Induction of Cyclooxygenase-2 by Activated Ha-rasOncogene in Rat-1 Fibroblasts and the Role of Mitogen-activated Protein Kinase Pathway*

Elevated cyclooxygenase-2 (COX-2) expression and activity have been observed in several different transformed cell types that express mutated ras genes. To investigate the mechanism of increased COX-2 expression following Ras-mediated transformation, Rat-1:iRas cell line was transfected with an Ha-Ras Val-12 cDNA expression vector that is under the transcriptional control of the lac operon and is inducible with isopropyl-1-thio-β-d-galactopyranoside (IPTG). IPTG treatment caused parallel increases in the levels of Ha-Ras and COX-2 proteins in Rat-1:iRas cells. The increased expression of COX-2 was accompanied by increased prostaglandin E2 production. Selective inhibition of COX-2 activity suppressed the production of prostaglandin E2 by >90% but did not alter the progress of the morphological transformation. The level of COX-2 mRNA was up-regulated by activated Ha-Ras. Induction of Ras increased the transcription of COX-2 by 44.3 ± 10.1% and increased the half-life of COX-2 mRNA by ∼3.5-fold. A specific mitogen-activated protein kinase/extracellular signal-regulated kinase (ERK) inhibitor (PD 98059) caused a delay in both the activation of ERK1/2 and the induction of COX-2 in IPTG-induced Rat-1:iRas cells. Inhibition of ERK activity by PD 98059 also suppressed the induction of COX-2 by epidermal growth factor in intestinal epithelial cells and significantly reduced the expression of COX-2 in Ha-Ras-transformed rat intestinal epithelial cells. ERK activity appears to be required for induction of COX-2 by Ras.

Ras proteins are GDP/GTP-regulated switch molecules that relay signals from receptor tyrosine kinases to the nucleus via activation of a number of signal transduction cascades. Ras activation stimulates both cell proliferation and differentiation (1)(2)(3). Receptor-mediated activation of Ras leads to the sequential activation of Raf-1 serine/threonine kinase, mitogen-activated protein (MAP) 1 kinase kinase (MEK1 and MEK2), and MAP kinases (4,5). Activated p42 and p44 MAP kinases (also referred to as extracellular signal-regulated kinases (ERKs)) translocate into the nucleus, modulate the phosphorylation of transcription factors, and ultimately lead to the expression of genes that are crucial for cell growth and differentiation (6,7). The stress-activated protein kinase (SAPKs)/c-Jun amino-terminal protein kinase (JNK) pathway is another kinase cascade distantly related to the MAPK pathway (8,9). This pathway can be activated in a Ras-dependent or Ras-independent manner by extracellular stimuli (10,11). MEKK1 (MEK kinase-1) phosphorylates SEK1 (SAPK/ERK kinase-1), a SAPK activator and in turn phosphorylates and activates SAPK (9).
Ras mutations are found in a wide variety of human malignancies, with the highest incidences observed in adenocarcinomas of the pancreas (90%), the colon (50%), and the lung (30%) (12). Oncogenic mutations in Ras result in constitutive activity of this small GTPase resulting in activation of downstream signaling proteins, including Raf and the downstream ERKs, as well as Raf-independent signaling proteins, including the Rho family proteins, RhoA, Rho B, and Rac1 among others, to modulate the expression of a specific subset of genes, and ultimately cause oncogenic transformation (13)(14)(15)(16)(17)(18). Forced overexpression of oncogenic Ras causes malignant transformation in multiple cell types including murine and rat fibroblasts (19,20), rat intestinal epithelial cells (21,22), and mammary epithelial cells (23).
Although much has been learned about Ras signal transduction pathways, the specific target genes and proteins that contribute to the transformed phenotype are largely unknown (24,25). Cyclooxygenase-2 (COX-2) is thought to be an important Ras target gene (23,26). Numerous studies have suggested that cyclooxygenase activity and/or prostaglandin synthesis may be pathogenic in numerous tumor types including colorectal carcinoma (27)(28)(29), breast cancer (23), and cancer of head and neck origin (30). Cyclooxygenase catalyzes the conversion of arachidonate to PGH 2 , the immediate precursor to prostaglandins, HETEs, and other eicosanoids. Two isoforms of cyclooxygenase, COX-1 and COX-2, have been identified (reviewed in Ref. 31). COX-1 is constitutively expressed in a variety of cells and tissues (32), whereas COX-2 is induced by cytokines, growth factors, and tumor promoters (31,33). Upregulation of COX-2 is a downstream effect of Ras-mediated transformation in intestinal epithelial cells (RIE-1) (26), mam-mary epithelial cells (C57/MG) (23), and non-small cell lung cancer cells (34). Selective inhibition of COX-2 activity leads to programmed cell death in non-transformed intestinal cells (35) and similar effects along with decreased tumorigenicity for transformed intestinal epithelial cells that express high levels of COX-2 (26,36,37) implying that COX-2 may play an important role in the survival and growth of transformed intestinal epithelial cells.
