Enhancement of colon carcinogenesis by prostaglandin E ₂ administration

Abstract

Although an accumulating body of evidence indicates that levels of prostaglandin E ₂ (PGE ₂ ) in human and rodent colon cancers are higher than those in surrounding normal tissues, the precise contribution of PGE ₂ to the process of colon cancer development has still been unclear. Therefore, we designed a study using a well-established azoxymethane (AOM)-induced colon carcinogenesis in male F344 rat model to investigate whether administration of exogenous PGE ₂ has a real impact on colon carcinogenesis. Intraperitoneal PGE ₂ injections (7.7 µg) once a week for 25 weeks significantly increased the AOM-induced colon tumor incidence (percent rats with tumors, 92 versus 53%, P < 0.05), especially adenocarcinomas (92 versus 47%, P < 0.05), and multiplicity (number of tumors per rat, 2.8 versus 1.0, P < 0.05). PGE ₂ treatment significantly increased 5-bromo-2′-deoxyuridine (BrdUrd) labeling index (11.8 versus 9.7%, P < 0.05) and reduced apoptotic index (0.34 versus 0.53%, P < 0.05) in colon cancers induced by AOM. PGE ₂ exhibits its physiological functions through binding to E-prostanoid (EP) membrane receptors EP _1-4 . All four types of EP receptors were detected in AOM-induced colon cancers using reverse transcription–polymerase chain reaction (RT–PCR). Our results provide evidence that PGE ₂ enhances colon carcinogenesis through induction of cell proliferation and reduction of apoptosis.

ACF, aberrant crypt foci, AOM, azoxymethane, BrdUrd, 5-bromo-2′-deoxyuridine, COX, cyclooxygenase, EP, E-prostanoid, NSAID, non-steroidal anti-inflammatory drug, PG, prostaglandin, RT–PCR, reverse transcription–polymerase chain reaction

Introduction

Individuals taking non-steroidal anti-inflammatory drugs (NSAIDs) demonstrate a 40–50% reduction in the relative risk of colorectal cancer and colorectal cancer-associated mortality ( 1 , 2 ), and there is a large body of evidence that inhibition of cyclooxygenase (COX) is one of the underlying mechanisms ( 3 , 4 ). There are two isoforms of COX, referred to as COX-1 and COX-2. COX-1 is constitutively expressed in most tissues that generate prostaglandins (PGs) for normal physiological functions and does not fluctuate due to stimuli, whereas COX-2 expression is induced by various agents, such as cytokines, growth factors and tumor promoters ( 5 , 6 ). As COX-2 up-regulation is observed in colon tumors in humans and rodents, this enzyme may play a pivotal role in colon carcinogenesis ( 7 , 8 ). In fact, COX-2 gene deficiency reduces intestinal polyp formation in Apc^Δ716 knockout mice ( 9 ) and overexpression in colorectal cancer cells increases their growth and invasiveness ( 10 , 11 ). Furthermore, selective COX-2 inhibitors cause significant reduction in adenoma burden in adenomatous polyposis coli patients ( 12 ) and azoxymethane (AOM)-induced colon cancer development in animals ( 13 ).

With regard to compounds produced by the actions of COX, such as PGs, data are relatively limited although elevation of serum and mucosal levels of PGE ₂ have been shown in colon cancers ( 14 , 15 ). Recently, two reports documented the effects of exogenous prostaglandin E ₂ (PGE ₂ ) on intestinal polyp formation in Min/ + mice, which possess a germ line mutation in the Apc gene and are employed as a model of familial adenomatous polyposis in humans ( 16 , 17 ). One report indicated that PGE ₂ exhibits inhibitory effects reflected reduced number and size of polyps although paradoxically PGE ₂ increased cell proliferation in polyps ( 16 ). The other report demonstrated that exogenous PGE ₂ administration increased the reduced number of polyp formation by NSAIDs through the elevation of the intracellular Ca ²⁺ concentration ( 17 ). These results confuse our ideas for effects of PGE ₂ on colon carcinogenesis and exact underlying mechanism is still unclear. Some of the reasons may be due to using Min/ + mice model because Min/ + mice induce intestinal polyps mainly in small intestine and cover the early stages of intestinal tumorigenesis. On the other hand, another animal model for colon cancer using F344 rats and AOM as a carcinogen can develop colon cancers. The spectrum of colonic lesions in this model is similar to the various types of neoplastic diseases in the human colon including adenoma and adenocarcinoma sequence. Biological behaviors of AOM-induced rat colon tumors have close similarity to those of human colon tumors ( 18 , 19 ).

