Increased expression of cyclooxygenase (COX) and overproduction of prostaglandins (PGs) have been implicated in the development and progression of colorectal cancer (CRC). Nonsteroidal anti-inflammatory agents (NSAIDS) inhibit growth of various CRC cell lines by both COX-dependent and COX-independent pathways. To specifically examine the effect of COX and PGs on proliferation in CRC cells, we introduced an antisense COX-2 cDNA construct under the control of a tetracycline (Tc)-inducible promoter into a CRC cell line, HCA-7, Colony 29 (HCA-7) that expresses COX and produces PGs. In the presence of Tc, PG production in COX-depleted cells was reduced 99.8% compared with either uninduced transfectants or parental HCA-7 cells. This decrease in PG production was associated with a concomitant 60% reduction in DNA replication. Subsequently, we examined the effects of various PGs to modulate cell growth in COX-depleted HCA-7 or COX-null HCT-15 cells by quantifying [3H]thymidine incorporation and/or growth in collagen gels. We report that J-series cyclopentenone PGs, particularly PGJ2 and 15-deoxy-Δ12,14-PGJ2, induce proliferation of these cells at nanomolar concentrations. Lipids extracted from parental HCA-7 cell conditioned medium stimulated mitogenesis in COX-depleted HCA-7 cells and COX-null HCT-15 cells. Using chromatographic and mass spectrometric approaches, we were able to detect PGJ2 in conditioned medium from parental HCA-7 cells. Taken together, these findings implicate a role for cyclopentenone PGs in CRC cell proliferation.

CRC3 is the second leading cause of cancer mortality in Western societies. Epidemiological studies have reported up to a 50% decrease in the relative risk of CRC in persons who regularly ingest aspirin or other NSAIDS, suggesting that these drugs exert their pharmacological effects by inhibiting COX, the enzyme responsible for PG synthesis (1). Two isoforms of COX, COX-1 and COX-2, have been identified (2). Although they both catalyze the formation of PGs, COX-1 and COX-2 are likely to have fundamentally different biological roles. COX-1 is constitutively expressed in many tissues and is thought to be involved in maintaining cellular homeostasis (3). In contrast, COX-2 is frequently undetectable at baseline in normal tissues but is readily expressed in gastrointestinal epithelial cells in response to inflammatory cytokines, lipopolysaccharide, mitogens, and reactive oxygen intermediates (4). Overexpression of COX-2 has been associated with CRC. In human studies, COX-2 is increased in 80–90% of CRC tumors and 40% of premalignant colorectal adenomas (5). Furthermore, COX-2 expression is markedly elevated in most colonic tumors in azoxymethane-treated rats and in intestinal adenomas from multiple intestinal neoplasia (Min) mice (6, 7).

COX enzymes catalyze the formation of PGH2, the unstable bicycloendoperoxide intermediate, which undergoes further metabolism to the parent eicosanoids PGD2, PGE2, PGF, PGI2, and TXA2(8). PGs exert a wide variety of biological activities including effects on cellular growth. Depending on the PG studied and the cell line used, parent eicosanoids have been shown to have both proliferative and antiproliferative effects (8, 9, 10). These varying results have been attributed to the interactions of different eicosanoids with various cell surface prostanoid receptors (11).

We have previously studied the relationship between COX-2 induction, PG formation, and proliferation in the CRC cell line, HCA-7, Colony 29 (HCA-7; Ref. 12). Treatment of this cell line with the growth factor transforming growth factor-α leads to a marked induction of COX-2 expression, increased formation of the eicosanoids PGD2, PGE2, PGF, and TxA2, and a significant increase in proliferation, suggesting a link between PG production and mitogenesis. On the other hand, HCA-7 cells, when treated with either selective COX-2 inhibitors (NS-398 or SC 58125) or a nonselective COX inhibitor (indomethacin), undergo a marked decrease in proliferation as quantified by [3H]thymidine incorporation, and this decrease cannot be overcome by the addition of exogenous parent eicosanoids. A possible explanation for this latter observation is that the effects of NSAIDS on cellular proliferation are independent of PG production (12, 13, 14, 15).

In addition to parent PGs, there has been significant interest in the role of another group of eicosanoids, termed cyclopentenone PGs, on cellular proliferation and differerentiation. Cyclopentenone PGs are dehydration products of either PGD2 or PGE2 (Fig. 1). Dehydration of PGD2 results in the formation of PGJ2, which can then rearrange to form Δ12-PGJ2(16). The latter compound can lose another molecule of water to form 15-deoxy-Δ12,14-PGJ2. PGE2 dehydrates to form PGA2. Interest in the cyclopentenone PGs has centered on their antiproliferative effects because they markedly inhibit cell growth and induce differentiation when applied to a variety of cell lines at concentrations in the micromolar range (17). In contrast, these compounds, at concentrations in the nanomolar range, have been reported to induce proliferation in the breast cancer cell line MCF-7 (18). This latter study is important because it is likely that endogenous PGs are produced in submicromolar concentrations, and these amounts are more biologically relevant (8, 11, 12).

Because of the interest in the effects of PGs on cell growth, we undertook studies to examine the ability of exogenous parent and cyclopentenone PGs to modulate proliferation in the CRC cell line HCA-7. For these studies, endogenous PG production was inhibited using a molecular approach in which a Tc-inducible antisense COX-2 construct was introduced into HCA-7 cells. We report that the cyclopentenone PGs, PGJ2 and 15-deoxy-Δ12,14-PGJ2, induce proliferation in antisense COX-2-transfected HCA-7 cells at concentrations in low nanomolar range. In addition, similar effects were observed in the CRC cell line HCT-15 that does not express COX or produce PGs. Furthermore, using chromatographic and mass spectrometric approaches, we detect formation of PGJ2 in conditioned medium derived from parental HCA-7 cells. Taken together, these studies suggest that cyclopentenone PGs play a role in CRC cellular proliferation.

Materials.

All cell culture reagents were purchased from Life Technologies, Inc. (Grand Island, NY). Chemicals were purchased from Sigma Chemical Co. (St. Louis, MO) unless otherwise stated. [3H]Thymidine and [α-32P]dCTP were purchased from Amersham Radiochemicals (Arlington Heights, IL). Polyclonal antibodies to COX-1 and COX-2 were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Parent and cyclopentenone PGs were obtained from Cayman Chemical Co. (Ann Arbor, MI).

Cell Lines.

HCA-7, Colony 29 cells were derived as a subpopulation from the parent human CRC adenocarcinoma cell line HCA-7 (19). HCT-15 cells were obtained from American Type Culture Collection (Bethesda, MD). Cells were maintained in DMEM supplemented with 10% FCS, glutamine, nonessential amino acids, 100 units/ml penicillin, and 100 μg/ml streptomycin in a 5% CO2 atmosphere with constant humidity at 37°C (12).