In this study, we have attempted to investigate the mechanism for the increased expression of COX-2 during conditional Ras-mediated transformation of Rat-1 fibroblasts. Rat-1 fibroblasts were stably transfected with an inducible mutant Ha-Ras Val-12 cDNA expression vector. In this system the oncogene is under the transcriptional control of the Lac operon and can be strongly induced by treatment of the cells with isopropyl-1thio-␤-D-galactopyranoside (IPTG). This conditionally transformed cell line (Rat-1:iRas) exhibits a non-transformed phenotype under normal culture conditions and undergoes rapid transformation after the induction of mutated Ha-Ras by treatment with IPTG (24,25). We found that COX-2 mRNA and protein expression and prostaglandin E 2 production were greatly increased in parallel with the induction of activated Ha-Ras. Increased expression of COX-2 was the result of a modest increase in the rate of COX-2 gene transcription combined with stabilization of COX-2 mRNA. Finally, biochemical analysis indicates that activation of the MAPK pathway (ERK1/2) appears to be necessary for Ras or EGF-receptor mediated induction of COX-2 expression.

EXPERIMENTAL PROCEDURES
Cell Culture-Rat-1:iRas cell line with an inducible activated Ha-Ras Val-12 cDNA was a generous gift from Dr. I. Hiroshi of Tokyo University of Technology (24,25). The Ha-Ras Val-12 cDNA is under the transcriptional control of the Lac operon in a eukaryotic expression system (Stratagene, La Jolla, CA). Rat-1:iRas cells were maintained in Dulbecco ' s modified Eagle medium (DMEM) containing 10% fetal bovine serum (FBS), 400 g/ml G418 (Life Technologies, Inc.), and 150 g/ml hygromycin B (Calbiochem) as described previously (25). IPTG (isopropyl-1-thio-␤-D-galactopyranoside, Life Technologies, Inc.) at a concentration of 5 mM was used to induce the expression of mutated Ha-Ras. A nontransformed rat intestinal epithelial (RIE-1) cell line was maintained in DMEM with 10% FBS. The mutated Ha-Ras-transformed RIE-1 cells were stably transfected with Ha-Ras Val-12 cDNA under the transcriptional control of RSV promoter and maintained in DMEM with 10% FBS and 500 g/ml of G418.
RNA Extraction and Northern Blot Analysis-Total cellular RNA was extracted according to Chirgwin et al. (38). RNA samples (20 g per lane) were separated on formaldehyde-agarose gels and blotted onto nitrocellulose membranes. The blots were hybridized with cDNA probes labeled with [␣-32 P]dCTP by random primer extension (Stratagene, La Jolla, CA). After hybridization and washes, the blots were subjected to autoradiography. 18 S rRNA signals were used as controls to determine integrity of RNA and equality of the loading. For determination of mRNA stability, Rat-1:iRas cells were treated with IPTG or vehicle for 8 h and then the transcription was stopped by addition of 40 M DRB (dichlorobenzimidazole riboside, Sigma). The RNA samples were isolated every 10 -15 min following the DRB treatment and analyzed for mRNA levels by Northern blotting.
Immunoblot Analysis-Immunoblot analysis was performed as described previously (39). Briefly, the cells were lysed for 30 min in radio-immunoprecipitation assay buffer (RIPA, 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 SDSpolyacrylamide gel electrophoresis. After electrophoresis, the proteins were transferred to nitrocellulose membrane. The filters were then probed with the indicated antibodies, developed by the enhanced chemiluminescence system (ECL, Amersham Pharmacia Biotech), and exposed to X-AR5 film (Kodak). Quantitation was by densitometry. The anti-COX-2 antibodies and anti-cdk4 antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-pan Ras antibody was purchased from Calbiochem. Anti-active MAPK antibody was purchased from Promega (Madison, WI).
