Introduction of full-length APC modulates cyclooxygenase-2 expression in HT-29 human colorectal carcinoma cells at the translational level

Evidence for the involvement of COXs in cancer was Mutation of the adenomatous polyposis coli (APC) gene is suggested from pharmacologic analysis of prostaglandins. associated with the earliest stages of colorectal tumoriVarious animal and human tumor tissues, including human genesis and appears to be responsible for the hereditary colon cancer, have been reported to contain high concentrations condition familial adenomatous polyposis (FAP). Evidence of prostaglandins (4–6). The increased levels of prostaglandins indicates that cyclooxygenase-2 (COX-2) is induced and at in tumors provided the rational for use of non-steroidal antielevated levels in human colorectal cancers and in the inflammatory drugs (NSAIDs) as potential chemoprevention polyps of mouse FAP models. We have used HT-29 cells, a agents. Accumulating evidence indicates that NSAIDs can human colorectal carcinoma cell line with a mutant carreduce the incidence of colorectal cancers in human and boxy-truncated APC gene, in which intact APC gene has experimental animals and can reduce the number and size been introduced under the control of an inducible promoter. of polyps in patients with the hereditary condition familial These HT-29-APC cells provide a suitable model system to adenomatous polyposis (FAP) (7). Recently, evidence has been examine how COX-2 expression becomes dysregulated after presented that COX-2 is induced and at elevated levels loss of APC function. Induction of full-length APC causes in human colorectal cancers, azoxymethane-induced mouse the HT-29-APC cells to undergo apoptosis. However, differtumors and in the polyps of mouse FAP models (8–12). entiation, as measured by alkaline phosphatase activity, is Moreover, the expression of COX-2 appears to be linked to not induced upon expression of full-length APC. Fullmutation of the APC gene since COX-2 expression was length APC protein has been shown to bind the intracellular observed in the intestinal tract of APC mutant mice (13). protein β-catenin and, as a result, the Lef/Tcf transcription Perhaps the most striking evidence implicating COX-2 in factors are down-regulated. Analysis of APC immunoprecarcinogenesis is the finding that a null mutation for COX-2 cipitates demonstrate a time-dependent increase of βmarkedly reduced the number and size of intestinal tumors in catenin interacting with full-length APC. Thus, the Lef/Tcf APC∆716knockout mice, a murine model of FAP (12). Induction signaling pathway is intact at this point in these cells. of COX-2 is a very early event in the sequence of polyp Furthermore, upon expression of full-length APC, COX-2 formation to colon carcinogenesis and suggests that COX-2 protein expression is down-regulated while COX-2 mRNA plays a critical role in polyp development itself. When the levels remain the same. These data indicate that APC plays COX-2 gene is inactivated in FAP-model mice, both the size a role, either directly or indirectly, in the translational and number of polyps are reduced drastically (12). In addition, regulation of COX-2. Treatment of the HT-29-APC cells selective inhibitors of COX-2 cause a reduction in the number with sodium butyrate, an inducer of apoptosis, does not and size of polyps similar to that caused by the COX-2 gene alter COX-2 protein expression. Thus, COX-2 down-regulaknockout mutations (12). These results provide direct genetic tion appears to be APC specific and not just due to evidence that COX-2 plays a key role in colorectal tumoriapoptotic induction. APC appears to uniquely regulate genesis resulting from mutation of the APC gene. COX-2 expression. The mechanism by which COX-2 proAlthough mutations of the APC gene are associated with tein expression is down-regulated in the HT-29-APC cells the earliest stages of colorectal carcinogenesis and the expresis under investigation. sion of COX-2, the molecular processes that lead to the

29-βgal cells were cultivated in the same media as the normal HT-29 cells plus the addition of 0.6 mg/ml hygromycin.HCT-116 cells were cultivated in McCoy's 5A media supplemented with 10% fetal bovine serum.The media for the cells also contained gentamicin (1 mg/100 ml; Life Technologies, Inc.).Unless otherwise indicated, HT-29-APC cells were induced with 100 µM ZnCl 2 (Sigma) for the times indicated.Unless otherwise indicated, HT-29-APC cells were treated with 5 mM sodium butyrate (NaBT) (Sigma) for the times indicated.

SDS-PAGE
For western analysis, remaining attached cells following treatment were harvested, lysed, normalized and then Laemmli sample buffer was added to the samples.The samples were then sheared through a 25 gauge needle and boiled.For APC, proteins were separated by 3% low melting point agarose gel and transferred onto polyvinyldifluoride membrane (Immobilon-P; Millipore) essentially as described previously (15).For COX-2, proteins were separated by 8% SDS-PAGE and transferred onto Hybond-enhanced chemiluminescence (ECL) nitrocellulose membrane (Amersham).

