogenized in Cytobuster reagent (Novagen, Madison, WI) with protease inhibitor cocktail (Roche, Indianapolis, IN) and protein concentration was measured by the method of Lowry (Bio-Rad, Hercules, CA). Fifty microgram of protein was loaded per lane, separated by 10% sodium dodecyl sulfate–polyacrylamide gel electrophoresis under reducing conditions and transferred onto nitrocellulose membrane (Amersham, Buckinghamshire, UK) by electroblotting. Membranes were blocked by 5% non-fat dry milk in Tris-buffered saline for 1 h. COX-2 protein was detected with a polyclonal antibody (1:1000; Santa Cruz Biotechnology, Santa Cruz, CA) for overnight at 4°C, followed by 1 h of anti-goat-horseradish peroxidase antibody (1:1000; Cell Signaling Danvers, MA) incubation. Enhanced chemiluminescence (Pierce, Rockford, IL) was determined by exposure to x-ray film (Fuji, Dusseldorf, Germany). Western blot for β-actin was also performed as an internal (sample loading) control using a monoclonal antibody (Cell Signaling).
Real-time reverse transcription–PCR
Total RNA was extracted from gastric tissues by using RNA Trizol reagent (Invitrogen, Carlsbad, CA) according to the manufacturer’s protocol. Five micrograms of total RNA was reverse transcribed into complementary DNA. RNA levels of inflammatory cytokines, tumor necrosis factor (TNF)-α, interferon (IFN)-γ, interleukin (IL)-1β, IL-6 and IL-10, were quantified by real-time reverse transcription–PCR using SYBR Green Master Mix (Applied Biosystems, Foster City, CA). The sequences of primers were listed in supplementary Table 1 (available at Carcinogenesis Online). β-Actin served as an internal control for total complementary DNA content. Samples were amplified using the ABI Prism 7700 Sequence Detection System (Applied Biosystems). All PCRs were run in triplicates to ensure reproducibility.
Numerical data were presented as mean ± SD unless otherwise specified. Multiple group comparison was made by one-way analysis of variance with Bonferroni’s adjustment. All statistical analysis was performed by GraphPad InStat (ver 3.0, GraphPad Software, San Diego, CA). A two-sided P value of <0.05 was considered statistically significant.
Verification of COX-2 expression in TG mice
The TG mice line with COX-2 overexpression in the gastric epithelium was selected for this study. There was a marked increase in COX-2 expression in the glandular stomach of TG mice as shown by immunohistochemistry ( Figure 1A and B ) and western blot ( Figure 1C ). In keeping with COX-2 expression, the level of PGE2 in the gastric mucosa of the TG mice was ∼4-fold higher than that of WT mice ( P = 0.016; Figure 1D ). There was no significant histological change in the gastric mucosa of TG mice.
Gastric cancer development in COX-2 TG mice
At the end of week 50, none of the control WT (Group A) or control COX-2 TG (Group D) littermates developed gastric cancer in the absence of MNU treatment. In contrast, gastric cancer was detected in 25% of WT mice treated with MNU alone (Group B). Additional treatment with 10% NaCl (Group C) further increased the rate of gastric cancer to 46.9% in WT mice ( P = 0.014 versus Group B; Figure 3 ). After MNU treatment, COX-2 TG mice (Group E) had a markedly higher gastric cancer rate than WT mice (47.5%; P = 0.007 versus Group B). Addition of 10% NaCl, however, did not further increase the gastric cancer incidence in COX-2 TG mice (48.1%, Group F).
Gastric inflammation scores
The pattern of chronic inflammatory infiltrates appears to parallel the rates of gastric cancer in different groups of mice ( Figure 4 ). While there were no chronic inflammatory infiltrates in WT and TG mice alone, the addition of MNU or 10% NaCl significantly induced chronic inflammatory cells infiltration ( P = 0.009). Similar trend was also observed in the atrophy and metaplasia scores but the difference did not reach statistical significance. The acute inflammatory scores did not appear to correlate with gastric cancer incidences in different groups of mice.