Development of HCA-7 Cell Lines Containing an Antisense COX-2 Construct.

The tetracycline-on gene expression plasmids, pTet/VP16 and pUHD.2neo, were provided by Dr. Dario Vignali (St. Jude’s Childrens Hospital, Memphis, TN; Ref. 20). A 1.93-kb fragment containing the human COX-2 cDNA was subcloned into the XbaI/EcoRV sites, and the orientation of the COX-2 sequence was verified by double-stranded DNA sequencing. HCA-7 cells were sequentially transfected with the above two plasmids using Cellfectin (Life Technologies, Inc.), according to the manufacturer’s instructions. After 48 h, cells were shifted to medium supplemented with 10 μg/ml puromycin and/or 1 mg/ml G418 (Life Technologies, Inc.) to select for transfected clones, and antibiotic-resistant cells were subcloned by limiting dilution. Loss of COX-2 expression (mRNA and protein) was determined after incubation with 2 μg/ml Tc for 24 h and Northern or Western blot analyses. Three independent positive clones were used for all subsequent assays.

Northern Blot Analysis.

Total cellular RNA from parental and transfected cells, grown in the presence or absence of Tc for 24 h, was isolated using the acid guanidinium thiocyanate-phenol-chloroform extraction method, followed by poly(A)+ mRNA selection using oligo(dT) (21). mRNA (3 μg) was separated by electrophoresis through 1% (w/v) agarose-formaldehyde gels and blotted onto 0.2-μm pore size nitrocellulose membranes. Hybridizations with human specific probes labeled by RNA polymerase-directed reverse transcription (1B15) or random primer extension (COX-1, COX-2) were performed in hybridization ovens as described previously (6, 12, 22). 1B15 is a constitutively expressed sequence used to assure equivalent loading and transfer of RNA (6).

Western Blot Analysis.

Parental or transfected HCA-7 cells were lysed in 50 mm Tris-HCl (pH 7.4), 200 mm NaCl, 2 mm EDTA, 0.5% NP40, 0.5 mm phenylmethylsulfonyl fluoride, 1 mg/ml aprotinin, 1 mg/ml pepstatin, and 2 mg/ml leupeptin. One hundred μg of extract (as determined by Bradford analysis) were applied to 10% SDS-PAGE gels and transferred to 0.2-μm pore size nitrocellulose membranes (Schleicher and Schuell). Blots were probed with antibodies raised against COX-1 or COX-2 at a final concentration of 0.1 μg/ml. After washing, blots were incubated with donkey-anti-rabbit or donkey-anti-goat IgG-horseradish peroxidase conjugates, and developed using Enhanced Chemiluminescence (Amersham, Arlington Heights, IL).

Effect of PGs on [3H]Thymidine Incorporation in Colon Cancer Cells.

Cells (1 × 106) were serum starved for 24 h in the presence or absence of Tc. Fresh serum-free medium containing various concentrations (10−12 to 10−5m) of PGs was added to cells for an additional 24 h. In some cases, cells were treated with either crude or HPLC-purified (see below) ethyl acetate-extracted lipids from HCA-7 or HCT-15 conditioned medium. For these latter experiments, cell medium (0.5 ml from ∼106 cells) was acidified to pH 3 with 1 m HCl, and lipids were extracted with two volumes of ethyl acetate. The organic layer was then dried under nitrogen and resuspended in a 0.5 ml of serum-free DMEM, and the entire volume was applied to various cell lines (∼106 cells).

At 21 h, cells were pulsed with [3H]thymidine (1 μCi/well) for an additional 3 h. The relative amount of radioactivity incorporated into trichloroacetic acid-insoluble material was determined by scintillation counting in an aqueous fluor (Amersham; Ref. 12). Results were normalized to cell number, which was determined in replicate wells for each condition, and all results are representative of at least three separate experiments.

Growth of HCA-7 Cells Containing an Antisense COX-2 Construct in Collagen Gels.

The Vitrogen collagen gel system was used (23). Briefly, collagen (2 mg/ml final volume) was added to DMEM containing 2% FCS with or without Tc (2 μg/ml). The mixture was added to each well of a 24-well plastic dish and allowed to solidify. Subsequently, 5 × 104 CRC cells grown in serum in the presence or absence of Tc were layered onto each well and covered with collagen, which was allowed to harden. One ml of medium containing a PG (concentrations of 10−11 to 10−5m) was applied over the collagen. Dishes were incubated at 37°C, and media were changed every other day for 6 weeks. Gels were subsequently stained with 0.2% (w/v) crystal violet and destained with 1% acetic acid; colony numbers were determined by light microscopy.

Endogenous PG Production by Parental HCA-7 Cells.

HCA-7 cells (2 × 107) grown on plastic dishes were incubated in serum-free DMEM containing 50 μCi [3H8]arachidonic acid and 10 μm unlabeled arachidonic acid for 18 h (12). Subsequently, PGs were extracted from conditioned medium with ethyl acetate after acidification to pH 3 and further analyzed by HPLC and MS.

HPLC Separation of PGs.

Aliquots of organic extracts from HCA-7 conditioned medium were dried under nitrogen, resuspended in acetonitrile, and chromatographed on reversed phase HPLC using a solvent system of acetonitrile:water:acetic acid (35:65:0.01, v/v/v), which was changed at 20 min to methanol:water:acetic acid (85:15:0.01,v/v/v).

Analysis of PGs by Mass Spectrometry.

PGD2 was quantified by GC/negative ion chemical ionization MS using stable isotope dilution techniques as described (24). PGJ2 was analyzed using a modification of methods developed for PGD2. Briefly, PGJ2 in the organic extracts from HCA-7 conditioned medium was converted to an O-methyloxime trimethylsilyl ether pentafluorobenzyl ester derivative and purified by TLC (Silica LKD60 plates; Waters, Maidstone, United Kingdom) using a solvent system of hexane:acetone (70: 30, v/v). Material migrating with a Rf identical to chemically pure PGJ2 (Rf = 0.31) was scraped from the plate and analyzed by GC/EI MS. GC was performed using a DB1701 fused silica capillary column (J and W Scientific, Folsom, CA; Ref. 24). EI MS was performed using a Finnegan Incos 50 mass spectrometer as described (25). In preliminary experiments, PGJ2 was not generated from the dehydration of exogenously added PGD2 during the purification and chromatographic procedures described.

Statistical Analyses.

Where appropriate, data were analyzed for significant differences using the unpaired t test. All tests were two sided, and differences were considered statistically significant when Ps were <0.05.

Characterization of HCA-7 Cell Lines Containing a Tetracycline-inducible Antisense COX-2 Construct.

