Prostaglandin E2 Increases Growth and Motility of Colorectal Carcinoma Cells*

Chronic use of nonsteroidal anti-inflammatory drugs results in a significant reduction of risk and mortality from colorectal cancer in humans. All of the mechanism(s) by which nonsteroidal anti-inflammatory drugs exert their protective effects are not completely understood, but they are known to inhibit cyclooxygenase activity. The cyclooxygenase enzymes catalyze a key reaction in the conversion of arachidonic acid to prostaglandins, such as prostaglandin E2 (PGE2). Here we demonstrate that PGE2 treatment of LS-174 human colorectal carcinoma cells leads to increased motility and changes in cell shape. The prostaglandin EP4 receptor signaling pathway appears to play a role in transducing signals which regulate these effects. PGE2 treatment results in an activation of phosphatidylinositol 3-kinase/protein kinase B pathway that is required for the PGE2-induced changes in carcinoma cell motility and colony morphology. Our results suggest that PGE2 might enhance the invasive potential of colorectal carcinoma cells via activation of major intracellular signal transduction pathways not previously reported to be regulated by prostaglandins.

There is a 40 -50% reduction in the relative risk of colorectal cancer and colorectal cancer-associated mortality in individuals taking nonsteroidal anti-inflammatory drugs (NSAIDs) 1 (1)(2)(3). Inhibition of cyclooxygenase-2 (COX-2) activity is thought to represent one of the mechanisms by which NSAIDs exert their anti-neoplastic effects (Refs. 4 and 5; reviewed in Ref. 6). In support of this hypothesis, lack of the COX-2 (prostaglandin endoperoxide synthase-2) gene results in a reduction of the number of tumors which develop in mice heterozygous for an APC ⌬716 mutation by more than 7-fold (7). Additionally, COX-2 expression in colorectal carcinoma cells provides a growth and survival advantage (5,8), and increases tumor cell invasiveness (9). Treatment with selective COX-2 inhibitors significantly reduces the adenoma burden in humans (10) and in animals (11). There are two isoforms of prostaglandin endoperoxide synthase, which are commonly referred to as COX-1 and COX-2. COX-1 is produced constitutively in many different cell types and tissues (12), but its expression can be regulated under some circumstances (13). COX-2 is induced by cytokines, growth factors, and tumor promoters (reviewed in Ref. 14). In studies of human colorectal cancer, COX-2 levels are increased in about 90% of cancers and ϳ50% of pre-malignant colorectal adenomas, but the enzyme is not usually detected in adult intestinal tissues (15,16). Cyclooxygenase catalyzes the conversion of arachidonic acid to prostaglandin (PG) G 2 and PGH 2 . PGH 2 is subsequently converted to a variety of prostaglandins, which include PGE 2 , PGD 2 , PGF 2␣ , PGI 2 , and thromboxane A 2 by each respective prostaglandin synthase. Prostaglandins are synthesized by a wide variety of human tissues and serve as autocrine or paracrine lipid mediators to signal changes within their immediate environment. PGs are involved in diverse biological processes, which include inflammation, blood clotting, ovulation, implantation, initiation of labor, bone metabolism, nerve growth, wound healing, kidney function, blood vessel tone, and immune responses (reviewed in Ref. 17).
The precise contribution of increased biosynthesis of prostaglandins by COX-2 to the progression of neoplasia is currently under evaluation. For example, PGE 2 generated in colorectal carcinomas may enhance cell survival and/or may affect other aspects of epithelial cell behavior such as cell-cell or cell-substrate adhesion (5). A link between the neoplastic effect of carcinogen treatment and prostaglandin signaling was recently made by the observation that genetic disruption of the E-prostanoid receptor subtype 1 (EP 1 ) results in a reduction in the number of aberrant crypt foci that develop in mice following carcinogen treatment (18). Based on these findings, we sought to determine the effects of PGE 2 on the biology of colorectal carcinoma cells. We found that PGE 2 stimulated an increase in the proliferation and motility of colorectal carcinoma cells.
