Recently, an inducible microsomal human prostaglandin E synthase (mPGES) was identified. This enzyme converts the cyclooxygenase (COX) product prostaglandin (PG) H2 to PGE2, an eicosanoid that has been linked to carcinogenesis. Increased amounts of PGE2 have been observed in many tumor types including colorectal adenomas and cancers. To further elucidate the mechanism responsible for increased levels of PGE2 in colorectal tumors, we determined the amounts of mPGES and COX-2 in 18 paired samples (tumor and adjacent normal) of colorectal cancer. With immunoblot analysis, mPGES was overexpressed in 83% of colorectal cancers. COX-2 was also commonly up-regulated in these tumors; marked differences in the extent of up-regulation of mPGES and COX-2 were observed in individual tumors. Immunohistochemistry revealed increased mPGES immunoreactivity in neoplastic cells in both colorectal adenomas and cancers compared with adjacent normal colonic epithelium. Cell culture was used to investigate the regulation of mPGES and COX-2. Chenodeoxycholate markedly induced COX-2 but not mPGES in colorectal cancer cells. Tumor necrosis factor-α induced both mPGES and COX-2, but the time course and magnitude of induction differed. As reported previously for COX-2, overexpressing Ras caused a several-fold increase in mPGES promoter activity. Taken together, our results suggest that overexpression of mPGES in addition to COX-2 contributes to increased amounts of PGE2 in colorectal adenomas and cancer. The mechanisms controlling the expression of these two enzymes are not identical.
There is a 40–50% reduction in the relative risk of developing colorectal cancer in individuals taking aspirin or other nonsteroidal antiinflammatory drugs (1, 2, 3). The knowledge that nonsteroidal antiinflammatory drugs inhibit COX3 made it possible to focus on the idea that PGs, the products of COX activity, are significant contributors to carcinogenesis. Increased amounts of PGE2 have been detected in colorectal adenomas and cancers (4, 5). Several lines of evidence, beyond the association between a cancer phenotype and increased levels of PGE2, suggest that PGE2 is a significant contributor to the development and progression of colorectal cancer. For example, PGE2 can stimulate angiogenesis (6) while enhancing the survival (7) and motility (8) of colon cancer cells. Immune surveillance is inhibited by PGE2 (9, 10). In addition to these mechanistic findings, studies with experimental animals suggest that PGE2 promotes tumorigenesis. In one study, treatment with anti-PGE2 monoclonal antibody retarded the growth of a transplantable tumor (11). In another study, genetic disruption of the E-prostanoid receptor subtype 2 was observed to decrease the number and size of intestinal polyps in ApcΔ716 mice, a model of human familial adenomatous polyposis (12). Given this background, it is important to define the enzymatic pathways that are dysregulated in colorectal tumors, leading to increased amounts of PGE2.
The synthesis of PGE2 from arachidonic acid requires two enzymes that act sequentially. COX catalyzes the synthesis of PGH2 from arachidonic acid. There are two isoforms of COX designated COX-1 and COX-2, respectively. COX-1 is constitutively expressed in most tissues including the colon (13, 14). By contrast, COX-2 is not expressed in the normal colon but can be induced by growth factors, cytokines, oncogenes, and tumor promoters (14, 15, 16, 17, 18, 19, 20). Elevated levels of COX-2 are detected in ∼50% of premalignant adenomas and 85% of colorectal cancers (21). Recently, an inducible, human mPGES was identified and characterized (22). This enzyme converts COX-derived PGH2 to PGE2. COX-2 and mPGES have been reported to be functionally linked (23), raising the possibility that aberrant mPGES expression could contribute to increased amounts of PGE2 in colorectal adenomas and cancer.
In this study, the expression of mPGES and COX-2 in colorectal tumors was evaluated. Overexpression of mPGES was detected in >80% of colorectal adenomas and cancers. Although COX-2 was also commonly up-regulated in colorectal tumors, mPGES and COX-2 do not appear to be regulated by precisely the same signaling mechanisms.