Induction of antisense COX-2 mRNA in transfected HCA-7 cells by Tc addition led to ablation of basal COX-2 expression, as demonstrated by both Northern and Western blotting (Fig. 2, A and B). Tc had no effect on COX-2 levels in parental or mock-transfected HCA-7 cells. In addition to the loss of COX-2, COX-1 levels were reduced, presumably due to sequence homology between the two mRNAs (Fig. 2,A). Levels of amphiregulin and transforming growth factor-α mRNA were, however, unaltered following Tc addition to the transfected cells (data not shown). These effects were reproduced in two other independently derived clones. HCA-7 cells produce large amounts of the eicosanoids PGD2, PGE2, PGF, and TxA2(12). In these studies, PG production was quantified by measuring PGD2 because in preliminary experiments, formation of other PGs paralleled PGD2 generation. When HCA-7 cells containing an antisense COX-2 construct were grown in the absence of Tc, PG production was identical to parental nontransfected cells (Fig. 2,C). In the presence of Tc, PGD2 levels were decreased 99.8% (P < 0.05). For comparison, treatment of these cells with the COX-2 inhibitor SC58125 (50 μm), in the absence of Tc, reduced PG production 90% (data not shown). To determine whether the reduction in PGD2 generation correlated with a decrease in mitogenesis, we measured [3H]thymidine incorporation in transfected cells grown in the presence or absence of Tc. In the absence of Tc, [3H]thymidine incorporation in HCA-7 cells was essentially the same as parental HCA-7 cells. Addition of Tc for 24 h led to a decrease in [3H]thymidine uptake in a dose-dependent manner (Fig. 2 D), suggesting that this near-complete reduction in PG generation is associated with a decrease in basal mitogenesis of transfected HCA-7 cells. The maximum decrease in proliferation was 60% (P < 0.05). For comparison, treatment of these cells with SC58125 in the absence of Tc reduced proliferation 43% (data not shown).

Effects of PGs on [3H]Thymidine Incorporation in COX-depleted HCA-7 Cells.

Because inhibition of PG synthesis in HCA-7 cells containing an antisense COX-2 construct was associated with a significant decrease in proliferation, we sought to determine whether addition of various PGs to these cells would restore proliferation. For these studies, we chose not only to examine the effects of parental PGs on mitogenesis but also cyclopentenone PGs because these latter compounds have been reported to modulate growth in a number of cell lines (17). Table 1 shows the effect of various PGs to restore mitogenesis in Tc-treated clones to levels observed in parental HCA-7 cells. Data are expressed as the EC50 for each compound. As is evident, all PGs tested enhanced mitogenesis to varying degrees in COX-depleted HCA-7 cells. In contrast, no PG induced proliferation in parental HCA-7 cells or antisense COX-2-transfected HCA-7 cells grown in the absence of Tc, although PGJ2 and 15-deoxy-Δ12,14-PGJ2, at concentrations of 10−6m and above, significantly reduced [3H]thymidine incorporation. Surprisingly, the cyclopentenone dehydration products of PGD2, most notably PGJ2 and 15-deoxy-Δ12,14-PGJ2, exhibited potent mitogenic effects at concentrations in the nanomolar range. In parallel experiments, we also quantified the number of antisense COX-2-transfected HCA-7 cells after a 24-h treatment with either PGJ2 or 15-deoxy-Δ12,14-PGJ2 (10−9m) compared with untreated cells. PGJ2 induced a mean 19% increase in cell number, whereas 15-deoxy-Δ12,14-PGJ2 effected a mean 14% increase (data not shown).

Fig. 3 shows the concentration-response curves for PGD2 and its dehydration products on proliferation in COX-depleted HCA-7 cells. As is evident, these compounds exhibit saturable sigmoidal-shaped concentration-response curves. The maximal response of 100% is the [3H]thymidine incorporation that was quantified in antisense COX-2-transfected cells grown in the absence of Tc. Data are presented as % maximal response so that a direct comparison between different PGs can be made because the incorporation of [3H]thymidine into antisense COX-2-transfected cells grown in the absence of Tc varied up to 40% between different experiments. At concentrations ≥10−6m, both PGJ2 and 15-deoxy-Δ12,14-PGJ2 significantly decreased proliferation in antisense COX-2-transfected HCA-7 cells by up to 85% compared with untreated cells, consistent with other studies (17). We have also found that in HCT-15 cells, a CRC cell line that does not express COX or produce PGs, PGJ2 (EC50 = 5.0 × 10−9m) and 15-deoxy-Δ12,14-PGJ2 (EC50 = 4.1 × 10−9m) increase proliferation in a concentration-dependent manner (Fig. 4). However, higher concentrations of these compounds (10−6 and 10−5m) significantly decreased [3H]thymidine incorporation by up to 70% compared with untreated cells. Taken together, these observations indicate that PGs, particularly PGJ2 and 15-deoxy-Δ12,14-PGJ2, stimulate mitogenesis at low concentrations in CRC cells that do not exhibit COX activity or produce endogenous PGs.

Effects of PGD2 and PGJ2 on Growth of Antisense COX-2 Transfected HCA-7 Cells in Collagen Gels.

To confirm that cyclopentenone PGs induce proliferation in COX-depleted HCA-7 cells, we carried out experiments in which Tc-treated transfected cells were plated on collagen gels (23) and then treated for 6 weeks with varying concentrations of either PGD2 or PGJ2 that was added to medium overlying the gel matrix. PGD2, the precursor of cyclopentenone PGs, was studied in addition to PGJ2, because the former compound readily dehydrates in aqueous solutions to J-series cyclopentenone PGs (16). Fig. 5,A shows the effect of various concentrations of PGD2 on COX-depleted HCA-7 cell colony formation. As is evident, no effect was observed at concentrations <10−9m. Above 10−9m PGD2, colony number increased in a concentration-dependent manner. Maximum increases in colony number in the presence of PGD2 were 170% of untreated transfectant cells at a concentration of PGD2 of 10−8m (P < 0.05). Fig. 5 B shows the effect of PGJ2 on colony formation in antisense COX-2-transfected HCA-7 cells grown in collagen gels in the presence of Tc. As is evident, at concentrations <10−10m, no effect on colony number was observed. Colony formation increased up to a concentration of 10−8m (P < 0.05). At concentrations >10−7m, colony number decreased significantly, reaching a maximum reduction of 60% below the colony number of untreated cells. These latter findings are consistent with previous studies showing a decrease in proliferation in various cell lines in the presence of higher concentrations of PGJ2(17). Neither PG increased colony formation in either COX-2 expressing transfected or parental HCA-7 cells (data not shown).

Effect of Lipids Extracted from Parental HCA-7 Conditioned Medium to Induce Proliferation in COX-2 Antisense HCA-7 and HCT-15 Cells.

