Prostaglandin H Synthase Expression Is Variable in Human Colorectal Adenocarcinoma Cell Lines

EXPERIMENTAL CELL RESEARCH ARTICLE NO.

236, 321–329 (1997)

EX973741

Prostaglandin H Synthase Expression Is Variable in Human Colorectal Adenocarcinoma Cell Lines Judy Parker,*,1 Mike K. Kaplon,* Consuelo J. Alvarez,* and Guha Krishnaswamy† *Division of Hematology/Oncology and †Division of Allergy and Immunology, Department of Internal Medicine, James H. Quillen College of Medicine, East Tennessee State University, Johnson City, Tennessee 37614-0622

The expression of prostaglandin H synthases can be induced by many stimuli and is likely to be important in control of the cell cycle. The analysis of prostaglandin H synthase-1 and -2 expression in colon adenocarcinoma cell lines is a useful model system for studying the function of the prostaglandin H synthases, especially with regard to proliferation and adhesion. Prostaglandin H synthase-1 protein is not found in any of eight human colon adenocarcinoma cell lines. Expression of prostaglandin H synthase-2 is variable for the eight cell lines: three constitutively expressed active protein, four did not express this gene at all, and one had mRNA but no active protein. Thus, five colorectal adenocarcinoma cell lines exhibit ‘‘null’’ expression of prostaglandin synthase-2. The three cell lines with constitutive expression of prostaglandin H synthase-2 produce PGE2 . Prostaglandin E2 production could be inhibited by aspirin and NS398 without inhibiting proliferation, while direct addition of prostaglandin E2 inhibits proliferation. Adhesion to collagen IV and fibronectin was stronger in those cell lines that expressed prostaglandin H synthase-2. The constitutive expression of prostaglandin H synthase-2 is associated with increased adhesion to extracellular matrix components and a potential inhibition of proliferation through the production of prostaglandin E2 . The absence of PGH synthase-2 expression in some cell lines may result from the original tumor’s need to inactivate these associated functions. Our evidence suggests that PGH synthase-2 is a possible candidate for a tumor suppressor gene at 1q23-qter. q 1997 Academic Press

INTRODUCTION

Prostaglandin synthase-2 is a recent addition to the group of cellular components that are involved in the control of cellular proliferation. The expression of the prostaglandin synthase-2 gene (also called TIS-10 and cyclooxygenase-2) is highly regulated and is transiently 1 To whom correspondence and reprint requests should be addressed. Fax: (423)-439-6387.

induced by growth factors and other stimuli in a variety of cell lines [1]. Prostaglandin H (PGH)2 synthase-2 is an immediate early response gene, i.e., this gene is activated without intervening protein synthesis in quiescent cells stimulated to begin proliferation anew [1]. Although the role PGH synthase-2 expression has in the control of proliferation is of great interest, this role is presently unclear. A homologous gene, PGH synthase-1, codes for a protein which is structurally related to PGH synthase2; however, the expression of the two genes is often independent [1]. The cellular levels of PGH synthase1 and -2 determine the amount of prostaglandin synthesized in conjunction with the additional regulation of the amount of substrate (primarily arachidonic acid) (see Ref. 2, for example). The prostaglandins produced upon activation of this pathway depend upon the cell type; thus, there may be different effects of PGH synthase expression in different cells. The study of PGH synthase-2 expression in colon epithelial cells may lead to an understanding of the function of an increase in PGH synthase-2 expression. Recent studies report that PGH synthase-2 expression is abnormally elevated in some colon adenocarcinomas [3–6]. If the role of altered PGH synthase-2 expression in such colon cancers could be defined we would learn about the normal role of PGH synthase-2 as well. However, half of colon adenocarcinoma biopsies are significantly infiltrated by lymphocytes, plasma cells, and mononuclear phagocytes [7, 8] and tumor biopsies cannot be used to study the function of the increased expression of PGH synthase-2. Colon cancer cell lines can be grown for extended periods in culture and thus provide a means used to study the function of PGH synthase-2 in vitro. (Normal colonic epithelial cells cannot be used because these cells rapidly undergo apoptosis in culture [9].) The expression of PGH synthase-2 has briefly been examined in colon cancer cell lines: five 2 Abbreviations used: PGH, prostaglandin H; DEDTC, diethyldithiocarbamic acid; DEPC, diethyl pyrocarbonate; DMSO, dimethyl sulfoxide; PCR, polymerase chain reaction; PVDF, polyvinyl difluoride; TPA, 12-tetradecanoylphorbol 13-acetate.

321

AID

ECR 3741

/

6i27$$$561

09-12-97 13:40:37

0014-4827/97 $25.00 Copyright q 1997 by Academic Press All rights of reproduction in any form reserved.

eca

322

PARKER ET AL.