In addition, PGE ₂ exerts biological activity through binding to its receptors, E-prostanoid (EP) _1-4 . Our previous studies indicated that EP ₁ and EP ₄ receptors are involved in colon carcinogenesis using knockout mice and specific antagonists ( 20 , 21 ). Therefore, we hypothesized that PGE ₂ is strongly associated with colon carcinogenesis.

In order to clarify the exact role of PGE ₂ on colon carcinogenesis, we designed the present study using AOM-induced colon carcinogenesis in vivo rodent model. Cell proliferation and apoptosis in colon cancers and mucosa tissues to elucidate mechanism of PGE ₂ were examined. Also, EP receptor profile in colon carcinogenesis induced by AOM was investigated using reverse transcription–polymerase chain reaction (RT–PCR).

Materials and methods

Animals and chemicals

Thirty-nine male F344 rats, 4 weeks of age, purchased from Charles River Japan (Atsugi, Japan), were quarantined for 1 week and then randomly assigned to one of four treatment groups. The treatment groups were: (i) AOM alone; (ii) AOM + PGE ₂ ; (iii) PGE ₂ alone; and (iv) saline treatment (negative controls). All the animals were housed three to a plastic cage in a holding room controlled at 23 ± 2°C, 50 ± 10% humidity, with a 12-h light/dark cycle. Powdered AIN-76A (Dyets, Bethlehem, PA) was available, together with tap water, ad libitum during the experiment. AOM and 5-bromo-2′-deoxyuridine (BrdUrd) were purchased from Sigma Chemical (St Louis, MO). PGE ₂ β-cyclodextrin clathrate including 7.7% of PGE ₂ was kindly provided from Ono Pharmaceutical (Osaka, Japan). As PGE ₂ is a very unstable form, PGE ₂ β-cyclodextrin clathrate was synthesized by Ono Pharmaceutical as a stabilized form of PGE ₂ and confirmed by radioimmunoassay and radioactivity that pharmacologic behavior of PGE ₂ β-cyclodextrin clathrate is the same as that of PGE ₂ ( 22 ). Enzyme Immunoassay (EIA) kit for detection of blood PGE ₂ level was purchased from Cayman Chemicals (Ann Arbor, MI). The study was performed with the approval of the Institutional Animal Care and Use Committee.

Experimental procedure

Starting at 5 weeks of age, all rats were given subcutaneous injections of AOM (15 mg/kg body wt) or saline vehicle once a week for 2 weeks ( Figure 1 ). At 1 week after the last dosing of AOM or saline, they received intraperitoneal (i.p.) injections of PGE ₂ β-cyclodextrin clathrate at a dose of 100 µg or saline vehicle once a week for 25 weeks. This dose was determined by being calculated from a previous study ( 23 ). All rats were monitored daily for their general health and weighed weekly. At 32 weeks of age, all animals were killed under ether euthanasia and blood was collected from rats in each group for PGE ₂ assay, 1 h after i.p. injection of BrdUrd in saline solution (50 mg/kg body wt) and complete autopsies were performed. After laparotomy, the entire stomach and intestines were resected and opened longitudinally, and the contents were flushed with saline. Using a dissection microscope, small and large intestinal tumors were noted grossly for their location, number, and size. The length, width and depth of each tumor were measured with calipers to allow calculation of tumor volume using the formula V = L · W · D · π/6, where V is volume, L is length, W is width and D is depth ( 24 ). All other organs, with particular attention to the kidneys, lungs and liver, were also grossly examined under the dissection microscope for any abnormalities. The colon was laid flat on a glass plate, and half of each tumor and 1 in sample from each part of normal appearing colon (proximal, middle and distal) were removed and fixed in 10% buffered formalin and routinely processed for embedding in paraffin blocks for sectioning. Histological diagnosis of intestinal tumors using H&E stain sections were performed according to the classification of Pozharisski ( 25 ). The remaining normal mucosa was scraped with a stainless steel, disposable, microtome-bladed knife, S35 (Feather Safety Razor, Osaka, Japan). The other half of each colon tumor and scraped normal mucosa from each group were snap frozen in liquid nitrogen for analysis of EP receptors expression.