PGE 2 levels and COX-2 expression
We next determined the PGE 2 levels in the gastric mucosa of different groups of mice. The PGE 2 levels in the normal non-cancerous gastric mucosa of different groups are shown in Figure 5 . There was a significant difference in PGE 2 production among different groups ( P = 0.003). While COX-2 TG mice have a marked induction in PGE 2 production when compared with WT mice, treatment with MNU increased the PGE 2 levels in WT by >5-fold. Similar induction of PGE2 was not detected in COX-2 TG mice treated with MNU (Group E). Notably, high salt further increased the PGE 2 production by ∼2-fold in both WT and COX-2 TG mice treated with MNU (Groups C and F). There was, however, no significant difference in the expression of COX-2 protein after MNU and/or NaCl treatment in both WT and TG mice (data not shown).
Inflammatory cytokine expressions
Apart from PGE 2 , we also determined the inflammatory cytokine expressions in the normal non-cancerous gastric epithelium of different groups of mice ( Figure 6 ). The levels of five inflammatory cytokines including TNF-α, IFN-γ, IL-1β, IL-6 and IL-10 were determined by quantitative reverse transcription–PCR. There was a marked difference in the expression levels of the five inflammatory cytokines among the six different groups ( P < 0.0001). Specifically, treatment with 10% NaCl significantly induced the expression of TNF-α, IFN-γ, IL-1β and IL-6 in WT mice treated with MNU (Group C; P < 0.001 versus Groups A, B, D and E). In COX-2 TG treated with MNU (Group F), addition of 10% NaCl enhanced all the five cytokines expression ( P < 0.01 versus Groups A, B, D and E).
Apoptosis and proliferation changes
We also compared the apoptosis and proliferation index in the non-cancerous stomach of the mice ( Figure 7A ). There was a significant difference in the apoptotic ( P = 0.003; Figure 7B ) and proliferation ( P = 0.03; Figure 7C ) indexes among different treatment groups. COX-2 TG (Group D) or treatment with salt (Groups C and F) did not appear to significantly alter the basal apoptotic and proliferation index of gastric epithelium. However, the addition of MNU to COX-2 TG mice resulted in the highest apoptotic ( P < 0.01 versus Group A control) and proliferation index ( P < 0.05 versus Group A control) among all six groups.
The current study showed that forced COX-2 expression in mice alone fails to induce gastric cancer development but rather enhances gastric cancer formation induced by chemical carcinogen MNU. This is the first piece of evidence to demonstrate the temporal effect of COX-2 on gastric carcinogenesis that appears to be more on promotion rather than initiation of cancer development. This is in keeping with our previous observations that COX-2 overexpression is detected in different stages of H.pylori -associated gastric carcinogenesis including chronic gastritis, glandular atrophy, intestinal metaplasia and cancer ( 10 , 11 ). Intuitively, forced overexpression of COX-2 promotes gastric cancer development in the multistep gastric carcinogenesis cascade.
On the other hand, high-salt intake has been suspected to be an important risk factor for gastric cancer. Based on epidemiological evidence, a Joint WHO/FAO Expert Consultation concluded that high-salt intake probably increase the risk of stomach cancer ( 20 ). In the current study, we showed that treatment with 10% NaCl for 10 weeks markedly promoted gastric cancer development in WT mice treated with MNU. The percentage of WT mice that developed gastric cancer increased from 25 to 46.9% after high-salt treatment. The effect of salt on cancer promotion in this mouse model is comparable with the effect of COX-2 overexpression, but there was no additive effect noted on the combination of high salt and COX-2 expression. In Mongolian gerbils, high-salt diet dose dependently promotes gastric chemical carcinogenesis in H.pylori -infected animals ( 16 ). However, similar additive effect with H.pylori infection was not seen in another animal model using B6129 mice ( 21 ). In that animal model, both mice fed with normal or high-salt diets developed gastric intraepithelial neoplasia. In our study, there is no synergistic or additive effect of salt and COX-2 on gastric cancer development, suggesting that the effect of salt on gastric development may be related to the host makeup. The cancer-promoting effect of salt is more pronounced in WT mice only.