human colorectal adenocarcinomas have been found to express mRNA for PGH synthase-2 [5]. However, further study is needed to demonstrate that an increase in mRNA leads to elevated protein and activity in the cell lines, since mRNA may sometimes be present for PGH synthase-2 without any protein product [10]. Elevated PGH synthase-2 would be expected to increase the PGE2 produced by colon epithelial cells, as it is known that PGE2 is the predominant prostaglandin produced by colonic epithelial cells [11]. PGE2 probably has a relationship to proliferation in colon epithelial cells; however, the available reports are contradictory. Thus, in colon epithelial cells in vitro, PGE2 either stimulates proliferation [12, 13] or inhibits proliferation [12, 14]. Studies in vivo report that prostaglandins of the E type reduce proliferation in colonic crypts [15, 16]. Overexpression of PGH synthase-2 in rat intestinal epithelial cells is associated with both a delay in the cell cycle in the G1 phase [17] and an increase in adhesion of the cells to components of the extracellular matrix [18]. However it is unclear if these effects also occur in human colon epithelial cells. Here we examine PGH synthase-2 expression in eight human colon adenocarcinoma cell lines requiring the presence of active protein product as the determinant of full expression. We find that some of the cell lines express PGH synthase-2 while others do not. We use this spontaneously occurring difference in PGH synthase-2 expression to address the relationship of expression of the PGH synthase-2 gene to cellular proliferation and adhesion to extracellular matrix components. MATERIALS AND METHODS Cells and cell culture. Human colon adenocarcinoma cell lines (Lovo, HT29, Caco2, SW48, SW948, SW1116, LS174T, and SW480) were purchased from the American Type Culture Collection, (Rockville, MD). All cell lines were routinely grown at 377C with 5% CO2 in RPMI 1640 supplemented with 10% fetal bovine serum. Cultures received one change in medium and one subdivision per week. Subdivision was achieved with exposure to 0.25% trypsin or 0.25% trypsin, 0.03% EDTA. Reagents. Indomethacin, acetylsalicylic acid (aspirin), nabumetone, 12-tetradecanoylphorbol 13-acetate, arachidonic acid, phenylmethylsulfonyl fluoride, dimethyl sulfoxide, sodium butyrate, diethyldithiocarbamic acid, a-mouse IgG–alkaline phosphatase conjugate, EcoRI, RNase A, and RPMI 1640 were purchased from Sigma Chemical Co. (St. Louis, MO). Fetal bovine serum, BamHI, biotin14–dCTP, and streptavidin–alkaline phosphatase conjugate were from GIBCO/BRL (Gaithersburg, MD). Dispase, Lumi-Phos 530, and CSPD were purchased from Boeringer-Mannheim (Indianapolis, IN). Prostaglandin E2 , monoclonal antibodies specific for PGH synthase1 and PGH synthase-2, NS-398 (N-[2-(cyclohexyloxy)-4-nitrophenyl]methanesulfonamide), and purified sheep PGH synthase-1 and -2 were purchased from Cayman Chemical (Ann Arbor, MI). Other reagents were from Fisher Scientific (Pittsburgh, PA). Biocoat dishes coated with laminin, collagen type IV, or fibronectin were purchased from Becton–Dickinson Labware (Bedford, MA). Determination of prostaglandin E2 production by adenocarcinoma cell lines. For the time course experiment, Caco2 cells were seeded

AID

ECR 3741

/

6i27$$$561

09-12-97 13:40:37

into 24-well plates and grown for 4 days. The medium was replaced with fresh growth medium containing TPA (100 nM) or DMSO (0.001%). Prostaglandin E2 in the culture medium at various times was then quantitated by enzyme immunoassay using a kit containing monoclonal a-prostaglandin E2 and a PGE2 –acetylcholinesterase tracer (Cayman Chemical). For experiments in which arachidonic acid was added, cells were grown for 4 days in 35-mm dishes and then fresh medium with or without 100 nM TPA was applied. After 6 h at 377C the medium was replaced with 1 ml of RPMI 1640 containing 20 mM arachidonic acid for 30 min at 377C followed by determination of prostaglandin E2 in the medium by enzyme immunoassay as above. Detection of PGH synthase-1 and -2 by Western blot analysis. Cell lines were seeded into 100-mm plastic dishes and grown for 4–5 days. Growth medium was replaced with RPMI 1640 supplemented with 0.5% fetal bovine serum for 24 h. TPA (80 nM) or 0.0008% DMSO was then directly added into the medium for 6 h. The medium was decanted and cell monolayers washed three times with phosphate-buffered saline. The cells were scraped off the dishes and collected by centrifugation at 500g for 5 min. The cells were resuspended in homogenization buffer (40 mM Tris–HCl, pH 8.0, 10 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 1 mM DEDTC) and broken using a Duall homogenizer. Intact nuclei and whole cells were removed by centrifugation at 500g for 5 min and the supernatant was then centrifuged at 100,000g for 1 h at 47C. The resultant membrane pellet was resuspended in 0.1 ml of dilute homogenization buffer and protein content determined by the bicinchoninic acid method using the BCA Protein Assay kit from Pierce (Rockford, IL). Aliquots of membrane protein were electrophoresed on an 8% SDS–polyacrylamide gel following the method of Laemmli [19]. Gels were transblotted to PVDF membrane (Micron Separations Inc.) for 1 h in 12.5 mM Tris–HCl, 96 mM glycine, pH 8.3. Blots were blocked in 5% bovine serum albumin, 25 mM Tris–HCl, pH 7.4, 0.15 M NaCl, stained with a 1/1000 dilution of monoclonal antibodies to either PGH synthase1 or PGH synthase-2, followed by staining with anti-mouse IgG conjugated to alkaline phosphatase. The position of stained bands was detected using Lumi-Phos 530 as the substrate and X-Omat AR film. RNA extraction and reverse transcriptase–PCR. Cells were seeded into 35-mm dishes and grown for 3–4 days. Medium was removed and fresh growth medium was applied containing 80 nM TPA or a solvent control (DMSO, 0.0008%) for 2 h. The cell monolayers were washed with PBS and the RNA extracted with RNazol B (Teltest Inc., Friendswood, TX). RNA was suspended in DEPCtreated water and absorbance was determined at 260 and 280 nm. Only RNA with a 260/280 ratio of ú1.6 was used. RNA was subjected to RT–PCR by established techniques. Briefly, first-strand cDNA was synthesized in the presence of murine leukemia virus reverse transcriptase (2.5 U/ml), 1 mM each dATP, dGTP, dCTP, and dTTP, RNase inhibitor (1 U/ml), 2 ml 101 PCR buffer (500 mM KCl, 100 mM Tris–HCl, pH 8.3), and MgCl2 (5 mM) using oligo(dT)16 (2.5 mM) as a primer. For reverse transcription, the preparation was incubated at 427C for 20 min, 997C for 10 min, and 57C for 5 min in a DNA thermocycler (Perkin–Elmer Corp., Norwalk, CT). PCR amplification was done on aliquots of the cDNA in the presence of MgCl2 (1.8 mM), dNTPs (0.2 mM each), and AmpliTaq polymerase (1 U/50 ml) and 0.2 nM of paired primers in a volume of 50 ml. Primers for the housekeeping gene, glyceraldehyde-3 phosphate dehydrogenase (GAPDH), were 5*-CCACCCATGGCAAATTCCATGGCA-3* (sense) and 5*-TCTAGACGGCAGGTCAGGTCCACC-3* (anti-sense), which results in a 593-bp product. Primers for PGH synthase-1 [20] were 5*-TGCCCAGCTCCTGGCCCGCCGCTT-3* (sense) and 5*GTGCATCAACACAGGCGCCTCTTC-3*, which produces a 303-bp product. Primers for PGH synthase-2 [20] were 5*-TTCAAATGAGAT-TGTGGGAAAATTGCT-3* (sense) and 5*-AGATCATCTCTGCCTGAGTATCTTT-3* (antisense), which produces a 305-bp product. PCR was carried out for 45 cycles as follows: initial denaturation at 957C for 90 s, denaturation 947C for 45 s, annealing 557C for 45 s,