Fig. 1.

Experimental protocol.

BrdUrd labeling and apoptotic indices

Two serial sections (3 µm in thickness) of colon tumors and normal mucosa were used for immunohistochemical examination. One section was used for immunohistochemical detection of BrdUrd incorporation with avidin–biotin complex method (Vectastain ABC kit, Vector Laboratories, Burlingame, CA). The other section was stained with in situ end-labeling of fragmented DNA using ApopTag in situ detection peroxidase kit (Intergen Co., Purchase, NY) according to the manufacturer's instructions. For detection of BrdUrd labeling and apoptotic indices, in normal colon samples, at least 20 well-oriented crypts, in which the base, lumen and top of each could be seen by a microscope, were selected in each section of three parts of the colon. All tumor cells were counted in each colon tumor. For analyses of BrdUrd and apoptosis, only well-defined and darkly stained cells were counted using a microscope. The percentage of labeled cells (labeling index) was determined by calculating the labeled cell number: total cell number × 100.

Blood levels of PGE ₂

At death, blood was collected from rats in each group. PGE ₂ levels were analyzed using an EIA kit according to the manufacturer's instructions. Briefly, whole blood samples were purified through a C-18 reverse phase cartridge (Bond Elut, Varian, CA) to remove organic solvents. Blood samples were applied to a plate pre-coated with mouse monoclonal antibody and incubated with PGE ₂ acetylcholinesterase tracer and PGE ₂ antiserum for 18 h at room temperature. Then, all the wells were emptied and rinsed five times, and incubated with Ellman's reagent for 60 min to produce 5-thio-2-nitrobenzoic acid, which has a strong absorbance at 412 nm; the plate was read at 412 nm. We calculated the results with negative and positive controls using the standard curve. The results were expressed as picogram per millilitre. All assays were performed in triplicate.

RT–PCR for EP receptors expressions

We analyzed EP receptors expressions using RT–PCR method in normal mucosa and tumors of the colon in each group. Total RNA was isolated from half of each tumor and normal mucosa of the colon using ISOGEN (Nippon Gene, Tokyo, Japan) according to the manufacturer's instructions. Reverse transcription with random 9mers was used to generate cDNAs from 0.8 µg of total RNA extracts using reverse transcriptase (AMV Reverse Transcriptase XL) and a Takara RNA LA PCR kit (Takara Biomedical, Japan). The following primers were used for PCR amplification of the resulting cDNA: EP ₁ (182 bp) forward primer 5′-TCGTGCATCTGCTGGAGCCC-3′ (nucleotides 1510–1529), and reverse primer 5′-AGGAGGCGAAGAAGTTGGCG-3′ (nucleotides 1672–1691); EP ₂ (233 bp) forward primer 5′-GTGCTGGCTTCTTATTCGAG-3′ (nucleotides 377–396), and reverse primer 5′-AGCAAGGAGACCCCATAGAT-3′ (nucleotides 590–619); EP ₃ (218 bp) forward primer 5′-ATCCTCGTGTACCTGTCGCA-3′ (nucleotides 348–367), and reverse primer 5′-AGCCACACACCCAGCAGTA-3′ (nucleotides 547–565); EP ₄ (220 bp) forward primer 5′-CTACTTCTACAGCCACTACG-3′ (nucleotides 372–391), and reverse primer 5′-GAG-GATGAGGAAGGAACTGA-3′ (nucleotides 572–591); β-actin (203 bp) forward primer 5′-TCCTCCCTGGAGAAGAGCTA-3′ (nucleotides 741–760), and reverse primer 5′-CCAGACAGCACTGTGTTGGC-3′ (nucleotides 924–943). PCR conditions were 94°C for 120 s and then 25–35 cycles of 94°C for 30 s, 60°C for 30 s and 72°C for 30 s. β-Actin was used for the internal control to normalize the sample amounts. Agarose gels (1.5%) were stained with ethidium bromide and the band intensity was quantified using an NIH image program. All assays were performed in triplicate.