We next elucidated the procarcinogenic effects of COX-2 and salt in this MNU model of gastric carcinogenesis. The rates of gastric cancer in different groups of mice appear to parallel the degree of chronic inflammatory infiltrates in the stomach, suggesting that chronic inflammation contributes to gastric carcinogenesis in this animal model. We also found that the addition of high salt significantly increased PGE 2 production in non-cancerous mucosa of both WT and TG mice treated with MNU. PGE 2 is generally believed to be one of the key prostanoids responsible for gastric carcinogenesis. TG mice expressing both COX-2 and microsomal prostaglandin E synthase (mPGES)-1 developed hyperplastic gastric tumors ( 12 ). It was further shown that increased PGE2 enhances macrophage infiltration and stimulation that was mediated through epithelial Toll-like receptor-4 in the gastric epithelial cells, resulting in gastric metaplasia and tumorous growth. In the present study, the elevated PGE 2 levels cannot be solely explained by COX-2 protein expression alone. We speculate that mPGES, particularly mPGES-1, may play a role in the PGE 2 expressions among different groups of animals. Since the COX-2–mPGES-1 pathway has been shown to induce neoplastic transformation of gastric epithelial cells ( 12 ), the expression and functional role of mPGES-1 in the COX-2 TG gastric cancer model warrant further investigation.
In addition to induction of PGE 2 synthesis, there was a marked induction of inflammatory cytokines in mice treated with salt. We found that there was a significant induction of TNF-α, IL-1β, IL-6 and IFN-γ in mice treated with salt. IL-10 was also significantly induced in TG mice treated with salt and MNU. There is an intense interest on the role of inflammatory cytokines on gastric carcinogenesis, particularly in relationship to H.pylori infection. Individuals with polymorphism in proinflammatory cytokines like IL-1β, IFN-γ and TNF-α are generally considered to be at higher risk of gastric cancer development after infection with H.pylori ( 22–27 ). Our previous studies also showed that the possession of proinflammatory cytokine IL-1β increased the risk of developing gastric cancer ( 28 ) as well as premalignant gastric lesions ( 29 ) even in high-risk population. While IL-6 signaling is disrupted by H.pylori CagA protein, it was recently found that IL-6 was increased in H.pylori -infected gastric biopsy samples as well as gastric cancer samples ( 30 ). In the AGS cell line, IL-6 induces cell motility and invasion via activation of the c-Src/RhoA/ROCK-signaling pathway ( 31 ). Hence, increased IL-6 production may play an important role on gastric cancer development promoted by high salt. We also demonstrated a significant increase in IL-10 of COX-2 TG mice treated with salt. While IL-10 is generally believed to be suppressive in nature, it was found that T-cells from peripheral blood and gastric mucosa of H.pylori -infected gastric cancer patients produced higher amounts of IL-10 ( 32 ). The higher production of IL-10 may lead to a diminished cytotoxic antitumor T-cell response in the stomach that contributes to gastric cancer progression.
In this study, there were only modest alterations in cell kinetics in COX-2 TG mice or mice treated with high-salt diet. Nonetheless, COX-2 TG mice treated with MNU had the highest apoptotic and proliferation index among all treatment groups. Based on this observation, disruption of cell kinetics appears to play a more important role on gastric cancer development mediated by COX-2.
This study has limitation. Due to the large number of mice involved, a control group treated with NaCl alone is lacking. However, there is no data in the literature to support that treatment with 10% NaCl without carcinogen could induce gastric cancer in mice. Future studies may be warranted to study the inflammatory changes in mice treated with NaCl only.
In conclusion, both forced COX-2 overexpression and high-salt treatment promoted chemical-induced gastric cancer development in mice. These cancer-promoting effects are possibly contributed by various factors including chronic inflammation, increased PGE 2 production, induction of inflammatory cytokines and disruption of cell kinetics.
Joint Research Project Scheme of the Natural Science Foundation of China; Research Grant Council of Hong Kong (N_CUHK420/03).
microsomal prostaglandin E synthase
polymerase chain reaction
- PGE 2
prostaglandin E 2
tumor necrosis factor
Conflict of Interest Statement: None declared.