eca

PGH SYNTHASE-2 AND COLON CANCER CELL LINES and extension at 727C for 90 s. Final extension was for 10 min at 727C. A 14-ml aliquot of each of the amplified products was subjected to electrophoresis on a 3% agarose gel and stained with ethidium bromide. Specific amplification was defined by the expected size of the amplified band and confirmed by restriction enzyme digestion. Equal RNA loading of the gel was determined by optical density measurements of added RNA and subsequently confirmed by similar intensity of the GAPDH bands in each sample. A biotinylated DNA probe for PGH synthase-2 was prepared by reverse transcriptioncoupled PCR (above) using RNA from Caco2 cells, with the exception that biotin-14–dCTP was added during the PCR amplification step, as instructed by the manufacturer. Effects of aspirin and NS398 on PGE2 production and proliferation. The effect of aspirin and NS398 on PGE2 production and proliferation was tested in the Caco2 cells. Cells were seeded into 96-well plates (5 1 104/well) in growth medium. After 24 h, the medium was replaced with fresh growth medium containing different concentrations of aspirin or NS398 in triplicate. The stock solution of NS398 was prepared in DMSO and aspirin was directly added in medium. The final concentration of DMSO in the medium did not effect proliferation. After 5 h, the medium in some wells was replaced by RPMI 1640 containing 20 mM arachidonic acid for 30 min at 377C, followed by determination of PGE2 in the medium by EIA (above). To determine proliferation after 24 h, 1 mCi of [3H-methyl]thymidine was added to each well for 1 h. Thymidine uptake was quantitated by a scintillation counter after the trypsinized monolayers were harvested onto glass filters. Results were compared by determining the concentration yielding 50% of the control thymidine uptake (IC50). Isolation of genomic DNA and Southern blots. Cells were resuspended in 100 mM NaCl, 10 mM Tris–HCl, pH 8.0, 25 mM EDTA, 0.5% SDS, and 0.1 mg/ml proteinase K and shaken for 18 h at 507C. The samples were extracted with phenol:chloroform:isoamyl alcohol (25:24:1) and DNA precipitated from the aqueous phase. After resuspension in 10 mM Tris–HCl, 1 mM EDTA, 0.1% SDS, pH 7.6, RNase A 0.6 mg was added for 1 h at 377C, followed by a repeat extraction and precipitation of the DNA. Forty micrograms of DNA was digested by 30 U of BamHI or EcoRI for 18 h at 377C, separated in a 0.7% agarose gel, and soaked for 15 min in 0.2 N HCl, 30 min in 1 M NaCl, 0.5 M NaOH, and 30 min in 0.5 M Tris–HCl, pH 8.0, 0.5 M NaCl. The gel was transferred to nylon membrane as instructed by the manufacturer. The membrane was prehybridized in 51 Denhardt’s, 0.5% SDS, 61 SSC, and 100 mg/ml salmon sperm DNA for 2 h at 657C (11 SSC is 0.15 M NaCl, 0.015 M sodium citrate). Hybridization to biotinylated DNA probe for PGH synthase-2 (see above) was for 20 h at 657C in 61 SSC, 0.5% SDS, and 100 mg/ml salmon sperm DNA. The membrane was washed twice in 51 SSC, 0.5% SDS at 657C (5 min each), for 30 min in 0.11 SSC, 1% SDS at 657C, and twice in 21 SSC at 18–207C. The membrane was blocked with 10% bovine serum albumin in 0.15 M NaCl, 25 mM Tris–HCl, pH 7.4, and then streptavidin–alkaline phosphatase in 1% BSA, 0.05% Tween 20 was added for 10 min at 207C. The membrane was washed in 0.1% bovine serum albumin, 0.05% Tween 20, 0.15 M NaCl, Tris–HCl, pH 7.4, and the chemiluminscent substrate CSPD was added according to the manufacturer’s instructions. Assay of adhesion to extracellular matrix proteins. Cells (5 1 105) were applied onto 35-mm plastic dishes coated with collagen type IV, laminin, or fibronectin in 1.5 ml of RPMI 1640, 10% fetal bovine serum. After incubation for 1 h at 377C, 5% CO2 , the dishes were washed three times with phosphate-buffered saline. Cells bound to the dishes were then released by 1-h exposure to dispase at 377C and counted in a hemocytometer. Single dishes were tested in multiple experiments.

RESULTS

Production of Prostaglandins by Human Colorectal Adenocarcinoma Cell Lines We initially searched for the presence of catalytically active PGH synthase-2 in human colorectal adenocarci-

AID

ECR 3741

/

6i27$$$561

09-12-97 13:40:37

323

FIG. 1. Time course of the effect of TPA on prostaglandin E2 production in Caco2 cells. Caco2 cells were exposed to TPA (100 nM) or a solvent control (DMSO, 0.001%) for varying times followed by determination of the PGE2 present in the medium by enzyme immunoassay as described under Materials and Methods. The average of duplicate wells is shown. (l) TPA; (s) DMSO.