Immunofluorescence-Rat-1:iRas 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-COX-2 antibody (Cayman Chemical, Ann Arbor, MI) 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 hour. The cells were washed with PBS, mounted, and observed under fluorescent microscopy with appropriate filters.
Quantitation of Eicosanoids-Subconfluent cell cultures were established; the cells were treated with 1 M SC-58125, a selective COX-2 inhibitor, {1-[(4-methylsulfonyl)phenyl]-3-trifluoromethyl-5-[(4-fluoro)phenyl]pyrazole}(Searle), or vehicle for 24 h. Serum-free DMEM with 15 M arachidonic acid (Cayman Chemical, Ann Arbor, MI) was replaced 1 h prior to collecting the conditioned medium. The PGE 2 formation in medium was quantified by utilizing stable isotope dilution techniques employing gas chromatography negative ion chemical ionization mass spectrometry. The results are expressed as nanograms of prostaglandin E 2 per ml of medium.
Run-on Transcription Assay-Nuclear run-on assays were performed as described previously (40). Briefly, Rat-1:iRas cells were stimulated with 5 mM IPTG or vehicle for 6 and 24 h. Nuclei were isolated. In vitro run-on transcription was carried out by using 1 ϫ 10 7 nuclei and 200 Ci of [[␣-32 P]UTP/assay at 30°C for 45 min. Labeled transcripts were purified by trichloroacetic acid precipitation. A total 1 ϫ 10 7 cpm elongated nascent RNA per assay were hybridized for 48 h at 65°C to filter-immobilized plasmid DNAs. The filters were washed with 2ϫ SSC at 65°C for 1 h, incubated with RNase A (10 g/ml) for 30 min at 37°C, and then for 1 h in 2ϫ SSC. The autoradiographs were subjected to densitometric analysis, and all data were normalized to the internal control. The results from four independent experiments were analyzed by Student's unpaired t test for statistical significance and expressed as mean Ϯ S.E.
Transfection of Reporter Constructs-Reporter constructs were generated from a plasmid containing 2.7 kb of rat COX-2 5Ј-flanking sequence (obtained as a generous gift from JoAnne Richards). Deletion constructs of the COX-2 promoter were generated by polymerase chain reaction using a blend of Pfu/Taq polymerase (Takara, Shiga, Japan). KpnI and XhoI restriction sites were engineered onto the tails of the primers. After digestion the polymerase chain reaction products were ligated into pGL3 Basic luciferase plasmid (Promega, Madison WI). Sequence validity was ascertained via direct sequencing. A total of five constructs were generated as follows: Ϫ2700 Cox-2/luc, Ϫ619 Cox-2/luc, Ϫ446 COX-2/luc, Ϫ289 COX-2/luc, Ϫ147 Cox-2/luc each containing 2700, 619, 446, 289, and 147 base pairs upstream of the transcriptional activation site of the rat cox-2 gene. For transient transfections, cells were plated in 24-well plates 24 h prior to transfection and then cotransfected with 0.5 g of one of the COX-2 promoter firefly luciferase plasmid constructs, pGL3 basic vector, or pGL3 control plasmid, and 0.25 g of pRL-TK plasmid, containing a herpes simplex virus thymidine kinase promoter upstream of the renilla luciferase gene (Promega, Madison WI), using the Lipofectin procedure (Life Technologies, Inc.) as described in the manufacturer's protocol. Transfected cells were cultured for 12 h in medium containing 10% FBS to allow recovery. The cells were switched to serum-free medium for 48 h, and then one plate was stimulated with 5 mM IPTG. 24 h later the cells were washed twice with 3 ml of Ca 2ϩ -and Mg 2ϩ -free PBS and lysed with passive lysis buffer (Promega, Madison WI). 20 l of lysate was used for both the firefly and renilla luciferase readings. Firefly and renilla luciferase activities were measured using a Dual-Luciferase Reporter assay system (Promega, Madison WI) and a model TD-20/20 Luminometer (Turner Design). Firefly luciferase values were standardized to renilla values.
Inhibition of MAPK Pathway-PD 98059 (Calbiochem), a specific inhibitor of mitogen-activated protein kinase kinase, was used for blocking the MAPK pathway activated by mutated Ha-Ras in Rat-1: iRas cells. PD 98059 at 75 M concentration in Me 2 SO was added to the culture media 1 h prior to induction of Ha-Ras by IPTG treatment. The same volume of Me 2 SO was added to the control cells. The media containing IPTG and PD 98059 were replaced daily. The morphology was observed on a daily basis. Protein lysates were collected for detection of COX-2. A similar protocol was followed for the epidermal growth factor (EGF, Sigma)-treated RIE-1 cells.