Statistical analysis

Body weights, blood PGE ₂ levels and BrdUrd labeling and apoptotic indices were analyzed among all groups. Tumor incidence, multiplicity and volume as well as apoptotic indices were compared between the animals treated with AOM alone and AOM plus PGE ₂ . Tumor incidence, expressed as the percentage of tumor-bearing animals, was analyzed by Fisher's exact probability test. Tumor multiplicity, expressed as the mean number of tumors per animal, tumor volume, blood PGE ₂ levels and BrdUrd labeling and apoptotic indices were analyzed by the unpaired Welch's or Student's t- test. Differences were considered statistically significant at P < 0.05.

Results

General observations

The body weights of animals treated with saline or AOM, with or without PGE ₂ , were comparable throughout the study, weight loss being consistently observed with PGE ₂ ( Figure 2 ). All animals were sacrificed at 26 weeks after the last AOM dosing and autopsy revealed no gross or histological changes in the colon, liver, kidneys, lungs and stomach of rats treated with saline, with or without PGE ₂ treatment. No neoplastic lesions other than colon tumors were found in rats treated with AOM.

Fig. 2.

Body weight changes during the study.

Colon tumor data

Table I summarizes data for effects of PGE ₂ on development of colon tumors induced by AOM. All tumors were classified as adenomas and adenocarcinomas through histopathological examination. There were no tumors in saline-treated animals injected with PGE ₂ or saline. Ninety-two percent of rats (11/12) treated with AOM and PGE ₂ developed colon tumors including 33% (4/12) with adenomas and 92% (11/12) with adenocarcinomas, whereas 53% (8/15) of rats treated with AOM alone had colon tumors including 7% (1/15) with adenomas and 47% (7/15) with adenocarcinomas. Regarding tumor multiplicity (number of tumors per animal), the figures were 2.8 for the AOM and PGE ₂ treatment group and 1.0 with AOM alone ( P < 0.05). Also, colon tumor volume in rats treated with AOM and PGE ₂ was 2-fold greater than that in rats treated with AOM (23 versus 11 mm ³ ), although the difference was not significant.

Table I.

Effects of PGE ₂ on the incidence, multiplicity and volume of AOM-induced colon tumors in male F344 rats

Treatment	Incidence (% of rats with colon tumors)					Multiplicity a (no. of tumors/rat)					Volume (mm 3 ) a
	Total tumors b	AD c	ADC c	Total tumors	AD	ADC
AOM-treated
Vehicle	8/15 (53)	1/15 (7)	7/15 (47)	1.0 ± 1.1	0.1 ± 0.3	0.9 ± 1.1	11 ± 6
PGE 2	11/12 (92) d	4/12 (33)	11/12 (92) d	2.8 ± 2.0 e	0.6 ± 1.0	2.2 ± 1.3 f	23 ± 21
Saline-treated
Vehicle	0/6 (0)	0/6 (0)	0/6 (0)	0	0	0	0
PGE 2	0/6 (0)	0/6 (0)	0/6 (0)	0	0	0	0

^a Values represent mean ± SD.

^b Include adenomas and adenocarcinomas.

^c AD indicates adenoma and ADC indicates adenocarcinoma.

^d Significantly different from vehicle group by Fisher's exact probability test ( P < 0.05).