MATERIALS AND METHODS
Rabbit polyclonal antihuman mPGES antiserum and mPGES blocking peptide were from Cayman Chemical (Ann Arbor, MI), and anti-COX-2 antiserum was from Oxford Biomedical Research (Oxford, MI). Anti-β-actin antiserum, TNF-α, CD, 3,3′-diaminobenzidine, Lowry Protein Assay kits, and secondary antibody to IgG conjugated to horseradish peroxidase were from Sigma Chemical Co. (St. Louis, MO). DMEM, Opti-MEM, and LipofectAMINE were from Life Technologies, Inc. (Rockville, MD). Streptavidin-horseradish peroxidase was from DAKO Corp. (Carpinteria, CA). Superfrost/Plus slides were from Fisher Scientific (Pittburgh, PA). Western blotting detection reagents (enhanced chemiluminescence) were from Amersham Pharmacia Biotech. Nitrocellulose membranes were from Schleicher & Schuell (Keene, NH). The Ras expression vector was a gift from Dr. Geoffrey Cooper (Harvard University, Cambridge, MA).
Specimens from patients with colorectal adenocarcinoma were procured at the time of surgery. Tissue samples were obtained from the resected colon specimen within 15 min of its removal. A viable portion of tumor and a sample of adjacent normal colonic mucosa at a minimum distance of 5 cm from the tumor were harvested by sharp dissection. Samples were immediately stored at −80°C or fixed in neutral buffered formalin until analysis. Colorectal adenomas for immunohistochemical analysis were obtained during endoscopy and fixed as described above. The study was approved by the Committee on Human Rights in Research at the participating institutions.
The HCA7 cell line was established from a moderately well-differentiated adenocarcinoma of the human colon (24). HCA7 cells were maintained in DMEM supplemented with 10% fetal bovine serum, 100 IU/ml penicillin, and 100 μg/ml streptomycin. Treatments with TNF-α or CD were carried out in serum-free medium. The vehicles for TNF-α and CD were 0.1% PBS supplemented with 0.1% BSA and 0.1% ethanol, respectively.
Frozen tissue was thawed in ice-cold lysis buffer [150 mm NaCl, 100 mm Tris (pH 8.0), 1% Tween 20, 50 mm diethyldithiocarbamate, 1 mm EDTA, 1 mm phenylmethylsulfonyl fluoride, 10 μg/ml aprotinin, 10 μg/ml trypsin inhibitor, and 10 μg/ml leupeptin]. Tissues were sonicated for 20 s on ice and centrifuged at 10,000 × g for 10 min at 4°C to remove the particulate material. The protein concentration of the supernatant was measured using the method of Lowry et al. (25). Cell lysates were prepared as described previously (17, 19). Immunoblot analyses for mPGES and COX-2 were performed using methods described in previous studies (17, 19, 26).
mPGES immunostaining was performed as described previously (26). Neutral buffered formalin-fixed tissue was embedded in paraffin. Tissue sections (4-μm) were prepared using a microtome and mounted on Superfrost/Plus slides. Immunohistochemical analysis was performed within 24 h of the sections being cut. Sections were deparaffinized in xylene, rehydrated in graded alcohols, and washed in distilled water. Antigen retrieval was performed by steaming the sections in 10 mm citric acid (pH 6.0) for 30 min. Subsequently, endogenous peroxidase activity was blocked with 3% hydrogen peroxide. The slides were washed three times in PBS and blocked for 20 min with 5% normal goat serum. Tissue sections were then incubated with antiserum to mPGES at a 1:1000 dilution (2% BSA in PBS) for 18 h at 4°C. Control sections were incubated with mPGES antiserum preabsorbed with a 100-fold excess of mPGES blocking peptide or preimmune serum. After being washed three times with PBS, the sections were incubated with biotinylated antirabbit antibody at a 1:500 dilution for 1 h at room temperature. The slides were then washed three times in PBS and labeled using 1:500 streptavidin-horseradish peroxidase for 1 h at room temperature. The reaction was visualized using 3,3′-diaminobenzidine. Subsequently, the slides were rinsed in tap water and counterstained with hematoxylin. The slides were then dehydrated with ethanol, rinsed with xylene, and mounted.