Immunoblot analysis
Blots were blocked with 10% skim milk in Tris-buffered saline containing 0.1% Tween-20 (TBS-T) and washed.The blots were then incubated in 1% milk in TBS-T with an anti-appropriate specific antibody.The following antibodies were used: PG27 (Oxford Biochemical Research) for COX-2, APC (Santa Cruz), β-catenin (Santa Cruz) or actin (Santa Cruz).After washing, blots were incubated with anti-rabbit IgG horseradish peroxidase-linked secondary antibody (Amersham) for COX-2 and APC or anti-goat IgG horseradish peroxidase-linked secondary antibody (Santa Cruz) for β-catenin and actin respectively.After reacting by chemiluminescence (Amersham ECL detection system), bands were detected by exposure to Hyperfilm-MP (Amersham).

Northern analysis
Total RNA from cells was extracted by using TRI Reagent (Sigma).RNA samples were normalized and separated by electrophoresis in a formaldehyde-1% agarose gel.The RNA was transferred in 10ϫ SCC by capillary action onto nylon membrane (Schleicher & Schull) and UV-cross-linked with a Stratalinker UV light source (Stratagene).Human COX-2 (Oxford Biochemical Research) and human glyceraldehyde-3-phosphate dehydrogenase (G3PDH; Clontech) cDNA probes were labeled with [α-32 P]dCTP (Amersham Pharmacia Biotech) using the Prime-IT-II random prime kit (Stratagene).After hybridization and washes, the blots were then exposed to X-ray film (Amersham) for autoradiography.) could be detected prior to induction (Figure 1).HCT-116 cells, a human colorectal cancer cell line with an control of an inducible promoter permits a regulated expression of full-length APC protein and induction of apoptosis (14).
intact APC gene that expresses full-length APC protein, were used as a positive control (Figure 1).HT-29 cells transfected HT-29-APC cell lysates were examined for expression of intact APC at various times after treatment with zinc.The expression with the β-galactosidase gene in place of the APC gene were also utilized as a control and demonstrate only mutant APC of full-length APC (300 kDa) was observed (Figure 1).The HT-29-APC cells express full-length APC after 12 h of zinc protein after treatment with zinc (Figure 1).These results confirm inducible regulation of full-length APC protein by induction and maintain such expression for at least 4 days (14).In contrast, only the endogenous mutant APC (100 and zinc treatment in the HT-29-APC cells. of cell differentiation (17).We found no increase in alkaline phosphatase activity over the time course from 0 to 96 h following induction (data not shown) indicating no increase in cellular differentiation.

Binding of APC to β-catenin
Intact APC has been shown to bind the intracellular protein β-catenin (18,19) and thus down-regulate the Lef/Tcf signaling pathway (20,21).Full-length APC was immunoprecipitated from cell lysates, before and after induction with zinc, with antibodies specific for the C-terminus of APC.The resulting APC immunoprecipitates were analyzed for the presence of β-catenin by western analysis with an antibody specific to βcatenin (Figure 2).Analysis of the APC immunoprecipitates demonstrates a time-dependent increase of β-catenin interacting with full-length APC protein.This shows that binding of fulllength APC to β-catenin does occur and that the Lef/Tcf signaling pathway is presumably active in these cells but is down-regulated as the full-length APC protein is expressed.

COX-2 expression in the HT-29-APC cells is altered following
Thirty-five micrograms of total protein were loaded per lane.Lane 1, COX-2 standard (10 µg); lanes 2-6, HT-29-APC cells collected at the induction of full-length APC.We performed western analysis indicated times after treatment with zinc (100 µM) and NaBT (5 mM).
of cell lysates both prior to and after treatment with zinc from 0 to 96 h.COX-2 protein was expressed prior to induction Cell growth and apoptosis and subsequently decreased at early time points following expression of full-length APC (Figure 3).By 48 h following The HT-29-APC cells do not reach confluence in the presence of induction, even after long periods of culture, while zinc has treatment with zinc, COX-2 protein was no longer observed.Furthermore, treatment of wild-type HT-29 cells with zinc has no apparent effect on cell density of the HT-29-βgal cells (14).The HT-29-APC cells gradually round up and detach.Upon no effect on COX-2 expression (data not shown) demonstrating that zinc itself is not the cause of the loss of COX-2.HCT-expressing full-length APC, HT-29-APC cells also undergo apoptosis as determined by cell counting, TUNEL assay and 116 cells, which do not express any detectable COX-2 protein, were used as negative control.Actin expression was used to Hoechst staining (14).The proportion of TUNEL-positive cells increased 10-fold to 3% positive cells after induction compared demonstrate similar protein loading.Thus, it appears that as full-length APC is expressed, COX-2 protein is down-with uninduced cells (14).We tested and confirmed that following induction of full-length APC, increased apoptosis regulated.
To test whether COX-2 down-regulation is due merely to does occur over a time course from 0 to 96 h (data not shown).Furthermore, we examined whether differentiation occurs fol-induction of apoptosis, we next treated the HT-29-APC cells with NaBT, a known inducer of apoptosis.NaBT-dependent lowing expression of full-length APC as determined by alkaline phosphatase assay.