The above studies show that the cyclopentenone PGs, PGJ2 and 15-deoxy-Δ12,14-PGJ2, induce proliferation in COX-depleted HCA-7 cells. These results led us to consider whether endogenous cyclopentenone PGs might stimulate the growth of parental HCA-7 cells. In this regard, it has been reported previously that the J-series cyclopentenone PG, Δ12-PGJ2, is formed in humans (26). Initially, we examined the possibility that lipids extracted from parental HCA-7 conditioned medium might stimulate growth in COX-depleted HCA-7 cells in a manner similar to chemically pure cyclopentenone PGs. COX-2-transfected HCA-7 cells were incubated with lipid extracts for 24 h, and [3H]thymidine incorporation was then measured. As shown in Fig. 6,A, lipids extracted from parental HCA-7 conditioned medium significantly increased DNA replication in COX-depleted HCA-7 cells. Similar effects were observed when lipids extracted from parental HCA-7 cell-conditioned medium were applied to HCT-15 cells (Fig. 6 B). Interestingly, lipids extracted from Tc-treated antisense COX-2-transfected HCA-7 or HCT-15 cell conditioned medium failed to significantly affect proliferation in COX-depleted HCA-7 cells or HCT-15 cells. Taken together, these studies support the concept that parental HCA-7 cells produce a lipid(s), possibly a cyclopentenone PG, that induces proliferation in CRC cell lines with negligible COX activity and that does not produce PGs.

Characterization of Cyclopentenone PGs Produced by Parental HCA-7 Cells.

We then determined whether parental HCA-7 cells produce cyclopentenone PGs. For these studies, cells were incubated in serum-free medium containing [3H8]arachidonic acid (50 μCi) and unlabeled arachidonic acid (10 μm) for 24 h. Subsequently, lipids were extracted from the medium, and PGs were identified by reversed phase HPLC using a solvent system that readily separates parent PGs from less polar eicosanoids. Eicosanoid products were detected by measuring radioactivity that eluted from the HPLC. Fig. 7 shows an HPLC chromatogram of the results from this experiment. A series of peaks that perfectly cochromatographed with chemically pure parental eicosanoids are shown eluting at early retention volumes. In addition, several less polar peaks (Peaks 1, 2, and 3) were present, eluting in a later retention volume. Peak 1 represents a compound with a retention time identical to PGJ2, whereas peak 2 eluted in a retention volume identical to both PGA2 and Δ12-PGJ2. Peak 3 eluted at a retention volume similar to 15-deoxy-Δ12,14-PGJ2.

Material from the incubation corresponding to Peak 1 in Fig. 7 was then further purified, derivatized to the O-methyloxime, trimethylsilyl ether, pentafluorobenzyl ester derivative and analyzed by GC/EI MS. Fig. 8 shows the results of this analysis. The mass spectrum shown is essentially identical to that obtained when chemically pure PGJ2 was analyzed in a similar manner (16, 27). A prominent molecular ion was present at m/z 615. Additional characteristic high molecular weight ions were present at m/z 600 (M-15, loss of .CH3); m/z 584 (loss of .OCH3); m/z 544 (M − 71, loss of .CH2—(CH2)3—CH3); m/z 525 (M − 90, loss of Me3SiOH); m/z 513 (M − 71 − 31); m/z 498 (M − 71 − 31 − 15), and m/z 454 (M − 71 − 90). Prominent low mass unit ions were also present and included m/z 181 (PFB ion, +CH2C6F5) and m/z 199. This latter ion represents the lower side chain of PGJ2 [CH⋕CH—CH(Me3SiOH)(CH2)4CH3] and is important because it distinguishes PGJ2 from Δ12-PGJ2(16).

In addition, material obtained from HPLC-purified Peak 1 (Fig. 7) was also examined for biological activity in comparison to chemically pure PGJ2. The ability of material contained in the peak (over a concentration range of 10−10 to 10−7m) to induce proliferation in antisense COX-2-transfected HCA-7 cells was virtually identical to that of PGJ2. The EC50 for the HPLC-purified material was determined to be 1.1 × 10−9m compared with an EC50 of 7.2 × 10−10m for PGJ2.

We also analyzed Peak 2 (Fig. 5) after derivatization to an O-methyloxime pentafluorobenzyl ester by EI MS and have found that the compound represented by the chromatographic peak was identical to PGA2 (data not shown). Taken together, these studies provide definitive evidence that cyclopentenone PGs are present in conditioned medium derived from HCA-7 cells.

These studies report that cyclopentenone PGs, particularly PGJ2 and 15-deoxy-Δ12,14-PGJ2, induce proliferation in a human CRC cell line, HCA-7, that has been transfected with a Tc-inducible antisense COX-2 construct. Similar findings were observed in HCT-15 cells that do not express COX or produce PGs and have not been genetically manipulated. HCA-7 cells containing an antisense COX-2 cDNA construct under the control of a Tc-inducible promoter were developed to directly examine the effect of PGs on CRC cell proliferation. It has been reported previously that inhibition of PG synthesis using a pharmacological approach is confounded by the fact that NSAIDS, in addition to inhibiting proliferation in CRC cells by decreasing PG production, also modulate cell growth by inhibiting PG-independent pathways (13, 14, 15, 28). For example, Chan et al. have reported recently that in CRC cells, NSAIDS induce apoptosis by stimulating ceramide production, and this effect is not directly dependent on PG production (13). Furthermore, Qiao et al.(14) have reported that sulindac sulfide inhibits proliferation in a CRC cell line that does not express COX and does not produce PGs. In the present studies, COX-depleted HCA-7 cells displayed marked inhibition of PG production, and this inhibition was associated with a significant reduction in basal proliferation compared with COX-2-expressing HCA-7 cells.

A large body of literature exists concerning the effects of PGs on cell growth. Depending on the particular cell type, parent PGs (PGD2, PGE2, PGF, PGI2, and TxA2) have been shown to exhibit either proliferative or antiproliferative effects (8, 9, 10). Very little information, however, exists concerning the ability of eicosanoids to modulate gastrointestinal epithelial cell growth. In one study, Qiao et al.(29) reported that PGE2, PGF, and PGI2, at submicromolar concentrations induced proliferation in the human CRC cell lines SW1116 and HT-29 in vitro and that 16,16-dimethyl-PGE2 stimulated colonic epithelial proliferation in vivo. On the other hand, Hanif et al.(28) reported recently that parental PGs do not modulate proliferation in CRC cells because treatment of cells with various NSAIDs reduces mitogenensis, and this inhibition cannot be overcome with exogenous administration of PGE2, PGF, or a PGI2 analogue. As reported previously, we have found that parental PGs cannot reverse NSAID-induced decreases in DNA replication in parental HCA-7 cells (12). This observation provided the impetus to develop HCA-7 cell lines containing an antisense COX-2 construct. In the present studies, parental PGs induced proliferation in COX-depleted HCA-7 cells but only at concentrations in the micromolar range. These relatively high concentrations of eicosanoids may not be physiologically relevant (11).

