Proliferation rate but not mismatch repair affects the long-term response of colon carcinoma cells to 5FU treatment.

The role of mismatch repair (MMR) in the response of colon carcinoma cells to 5-fluorouracil (5FU) is not well understood. In most of the in vitro studies only short-term response was investigated. We focussed here on the influence of MMR status on the mechanism of the short- and long-term response to clinically relevant 5FU concentrations by using isogenic or semiisogenic cell line pairs expressing/nonexpressing the hMLH1 protein, an important component of the MMR system. We show that the lower survival of MMR-proficient than of MMR-deficient cells in the clonogenic survival assay is due to a more frequent early cell arrest and to subsequent senescence. By contrast, the long-term cell growth after treatment, which is also affected by long-term arrest and senescence, is independent from the MMR status. The overall effect on the long-term cell growth is a cumulative result of cell proliferation rate-dependent growth inhibition, apoptosis and necrotic cell death. The main long-term cytotoxic effect of 5FU is the inhibition of growth while apoptosis and the necrotic cell death are minor contributions.

UDCA slows down intestinal cell proliferation by inducing high and sustained ERK phosphorylation.

Ursodeoxycholic acid (UDCA) attenuates colon carcinogenesis in humans and in animal models by an unknown mechanism. We investigated UDCA effects on normal intestinal epithelium in vivo and in vitro to identify the potential chemopreventive mechanism. Feeding of mice with 0.4% UDCA reduced cell proliferation to 50% and suppressed several potential proproliferatory genes including insulin receptor substrate 1 (Irs-1). A similar transcriptional response was observed in the rat intestinal cell line IEC-6 which was then used as an in vitro model. UDCA slowed down the proliferation of IEC-6 cells and induced sustained hyperphosphorylation of ERK1/ERK2 kinases which completely inhibited the proproliferatory effects of EGF and IGF-1. The hyperphosphorylation of ERK1 led to a transcriptional suppression of the Irs-1 gene. Both, the hyperphosphorylation of ERK as well as the suppression of Irs-1 were sufficient to inhibit proliferation of IEC-6 cells. ERK1/ERK2 inhibition in vitro or ERK1 elimination in vitro or in vivo abrogated the antiproliferatory effects of UDCA. We show that UDCA inhibits proliferation of nontransformed intestinal epithelial cells by inducing a sustained hyperphosphorylation of ERK1 kinase which slows down the cell cycle and reduces expression of Irs-1 protein. These data extend our understanding of the physiological and potentially chemopreventive effects of UDCA and identify new targets for chemoprevention.

Equivalent effect of DNA damage-induced apoptotic cell death or long-term cell cycle arrest on colon carcinoma cell proliferation and tumour growth.

Knowledge of the type of biological reaction to chemotherapy is a prerequisite for its rational enhancement. We previously showed that irinotecan-induced DNA damage triggers in the HCT116p53(wt) colon carcinoma cell line a long-term cell cycle arrest and in HCT116p53(-/-) cells apoptosis (Magrini et al., 2002). To compare the contribution of long-term cell cycle arrest and that of apoptosis to inhibition of cell proliferation after irinotecan-induced DNA damage, we used this isogenic system as well as the cell lines LS174T (p53(wt)) and HT-29 (p53(mut)). Both p53(wt) cell lines responded to damage by undergoing a long-term tetraploid G1 arrest, whereas the p53(mut) cell lines underwent apoptosis. Cell cycle arrest as well as apoptosis caused a similar delay in cell proliferation. Irinotecan treatment also induced in mouse tumours derived from the p53(wt) cell lines a tetraploid G1 arrest and in those derived from the p53-deficient cell lines a transient G2/M arrest and apoptosis. The delay of tumour growth was in the same range in both groups, that is, arrest- and apoptosis-mediated tumour growth inhibition was comparable. In conclusion, cell cycle arrest as well as apoptosis may be equipotent mechanisms mediating the chemotherapeutic effects of irinotecan.