Immunoblot Analysis-Immunoblot analysis was performed as described previously (19). Cells were lysed for 30 min in radioimmunoprecipitation assay buffer (1ϫ PBS, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, 10 mg/ml phenylmethylsulfonyl fluoride, 10 g/ml aprotinin, 1 mM sodium orthovanadate) and then clarified cell lysates were denatured and fractionated by SDS-polyacrylamide gel electrophoresis. After electrophoresis, the proteins were transferred to nitrocellulose membranes and the filters were incubated with the antibodies indicated and developed by the enhanced chemiluminescence system (ECL, Amersham Pharmacia Biotech). The anti-phosphorylated Akt antibody was purchased from New England Biolabs (Beverly, MA), and the anti-active ERK1/2 antibody was from Promega (Madison, WI).
Cell Growth in Matrigel ® -1 ϫ 10 4 cells were suspended in 0.5 ml of 1:2 diluted Matrigel ® (Collaborative Biomedical Products, Bedford, MA), and the mixture was plated into 24-well plates. PGE 2 in fresh medium was added to the cell culture every 2 days. After the plates were incubated for 10 -15 days, they were photographed using a camera attached to an inverted microscope. Relative colony size was determined by measuring 10 random colonies in each slide (50 measurements/well). The mean for each treatment set was calculated and compared with controls.
ERK Kinase Assay-p42/p44 MAP kinase activity was measured by determining the transfer of the phosphate group of adenosine 5Јtriphosphate to a peptide that is a highly specific substrate for p42/44 MAP kinase (Biotrak system, Amersham Pharmacia Biotech).
Akt Kinase Assay-For determination of Akt kinase activity we used the Akt kinase assay kit made by New England Biolabs (Beverly, MA) according to the manufacturer's instructions. Serum-starved cells were treated with PGE 2 and then lysed at the indicated times. Akt was immunoprecipitated using a monospecific Akt antibody. The immunoprecipitate was then incubated with a GSK-3 fusion protein in the presence of ATP. Phosphorylation of GSK-3 was measured by Western blotting using an anti-phospho-GSK-3␣/␤ (Ser21/9) antibody.
Immunofluorescence-LS-174 cells were grown in 35-mm tissue culture plates and fixed in methanol/acetone at Ϫ20°C for 10 min. Fixed cells were incubated with 10% normal donkey serum for 1 h and then with anti-FAK or anti-paxillin antibody (Transduction Laboratories, Lexington, KY) for 2 h at room temperature. After washing the cells three times with PBS, they were incubated with Cy3-conjugated donkey anti-mouse IgG (Jackson ImmunoResearch, West Grove, PA) for an additional 1 h. The cells were then washed with PBS, mounted, and observed under fluorescent microscopy with appropriate filters. For direct immunofluorescence, cells were fixed with formaldehyde-Triton solution and then incubated with 10 nM fluorescent phalloidin for 30 min.
Cell Migration and Invasion Assays-Cell migration and invasion assays were carried out using Transwell chambers (8 m, Corning Costar Co., Cambridge, MA). 5 ϫ 10 4 cells suspended in 400 l of serum-free McCoy's 5A medium were placed in the uncoated (migration assay) or 1:10 diluted Matrigel ® -coated (invasion assay) upper chamber. The lower chamber was filled with 1 ml of McCoy's 5A medium containing vehicle or 0.1 M PGE 2 . After an incubation period of 20 h at 37°C, the cells on the upper surface of the filter were removed with a cotton swab. The filters were fixed and stained with 0.5% crystal violet solution. Cells adhering to the undersurface of the filter were counted. Three independent experiments were carried out, and the data are expressed as the mean Ϯ S.E. of assays performed in triplicate.

Reverse Transcription-Polymerase Chain Reaction (RT-PCR)-RT-
PCR was carried out using the RNA PCR kit from PerkinElmer Life Sciences.
One g of total RNA was reverse-transcribed and amplified with 35 PCR cycles. The amplified products were visualized on 1.5% agarose gels.

Alterations in the Phenotype of LS-174 Cells following PGE 2
Treatment-Constitutive expression of COX-2 has been reported in 85-90% of colorectal carcinomas (15,16). COX-2 is expressed in both carcinoma and stromal cells (21). Therefore, it is possible that carcinoma cells that do not express COX-2 could receive paracrine signals from PGE 2 produced by neighboring stromal cells. In order to elucidate whether PGE 2 might exert any effect on the phenotype of colon cancer cells, LS-174 human colon cancer cells were treated with PGE 2 . LS-174 cells do not generate detectable prostaglandins, although COX-2 protein is detected in this cell line (22). LS-174 cells are able to form "crypt-like" aggregates when they are cultured in Matrigel ® . We found that exogenously added PGE 2 exerted a growthstimulatory effect on LS-174 cells (Fig. 1A). The size of LS-174 colonies in Matrigel ® increased following PGE 2 treatment in a dose-dependent manner (Fig. 1B). Treatment with 10 nM PGE 2 resulted in optimal stimulation of LS-174 cell growth, causing a 2-fold increase in colony diameter.