In addition to parental PGs, there has been considerable recent interest in the role of cyclopentenone PGs, which are dehydration products of PGD2 and PGE2, on cellular proliferation and differentiation. Interest in cyclopentenone PGs has centered on their antiproliferative activity because they markedly inhibit cell growth when applied to a variety of tumor cell lines including breast, lung, and myeloproliferative cells (17). These effects, however, are observed only at concentrations in the micromolar range or higher. Santoro and others have shown that in association with these antiproliferative effects, there is activation of genes, including those encoding various heat shock proteins and heme oxygenase (30, 31). This activation occurs in association with the modulation of nuclear transcription factors including nuclear factor-κB (31). In contrast, there is at least one report noting that cyclopentenone PGs induce proliferation in the breast cancer cell line MCF-7 at concentrations in the low nanomolar range (18). Interestingly, at higher concentrations, proliferation is inhibited in this cell line. In the present studies, we report that cyclopentenone PGs, particularly PGJ2 and 15-deoxy-Δ12,14-PGJ2, induce proliferation in COX-depleted HCA-7 cells at concentrations in the nanomolar range. These effects were quantified both by measuring DNA replication and growth in collagen gels. In addition, similar effects on DNA replication were observed in the CRC cell line HCT-15 that does not express COX or produce PGs. Consistent with other reports, proliferation in both COX-depleted HCA-7 and HCT-15 cells decreased at high concentrations of cyclopentenone PGs (17, 18).

The mechanism by which cyclopentenone PGs induce growth in CRC cells is presently unknown. The fact that cyclopentenone PGs increase proliferation in antisense COX-2-transfected HCA-7 cells in a saturable, concentration-dependent manner suggests that this effect might be receptor mediated. Physiological effects of PGs are thought to be mediated by interactions with receptors. Receptors for all of the parent PGs have been identified, and amino acid sequences have been reported (11). In addition, the cyclopentenone PG 15-deoxy-Δ12,14-PGJ2 has been found to be a ligand for the nuclear receptor, PPARγ (32, 33). A recent report has noted that activation of this receptor in various colon cancer cells in vitro inhibits growth (34). In the same study, Sarraf et al.(34) also found that the PPARγ agonist troglitazone decreases the growth of human tumor xenografts in nude mice. In contrast, Saez et al.(35) have reported that APCmin mice treated with troglitazone develop an increased number of intestinal tumors compared with untreated animals. HCA-7 cells have been shown to express PPARγ (36). Thus, whether cyclopentenone PGs enhance proliferation in HCA-7 cells via interaction with the PPARγ receptor or another receptor(s) deserves further study.

Evidence has been obtained previously that J-series cyclopentenone PGs derived from PGD2 are produced in vivo. Hirata et al.(26) reported, using mass spectrometric methodologies, that humans generate significant amounts of Δ12-PGJ2 that can be detected in the urine of healthy volunteers. In addition, they reported that infusion of PGD2 into monkeys led to large increases in the urinary excretion of Δ12-PGJ2. Furthermore, administration of NSAIDS in vivo markedly reduced urinary levels of Δ12-PGJ2. The finding that cyclopentenone PG products of PGD2 are formed in humans provided a rationale to determine whether these compounds are generated by HCA-7 cells in culture. Initial studies revealed that lipids extracted from conditioned medium from parental HCA-7 cells could stimulate proliferation in both COX-depleted HCA-7 and HCT-15 cells. Subsequent experiments using mass spectrometry provided definitive evidence for the generation of one of these compounds, i.e., PGJ2, by parental HCA-7 cells. For these studies, we used methods to maximize generation of cyclopentenone PGs by incubating cells in the presence of exogenous arachidonic acid. In subsequent experiments, we have also identified PGJ2 in conditioned medium from parental HCA-7 cells not stimulated with exogenous arachidonic acid. Levels of this compound are in the nanomolar range when quantified by a mass spectrometric assay (data not shown), and thus PGJ2 is produced by parental HCA-7 cells in amounts capable of stimulating proliferation in COX-deficient cells in culture. For comparison, the parental eicosanoids PGD2, PGE2, PGF, and thromboxane are also produced by parental HCA-7 cells at concentrations ∼10-fold higher than PGJ2. These concentrations, however, are significantly lower than those required to induce proliferation in antisense COX-2-transfected HCA-7 cells in culture. COX-2 protein as determined by immunostaining is heterogeneous in cultured HCA-7 cells and in CRC tumors in situ(6, 7, 12). We propose, therefore, that cyclopentenone PGs may stimulate proliferation of CRC cells in an autocrine and paracrine manner.

In conclusion, we report that cyclopentenone PGs are inducers of proliferation in cyclooxygenase-depleted CRC cells at concentrations in the nanomolar range. Studies aimed at elucidating the mechanism by which cyclopentenone PGs induce proliferation will allow for a further understanding of how PGs modulate cellular proliferation.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

      
1

This work was supported by NIH Grants DK48831 (to J. D. M.), CA77839 (to J. D. M.), GM15431 (to J. D. M.), CA46413 (to R. J. C.), and CA68485 (Vanderbilt Cancer Center). R. J. C. acknowledges the generous support of the Joseph and Mary Keller Foundation. J. D. M. is the recipient of a Burroughs Wellcome Fund Clinical Scientist Award in Translational Research.

            
3

The abbreviations used are: CRC, colorectal cancer; NSAIDS, nonsteroidal anti-inflammatory drugs; COX, cyclooxygenase; PG, prostaglandin; Tx, thromboxane; Tc, tetracycline; HPLC, high-pressure liquid chromatography; GC, gas chromatography; MS, mass spectrometry; EI, electron ionization; PPARγ, peroxisome proliferator-activated receptor γ.

Fig. 1.

Pathway of formation of cyclopentenone PGs from PGD2 and PGE2. PGD2 loses a molecule of water to form PGJ2, which can rearrange to form Δ12-PGJ2. Δ12-PGJ2 can then dehydrate to form 15-deoxy-Δ12,14-PGJ2. PGE2 can lose a molecule of water to form PGA2.

Fig. 1.

Pathway of formation of cyclopentenone PGs from PGD2 and PGE2. PGD2 loses a molecule of water to form PGJ2, which can rearrange to form Δ12-PGJ2. Δ12-PGJ2 can then dehydrate to form 15-deoxy-Δ12,14-PGJ2. PGE2 can lose a molecule of water to form PGA2.