noma cell lines. Thus, we assessed prostaglandin E2 production in the eight cell lines in response to TPA, a known stimulant of prostaglandin production. TPA stimulates PGE2 production only in the Caco2 cells, with a maximum effect at 20–30 nM (not shown), and requires 3 h or more to be observed (Fig. 1). The other seven colorectal adenocarcinoma cell lines do not accumulate detectable PGE2 under these conditions (not shown). Since the availability of substrate is known to restrict prostaglandin production, we supplied exogenous arachidonic acid to the cells. Upon addition of arachidonic acid, prostaglandin E2 production occurs in the absence of TPA in the Lovo, HT29, and Caco2 cells (Fig. 2). The other five cell lines failed to produce PGE2 upon addition of arachidonic acid (Fig. 2). TPA stimulates greater PGE2 production in the Lovo, HT29, and Caco2 cells but has no effect on the other five cell lines (Fig. 2). The Caco2 cells produce 10–20 times more PGE2 per cell when treated with TPA than the Lovo and HT29 cells. Addition of a preferential inhibitor of PGH synthase-2, NS398 [21], completely blocks PGE2 production (Fig. 2). Detection of PGH Synthase Proteins and mRNA Transcripts for PGH Synthases in Human Colon Cancer Cell Lines Since PGE2 production occurs only in three of the eight cell lines, we next looked at the expression of the two isoforms of PGH synthase in the human colon cancer cell lines. In order to improve the possibility of detecting low levels of PGH synthase-1 or -2, we stimulated cells with TPA to induce PGH synthases and isolated a membrane fraction from the cells to con-

eca

324

PARKER ET AL.

FIG. 4. Reverse transcriptase-coupled PCR analysis of PGH synthase-1 and -2 gene expression. Cells were exposed to TPA or DMSO for 2 h in RPMI 1640, 10% FBS. RNA was isolated from the cells and used for RT–PCR as described under Materials and Methods. Marker bands from a DNA ladder are indicated at the left. C, DMSO control (0.0008%); T, 80 nM TPA. (A) PGH synthase-2 product; (B) PGH synthase-1 product; (C) GAPDH product. FIG. 2. Prostaglandin E2 production in colon adenocarcinoma cell lines. Cells in 35-mm dishes were exposed to standard medium, medium plus TPA (100 nM), or TPA plus 25 mM NS398 for 6 h at 377C. The medium was replaced with 1 ml of RPMI 1640 containing 20 mM arachidonic acid for 30 min at 377C followed by determination of the amount of PGE2 in the medium as described under Materials and Methods. The results are presented as ng of PGE2 produced per 106 cells. The actual number of cells/dish in millions follows each cell type in parentheses. (A) Lovo (2.3); (B) HT29 (3.1); (C) Caco2 (0.8); (D) SW48 (2.0); (E) SW1116 (1.0); (F) LS174T (1.0); (G) Sw480 (1.0); (H) SW948 (2.5). Open bars, control; solid bars, TPA; hatched bars, TPA / NS398.

centrate the PGH synthases. Western blot analysis indicates that PGH synthase-2 protein is present in the Lovo, Caco2, HT29, and LS174T cell lines (Fig. 3) West-

ern blot analysis under conditions sufficient to detect purified PGH synthase-1 failed to find PGH synthase1 protein in any of the cell lines (not shown). Purified PGH synthase-1 and -2 were similarly stained by their respective specific antibodies, which suggests that poor antibody sensitivity is not the reason for the absence of PGH synthase-1 protein. To better understand the variable expression of the PGH synthase isoforms, we then searched for mRNA. The presence of mRNA for PGH synthase-1 and -2 in the eight different human colon cancer cells was monitored by RT–PCR. Lovo, Caco2, HT29, and LS174T cells contain mRNA for PGH synthase-2 both with and without TPA treatment (Fig. 4A). mRNA for PGH synthase-2 is not found in the SW1116, SW480, SW948, or SW48 cells with or without TPA (Fig. 4A). mRNA for PGH synthase-1 was present in the SW948, SW48, HT29, and Lovo cell lines but not the other four cell lines (Fig. 4B). Southern Blot Analysis of PGH Synthase-2 Gene in Colon Cancer Cell Lines

FIG. 3. Detection of PGH synthase-2 in human colon adenocarcinoma cell lines. Cells grown in 100-mm dishes were cultured for 24 h in RPMI 1640, 0.5% fetal bovine serum. TPA (80 nM) or a solvent control (DMSO, 0.0008%) was added for 6 additional h. Aliquots of the membrane fraction (100 mg protein) of the cells were loaded to 8% Laemmli gels. The gel was transferred to PVDF membrane and the presence of PGH synthase-2 protein was detected using antibodies to PGH synthase-2 and chemiluminescence as described under Materials and Methods. The pattern of the light-producing (positive) bands on X-ray film is shown. C, DMSO control; T, 80 nM TPA. Lane 1, SW48 (120 mg); lane 2, SW480; lane 3, LS174T; lane 4, SW1116; lane 5, Sw948; lane 6, Caco2; lane 7, Lovo; lane 8, HT29. Lane A, authentic PGH synthase-2,50 ng (marked COX-2).

AID

ECR 3741

/

6i27$$$561

09-12-97 13:40:37

The failure of some cell lines to express PGH synthase-2 could result from many possible mechanisms. However, since tumor cells often have deletions in certain parts of the genome, we examined the PGH synthase-2 gene by Southern blot analysis. We prepared a biotin-labeled probe by reverse transcription-coupled PCR with the primers for PGH synthase-2 (see Materials and Methods). For all the cell lines, the probe hybridizes to a single fragment of about 15,000 bp after digestion of genomic DNA by BamHI (Fig. 5A). This same fragment is the only one detected in normal human peripheral blood mononuclear cells (Fig. 5A). Two fragments of 7000 and 7800 bp are detected after diges-

eca

PGH SYNTHASE-2 AND COLON CANCER CELL LINES

325

tion by EcoRI (Fig. 5B). There is variation in the intensity of the stained fragments between the different cell lines, i.e., the darkest staining is seen in the HT29 and Caco2 cells and the lightest in the LS174T cells. The variation in staining intensity is generally consistent for the cell lines whether digested with BamHI or EcoRI. Impact of PGH Synthase-2 Expression on Biological Parameters The cell lines we studied were found to spontaneously exhibit either constitutive or null expression of PGH synthase-2. These cell lines thus provide an opportunity to assess the functional aspects of these two variants of PGH synthase-2 expression. Constitutive expression of PGH synthase-2 would increase the production of PGE2 since this is the major prostaglandin produced in colonic epithelium [11]. To ask what would be the effect of constitutive expression of PGH synthase-2 we directly added PGE2 and monitored proliferation. We find that PGE2 fails to stimulate proliferation of HT29 cells over a wide range of concentrations and at concentrations ú20 mM PGE2 inhibits proliferation (Fig. 6). Two methods to assess cell number, direct cell counting and the MTT technique, gave the same result (Fig. 6). The SW480, SW948, Lovo, Caco2, and LS174T cells examined over the same concentration range also exhibit only inhibition of proliferation by PGE2 (not shown). A second approach was also used to ask again if the constitutive PGH synthase-2 expression in the Caco2 cells is able to enhance proliferation. Thus we asked if