^e Significantly different from vehicle group by Welch's t -test ( P < 0.05).

^f Significantly different from vehicle group by Student's t -test ( P < 0.01).

Blood PGE ₂ levels

To confirm the effects of PGE ₂ , we examined blood levels of PGE ₂ in each group ( Table II ). The mean blood PGE ₂ level in rats treated with AOM and PGE ₂ was greater than those of the other groups. As expected, PGE ₂ blood levels in rats treated with PGE ₂ were significantly higher than those of rats without PGE ₂ treatment with/without AOM ( P < 0.01). In groups without PGE ₂ treatment, blood PGE ₂ level in rats treated with AOM alone was significantly higher than that of rats treated with saline (13.6 versus 5.7 pg/ml, P < 0.05).

Table II.

Blood levels of PGE ₂ in each group

Treatment	No. of rats examined	Blood levels of PGE 2a (pg/ml)
AOM-treated
Vehicle	15	13.6 ± 6.3 b
PGE 2	12	27.1 ± 10 c
Saline-treated
Vehicle	6	5.7 ± 4.5
PGE 2	6	23.3 ± 7.1 c

^a Values represent mean ± SD.

^b Significantly different from saline-treated vehicle group by Student's t -test ( P < 0.05).

^c Significantly different from vehicle group by Student's t -test ( P < 0.01).

Cell proliferation and apoptosis

To elucidate mechanism of PGE ₂ on AOM-induced colon carcinogenesis, BrdUrd labeling and apoptotic indices were examined. The results are shown in Table III . Representative figures of BrdUrd and ApopTag staining are shown in Figure 3 . PGE ₂ treatment significantly increased BrdUrd labeling indices not only in colon cancers (11.8 versus 9.7%, P < 0.05) but also in normal mucosa (8.5 versus 5.7%, P < 0.05) of rats receiving AOM. Interestingly, PGE ₂ treatment alone did not show any effects on BrdUrd labeling index in normal mucosa of rats receiving saline (3.3 versus 3.1%), while AOM treatment significantly increased cell proliferation than saline treatment in normal mucosa (5.7 versus 3.1%, P < 0.05). Apoptotic cells in cancers of rats treated with AOM and PGE ₂ were significantly lower than those in rats treated with AOM alone (0.34 versus 0.53, P < 0.05). In normal colon mucosa, only very few apoptotic cells in rats treated with/without AOM and PGE ₂ or saline injections were observed. Since their values were <0.1%, the results did not allow us to analyze effects of PGE ₂ on apoptosis in colon normal mucosa.

Table III.

Effects of PGE ₂ on cell proliferation and apoptosis in the colon

Treatment	No. of rats examined	BrdUrd labeling index (%) a			Apoptotic index (%) a
		Normal mucosa	Cancer	Cancer
AOM-treated
Vehicle	15	5.7 ± 2.1 b	9.7 ± 2.3	0.53 ± 0.21
PGE 2	12	8.5 ± 3.3 c	11.8 ± 2.9 c	0.34 ± 0.18 c
Saline-treated
Vehicle	6	3.1 ± 2.7	N/A	N/A
PGE 2	6	3.3 ± 2.7	N/A	N/A

^a Values represent mean ± SD.

^b Significantly different from saline-treated vehicle group by Student's t -test ( P < 0.05).

^c Significantly different from vehicle group by Student's t -test ( P < 0.05).

Fig. 3.

Immunohistochemical stainings. Figures in left column indicate samples from rats treated with AOM alone and those in right column indicate samples from rats treated with AOM and PGE ₂ . Immunohistochemical stainings with BrdUrd are shown in figures in the upper two rows ( A ) and staining with ApopTag are shown in figures in the lower row ( B ). Figures in the upper row indicate normal colon mucosa with BrdUrd staining. Figures in the middle row indicate colon cancers with BrdUrd staining. Figures in the lower row indicate colon cancers with ApopTag staining.