Close modal
Fig. 2.

Characterization of antisense COX-2 HCA-7 transfectants. A, Northern blot analysis showing COX-1 and COX-2 mRNA expression in mock-transfected (empty pTet/VP16 and pUHD.2neo) and antisense COX-2 HCA-7 cells grown in the presence (2 μg/ml) or absence of Tc for 24 h in serum-free DMEM. Similar changes in COX mRNA levels were observed in two other independently derived antisense COX-2 clones. IB15 is shown as a control for equivalent loading and transfer of mRNA. B, Western blot analysis showing COX-1 and COX-2 protein levels in the transfected HCA-7 cells grown as described above. C, effect of antisense COX-2 expression on PGD2 production in mock-transfected and antisense COX-2-transfected HCA-7 cells grown in the presence (1 or 2 μg/ml) or absence of Tc. Triplicate cultures of transfected cells (∼1 × 106) were grown for 24 h in serum-free DMEM containing Tc (0, 1, or 2 μg/ml); medium was collected, and PGD2 production was analyzed by MS. Cell number was determined using a hemacytometer. Data are expressed as means for three experiments carried out in triplicate; bars, SE. D, effect of COX depletion on DNA synthesis in antisense COX-2-transfected HCA-7 cells. Transfected HCA-7 cells were grown in the presence or absence of Tc for 24 h, and DNA synthesis was determined by [3H]thymidine incorporation. Duplicate cultures were counted in triplicate using a hemacytometer. Data are expressed as means; bars, SE.

Fig. 2.

Characterization of antisense COX-2 HCA-7 transfectants. A, Northern blot analysis showing COX-1 and COX-2 mRNA expression in mock-transfected (empty pTet/VP16 and pUHD.2neo) and antisense COX-2 HCA-7 cells grown in the presence (2 μg/ml) or absence of Tc for 24 h in serum-free DMEM. Similar changes in COX mRNA levels were observed in two other independently derived antisense COX-2 clones. IB15 is shown as a control for equivalent loading and transfer of mRNA. B, Western blot analysis showing COX-1 and COX-2 protein levels in the transfected HCA-7 cells grown as described above. C, effect of antisense COX-2 expression on PGD2 production in mock-transfected and antisense COX-2-transfected HCA-7 cells grown in the presence (1 or 2 μg/ml) or absence of Tc. Triplicate cultures of transfected cells (∼1 × 106) were grown for 24 h in serum-free DMEM containing Tc (0, 1, or 2 μg/ml); medium was collected, and PGD2 production was analyzed by MS. Cell number was determined using a hemacytometer. Data are expressed as means for three experiments carried out in triplicate; bars, SE. D, effect of COX depletion on DNA synthesis in antisense COX-2-transfected HCA-7 cells. Transfected HCA-7 cells were grown in the presence or absence of Tc for 24 h, and DNA synthesis was determined by [3H]thymidine incorporation. Duplicate cultures were counted in triplicate using a hemacytometer. Data are expressed as means; bars, SE.

Close modal
Fig. 3.

Concentration-response curves for the effect of either PGD2, PGJ2, Δ12-PGJ2, or 15-deoxy-Δ12,14-PGJ2 to induce proliferation in COX-depleted HCA-7 cells. Proliferation was quantified by measuring [3H]thymidine incorporation in cells 24 h after treatment with PG. The maximal response of 100% is the [3H]thymidine incorporation that was quantified in antisense COX-2-transfected cells grown in the absence of Tc. Concentration-response curves for each compound are: A, PGD2; B, Δ12-PGJ2; C, 15-deoxy-Δ12,14-PGJ2; and D, PGJ2. Results represent data from at least four experiments performed in triplicate.

Fig. 3.

Concentration-response curves for the effect of either PGD2, PGJ2, Δ12-PGJ2, or 15-deoxy-Δ12,14-PGJ2 to induce proliferation in COX-depleted HCA-7 cells. Proliferation was quantified by measuring [3H]thymidine incorporation in cells 24 h after treatment with PG. The maximal response of 100% is the [3H]thymidine incorporation that was quantified in antisense COX-2-transfected cells grown in the absence of Tc. Concentration-response curves for each compound are: A, PGD2; B, Δ12-PGJ2; C, 15-deoxy-Δ12,14-PGJ2; and D, PGJ2. Results represent data from at least four experiments performed in triplicate.

Close modal
Fig. 4.

Effect of various concentrations of PGJ2 on proliferation in HCT-15 cells quantified by measuring [3H]thymidine incorporation at 24 h after treatment with PG. ∗, P < 0.05 for a particular PG concentration inducing a significant increase compared with vehicle only. Results represent data from at least four experiments performed in triplicate; bars, SE.

Fig. 4.

Effect of various concentrations of PGJ2 on proliferation in HCT-15 cells quantified by measuring [3H]thymidine incorporation at 24 h after treatment with PG. ∗, P < 0.05 for a particular PG concentration inducing a significant increase compared with vehicle only. Results represent data from at least four experiments performed in triplicate; bars, SE.

Close modal
Fig. 5.

Effect of various concentrations of PGD2 (A) and PGJ2 (B) on COX-depleted HCA-7 colony number when grown in collagen gels. PGs were administered to cells every other day. Colony number was quantified after 6 weeks. ∗, P < 0.05 for a particular PG concentration inducing a significant increase compared with vehicle alone. Results represent data from four separate experiments performed in quadruplicate; bars, SE.

Fig. 5.

Effect of various concentrations of PGD2 (A) and PGJ2 (B) on COX-depleted HCA-7 colony number when grown in collagen gels. PGs were administered to cells every other day. Colony number was quantified after 6 weeks. ∗, P < 0.05 for a particular PG concentration inducing a significant increase compared with vehicle alone. Results represent data from four separate experiments performed in quadruplicate; bars, SE.

Close modal
Fig. 6.

Effect of lipids extracted from CRC cell conditioned medium on proliferation in CRC cells quantified by [3H]thymidine incorporation 24 h after treatment. A, effect on COX-depleted HCA-7 cells. i, buffer alone; ii, lipids from parent HCA-7 cells; iii, lipids from antisense COX-2-transfected HCA-7 cells grown in the presence of Tc; iv, lipids from HCT-15 cells. B, effect on HCT-15 cells. ∗, P < 0.05 for lipid-treated compared with untreated cells; bars, SE.

Fig. 6.

Effect of lipids extracted from CRC cell conditioned medium on proliferation in CRC cells quantified by [3H]thymidine incorporation 24 h after treatment. A, effect on COX-depleted HCA-7 cells. i, buffer alone; ii, lipids from parent HCA-7 cells; iii, lipids from antisense COX-2-transfected HCA-7 cells grown in the presence of Tc; iv, lipids from HCT-15 cells. B, effect on HCT-15 cells. ∗, P < 0.05 for lipid-treated compared with untreated cells; bars, SE.