MAP Kinase Activity Assay-p42/p44 MAP kinase activity was measured by determining the transfer of the phosphate group of adenosine 5Ј-triphosphate to a peptide that is highly specific substrate for the p42/44 MAP kinase (Biotrak System, Amersham Pharmacia Biotech).

Rapid Induction of COX-2 Protein by Activated Ha-Ras-We
previously reported that COX-2 expression is increased more than 12-fold in an Ha-Ras-transformed rat intestinal epithelial cell line (26). To determine whether induction of COX-2 is an important early event in activated Ras-induced transformation, we used Rat-1 cells stably transfected with an inducible Ha-Ras expression construct (Rat-1:iRas). Previous studies have shown that addition of IPTG into the culture medium rapidly induces expression of mutated Ha-Ras and subsequent morphological transformation in these cells (24,25). We used these cells to determine the temporal expression of COX-2 in the context of activated Ha-Ras. As shown in Fig. 1A, COX-2 was expressed at very low levels in the parental Rat-1 fibroblasts. IPTG treatment for 24 h slightly reduced the expression of COX-2 in parental Rat-1 cells. In Rat-1:iRas cells, addition of IPTG into the culture medium induced activated Ha-Ras protein by 2 h. Thereafter, the level of Ras protein was continuously elevated for the duration of IPTG treatment (Fig. 1B). The effect of Ras induction on the expression of COX-2 is demonstrated in Fig. 1, B and C. The elevation of COX-2 protein was detected 4 h after addition of IPTG. The induction of COX-2 temporally coincided with the induction of mutated Ha-Ras protein. The expression of cyclin-dependent kinase 4 (cdk4) was unchanged during induction of Ha-Ras and is shown as a loading control (Fig. 1, B and C). Immunofluorescence staining demonstrates that the COX-2 protein predominantly was located in perinuclear cytoplasm, and COX-2 immunoreactivity was increased by 8 h after addition of IPTG. Within 24 h after removal of IPTG, COX-2 immunoreactivity returned to the uninduced basal level (Fig. 1D).
Inhibition of COX-2 Activity-Analysis of eicosanoid production revealed that PGE 2 was the predominant prostanoid pro-duced in both transformed and non-transformed Rat-1:iRas cells. Upon induction of mutated Ras with IPTG, PGE 2 levels increased 1.4-fold by 4 h, 3.0-fold at 8 h, and 3.1-fold at 24 h ( Fig. 2A). A selective COX-2 inhibitor, SC-58125 (42,43), at 1 M inhibited the conversion of prostaglandin E 2 from arachidonic acid by Ͼ 90%, confirming that the elevated prostanoid production was caused by the increased COX-2 activity.
To determine the role of COX-2 activity in Ras-mediated transformation, Rat-1:iRas cells were treated with IPTG in the presence or absence of SC-58125 (1, 10, and 25 M). The morphological alteration was recorded at 24 and 48 h. As demonstrated in Fig. 2B, the morphological transformation of the Rat-1:iRas cells was observed within 24 h of IPTG treatment. Cell-cell contact inhibition was lost, and the cells acquired a spindly appearance and grew in overlapping clusters. After 48 h of IPTG treatment the cells formed large foci. Addition of SC-58125 at 25 M neither delayed the progress nor reduced the degree of morphological transformation. Increased prostanoid production by COX-2 did not appear to be a necessary factor for Ha-Ras-mediated morphological transformation in Rat-1 fibroblasts.