EP receptor expressions in colon carcinogenesis induced by AOM

PGE ₂ activates seven transmembrane receptors EP _1-4 as a ligand. Using RT–PCR, EP receptor expressions were investigated in colon normal mucosa and cancers in each group ( Figure 4 ). PCR products were confirmed by sequencing analyses (data not shown). All four types of EP receptor were expressed in AOM-induced colon cancers. EP ₁ and EP ₂ receptors mRNA were overexpressed in colon cancers as compared to the levels of normal mucosa. On the other hand, EP ₃ and EP ₄ receptors mRNA were found to be similar expression levels in both colonic normal mucosa and cancers.

Fig. 4.

RT–PCR results of EP receptor expression in colonic normal mucosa and cancers induced by AOM in male F344 rats. Lanes 1–4, normal mucosa samples from rats treated with saline, PGE ₂ , AOM and AOM + PGE ₂ , respectively; lanes 5–7, adenocarcinomas from rats treated with AOM; lanes 8 and 9, adenocarcinomas from rats treated with AOM + PGE ₂ .

Discussion

The present study provided evidence that exogenous PGE ₂ administration enhances AOM-induced colon cancer development. All four types of EP receptor were expressed in AOM-induced colon cancers. Its enhancing effects are associated with induction of cell proliferation and reduction of apoptosis.

It is clear that COX-2 plays a role in the promotion of colorectal cancer ( 26 ). However, a recent report demonstrated that deficiency of either COX-1 or COX-2 causes a decrease in intestinal tumorigenesis in the Min/ + mouse with both COX-1 and COX-2 contributing to PGE ₂ production in polyps ( 27 ). It is therefore probable that the PGE ₂ level is important for tumor development, and that decreased production attributable to the loss of either isoform significantly reduces development of the lesions. DuBois and his group indicated that decreased cell death caused by PGE ₂ through up-regulated Bcl-2 expression would enhance the tumorigenic potential in human colon cancer cells ( 28 ) and PGE ₂ can induce phenotypic alterations including increased motility, changes in cell shape and stimulation of cell growth in colon cancer cells through the phosphatidylinositol 3-kinase/Akt signaling pathway ( 29 ). In the present study, to evaluate the effects of PGE ₂ on the behavior of in vivo colon carcinogenesis, we employed an AOM-induced colon cancer development model. Because it is a well-established rodent model, and it takes around 16 weeks and 50 weeks to develop adenomas and adenocarcinomas in the colon, respectively. This model allows us to investigate sequential analysis of colon carcinogenesis, such as ACF, adenomas and adenocarcinomas. Short life span of Min/ + mouse may make it very difficult to identify effects of PGE ₂ on intestinal tumorigenesis. Because the number of intestinal polyps may reach the maximum and plateau in Min/ + mouse at 16 weeks of age. In addition, neoplasms arise when there is abnormal accumulation of altered cells, characterized by excessive proliferation and decreased cell death. The results in this study clearly indicated that exogenous administration of PGE ₂ significantly enhances AOM-induced colon cancer development through induction of cell proliferation and reduction of apoptosis.