FIG. 5. Southern blot analysis of the PGH synthase-2 gene. Genomic DNA was isolated from each cell line and from human peripheral blood mononuclear cells. 40 mg DNA was digested by BamHI or EcoRI and electrophoresed at 4 V in a 0.7% agarose gel. Hybridization to a biotinylated probe for the PGH synthase-2 gene was as described under Materials and Methods. The probe was detected by streptavidin–alkaline phosphatase and a chemiluminescent substrate (CSPD). (A) BamHI, (B) EcoRI. Lanes: 1, Caco2; 2, HT29; 3, Lovo; 4, LS174T; 5, SW1116; 6, SW480; 7, SW48; 8, SW948; 9, blood mononuclear cells.

AID

ECR 3741

/

6i27$$$561

09-12-97 13:40:37

FIG. 6. Effect of prostaglandin E2 addition on proliferation of HT29 cells. HT29 cells (2 1 104/well) were grown in RPMI 1640, 10% fetal bovine serum containing PGE2 at different concentrations in quadruplicate. After 72 h, cell number was assessed. (A) After 4 h of MTT exposure, A540 nm was read in a multiwell plate reader. The results were converted to percentage of control for each experiment. The average { SD for three experiments is shown. (B) Cells were released from the wells by trypsinization and counted in a hemocytometer. The average of quadruplicates from a single experiment is shown.

proliferation would decrease in the Caco2 cells after blocking PGE2 production with aspirin or NS398. Full inhibition of PGE2 production occurs at concentrations of aspirin and NS398 which do not inhibit proliferation (Fig. 7). Thus, the observed IC50 Å 0.07 mM for aspirin inhibition of prostaglandin E2 is less than the IC50 Å 5 mM for the inhibition of proliferation. Similarly, the IC50 Å 4 mM for NS398 inhibition of prostaglandin E2 was less than the IC50 Å 130 mM for the inhibition of proliferation by this NSAID. We also tested whether adherence to subendothelial matrix components would correlate to PGH synthase2 expression. All cell lines bind equally well to laminincoated dishes except for the SW48, which exhibits less binding than the others (Fig. 8A). The three cell lines with active PGH synthase-2 (HT29, Lovo, and Caco2) exhibited strong binding to fibronectin or collagen type IV dishes. The cell lines which lack active PGH synthase-2 bound less or not at all to these proteins (Figs. 8B and 8C). However, the SW480 exhibits high binding

eca

326

PARKER ET AL.

DISCUSSION

We find PGH synthase-2 protein is expressed in three cell lines (Lovo, Caco2, and HT29) even in the absence of a stimulus. Nonetheless, such constitutive expression of PGH synthase-2 results in minimal PGE2 synthesis, especially for the Lovo and HT29 cells, unless exogenous arachidonic acid is added. Surprisingly, we found that four colon adenocarcinoma cell lines do not contain mRNA, protein, or activity of PGH synthase-2 even after induction by TPA. In addition, the

FIG. 7. Effect of aspirin and NS398 on PGE2 production and proliferation. Caco2 cells were seeded at 5 1 104 per well in a 96-well plate. After overnight incubation, the cells were exposed to different concentrations of aspirin or NS398 in growth medium. For determination of prostaglandin E2 synthesis, medium was replaced with RPMI 1640 containing 20 mM arachidonic acid after 6 h and the amount of PGE2 in the medium after 30 min was determined by EIA as described under Materials and Methods. For determination of proliferation, a 1-h pulse of [3H-methyl]thymidine (1 mCi/well) was performed at 24 h. The thymidine uptake was then determined and is presented as mean { SD, n Å 3. Production of PGE2 was normalized to ng/106 cells to correspond to Fig. 2. PGE2 , (l); [3H]thymidine, (s). (A) Aspirin, (B) NS398.

to fibronectin and collagen IV even though it lacks PGH synthase-2. Thus, with the exception of the SW480, binding to fibronectin and collagen type IV is correlated with the presence of active PGHS-2 in this group of cell lines. We also looked for a correlation of the functional expression of PGH synthase-2 with characteristics of the tumors from which the cell lines were derived (Table 1). The presence of active PGH synthase-2 is not related to the age or gender of the patient in which the tumor arose. The doubling time for each cell line was determined under standard culture conditions in vitro and is unrelated to the presence of active PGH synthase-2 (not shown) and the stage of the original tumor did not appear to be related to PGH synthase-2. However, the presence of functional PGH synthase-2 is associated with the cell lines derived from the more differentiated tumors.