Transient Transfection Assays.
HCA7 cells were seeded at a density of 5 × 104 cells/well in 6-well dishes and grown to 50–60% confluence. The −651/−20 and −190/−20 human mPGES promoter constructs have been described previously (27). For each well, 2 μg of plasmid DNA were introduced into cells using 4 μg/ml of LipofectAMINE in Opti-MEM medium as per the manufacturer’s instructions. After 16 h of incubation, the medium was replaced with culture medium. The activities of luciferase and β-galactosidase were measured in cellular extracts as described previously (28).
Comparisons between groups were made by using the Student’s t test. A difference between groups of P < 0.05 was considered significant.
mPGES Is Overexpressed in Colorectal Cancer.
Immunoblot analysis for mPGES was performed on 18 cases of colorectal cancer. Overall, increased expression of mPGES was detected in 15 of 18 cases (83%) of colorectal cancer (Fig. 1). COX-2 was also overexpressed in ∼80% of these tumors. Interestingly, marked differences in the extent of up-regulation of mPGES and COX-2 were observed. In cases 4 and 5, for example, expression of mPGES was markedly increased whereas COX-2 was nearly undetectable (Fig. 1). This pattern was reversed in case 6. In this instance, high levels of COX-2 but not mPGES were observed in tumorous tissue. This difference in the relative magnitude of overexpression of mPGES and COX-2 suggests that the two enzymes are regulated differently.
Immunohistochemistry revealed finely granular cytoplasmic staining for mPGES in epithelial cells of colorectal adenomas and carcinomas (Fig. 2). Immunoreactivity was distributed throughout the tumor. Consistent with the immunoblot findings, markedly increased mPGES staining was observed in 9 of 10 colon cancers (Fig. 2,C) versus adjacent normal epithelium (Fig. 2,A). In comparison with normal epithelium, increased mPGES immunoreactivity was also detected in colorectal adenomas (10 of 10 samples; Fig. 2,B) and metastatic colon cancer from lymph nodes (3 of 4 samples; Fig. 2 D) and liver (2 of 3 samples). This staining was specific for mPGES because immunoreactivity was lost when the antiserum to mPGES was preincubated with a mPGES-blocking peptide.
Comparative Effects of CD, TNF-α, and Ras on mPGES and COX-2 in Colon Cancer Cells.
To determine whether mPGES and COX-2 are regulated by similar mechanisms, we investigated whether several known inducers of COX-2 also stimulated the expression of mPGES. Bile acids promote colorectal cancer and induce COX-2 (19, 29). Treatment of HCA7 cells with CD caused dose-dependent induction of COX-2 but had no effect on mPGES (Fig. 3,A). In contrast, treatment with TNF-α induced both COX-2 and mPGES (Fig. 3,B). However, the inductive effects of TNF-α were greater and more rapid for COX-2 than mPGES. Previously, overexpression of Ras was reported to induce COX-2 (17, 20, 26). Transient transfections were performed to investigate whether Ras could stimulate mPGES promoter activity in HCA7 cells. As shown in Fig. 4, overexpressing Ras led to a marked increase in mPGES promoter activity regardless of whether the −651/−20 or −190/−20 mPGES constructs were used.