PGE ₂ activates seven transmembrane receptors EP _1-4 as a ligand. Our results that all four types of EP receptor were detected by RT–PCR in AOM-induced colon cancers in male F344 rats confirm that PGE ₂ signaling pathways exist in this model. We found that mRNA of EP ₁ and EP ₂ receptors are up-regulated in colon carcinogenesis-induced by AOM in male F344 rats. These results support our previous conclusion that EP ₁ and EP ₄ receptors are involved in the early event of AOM-induced colon carcinogenesis ( 20 , 21 , 30 ) and the report from Taketo and his group that EP ₂ is the major receptor mediating the PGE ₂ signaling pathway in intestinal polyp formation in Apc^Δ716 mice ( 31 ). According to binding assays, PGE ₂ has affinity for all four types of EP receptor in the rat ( 32 ) as well as the human ( 33 ), while 16,16-dimethyl PGE ₂ demonstrates binding to EP ₂ , EP ₃ and EP ₄ , but not EP ₁ and 17-phenyl-trinor-PGE ₂ plays as an agonist for EP ₁ and EP ₃ ( 34 ). In the present study, we employed PGE ₂ β-cyclodextrin clathrate for PGE ₂ because of its stability and solubility in saline. PGE ₂ β-cyclodextrin clathrate acts as PGE ₂ after solution in saline, in fact, we confirmed high blood levels of PGE ₂ in rats treated with PGE ₂ . There were two contrary results using Min/ + mouse intestinal polyp formation model. A report from Wilson and Potten indicated that 16,16-dimethyl PGE ₂ reduces both number and size of intestinal polyps in Min/ + mice ( 16 ). One of the reasons for inhibitory effect of 16,16-dimethyl PGE ₂ on intestinal polyp formation may be related to its binding affinity to EP receptors. Because 16,16-dimethyl PGE ₂ cannot activate EP ₁ receptor which plays an important role in intestinal polyp formation in Min/ + mice since a selective EP ₁ antagonist significantly inhibits intestinal polyp formation in this model ( 20 ). Another study using the Min/ + mouse model by Hansen-Petrik et al . indicated that administration of a mixture of 16,16-dimethyl PGE ₂ and 17-phenyl-trinor PGE ₂ attenuates the reduced intestinal polyp formation by NSAIDs treatment ( 17 ). Based on binding assays, a mixture of 16,16-dimethyl PGE ₂ and 17-phenyl-trinor PGE ₂ can activate all four types of EP receptor. In addition, they found that NSAIDs decrease intracellular Ca ²⁺ concentration in tumors of Min/ + mice and the effects are attenuated by treatment of 17-phenyl-trinor PGE ₂ , which can activate EP ₁ and EP ₃ receptors. As stimulation of EP ₁ receptor results in elevation of intracellular Ca ²⁺ concentration, they suggested that intracellular Ca ²⁺ plays a role to maintain colon tumor integrity. Taken together, it is possible that PGE ₂ plays a role in colon carcinogenesis and its effects may be dependent on EP receptors, especially EP ₁ receptor. In down-stream of EP receptors signaling pathways, activation of EP ₃ receptor is coupled with G _i protein, leading to an inhibition of adenylate cyclase activity and resulting in a decrease of cAMP concentration. EP ₂ and EP ₄ receptors, in contrast, are coupled with G _s protein, increasing intracellular cAMP concentration through activation of adenylate cyclase. Protein kinase A activity appears to alter proliferation and differentiation in several cancer cell lines ( 35 ). Therefore, there may be crosstalk of second messengers after ligand binding to EP receptors and all four types of EP receptor may be involved in colon carcinogenesis.

In conclusion, the present study provides additional evidence that exogenous administration of PGE ₂ indeed enhanced development of AOM-induced colon cancers in in vivo rodent model through up-regulation of cell proliferation and down-regulation of apoptosis. All four types of EP receptors express in AOM-induced colon cancers. Additional studies to investigate the effects of EP receptor deficiency in mice on colon carcinogenesis using cancer as an end point are now ongoing in our laboratory.

¹ Present address: Department of Pathology and Laboratory Medicine, Medical University of South Carolina, 165 Ashley Avenue, Suite 309, Charleston, SC 29425, USA

We thank Ono Pharmaceutical Co. Ltd. for generously supplying PGE ₂ β-cyclodextrin clathrate. We thank Dr Masahiko Watanabe for his helpful discussion on RT–PCR method and Ms Yurika Teramoto for her excellent technical assistance. This work was supported in part by Grants-in-Aid for Cancer Research from the Ministry of Health, Labour and Welfare, Japan, and for Scientific Research (C) from the Japan Society for the Promotion of Science, the Ministry of Education, Culture, Sports, Science and Technology, Japan (12670226 to T.K.). This study was also supported by grants from the Second Term Comprehensive 10-year Strategy for Cancer Control and a Research Grant of the Princess Takamatsu Cancer Research Fund (99-23104 to T.K.).