AID

ECR 3741

/

6i27$$$561

09-12-97 13:40:37

FIG. 8. Adhesion of colon cancer cell lines to various components of the extracellular matrix. Cells (5 1 105) were applied onto dishes coated with various proteins in growth medium and allowed to adhere for 1 h at 377C. The unbound cells were removed by washing and the bound cells determined after release by dispase as described under Materials and Methods. Results from multiple experiments are presented as the percentage of initial cells which bound [(cells bound/total initial cells) 1 100]. The average { SD is shown for each cell line (n Å 3–9). For each matrix component, significant differences relative to the HT29 binding were determined using Student’s t test and are indicated: *P £ 0.005, **P õ 0.01, #P õ 0.025, / P õ 0.05. (A) Binding to laminin-coated dishes. (B) Binding to collagen type IV-coated dishes. (C) Binding to fibronection-coated dishes.

eca

PGH SYNTHASE-2 AND COLON CANCER CELL LINES

TABLE 1 Characteristics of Colon Cancer Cell Lines Cell line

Age, sex

Caco2 HT29 Lovo LS174T Sw1116 Sw480 Sw48 Sw948

72, 44, 56, 58, 73, 50, 83, 81,

? / ? / ? ? / /

Stage

Grade

PGE2 production

Adhesion to ECM

Ba nd C B B B C C

IIb II IIc IIc II III/IV IV III

/ / / 0 0 0 0 0

/ / / 0 0 / 0 0

Note. nd, not determined. a Duke’s stage of disease: B, tumor into muscularis mucosa; C, metastasis to lymph nodes. b Degree of histological differentiation: the higher the number the less differentiated. c See [22, 23].

LS174T cell line contains small amounts of PGH synthase-2 mRNA and PGH synthase-2 protein but no detectable prostaglandin E2 production. Therefore, the LS174T cells are unable to express functional PGH synthase-2 and belong in the ‘‘null’’ group. Our results extend those of a previous report [5] by demonstrating that constitutive expression of mRNA for PGH synthase-2 in colon cancer cell lines can lead to constitutive, active, protein product. We failed to detect PGH synthase-1 protein in any of the cell lines. We also found that all the PGE2 produced by the HT29, Lovo, and Caco2 cells was inhibited by a NS398, a preferential inhibitor of PGH synthase2 [21]. These two observations indicate that the PGE2 produced by the HT29, Lovo, and Caco2 cells is generated through the activity of PGH synthase-2. Thus, the mRNA for PGH synthase-1 observed in four cell lines by the highly sensitive PCR technique does not result in a functional product. The absence of PGH synthase1 protein is in agreement with reports of low expression of PGH synthase-1 in human colorectal adenocarcinomas [3–5] and normal mucosa [3, 4]. The cell line we found to have the strongest expression of PGH synthase-2 (Caco2) retains many differentiated functions of intestinal epithelium [24, 25] and is unable to form tumors in nude mice [26, 27]. Therefore, expression of active PGH synthase-2 is most strongly associated with a cell line derived from a colorectal adenocarcinoma in which differentiation is largely retained, i.e., an early stage. The other cells expressing active PGH synthase-2 (HT29, Lovo) are also derived from moderately well-differentiated tumors (Table 1). However, it must be noted that the HT29 and Lovo cells produce significantly less PGE2 than the Caco2. Our results indicate that the presence of active PGH synthase-2 does not stimulate proliferation. Thus, in

AID

ECR 3741

/

6i27$$$561

09-12-97 13:40:37

327

the Caco2 cells, production of PGE2 could be blocked by aspirin or NS398 without inhibiting proliferation. (High concentrations of aspirin or NS398 do, however, inhibit proliferation through an independent mechanism.) Furthermore, we find that PGE2 fails to stimulate proliferation over a wide range of concentrations in the HT29 and several other cell lines. Although we also used HT29 cells, we are unable to confirm previous reports of a stimulation of proliferation by PGE2 [12, 13]. This discrepancy is probably because the earlier reports restricted their analysis to measurement of thymidine uptake [12] and did not estimate the error of cell counting [13]. Significantly, we find that direct addition of the major end product of the PGH synthase-2 pathway (PGE2) inhibits the proliferation of all six colon cancer cell lines which have been tested. This result is in agreement with other reports that prostaglandins of the E type inhibit the proliferation of both normal colonic mucosa and colorectal adenocarcinoma cells [9, 11–13]. Relatively high concentrations of exogenous PGE2 are required to inhibit proliferation. Such high concentrations may be necessary because PGE2 cannot easily cross the plasma membrane [28] and/or due to significant loss of the added PGE2 over time [29, 30]. The amount of PGE2 produced by cell lines containing potentially active PGH synthase-2 is very low under standard culture conditions (Fig. 1). This could explain why there is no detrimental effect of the simple presence of PGH synthase-2 on the doubling time under standard culture conditions. However, addition of TPA can stimulate the production of PGE2 in colon cancer cell lines containing PGH synthase-2. It is likely that in vivo there are physiological or pathological signals which enhance PGE2 production and that more substrate will be available. Indeed, some colon adenocarcinoma biopsies produce large amounts of PGE2 [31–33]. There is other evidence to support the conclusion that generation of PGE2 via PGH synthase-2 activity may lead to an inhibition of proliferation. For example, the overexpression of PGH synthase-2 causes a delay in the G1 phase of the cell cycle in rat intestinal epithelial cells [17] and PGE2 inhibits rasp21 activity [34]. Also, there are several previous reports that stimulation of prostaglandin production by TNFa or IL-1b inhibits cellular proliferation (for example, Refs. 30, 35, 36). Nonetheless, the mechanism by which PGE2 inhibits cellular proliferation is incompletely defined. There is an association between the presence of active PGH synthase-2 and adhesive properties of the colon cancer cell lines. We found that most of the colorectal adenocarcinoma cell lines bind to laminin. In contrast, cells containing PGH synthase-2 (HT29, Lovo, Caco2) were significantly more able to bind to collagen type IV and fibronectin than those without PGH synthase-2. The correlation between PGH synthase-2 ex-

eca

328

PARKER ET AL.