To our surprise, treatment with PGE 2 caused a dramatic change in the morphology of the LS-174 colonies. When grown in extracellular matrix components (Matrigel ® ), LS-174 cells formed well organized structures consisting of an outside layer of cells with an acellular center ( Fig. 2A, panel a). Positive Alcian Blue staining indicated that the LS-174 colonies were filled with colonic type mucin (data not shown). In contrast, the LS-174 cells exposed to PGE 2 formed irregular solid clumps of cells with a poorly organized structure ( Fig. 2A, panel b). When grown on plastic culture dishes, LS-174 cells formed in "nonspreading" round clumps ( Fig. 2A, panel c). Addition of 10 nM PGE 2 led to a rapid change in phenotype, which included increased spreading of cells within 2-4 h ( Fig. 2A, panel d).
Fluorescent staining with rhodamine-phalloidin demonstrated that PGE 2 treatment for 24 h resulted in protruding actin filaments from the cell periphery in the form of microspikes (Fig. 2B, panel b, white arrows) and an increase in the number of stress fibers (Fig. 2B, panel b, black arrows). PGE 2 treatment also increased focal adhesion complexes as determined by immunostaining for focal adhesion kinase (FAK) and paxillin. Normally, FAK and paxillin are localized to the cytoplasm in LS-174 cells (Fig. 2B, panels c and e), but following PGE 2 treatment the proteins accumulated into focal adhesions at the ends of actin stress fibers (Fig. 2B, panels d and f, arrows).
To further examine the spreading behavior induced by PGE 2 , we carried out experiments using a modified Boyden chamber.
Treatment of cells with PGE 2 resulted in a significant increase (2-3-fold) in cell motility (Fig. 3A). Addition of 0.1 M PGE 2 also promoted the movement of LS-174 cells through a Matrigel ®coated polycarbonate membrane by 2-3-fold (Fig. 3B). Therefore, PGE 2 altered the behavior of LS-174 cells by stimulating an increase in their motility, which could explain, in part, their change in cellular organization when grown as multicellular colonies.
Evaluation of EP Receptor Subtypes-We next determined if LS-174 cells express EP receptors, which are known to bind PGE 2 with a high affinity (reviewed in Ref. 23). The expression of EP receptors in LS-174 cells was determined by RT-PCR using specific oligonucleotide primers. EP 2 , EP 3 , and EP 4 were clearly expressed in LS-174 cells (Fig. 4A), but mRNA for the EP 1 receptor was barely detectable. To elucidate the functional role of EP receptor subtypes in LS-174 cells, we treated the cells with butaprost (1 M, a selective EP 2 receptor agonist), sulprostone (5 M, a selective EP 3 receptor agonist), and PGE 1 alcohol (10 nM, a selective EP 4 receptor agonist). Treatment with butaprost or sulprostone did not cause significant changes in cell morphology (data not shown). However, treatment with the PGE 1 alcohol (10 nM) resulted in more rapid and significant cell spreading when compared with the effect of PGE 2 alone. Alterations in LS-174 cell spreading were seen within 1 h following addition of the PGE 1 alcohol (Fig. 4B). Thus, LS-174 cell spreading and migration, stimulated by PGE 2 , may be predominantly mediated through the EP 4 signaling pathway.
Regulation of ERK and Akt Activity by PGE 2 -A number of signaling pathways is known to regulate cell growth and motility. The MAP kinase/ERK kinase/extracellular signal-regulated kinase (ERK) and phosphatidylinositol 3-kinase (PI3K)/ Akt signaling pathways were evaluated following PGE 2 treatment. Treatment of LS-174 cells with PGE 2 (100 nM) only had a modest effect on the activity of ERK1/2. PGE 2 treatment slightly increased the levels of phosphorylated ERK1/2 as determined by Western blotting analysis (Fig. 5A, upper panel). The results of an ERK kinase assay confirmed this finding (Fig.  5A, lower panel).