Close modal
Fig. 7.

Reversed phase HPLC chromatogram of PGs from conditioned medium of parental HCA-7 cells. PGs were detected by radioactivity. Parent PGs (PGD2, PGE2, and PGF) were identified by cochromatography with chemically pure standards. A series of less polar radioactive compounds are also present. Peak 1 coelutes with chemically pure PGJ2. Peak 2 coelutes with both PGA2 and Δ12-PGJ2. Peak 3 elutes with 15-deoxy-Δ12,14-PGJ2. The solvent systems used for this separation are described in “Materials and Methods.”

Fig. 7.

Reversed phase HPLC chromatogram of PGs from conditioned medium of parental HCA-7 cells. PGs were detected by radioactivity. Parent PGs (PGD2, PGE2, and PGF) were identified by cochromatography with chemically pure standards. A series of less polar radioactive compounds are also present. Peak 1 coelutes with chemically pure PGJ2. Peak 2 coelutes with both PGA2 and Δ12-PGJ2. Peak 3 elutes with 15-deoxy-Δ12,14-PGJ2. The solvent systems used for this separation are described in “Materials and Methods.”

Close modal
Fig. 8.

EI MS analysis of a presumed cyclopentenone PG purified from parental HCA-7 cell conditioned medium. Material was analyzed as the O-methyloxime, pentafluorobenzyl ester, trimethylsilyl ether derivative. This mass spectrum was essentially identical to that obtained when chemically pure PGJ2 was analyzed in the same manner.

Fig. 8.

EI MS analysis of a presumed cyclopentenone PG purified from parental HCA-7 cell conditioned medium. Material was analyzed as the O-methyloxime, pentafluorobenzyl ester, trimethylsilyl ether derivative. This mass spectrum was essentially identical to that obtained when chemically pure PGJ2 was analyzed in the same manner.

Close modal
Table 1

EC50 values of various eicosanoids to induce proliferation in COX-depleted HCA-7 cells