De novo expression of the Muc2 gene in pancreas carcinoma cells is triggered by promoter demethylation.

It has been established that mucin-producing variants of different subtypes of pancreatic carcinomas, including the intraductal papillary and ductal mucinous tumors, have usually a more favorable prognosis. Intraductal papillary and ductal mucinous tumors have also been shown to ectopically express the intestinal mucin gene MUC2. The mechanism of the de novo expression of this gene in tumors may have potential implications for the modulation of its behavior. We studied, therefore, the mechanism of the de novo expression of MUC2 in pancreas carcinoma cells in vitro. The MUC2 gene promoter is methylated in the nonexpressing pancreatic cell line PANC-1 and is not methylated in the expressing cell line BxPC-3. The promoter is silenced by methylation as shown by reporter expression assays. De novo expression of MUC2 in PANC-1 cells is triggered by treating the cells with a pharmacological inhibitor of DNA methylation (5-aza-2'-deoxycytidine). There was no decrease or loss of expression of the methyltransferase DNMT1 in the MUC2-producing cells. These data show that the de novo expression of the MUC2 gene in pancreas carcinoma cells is associated with promoter demethylation. They warrant further investigations on the relationship between MUC2 promoter demethylation in pancreatic cancer and the prognosis of carcinoma patients.

Regulation of the intestinal mucin MUC2 gene expression in vivo: evidence for the role of promoter methylation.

In the present work we investigated the in vivo regulation of the mucin gene MUC2, which is overexpressed in all mucinous colorectal carcinomas. The inhibition of methylation by 5-azadeoxycytidine induces de novo expression of MUC2 in the colon carcinoma cell line COLO 205. The expression is retained in xenograft tissue and the cells give rise to MUC2-expressing tumours in nude mice. The strong expression of MUC2 in the normal human goblet cells and in the tissue of human mucinous colorectal carcinomas is associated with the average methylation of about 50% at every investigated CpG site of the MUC2 promoter. In contrast, MUC2 promoter in the non-expressing normal columnar cells and in the non-mucinous carcinoma tissue is methylated to nearly 100%. These data show that (i) low methylation of MUC2 promoter is associated with MUC2 expression in vivo and (ii) the pattern of MUC2 promoter methylation in the normal goblet or columnar cells most closely resembles that in mucinous or non-mucinous colorectal carcinomas, respectively. They indicate that MUC2 expression in vivo is regulated by promoter methylation and support the hypothesis that cells with goblet-like differentiation give rise to mucinous colonic carcinomas.

Novel adapter protein AP162 connects a sialyl-Le(x)-positive mucin with an apoptotic signal transduction pathway.

Glycoproteins modified with a sialyl-Le(x)-moiety are important sensors for extracellular signals regulating cellular recognition, adhesion and migration. The transduction pathways and signals mediated by these glycoproteins within the cell are largely unknown. In search of novel glycoproteins modified with sialyl-Le(x)-moiety, we screened a human colonic cDNA expression library with a rabbit antiserum produced against sialyl-Le(x)-positive mucins. The antiserum detected a new protein, named B2, which was cloned and characterised in detail. The analysis of the B2 gene revealed a 5.7 kb RNA transcript detectable in all investigated tissues and a complete coding sequence of 2778 bp. The B2 protein exhibited two putative PH (pleckstrin homology) domains and a leucine zipper motif but no homology to any known proteins. Monospecific antibodies against the B2-protein precipitated from the solubilised membrane fraction of the colon carcinoma cell line LS 174T a protein with an apparent Mr = 162 kDa and, additionally, a mucin-like glycoprotein with an apparent Mr = 220 kDa. Protein fractionation on a CsCl gradient, Western blots and sandwich ELISA showed that the 220 kDa mucin carries the sialyl-Le(x) moiety and is tightly bound to the 162 kDa protein. The expression of the recombinant B2-protein enhanced staurosporine-induced apoptosis in epithelial cancer cell lines. These data indicate that B2 is a novel, ubiquitously expressed protein with a putative adapter function. The protein has been named AP162 (adapter protein 162). In colon carcinoma cells B2-protein is tightly associated with a sialyl-Le(x)-positive mucin and has a potential for involvement in sialyl-Le(x)-mediated transduction of apoptotic signals.