In this study, we show that mPGES was overexpressed in >80% of colorectal tumors. COX-2 is up-regulated in the majority of colorectal tumors (21, 30) and is known to be functionally linked to mPGES (23). It seems likely, therefore, that enhanced expression of mPGES in addition to COX-2 contributes to the increased amounts of PGE2 observed in colorectal adenomas and cancers (4, 5). In addition to colorectal cancer, mPGES was recently found to be overexpressed in non-small cell lung cancer (26). Whether mPGES, like COX-2, is overexpressed in a spectrum of premalignant and malignant conditions should be evaluated (21, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42). If these two enzymes are up-regulated in most tumor types, it would explain the longstanding observation that PGE2 levels are increased in multiple epithelial malignancies (4, 5, 40, 43, 44, 45). Increased amounts of COX-2 and PGE2 have also been observed in ulcerative colitis and Crohn’s disease, chronic inflammatory diseases that predispose to colon cancer (46, 47). On the basis of the results of this study, it will be worthwhile to evaluate whether mPGES is overexpressed in these inflammatory conditions.
Marked differences in the extent of up-regulation of mPGES and COX-2 were observed in some tumors. This suggests that the mechanisms controlling the expression of these two enzymes are not identical. COX-2 is regulated by transcriptional and posttranscriptional mechanisms (15, 16, 17, 18, 19, 20, 28, 29, 48). Both the cyclic AMP response element and nuclear factor for interleukin-6 regulatory element are important for regulating COX-2 expression in colon cancer cells (49). On the basis of the known sequence of the mPGES promoter (27), it is uncertain whether either of these elements plays a significant role in regulating mPGES transcription. Furthermore, the mPGES promoter lacks a TATA box, whereas one is present in the COX-2 promoter (50). The 3′-untranslated region of COX-2 mRNA contains a series of AU-rich elements that are important for mRNA stability (29, 48). Whether the stability of mPGES mRNA is tightly regulated is unknown. Ras stimulated mPGES promoter activity in HCA7 cells, a finding reported previously for COX-2 in intestinal epithelial cells (51) and mPGES in bronchial epithelial cells (26). This is relevant because of the well-established link between Ras signaling and the pathogenesis of colorectal cancer (51). By contrast, bile acids induced COX-2 but not mPGES in HCA7 cells. Bile acids induce COX-2 by stimulating transcription (19) and stabilizing mRNA (29). The fact that mPGES was unaffected by treatment with CD is consistent with there being marked differences in the regulation of these two enzymes. Interestingly, TNF-α induced both mPGES and COX-2, but the timing for induction varied. A similar difference in timing of induction was recently observed following interleukin-1β treatment of synoviocytes (52). Clearly, additional studies are warranted to further delineate similarities and differences in the regulation of mPGES and COX-2.
As detailed above, there is ample evidence that overproduction of PGE2 is protumorigenic. Selective COX-2 inhibitors prevent the synthesis of PGE2 and possess anticancer properties (53). The results of initial studies suggest that mPGES may represent a therapeutic target. More specifically, Murakami and colleagues (23) showed that cells overexpressing mPGES and COX-2 produced more PGE2, grew faster, and exhibited abnormal morphology compared with cells in which either mPGES or COX-2 were overexpressed. To more fully evaluate the role of mPGES in intestinal tumorigenesis, an experimental approach analogous to that used to investigate the link between COX-2 and intestinal tumorigenesis should be used. It will be important, for example, to determine whether knocking out one or both alleles of mPGES inhibits tumor formation in the Min mouse (54). The development of selective inhibitors of mPGES would permit the same question to be evaluated. Importantly, our observation that mPGES is overexpressed in both benign and malignant colorectal tumors provides the basis for future studies that will evaluate whether mPGES is a bona fide therapeutic target.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by the Tokyo Association for Clinical Surgery, Pharmacia Corp., the New York Crohn’s Foundation, and the James E. Olson Foundation.
The abbreviations used are: COX, cyclooxygenase; PG, prostaglandin; mPGES, microsomal prostaglandin E synthase; TNF-α, tumor necrosis factor-α; CD, chenodeoxycholate.