Mechanism of Induction of COX-2 by Activated Ha-Ras-We next investigated the mechanism by which Ras activation increased levels of COX-2 protein in Rat-1:Ras cells. Northern analyses for COX-1 and COX-2 messages are demonstrated in Rat-1:iRas cells were treated with IPTG and lysed in RIPA 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) at the indicated time points. 50 g of each cell lysate was fractionated by SDSpolyacrylamide gel electrophoresis. After electrophoresis, the proteins were transferred to nitrocellulose membranes. The filters were blotted with the indicated antibodies, developed by ECL chemiluminescence system. C, densitometric analysis of the autoradiogram. D, immunofluorescence staining for COX-2. Rat-1:iRas cells (1, untreated; 2, IPTG-treated for 8 h; 3, IPTG-treated for 24 h; 4, IPTG for 24 h and then removal of IPTG for 24 h) were grown on 35-mm tissue culture plates and fixed by methanol/acetone, incubated with 10% normal donkey serum for 1 h, and then with anti-COX-2 antibody for 2 h at room temperature. After the cells were washed with PBS they were incubated with CY3-conjugated donkey anti-mouse IgG for an additional hour. The cells were washed with PBS, mounted, and observed under a fluorescent microscope. hibit transcription. Rat-1:iRas cells were first treated with vehicle or IPTG for 8 h and then subjected to DRB treatment. Serial RNA samples were isolated immediately following the addition of DRB, and Northern analysis for COX-2 was performed. Because IPTG-treated (Ha-Ras-induced) Rat-1:iRas cells expressed higher levels of COX-2 mRNA as compared with untreated Rat-1:iRas cells prior to the DRB treatment, the autoradiograms of control RNA blots were exposed 6 times longer than the autoradiograms of IPTG-treated RNA blots. As demonstrated in Fig. 3B, the Ha-Ras induction caused a decrease in the rate of COX-2 mRNA degradation as compared with the degradation rate in untreated Rat-1:iRas cells. The level of COX-2 mRNA was decreased Ͼ50% by 10 min after DRB treatment of Rat-1:iRas cells in the absence of Ras induction. The t1 ⁄2 of COX-2 mRNA was estimated to be 6 min in the absence of Ras induction. Induction of activated Ha-Ras oncogene prolonged the half-life of COX-2 mRNA to Ͼ 20 min (ϳ3.5-fold). An immediate early gene jun B was induced by Ha-Ras at mRNA levels (Fig. 3A). 50% of Jun B mRNA was degraded by 20 min, and Ha-Ras induction increased the t1 ⁄2 of Jun B mRNA to slightly over 20 min (Fig. 3B).
Increased COX-2 mRNA stabilization may not be the sole mechanism whereby COX-2 message is increased. In order to determine whether activated Ha-Ras increased the transcrip-tion of COX-2, nuclear run-on transcription assays were conducted. Rat-1:iRas cells were treated with IPTG or vehicle for 6 or 24 h, and radioactive labeled nascent transcripts were analyzed by hybridization to immobilized DNAs. cox-2 gene transcription at 6 and 24 h after induction of Ha-Ras was increased by 44.3 Ϯ 10.1% (mean Ϯ S.E., n ϭ 4, p Ͻ 0.005) by activated Ha-Ras, whereas the transcription of cox-1 was not altered by activation of the Ras signaling pathway, as determined by four independent experiments. Transcription of the jun B gene was increased 2.9-fold at 6 and 24 h after induction of oncogenic Ha-Ras. Fig. 4A represents a representative autoradiogram of the experiment.
To evaluate further the role of transcriptional induction in the COX-2 mRNA increase, COX-2 promoter/luciferase constructs were transfected into Rat-1:iRas cells. Luciferase activity was determined after either no stimulation or 24 h of IPTGinduced Ha-Ras expression. A slight transcriptional effect was observed, with a 2.7-kb COX-2 luciferase construct showing a 1.3-fold induction and a Ϫ619 COX-2 construct showing a maximal induction of 2.5-fold (Fig. 4B). These results suggest that activation of Ha-Ras oncogene induces an increase in COX-2 in Rat-1:iRas cells and is caused by a combination of mRNA stabilization and a modest transcriptional induction.
The Role of MAPK Pathway-Transformation by activated The RNA samples were isolated at the indicated time points following DRB treatment for detection of COX-2 and Jun B mRNA levels. Since IPTG-treated Rat1: iRas cells expressed higher levels of COX-2 mRNA as compared with uninduced Rat1:iRas cells prior to the DRB treatment, the autoradiograms of control RNA blots for COX-2 were exposed 6 times longer than the autoradiograms of IPTG-treated RNA blots.
Ras involves a phosphorylation cascade that leads to activation of multiple signaling pathways including mitogen-activated protein kinase (MEK/ERK) pathway. PD 98059, a specific inhibitor of mitogen-activated protein kinase kinase (MEK), has been shown to inhibit the activation of MEK both in vitro and in vivo (44,45). It was of interest to study the effect of PD 98059 on the transformation of Rat-1:iRas cells and associated induction of cyclooxygenase-2. The morphological transformation of the Rat-1:iRas cells was observed within 24 h of IPTG treatment. Addition of PD 98059 at 75 M prevented the cells from undergoing morphological transformation by 24 h but by 48 -72 h the cells appeared transformed despite the continuous presence of PD 98059 (Fig. 5A).