pression and adhesion is not perfect, however, as the SW480 cells are good binders but do not express PGH synthase-2. Our results are in agreement with those of Tsujii and DuBois [18], who found binding to Matrigel (contains collagen IV and laminin) was increased upon enhanced PGH synthase-2 expression in rat intestinal epithelial cells. Yet we find the relationship between PGH synthase-2 and adhesion to laminin and fibronectin to be the opposite of that reported by Tsujii and DuBois, perhaps because different species were studied and different prostaglandins were produced. Further study will be required to fully determine if and how PGH synthase-2 expression moderates the adhesion of colorectal adenocarcinoma cells to components of the subendothelial matrix. The mechanism(s) which results in the absence of PGH synthase-2 expression is incompletely defined. It is unlikely that the activation mechanism employed by TPA is defective in these ‘‘null’’ cells since TPA is able to stimulate production of IL-8 in at least two of these cells, SW480 and SW948 (not shown). Southern blot analysis yielded fragments of the genomic DNA of the expected length for PGH synthase-2 [37] in cells that express PGH synthase-2 as well as those unable to express PGH synthase-2. Yet the restriction fragments were more intensely stained for the HT29 and Caco2, which are also the cell lines with the greatest ability to produce PGE2 . Only HT29 and Caco2 exhibited staining of the BamHI fragment with an intensity similar to that for normal human peripheral blood mononuclear cells. These results suggest that the HT29 and Caco2 have close to the normal copy number of the PGH synthase-2 gene, while the other cell lines have fewer copies. However, the staining of the Southern blots includes an amplification step and the results are thus only qualitative. Indeed, the lack of expression of the PGH synthase-2 gene could result from several different mechanisms and further investigation will be required to determine the explanation for the lack of PGH synthase-2 expression in each case. The failure in five of eight cell lines to produce PGE2 through either PGH synthase-1 or -2 is very striking. Others have reported that another human colorectal adenocarcinoma cell line, HCT15, also lacks mRNA for PGH synthase-2 and is unable to produce PGE2 [38]. For a total of six human colon cancer cell lines all to have lost the ability to produce PGE2 through the apparent loss of PGH synthase-2 expression must reflect a change that is important to the biology of the tumors in vivo or to the long-term culture of cell lines. The latter is unlikely since three cells lines (Caco2, Lovo, and HT29) retain PGH synthase-2 expression. Therefore, we hypothesize that the cell lines which are ‘‘null’’ expressors of PGH synthase-2 were derived from tumors in which an alteration occurred that led to a block in the expression of PGH synthase-2. There is support

AID

ECR 3741

/

6i27$$$561

09-12-97 13:40:37

for this hypothesis since colon adenocarcinomas with little expression of PGH synthase-2 [4–6] and PGE2 synthesis [31, 32] are reported to occur. In addition, many other cell lines derived from various types of tumors are also unable to express the PGH synthase-2 gene [39–41]. We find that the expression of PGH synthase-2 is correlated with a more adhesive, more differentiated phenotype and may cause inhibition of proliferation through an end product of the pathway, PGE2 . Should one or more of these aspects of PGH synthase-2 expression be functional in vivo, then tumors would gain an advantage from the inactivation of PGH synthase-2. Indeed, the PGH synthase-2 gene is located on chromosome 1 (1q25.2–25.3, Ref. 42), where frequent deletions have been observed in colorectal adenocarcinoma [43, 44] and other cancers (for example, Refs. 45–47). Evidence that a tumor suppressor gene lies within the region 1q23-qter has also been reported [43, 48]. We have shown a loss of function of the PGH synthase-2 gene in five of eight colon cancer cell lines. It is possible that the high rate of inactivation of the PGH synthase2 gene product is a by-product of its close proximity to an unknown tumor suppressor gene. However, we and others have shown that an active PGH synthase-2 which is generating prostaglandins in response to stimulation could cause an inhibition of proliferation and may perhaps also increase adhesion to the extracellular matrix. Thus, the PGH synthase-2 gene may be the tumor suppressor gene at 1q23-qter. This work was supported by a grant from the Medical Education Assistance Corporation (Johnson City, TN) (J.P., M.K.K.) and in part by the Paul Dishner Chair of Excellence, State of Tennessee Grant 20233 (G.K.). We thank Kenton Hall for excellent technical assistance in the reverse transcriptase–polymerase chain reaction experiments and Dr. David Chi for the gift of the U937 cells.

REFERENCES 1. Herschman, H. R. (1996) Biochim. Biophys. Acta 1299, 125– 140. 2. Chepenik, K. P., Diaz, A., and Jimenez, S. A. (1994) J. Biol. Chem. 269, 21786–21792. 3. Sano, H., Kawahito, K., Wilder, R. L., Hashiramoto, A., Mukai, S., Asai, K., Kimura, S., Kato, H., Kondo, M., and Hla, T. (1995) Cancer Res. 55, 3785–3789. 4. Kargmann, S. L., O’Neill, G. P., Vickers, P. J., Evans, J. F., Manicini, J. A., and Jothy, S. (1995) Cancer Res. 55, 2556–2559. 5. Kutchera, W., Jones, D. A., Matsunmi, N., Groden, J., McIntyre, T. M., Zimmerman, G. A., White, R. L., and Prescott, S. M. (1996) Proc. Natl. Acad. Sci. USA 93, 4816–4820. 6. Eberhart, C. E., Coffey, R. J., Radhika, A., Giardiello, F. M., Ferrenbach, S., and DuBois, R. N. (1994) Gastroenterology 107, 1183–1188. 7. Murray, D., Hreno, A., Dutton, J., and Hampson, L. G. (1975) Arch. Surg. 110, 908–913. 8. Carlon, C. A., Fabris, G., Arslan-Pagnini, C., Pluchinotta, A. M.,

eca

PGH SYNTHASE-2 AND COLON CANCER CELL LINES Chinelli, E., and Carniato, S. (1985) Dis. Colon Rect. 28, 47– 50. 9. Bedi, A., Pasricha, P., Akhtar, A. J., Barber, J. P., Bedi, G. C., Biardiello, F. M., Zehnbauer, B. A., Hamilton, S. R., and Jones, R. J. (1995) Cancer Res. 55, 1811–1816. 10. Lee, S. H., Soyoola, E., Chanmugam, P., Hart, S., Sun, W., Zhong, H., Liou, S., Simmons, D., and Hwang, D. (1992) J. Biol. Chem. 267, 25934–25938. 11. Boughton-Smith, N. K., Hawkey, C. J., and Whittle, B. J. R. (1985) Gut 24, 1176–1182. 12. Mengeaud, V., Nano, J. L., Fournel, S., and Rampal, P. (1992) Prostaglandins, Leukotriennes, Essent. Fatty Acids 47, 313– 319. 13. Qiao, L., Kozoni, V., Tsioulias, G. J., Koutsos, M. I., Hanif, R., Shiff, S. J., and Rigas, B. (1995) Biochim. Biophys. Acta 1258, 215–223. 14. Patrice, T., Harb, J., Foultier, M. T., Berrada, A., Robillard, N., Galmiche, J. P., Meflah, K., and Levodic, L. (1989) Dig. Dis. Sci. 34, 1681–1685. 15. Craven, P. A., Saito, R., and DeRubertis, F. R. (1983) J. Clin. Invest. 72, 1365–1375.