On the other hand, treatment with PGE 2 led to a marked activation of the PI3K/Akt pathway. The levels of phosphorylated (Ser-473) Akt/PKB were elevated following treatment  with PGE 2 in LS-174 cells (Fig. 5B, upper panel). Kinase assays, which measure the capacity to phosphorylate GSK-3 kinase, demonstrated that Akt kinase activity was greatly increased following PGE 2 treatment of LS-174 cells (Fig. 5B,  lower panel). It is known that Akt/PKB can be activated by G protein-coupled signaling in both a PI3K-dependent and -independent manner. Kinase assays failed to detect any PI3K activity in serum-starved LS-174 cells, and treatment with PGE 2 resulted in rapid induction of PI3K activity, as determined by the conversion of phosphatidylinositol 4-phosphate to phosphatidylinositol 3,4-bisphosphate (Fig. 5C). To confirm the involvement of PI3K in PGE 2 activation of Akt/PKB, we evaluated two inhibitors of this pathway (wortmannin (0.1 M) and LY 294002 (10 M)) and found that they both completely blocked PGE 2 -induced phosphorylation of Akt/ PKB (Fig. 5D).  (Fig. 6, C and D).

The Role of PI3K/Akt in PGE 2 -induced Pro-neoplastic Effects-To
Since PGE 2 treatment dramatically altered the growth and morphology of LS-174 colonies in Matrigel ® , it was of interest to determine the effects of PI3K/Akt activity on LS-174 cells grown in Matrigel ® . As demonstrated in Fig. 7A, LY 294002 impaired the ability of LS-174 cells to grow in Matrigel ® whereas PGE 2 significantly increased the size and altered the morphology of LS-174 colonies. Interestingly, addition of LY 294002 completely blocked the PGE 2 effects on cells grown in Matrigel ® by inhibiting colony growth and the invasive morphology. Wortmannin exerted similar effects but to a lesser degree on LS-174 cells grown in Matrigel ® (Fig. 7B and data not shown). DISCUSSION It is now clear that COX-2 plays a role in the promotion of colorectal cancer (6). However, the effects of prostaglandins generated by COX-2 have largely been unexplored. Here we provide evidence that prostaglandin-mediated signaling affects cell proliferation, motility, and morphogenesis and that activa- FIG. 4. Evaluation of the role of EP receptors. A, expression of PGE 2 receptors. One g of total RNA extracted from LS-174 was reverse-transcribed by using random hexamers. The fragment was amplified by specific primers for EP 1 , EP 2 , EP 3 , and EP 4 for 35 PCR cycles. The amplified products were visualized on 1.5% agarose gels. M, molecular weight marker. B, the effect of EP receptor agonists. LS-174 cells were serum-starved for 48 h prior to the treatment with vehicle (Control) or PGE 1 alcohol (0.1 M). The pictures were taken 24 h after initiation of treatment. The morphology of cell clumps is shown (original magnification, ϫ200). tion of the PI3K/Akt pathway is essential for the PGE 2 -induced changes in neoplastic potential.
To evaluate the effect of prostaglandins on the behavior of colorectal carcinoma cells, we employed several approaches. Treatment with PGE 2 stimulated DNA synthesis and cell spreading in LS-174 cells grown on plastic cultures. LS-174 cells form well differentiated multicellular colonies in Matrigel ® , mimicking tumor growth in animals. Treatment of LS-174 colonies with PGE 2 led to a significant disruption of their cellular organization with increased motility. The stimulation of cell migration by PGE 2 has been observed previously in mesangial, endothelial, and T cells (25)(26)(27). Forced expression of COX-2 in colon carcinoma cells results in increased invasiveness compared with the parental cells (9). These findings suggest that a prostaglandin product, such as PGE 2 , might stimulate cell motility and invasiveness under certain circumstances. In the present study, we show that addition of PGE 2 to serum-deprived LS-174 cells results in increased cell spreading accompanied by polymerization of actin and assembly of stress fibers, indicating that PGE 2 induced cytoskeletal reorganiza- tion. A role for the actin cytoskeleton has been implicated in many cellular functions, including motility, chemotaxis, cell division, endocytosis, and secretion (28 -30). Our data also demonstrate that PGE 2 treatment caused aggregation of FAK and paxillin, promoting the formation of focal adhesion complexes, which are known to be essential for cell migration (31,32). PGE 2 acts via specific transmembrane G protein-coupled receptors (EP receptors) (23). Four EP receptor subtypes have been identified and are designated EP 1 , EP 2 , EP 3 , and EP 4 . EP 1 signals via increased Ca 2ϩ , which leads to vasoconstriction. EP 3 can also serve to stimulate vasoconstriction and inhibits the generation of cAMP, whereas EP 2 and EP 4 are known to mediate vasorelaxation by stimulating an increase in cAMP levels. Our results show that both sulprostone (EP 3 agonist) and butaprost (EP 2 agonist) (33, 34) did not mimic the effect of PGE 2 to increase cell spreading. However, both the PGE 1 alcohol and misoprostol (relatively selective EP 4 agonist, data not shown) (33-35) induced significant cell spreading. These findings suggest that signaling via the EP 4 receptor is, at least in part, responsible for the PGE 2 -induced changes in LS-174 cell behavior.