Interestingly, PD 98059 treatment also delayed the induction of COX-2 at both mRNA and protein levels. As shown in  (Fig. 5C). PD 98059 treatment also prevented the Ras-induced stabilization of COX-2 mRNA (Fig. 5D). Treatment of the IPTG-induced Rat-1:iRas cells with PD 98059 resulted in a less stable COX-2 mRNA (t1 ⁄2 Ͻ10 min) similar to that in Rat-1:iRas cells in which Ha-Ras has not been induced. Therefore, pharmacologic inhibition of the MEK/ERK pathway delayed both transcriptional activation and stabilization of COX-2 mRNA in Rat-1:iRas cells.
Next, immunoblot analyses were performed to determine whether ERK activation correlated with the expression of COX-2. As demonstrated in Fig. 6A, the lysates from uninduced Rat-1:iRas cells contained very low levels of activated 42-and 44-kDa ERK1/2, and PD 98059 treatment, as expected, further reduced the levels of activated ERK1/2. IPTG treatment significantly increased the levels of activated ERK1/2 from 8 through 72 h. The addition of 75 M PD 98509 completely blocked the activation of ERK1/2 by MEK for at least 24 h. Despite the continued presence of PD 98509, the Ras-activated ERK levels had begun to recover from the inhibitory effect of PD 98509 by 48 h. The expression of Cox-2 appeared to increase in parallel with levels of active ERK1/2. PD 98059 inhibited the induction of COX-2 protein at 8 and 24 h following IPTG treatment; however, increased levels of COX-2 protein were observed by 48 and 72 h following Ras induction even in the presence of PD 98509 treatment. These increases in COX-2 levels correlated with the increases in activated ERK1/2 (Fig. 6A). In order to confirm these findings, MAPK activity was evaluated. Activation of Ha-Ras increased ERK1/2 kinase activity by 6 -10-fold from 8 to 72 h following addition of IPTG. Addition of PD 98059 completely inhibited the ERK activity at 8 and 24 h; however, partial recovery of ERK activity occurred by 48 and 72 h with respective activity levels of 44.7 and 55.6% of the levels achieved by IPTG exposure in the absence of PD 98509 (Fig.  6B). The recovery of ERK1/2 activities from PD 98059 inhibitory effect coincided with the induction of COX-2 and morphological transformation in Rat-1:iRas cells. These results suggest that COX-2 induction is temporally related to the process of transformation but do not establish a causal relationship between COX-2 and transformed phenotype. We further examined this question by determining whether the COX-2 inhibition was able to prevent the morphological transformation of the Rat-1:iRas cells after IPTG treatment. We found that addition of COX-2 inhibitor SC-58125 at 10 M did not delay or inhibit Ras-mediated morphological transformation in Rat-1: iRas cells.
COX2 and ERK in Intestinal Epithelial Cells-Colorectal cancers often bear Ras mutations and have elevated expression of cyclooxygenase-2. In order to determine whether the above observations in fibroblasts were applicable to intestinal epithelial cells, further experiments were performed to determine whether activation of ERK1/2 is necessary for induction of COX-2 in intestinal epithelial cells. Growth factors known as ERK stimulators were used to activate the MEK/ERK pathway in RIE cells. Both EGF and TGF-␣ have been shown to stimulate the proliferation of rat intestinal epithelial (RIE-1) cells by activating the MEK/ERK pathway and to a lesser degree the SAPK/JNK pathway (8). Both EGF and TGF-␣ can also induce COX-2 expression in intestinal epithelial cells (30,46). Active 42-and 44-kDa ERK1/2 bands were abundantly detected by immunoblot analysis at 1, 3, and 6 h following addition of EGF (100 ng/ml) to the culture media (Fig. 7A). Associated with the increase in ERK1/2 activity, elevated expression of COX-2 protein was observed at 1, 3, and 6 h following EGF treatment. Addition of PD 98509 abrogated both increases in active ERK1/2 and induction of COX-2 after EGF stimulation. Similar results were obtained in the RIE-1 cells treated with TGF-␣ (data not shown).
It has been previously shown that COX-2 is expressed at high levels in RIE cells that are permanently transfected by constitutively expressed Ha-Ras (26). To determine whether the overexpression of COX-2 in Ras-transformed RIE cells de-pends upon MEK/ERK activation, Ha-Ras Val-12 -transfected RIE cells were examined. As shown in Fig. 7B, RIE-Ha-Ras-Val-12 cells abundantly express COX-2 protein. Addition of PD 98509 reduced the levels of COX-2 by 41.9% at 6 h and 67.8% at 24 h and decreased the amount of active ERK1/2 by Ͼ90%.