Proliferation was quantified by measuring [3H]thymidine incorporation in cells 21 h after treatment with PGs. The EC50 represents the molar concentration of a particular PG required to restore proliferation to 50% of that measured in antisense COX-2-transfected HCA-7 cells grown in the absence of Tc.
Compound EC50 (m
PGH2 4.0 × 10−7 
U46619 (TxA2/PGH2 agonist) 9.2 × 10−8 
PGF 1.0 × 10−6 
PGE2 4.5 × 10−7 
PGA2 3.8 × 10−8 
PGB2 1.0 × 10−7 
PGD2 3.7 × 10−7 
PGJ2 7.2 × 10−10 
Δ12-PGJ2 1.2 × 10−8 
15-Deoxy-Δ12,14-PGJ2 9.2 × 10−10 
Proliferation was quantified by measuring [3H]thymidine incorporation in cells 21 h after treatment with PGs. The EC50 represents the molar concentration of a particular PG required to restore proliferation to 50% of that measured in antisense COX-2-transfected HCA-7 cells grown in the absence of Tc.
Compound EC50 (m
PGH2 4.0 × 10−7 
U46619 (TxA2/PGH2 agonist) 9.2 × 10−8 
PGF 1.0 × 10−6 
PGE2 4.5 × 10−7 
PGA2 3.8 × 10−8 
PGB2 1.0 × 10−7 
PGD2 3.7 × 10−7 
PGJ2 7.2 × 10−10 
Δ12-PGJ2 1.2 × 10−8 
15-Deoxy-Δ12,14-PGJ2 9.2 × 10−10 
1
Levy G. N. Prostaglandin H synthases, nonsteroidal anti-inflammatory drugs and colon cancer.
FASEB J.
,
11
:
234
-247,  
1997
.
2
Eberhart C. E., DuBois R. N. Eicosanoids and the gastrointestinal tract.
Gastroenterology
,
109
:
285
-301,  
1995
.
3
O’Neill G., Hutchinson A. F. Expression of mRNA for cyclooxygenase-1 and cyclooxygenase-2 in human tissues.
FEBS Lett.
,
330
:
156
-160,  
1993
.
4
Williams C. W., DuBois R. N. Prostaglandin endoperoxide synthase: why two isoforms?.
Am. J. Physiol.
,
270
:
G393
-G400,  
1996
.
5
Sano H., Kawahito Y., Wilder R., Hashiramoto A., Mukai S., Asai K., Kimura S., Kato H., Kondo M., Hia T. Expression of cyclooxygenase-1 and -2 in human colorectal cancer.
Cancer Res.
,
55
:
3785
-3789,  
1995
.
6
DuBois R. N., Radhika A., Reddy B. S., Entingh A. J. Increased cyclooxygenase-2 levels in carcinogen-induced rat colonic tumors.
Gastroenterology
,
110
:
1259
-1262,  
1996
.
7
Williams C. S., Luongo C., Radhika A., Zhang T., Lamps L. W., Nanney L. B., Beauchamp R. D., DuBois R. N. Elevated cyclooxygenase-2 levels in Min mouse adenomas.
Gastroenterology
,
111
:
1134
-1140,  
1996
.
8
Campbell W. B., Halushka P. V. Lipid derived autacoids. Eicosanoids and platelet-activating factor Ed. 9 Hardman J. G. Limbird L. E. eds. .
Goodman & Gilman’s The Pharmacological Basis of Therapeutics
,
:
601
-616, McGraw-Hill New York  
1996
.
9
Santoro M. G., Philpott G. W., Jaffee B. M. Inhibition of tumour growth in vivo and in vitro by prostaglandin E.
Nature (Lond.)
,
263
:
777
-779,  
1976
.
10
Fukunaga M., Makita N., Roberts L. J., Morrow J. D., Takahashi K., Badr K. F. Evidence for the existence of F2-isoprostane receptors on rat vascular smooth muscle cells.
Am. J. Physiol.
,
264
:
C1619
-C1624,  
1993
.
11
Breyer M. D., Jacobson H. R., Breyer R. M. Functional and molecular aspects of renal prostaglandin receptors.
J. Am. Soc. Nephrol.
,
7
:
8
-17,  
1996
.
12
Coffey R. J., Hawkey C. J., Damstrup L., Graves-Deal R., Daniel V. C., Dempsey P. J., Chinery R., Kirkland S. C., DuBois R. N., Jetton T. L., Morrow J. D. Epidermal growth factor receptor activation induces nuclear targeting of cyclooxygenase-2, basolateral release of prostaglandins, and mitogenesis in polarizing colon cancer cells.
Proc. Natl. Acad. Sci. USA
,
94
:
657
-662,  
1997
.
13
Chan T. A., Morin P. J., Vogelstein B., Kinzler K. W. Mechanisms underlying nonsteroidal antiinflammatory drug-mediated apoptosis.
Proc. Natl. Acad. Sci. USA
,
95
:
681
-686,  
1998
.
14
Qiao L., Shiff S. J., Rigas B. Sulindac sulfide inhibits the proliferation of colon cancer cells: diminished expression of the proliferation markers PCNA and Ki-67.
Cancer Lett.
,
115
:
229
-234,  
1997
.
15
Shiff S. J., Koutsos M. I., Qiao L., Rigas B. Nonsteroidal antiinflammatory drugs inhibit the proliferation of colon adenocarcinoma cells: effects on cell cycle and apoptosis.
Exp. Cell Res.
,
222
:
179
-188,  
1996
.
16
Fitzpatrick F. A., Wynalda M. A. Albumin-catalyzed metabolism of prostaglandin D2.
J. Biol. Chem.
,
258
:
11713
-11718,  
1983
.
17
Fukushima M. Biological activities and mechanisms of action of PGJ2 and related compounds: an update. Prostaglandins Leukotrienes Essent.
Fatty Acids
,
47
:
1
-12,  
1992
.
18
Shahabi N. A., Chengini N., Wittliff J. L. Alterations of MCF-7 human breast cancer cell after prostaglandins PGA1 and PGF treatment.
Exp. Cell Biol.
,
55
:
18
-27,  
1987
.
19
Marsh K. A., Stamp G. W. H., Kirkland S. C. Isolation and characterization of multiple cell types from a single human colonic carcinoma: tumourigenicity of these cell types in a xenograft system.
J. Pathol.
,
170
:
441
-450,  
1993
.
20
Gossen M., Freundlieb S., Bender G., Muller G., Hillen W., Bujard H. Transcriptional activation by tetracyclines in mammalian cells.
Science (Washington DC)
,
268
:
1766
-1769,  
1995
.
21
Schwab M. K., Alitalo H. E., Varmus H. E., Bishop J. M., George D. A cellular oncogene (c-Ki-ras) is amplified, overexpressed, and located within karyotypic abnormalities in mouse adrenocortical tumour cells.
Nature (Lond.)
,
303
:
497
-501,  
1983
.
22
Barnard J. A., Beauchamp R. D., Coffey R. J., Moses M. L. Regulation of intestinal epithelial cell growth by transforming growth factor type β.
Proc. Natl. Acad. Sci. USA
,
86
:
1578
-1582,  
1989
.
23
Al-Tubuly A. A., Spijker R., Pignatelli M., Kirkland S. C., Ritter M. A. Inhibition of growth and enhancement of differentiation of colorectal carcinoma cell lines by Mab and IL-4.
Int. J. Cancer
,
71
:
605
-611,  
1997
.
24
DuBois R. N., Awad J., Morrow J., Roberts L. J., II, Bishop P. R. Regulation of eicosanoid production and mitogenesis in rat intestinal epithelial cells by transforming growth factor-α and phorbol ester.
J. Clin. Invest.
,
93
:
493
-498,  
1994
.
25
Morrow J. D., Minton T. A., Mukundan C. R., Campbell M. D., Zackert W. E., Daniel V. C., Badr K. F., Blair I. A., Roberts L. J., II Free radical-induced generation of isoprostanes in vivo, evidence for the formation of d-ring and E-ring isoprostanes.
,
269
:
4317
-4326,  
1994
.
26
Hirata Y., Hayashi H., Ito S., Kiwawa Y., Ishibashi M., Sudo M., Miyazaki H., Fukushima M., Narumiya S., Hayaishi O. Occurrence of 9-deoxy-Δ912-13, 14-dihydroprostaglandin D2 in human urine.
J. Biol. Chem.
,
263
:
16619
-16625,  
1988
.
27
Asciak-Pace R. R. Advances in Prostaglandin, Thromboxane, and Leukotriene Research. Mass Spectra of Prostaglandins and Related Products Samuelsson B. Paoletti R. eds. .
,
18
:
382
-386, Raven Press New York  
1989
.
28
Hanif R., Pittas A., Feng Y., Kuotsos M., Qiao L., Staiano-Coico L., Shiff S. I., Rigas B. Effects of nonsteroidal anti-inflammatory drugs on proliferation and on induction of apoptosis in colon cancer cells by a prostaglandin-independent pathway.
Biochem. Pharmacol.
,
52
:
237
-245,  
1996
.
29
Qia L., Kozoni V., Tsioulias G. J., Koutsos M. I., Hanif R., Shiff S. J., Rigas B. Selected eicosanoids increase the proliferation rate of human colon carcinoma cell lines and mouse colonocytes in vivo.
Biochem. Biophys. Acta
,
1258
:
215
-223,  
1995
.
30
Rossi A., Santoro M. G. Induction by prostaglandin A1 of haem oxygenase in myoblastic cells: an effect independent of expression of the 70 kDa heat shock protein.
Biochem. J.
,
308
:
455
-463,  
1995
.
31
Rossi A., Elia G., Santoro M. G. Inhibition of nuclear factor KB by prostaglandin A1: an effect associated with heat shock transcription factor activation.
Proc. Natl. Acad. Sci. USA
,
94
:
746
-750,  
1997
.
32
Forman B. M., Tontonoz P., Chen J., Brun R. P., Spiegelman B. M., Evans R. M. 15-Deoxy-Δ12,14-prostaglandin J2 is a ligand for the adipocyte determination factor PPARγ.
Cell
,
83
:
803
-812,  
1995
.
33
Kliewer S. A., Lenhard J. M., Willson T. M., Patel I., Morris D. C., Lehmann J. M. A Prostaglandin J2 metabolite binds peroxisome proliferator-activated receptor γ and promotes adipocyte differentiation.
Cell
,
83
:
813
-819,  
1995
.
34
Sarraf P., Mueller E., Jones D., King F. J., DeAngelo D. J., Partridge J. B., Holden S. A., Chen L. B., Singer S., Fletcher C., Spiegelman B. M. Differentiation and reversal of malignant changes in colon cancer through PPARγ.
Nat. Med.
,
4
:
1046
-1052,  
1998
.
35
Saez E., Tontonoz P., Nelson M. C., Alvarez J. G. A., T. M., Baird S. M., Thomazy V. A., Evans R. M. Activators of the nuclear receptor PPARγ enhance colon polyp formation.
Nat. Med.
,
4
:
1058
-1062,  
1998
.
36
Brockman J. A., Gupta R. A., DuBois R. N. Activation of PPARγ leads to inhibition of anchorage-independent growth of human colorectal cancer cells.
Gastroenterology
,
115
:
1049
-1055,  
1998
.