Evidence suggests that the PI3K/Akt pathway promotes growth factor-mediated cell survival and inhibits apoptosis (36). PI3K/Akt also plays a key role in the regulation of cell adhesion and actin rearrangement (37,38). These observations suggest that the PI3K/Akt pathway is oncogenic and involved in the neoplastic transformation of mammalian cells. PI3K can be activated by growth factors, oncogenes, and is involved in the transmission of signals from certain G protein-coupled receptors (39 -41). Akt is stimulated by a variety of agonists acting on G protein-coupled receptors (42)(43)(44). Murga et al. (41) recently reported that PI3K␤ is necessary and sufficient to transmit signals from G proteins to Akt/PKB. Akt/PKB may also be activated by cyclic AMP-dependent protein kinase in a wortmannin-insensitive manner (42,45). Here, we found that treatment with PGE 2 rapidly increased the kinase activity of Akt/PKB and that wortmannin and LY 294002 blocked PGE 2induced phosphorylation of Akt/PKB, suggesting the involvement of PI3K. Thus far, the mechanism by which prostaglandin activates Akt/PKB is not clear, and, to our knowledge, this represents the first report of Akt/PKB modulation by PGE 2 . However, we have not established a direct link between the EP receptor and PI3K activation in the present study.
Our data further demonstrate the involvement of Akt/PKB activity in the PGE 2 -induced increase in cell proliferation and motility. LY 294002, at low concentrations (5-20 M), specifically targets PI3K activity (24). The observation that both LY-294002 and wortmannin (structurally unrelated PI3K inhibitors) exerted similar effects on LS-174 cells indicates that PI3K is the likely target of these compounds (reviewed in Refs. 46 and 47). We found that the growth of LS-174 cells (either on plasic or in Matrigel ® ) was significantly impaired by LY294002 (5 M), suggesting that basal levels of PI3K/Akt activity are required for continuous growth of LS-174 cells. Specific inhibitors of PI3K that blocked the activation of Akt/PKB did inhibit PGE 2 -induced changes in cell behavior, suggesting that both PGE 2 -induced growth stimulation and cytoskeletal reorganization involve the activation of the PI3K/Akt pathway. Several studies demonstrated that Akt/PKB pathway plays an extremely important role in cell cycle progression via modifying the expression of cell cycle proteins, such as cyclin D1 and p27 kip (48 -51). Activation of the PI3K/Akt pathway is thought to be essential for cytoskeletal reorganization under certain circumstances (37). These previous studies strongly support our findings that the PI3K/Akt activity is required for PGE 2 -induced increases in the growth and invasiveness of LS-174 cells.
Although COX enzyme activity is proposed to play a proneoplastic role in colorectal carcinogenesis, the downstream signaling that mediates these effects is poorly understood. Our results demonstrate that PGE 2 can induce significant phenotypic alterations in colorectal carcinoma cells. These changes include increased motility, changes in cell shape, and stimulation of cell growth. We found the PI3K/Akt signaling pathway to be involved in the regulation of morphogenic and proliferative changes. This work establishes a role for PGE 2 in the stimulation of tumor cell motility and reveals an additional cellular target, the EP 4 receptor, which appears to be involved in this process.