1 .
Cells were lysed in MEBC lysis buffer [50 mM Tris-HCl (pH 7.5), 100 mM Fig. Expression of full-length APC in HT-29 cells.Western analysis of NaCl, 0.5% Nonidet P-40] containing protease inhibitors [0.2% (w/v) 4-(2cell lysates demonstrating HT-29-APC cells inducible expression of fullamino-ethyl)benzenesulfonylfluoride (Calbiochem) and 0.01 mg/ml each of length APC (F.L.-APC) when treated for the indicated times with zinc chymostatin, leupeptin, antipain and pepstatin A (all from Sigma)] and (100 µM) and endogenous mutant APC (MT.-APC).The data shown phosphatase inhibitors (0.1 mM Na 3 VO 4 , 50 mM NaF).The samples were represent one of three separate experiments with similar results.Forty equalized for the amount of protein.Samples were pre-cleared with normal micrograms of total protein were loaded per lane.Lanes 1 and 8, HCT-116 mouse or rabbit serum (Accurate Chemical and Scientific Corp.).The samples cells; lanes 2-4, HT-29-APC cells collected at the indicated times, were then immunoprecipitated with either protein A-Sepharose (Pharmacia respectively, after treatment with zinc (100 µM); lanes 5-7, HT-29-βgal Biotech) or protein G-Sepharose (Santa Cruz) beads with the appropriate cells collected at the indicated times after treatment with zinc (100 µM).antibody.The following antibodies were used: mouse monoclonal FE-9 and IE1 (Oncogene Science) for the N-and C-termini of APC, respectively, or rabbit polyclonal APC (Santa Cruz).

Fig. 5 .
Fig. 5. Expression of COX-2 mRNA in HT-29-APC cells.Northern blot of COX-2 (4.6 kb) and G3PDH.The data shown represent one of three separate experiments with similar results.Each lane contains 20 µg of total isolated RNA.G3PDH was used as a control for the amount of RNA Fig. 3. Expression of COX-2 in HT-29-APC cells.Western analysis of loaded.Lanes 1-5, RNA isolated from HT-29-APC cells collected at the HT-29-APC cells treated with zinc (100 µM) and harvested at the times indicated times after treatment with zinc (100 µM); lane 6, RNA isolated indicated.Top, COX-2; bottom, actin.The data shown represent one of four from HT-29 cells treated with TPA (50 ng/ml) for 24 h as a positive control separate experiments with similar results.Thirty micrograms of total protein for COX-2.Dositometry of the bands was performed and expressed as the loaded per lane.Lane 1, COX-2 standard (10 µg); lanes 2-6, HT-29-APC ratio COX-2/G3PDH.cells collected at the indicated times after treatment with zinc (100 µM); lane 7, HCT-116 cells.

Fig. 4 .
Fig. 4. Expression of COX-2 in NaBT-treated HT-29-APC cells.Western analysis of HT-29-APC cells treated with zinc (100 µM) and NaBT (5 mM) Alkaline phosphatase activity is a marker induced apoptosis and APC-dependent apoptosis were similar in extent and time course (data not shown).Interestingly, not been extensively investigated.Kutchera et al. (26) report that HCT-116 cells have increased mRNA for COX-2 as a NaBT did not cause down-regulation of COX-2 expression (Figure 4).Actin expression was used to demonstrate similar result of constitutive transcription controlled by the COX-2 promoter.They also transfected their HCT-116 cells to protein loading.APC appears to uniquely regulate COX-2 expression.overexpress the APC gene and found no change in the increased level of COX-2 mRNA.COX-2 protein levels, however, were COX-2 mRNA expression not examined in this study.Kutchera et al. (26) do warn that We next measured COX-2 mRNA levels both prior to and HCT-116 cells may represent a minority of colon carcinomas following expression of full-length APC in the HT-29-APC as it has an intact APC gene.In HT-29 cells, the APC gene is cells from 0 to 96 h.The COX-2 mRNA levels remained mutated and expresses truncated APC protein, similar to that relatively the same despite induction of full-length APC (Figure observed in most colorectal carcinomas.Our results suggest 5).G3PDH was used as a control for the amount of mRNA that COX-2 mRNA is constitutively expressed in HT-29 cells loaded.The ratio of COX-2 to G3PDH was unchanged as Kutchera et al. observed in HCT-116 cells (26), but downindicating the level of COX-2 mRNA expression was not regulation of the β-catenin pathway by full-length APC protein altered as full-length APC protein was expressed.Thus, the disrupts translation of the COX-2 mRNA to protein.transcriptional control of COX-2 is not effected by expression Our findings in HT-29-APC cells indicate that as full-length of full-length APC.APC is expressed, COX-2 protein is down-regulated while COX-2 mRNA levels remain unchanged.Translational regula-Discussion tion of COX-2 is defective in this human colon cancer cell line.The mechanism by which COX-2 expression is lost is Previous studies on the regulation of COX-2 expression have indicated that COX-2 is regulated at the transcriptional level.currently under investigation.It appears that APC uniquely regulates, either directly or indirectly, COX-2 expression.This Many growth factors, cytokines and inflammatory agents all appear to enhance the expression of COX-2 by interacting phenomenon is likely to occur in other human colon cancer cell lines which have a mutant APC gene.with various regulatory sequences in the promoter region of this gene (3).However, the regulation of COX-2 expression in colorectal tissue and colorectal cells in culture as a result