DISCUSSION
Identifying the target genes of Ras signaling and exploring their regulation by activated Ras is a key step in understanding the mechanism(s) whereby Ras gene mutations contribute to malignant transformation. In this study, using inducible activated Ha-Ras Val-12 -transfected Rat-1 cells, we found that activation of Ha-Ras results in rapid induction of COX-2 and prostaglandin E 2 production. In addition, studies in non-small cell lung cancer cells have demonstrated that activated Ras increased expression of cytosolic phospholipase A 2 , an enzyme that releases arachidonic acid from cellular phospholipid stores, thus making it available for the cyclooxygenase enzymes. These two observations suggest that increased prostanoid production may be an important phenotypic consequence of Ras activation (34).
Although the expression of cyclooxygenase-2 is often associated with neoplastic transformation of epithelial cells, the precise role of COX-2 in transformation is not clear. We have previously reported that increased COX-2 expression is not sufficient for neoplastic transformation of rat intestinal epithelial cells (35), and selective inhibition of COX-2 activity did not reverse the transformed phenotype of Ha-Ras-transformed RIE cells (RIE-Ras) (26) or COX-2 expressing human colon cancer cells (HCA-7) (36). In this study, we found that selective inhibition of COX-2 activity by treatment of SC-58125 reduced the production of PGE 2 by Ͼ90% but did not alter the progress of morphological transformation. These results suggest that induction of COX-2 is an early consequence of Ras activation in Rat-I fibroblasts but is not required for morphological alterations that are indicative of the transformed phenotype. What role might induction of COX-2 play in tumor progression? In previous studies, forced expression of COX-2 in non-transformed intestinal epithelial cells (RIE-1) resulted in resistance to apoptosis (35). Selective inhibition of COX-2 activity suppressed the growth of RIE-Ras (26) and HCA-7 cells (36) primarily via the induction of apoptosis. Addition of prostaglandin E 2 , the predominant eicosanoid product of COX-2 in HCA-7 cells, significantly increased the clonogenicity of HCA-7 cells and abrogated the induction of apoptosis caused by treatment with a COX-2 antagonist (29). These findings suggest that increases in COX-2 and prostaglandins that occur during cell transformation may enhance cell survival and provide a selective advantage for the transformed cells that overexpress COX-2. Interestingly, the induction of COX-2 by Ha-Ras was not sufficient to prevent the Rat-1:iRas cells undergoing apoptosis beginning 72 h after IPTG treatment. Further studies will be required to determine whether COX-2 expression and prostaglandins provide selective protection for epithelial cells as compared with fibroblasts.
Previously, it was reported that COX-2 expression was regulated at the transcriptional level by activated Ha-Ras in mammary cells (23,30) and by oncogene v-src (47) and growth factors (48) in NIH 3T3 cells. Both human and murine COX-2 promoters contain a consensus cyclic AMP response element (CRE). The CRE element in the murine COX-2 promoter is essential for optimal COX-2 gene expression in response to v-src and platelet-derived growth factor or serum (47,48). The rat COX-2 promoter does not contain this CRE sequence implying that different regulatory pathways may be utilized for induction of COX-2 between murine and rat cells. Analysis of nuclear run-on experiments revealed that the basal transcrip-tion rate of COX-2 in Rat-1 fibroblasts was relatively high and was increased only modestly after Ras induction. In addition to modest transcriptional activation of COX-2, we also observed post-transcriptional stabilization of COX-2 mRNA after Ras induction. In non-transformed Rat-1:iRas cells, the COX-2 mRNA was very unstable and was quickly degraded, thus steady-state levels of COX-2 mRNA and protein were barely detectable. There was a ϳ3.5-fold reduction in the COX-2 mRNA decay rate in Ras-transformed Rat-1 cells as compared with non-transformed Rat-1:Ras cells. These results suggest that activation of the Ras signal transduction pathway either induces factor(s) that stabilize or inhibits factors that destabilize COX-2 mRNA. Recent studies suggest that under selective conditions post-transcriptional regulation of mRNA stability may significantly contribute to the regulation of COX-2 mRNA levels. Interleukin-1␣ induced rapid but transient activation of COX-2 transcription and also prolonged the half-life of the COX-2 mRNA (49). Post-transcriptional regulation of cytokineinduced cyclooxygenase-2 transcript isoforms by dexamethasone has also been reported (50).