28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38.

16. DeRubertis, F. R., Craven, P. A., and Saito, R. (1985) Gastroenterology 89, 1054–1063. 17. DuBois, R. N., Shao, J., Tsujii, M., Sheng, H., and Beauchamp, R. D. (1996) Cancer Res. 56, 733–737.

39.

18. Tsujii, M., and DuBois, R. N. (1995) Cell 83, 493–501.

40.

19. Laemmli, U. K. (1970) Nature 227, 680–685.

41.

20. O’Neill, G. P., and Ford-Hutchinson, A. W. (1993) FEBS Lett. 330, 156–160.

42.

21. Gierse, J. K., Hauser, S. D., Creely, D. P., Koboldt, C., Rangwala, S. H., Isakson, P. C., and Seibert, K. (1995) Biochem. J. 305, 479–484.

43.

22. Drewinko, B., Romsdahl, M. M., Yang, L. Y., Ahearn, M. J., and Trujillo, J. M. (1976) Cancer Res. 36, 67–75.

44.

23. Tom, B. H., Rutzky, L. P., Jakstys, M. M., Oyasu, R., Kaye, C. I., and Kahan, B. D. (1976) In Vitro 12, 180–191.

45.

24. Grasset, E., Pinto, M., Dussaulx, E., Zwiebaum, A., Desjeux, J. (1984) Am. J. Physiol. 247, C260–267. 25. Jumarie, C., and Malo, C. (1991) J. Cell Physiol. 149, 24–33.

46.

26. de Bruı¨ne, A. P., de Vries, J. E., Dinjens, W. N. M., Moerkerk, P. T., v. d. Linden, E. P. M., PijM. M. J., ten Kate, J., and Bosman, F. T. (1993) Differentiation 53, 51–60.

47.

27. de Vries, J. E., Dinjens, W. N. M., De Bruyne, G. K., Verspaget, H. W., v. d. Linden, E. P. M., de Bruı¨ne, A. P., Mareel, M. M.,

48.

Bosman, F. T., and ten Kate, J. (1995) Br. J. Cancer 71, 271– 277. Bito, L. Z., and Baroody, R. A. (1975) Am. J. Physiol. 229, 1580– 1584. Bandyopadhyay, G. K., Imagawa, W., Wallace, D., and Nandi, S. (1987) J. Biol. Chem. 262, 2750–2756. Stahl, R. A. K., Thaiss, F., Haberstroh, U., Kahf, S., Shaw, A., and Schoeppe, W. (1990) Am. J. Physiol. 259, F419–F424. Rigas, B., Goldman, I. S., and Levine, L. (1993) J. Lab. Clin. Med. 122, 518–523. Bennett, A., Del Tacca, M., Stamford, I. F., and Zebro, T. (1977) Br. J. Cancer 35, 881–884. Bennett, A., Civier, A., Hensby, C. N., Melhuish, P. B., and Stamford, I. F. (1987) Gut 28, 315–318. Han, J-W, McCormick, F., and Macara, I. G. (1991) Science 252, 576–579. Shimakura, S., and Boland, C. R. (1992) Cancer Res. 52, 1744– 1749. Hla, T., and Maciag, T (1991) J. Biol. Chem. 266, 24059–24063. Tazawa, R., Xu, X-M., Wu, K. K., and Wang, L-H. (1994) Biochem. Biophys. Res. Commun. 203, 190–199. Hanif, R., Pittas, A., Feng, Y., Koutsos, M. I., Qiao, L., StaianoCoico, L., Shiff, S. I., and Rigas, B. (1996) Biochem. Pharmacol. 52, 237–245. Hla, T., and Neilson, K. (1992) Proc. Natl. Acad. Sci. USA 89, 7384–7388. Kubuju, D. A., Fletcher, B. S., Varnum, B. C., Lim, R. W., and Herschman, H. R. (1991) J. Biol. Chem. 266, 12866–12872. Arenander, A. T., Lim. R. W., Varnum, B. C., Cole, R., de Vellis, J., and Herschman, H. R. (1989) J. Neurosci. Res. 23, 247–256. Kosaka, T., Miyata, A., Ihara, H., Hara, S., Sugimoto, T., Takeda, O., Takahashi, E., and Tanabe, T. (1994) Eur. J. Biochem. 221, 889–897. Young, J., Searle, J., Stitz, R., Cowen, A., Ward, M., and Chenevix-Trench, G. (1992) Cancer Res. 52, 285–289. Reichmann, A. Martin, P., and Levin, B. (1984) Cancer Genet. Cytogenet. 12, 295–301. Miele, M. E., Robertson, G., Lee, J. H., Coleman, A., McGary, C. T., Fisher, P. B., Lugo, T. G., and Welch, D. R. (1996) Mol. Carcinog. 15, 284–299. Callahen, R., Cropp, C. S., Merlo, G. R., Liscia, D. S., Cappa, A. P., and Lidereau, R. (1992) Cancer 69, 1582–1588. Uchino, S., Tsuda, H., Noguchi, M., Yokota, J., Terada, M., Saito, T., Kobayashi, M., Sugimura, T., and Hirohashi, S. (1992) Cancer Res. 52, 3099–3102. Klein, K. G., and Bouck, N. P. (1994) Cancer Genet. Cytogenet. 73, 109–121.

Received May 12, 1997 Revised version received July 22, 1997

AID

ECR 3741

/

6i27$$$561

09-12-97 13:40:37

329

eca