Differential gene expression in colon carcinoma cells and tissues detected with a cDNA array.

Expression of selected genes coding for proteins with defined cellular functions was analysed in human cell lines derived from normal colonic mucosa, non-mucinous colonic carcinomas and mucinous colonic carcinomas. Altered expression of 10 genes in colon carcinoma cells was found by using a cDNA array; 6 of these alterations (60%) were confirmed by Northern blotting or semi-quantitative reverse transcription-polymerase chain reaction (RT-PCR). Among these 6 genes, 3 transcription factors as well as the topoisomerase II alpha and the mitosis inhibitor WEE1Hu gene were significantly suppressed in the tumour cell lines. In addition, the gene coding for the cell cycle inhibitor p21 was overexpressed only in cell lines derived from mucinous carcinomas. The significant suppression of the kinase WEE1Hu gene in carcinoma cells of both phenotypes and the tendency of the mucinous phenotype to overexpress p21 protein were confirmed in human colon carcinoma tissues. Our data show that the cDNA array method permits a correct identification of changes in gene expression with a relatively high accuracy. The different expression of the p21 gene in the non-mucinous and mucinous carcinoma cells supports the hypothesis that these phenotypes may develop along different genetic pathways. The detection of WEE1Hu gene suppression in colon carcinoma cells and tissues suggests its potential role in tumourigenesis.

FasL is more frequently expressed in liver metastases of colorectal cancer than in matched primary carcinomas.

Colorectal carcinoma cells have recently been shown to express Fas ligand (FasL). This ligand could allow the tumour cells to evade activated tumour-infiltrating lymphocytes (TILs) by inducing their apoptosis and would thus promote tumour survival and possibly metastasis formation. To test this hypothesis in vivo we analysed the expression of FasL mRNA and protein in paired tissue samples of normal colonic mucosa (N), primary colorectal carcinomas (T) and their metastases (M) from a total of 21 patients by four different methods. Additionally, the presence and activation status of infiltrating lymphocytes, which might contribute to the total amount of FasL in the tissue, was determined by semiquantitative reverse transcription-polymerase chain reaction (RT-PCR) in the same samples. The frequency of FasL detection was 30-40% in T and was 60-100% in M, depending on the sensitivity of the method. Simultaneously, the amount of CD25 mRNA, used as a measure of the number of activated TILs, was in 90% of patients lower in M than in T. The increased frequency of FasL detection in liver metastases was therefore not due to the presence of activated TILs. We conclude that metastasizing subpopulations of colorectal tumour cells express FasL more frequently than the primary carcinomas and may be able to eliminate activated TILs in vivo via Fas/FasL-induced apoptosis or other hitherto unknown mechanisms.

Target genes of beta-catenin-T cell-factor/lymphoid-enhancer-factor signaling in human colorectal carcinomas.

Mutations in the adenomatous polyposis coli or beta-catenin gene lead to cytosolic accumulation of beta-catenin and, subsequently, to increased transcriptional activity of the beta-catenin-T cell-factor/lymphoid-enhancer-factor complex. This process seems to play an essential role in the development of most colorectal carcinomas. To identify genes activated by beta-catenin overexpression, we used colorectal cell lines for transfection with the beta-catenin gene and searched for genes differentially expressed in the transfectants. There are four genes affected by beta-catenin overexpression; three overexpressed genes code for two components of the AP-1 transcription complex, c-jun and fra-1, and for the urokinase-type plasminogen activator receptor (uPAR), whose transcription is activated by AP-1. The direct interaction of the beta-catenin-T cell-factor/lymphoid-enhancer-factor complex with the promoter region of c-jun and fra-1 was shown in a gel shift assay. The concomitant increase in beta-catenin expression and the amount of uPAR was confirmed in primary colon carcinomas and their liver metastases at both the mRNA and the protein levels. High expression of beta-catenin in transfectants, as well as in additionally analyzed colorectal cell lines, was associated with decreased expression of ZO-1, which is involved in epithelial polarization. Thus, accumulation of beta-catenin indirectly affects the expression of uPAR in vitro and in vivo. Together with the other alterations, beta-catenin accumulation may contribute to the development and progression of colon carcinoma both by dedifferentiation and through proteolytic activity.