In the present study, we observed no change in the levels of COX-1 mRNA after Ha-Ras activation in Rat-1:iRas cells. Although COX-1 and COX-2 are very similar at the amino acid level, and even have comparable enzymatic activities, their expression pattern is markedly different. COX-1, often referred to as the "constitutive" cyclooxygenase, is present in most tissues. In contrast, COX-2 is normally absent but can be rapidly induced by a wide range of stimuli. The 3Ј-untranslated region of the COX-2 transcript differs from COX-1 and has two polyadenylation signals, which potentially explain different mRNA sizes observed by Northern analysis (2.8 -4.5 kb). This region is extremely AT-rich, and contains 9 copies of the Shaw-Kamens sequence (ATTTA) otherwise known as A ϩ U-rich elements (AREs) (51,52). This motif is present in many immediate early genes and is thought to be involved in regulating the rate of mRNA degradation. For example, if AREs are added to a normally stable transcript, such as ␤-globin, the transcript is rapidly degraded (53). It is of interest to note that these ARE sequences are conserved between murine, rat, and human COX-2 3Ј-untranslated regions, suggesting an important role for their presence.
The mechanism by which mutated Ras induces transformation is complex and incompletely understood. Different Ras effectors may be required for transformation of different cell types. Activation of the Raf-1/MAP kinase cascade alone is not sufficient for Ras transformation of rat intestinal epithelial cells but appears to be sufficient for the transformation of rodent fibroblasts (54). A selective MEK inhibitor PD 98059 reverses the phenotype of Ras-transformed BALB 3T3 mouse fibroblasts and rat kidney cells (44). Activated Ras promotes the translocation of Raf-1 to the plasma membrane (55, 56), additional factors complete the activation of Raf-1 kinase activity (57), and then Raf-1 in turn phosphorylates MEK1 and -2. Activated Ras also activates Rho family proteins (RhoA, Rho B, and Rac1), and the Rho proteins phosphorylate and activate PAKs and MEKK1-3. MEKK1-3 are able to activate both JNK/SAPK and ERK (13,15). We observed that morphological transformation of Rat-1:iRas cells was accompanied by an increase in ERK activity indicating the importance of the Ras-Raf1-MEK-ERK pathway in oncogenic Ha-Ras-induced transformation of Rat-1 cells. Treatment with the MEK inhibitor, PD 98059, delayed both the induction of active ERK and the morphological transformation of IPTG induced Rat-1:iRas cells, but after 24 h the ERK activity recovered and the cells acquired the transformed phenotype despite repeatedly adding PD 98059. The PD 98059 compound does not inhibit phosphorylated MEK but inhibits both activation and phosphorylation of MEK by Raf kinase and MEKK. Whereas PD 98059 potently inhibits the activation of MEK1, it has much less inhibitory activity for MEK2 (45). The incomplete transient inhibition of ERK activation by PD98059 may be caused either by incomplete suppression of Raf and MEK activities or by other Rafindependent Ras signaling pathways (58) that may not be suppressed by PD 98059 treatment.
Previously Xie and Herschman (47,48) reported that induction of COX-2 by the oncogene v-src (47) and platelet-derived growth factor or serum (48) in NIH 3T3 cells requires activation of both Ras/Rac1/MEKK1/JNK and Ras/Raf-1/ERK signal transduction pathways. Our studies suggest that both the transcriptional and post-transcriptional regulation of COX-2 expression by activated Ha-Ras is dependent on the MAPK. Complete inhibition of MEK/ERK activity by PD 98059 between 8 and 24 following treatment abolished both the increased transcriptional activity and the stabilization of COX-2 mRNA. When the activation of ERK escaped from PD 98059 inhibitory effect, the levels of COX-2 mRNA and protein increased. Our results do not exclude the possibility that JNK or other signaling pathways are also required for full expression of COX-2 in Ras-transformed Rat-1 cells (47,48). In addition, we also found that MAPK activity is necessary for the high level expression of COX-2 in intestinal epithelial cells that have been transfected with activated Ha-Ras or in EGF-stimulated non-transformed rat intestinal epithelial cells. Our observations suggest that induction of MAPK activity is essential for induction of COX-2 that occurs after activation of the Ras pathway, whether by activation of the EGF receptor tyrosine kinase or by activation of mutant Ras.