A modification of the JAM test is necessary for a correct determination of apoptosis induced by FasL+ adherent tumor cells.

Tumor cells from several organs including colon have recently been shown to express Fas ligand (FasL) in vitro and in vivo. The expression, which in some tumours occurs de novo, was suggested to facilitate immune escape of malignant cells by killing tumor-infiltrating lymphocytes via Fas-FasL-induced apoptosis. An argument to support this hypothesis is the detection of tumor cell-induced apoptosis in Jurkat cells (as model T cells) by means of the widely used JAM test. In the present work the validity of this test for the analysis of colon carcinoma cell-mediated apoptosis in Jurkat cells was scrutinized in detail. The presented data show that the JAM test as described previously is prone to false-positive detection of apoptosis, when adherent epithelial cells are used as effectors. Furthermore, three lines of evidence indicated that several FasL+ colon carcinoma cell lines did not induce detectable apoptosis in Jurkat cells in vitro. We conclude that: (1) The JAM test must be modified for testing DNA fragmentation induced through adherent effector cells and (2) FasL+ colon carcinoma cells may be unable to induce apoptosis in vitro.

Identification of HLA-A2-restricted epitopes of the tumor-associated antigen MUC2 recognized by human cytotoxic T cells.

Previous studies have shown that self-antigens overexpressed in malignant tissue can provide a basis for a tumor-specific immune response. The mucin MUC2 is strongly overexpressed in all mucinous tumors of colon, breast, ovary and pancreas. In the corresponding normal tissue it is either not expressed (breast, ovary, pancreas) or it is expressed at considerably lower levels than in the mucinous tumors (colon). We therefore investigated whether the MUC2 molecule comprises HLA-A2-binding epitopes recognized by human cytotoxic T cells. Four MUC2 peptides with high affinity and stable binding to HLA-A2 were identified. Those peptides and additionally 3 peptides with moderate binding to HLA-A2 were loaded onto dendritic cells, which were used for stimulation of autologous T cells from healthy donors. Two MUC2 peptides, which belonged to the group of stable binders, induced specific cytotoxic T-cell lines. Target cells loaded with these peptides were strongly lysed in a concentration-dependent and HLA-A2-restricted manner. Our data show that the tumor-associated mucin MUC2 has potential as a target antigen for cytotoxic T cells in patients with mucinous carcinomas.

Overexpression or ectopic expression of MUC2 is the common property of mucinous carcinomas of the colon, pancreas, breast, and ovary.

Mucinous carcinomas of the colorectum have been reported to overexpress the intestinal mucin MUC2. The purpose of this study was to determine whether this alteration is shared by mucinous tumours of the ovary, breast, and pancreas. A total of 40 breast carcinomas (22 of mucinous and 18 of ductal invasive type), 39 ovarian adenocarcinomas (16 mucinous, 23 serous), 47 colorectal carcinomas (25 mucinous and 22 non-mucinous), and 41 pancreatic adenocarcinomas (14 mucinous, 27 non-mucinous) were investigated by immunohistochemistry with the anti-MUC2 monoclonal antibody 4F1 and the expression pattern was ranked. MUC2 mucin is expressed in the normal colonic epithelium; in the normal epithelium of the breast, ovary, and pancreas, it was not detectable by immunohistochemistry or by reverse transcriptase-polymerase chain reaction (RT-PCR). In agreement with previous reports, the colonic mucinous carcinomas differed significantly from the non-mucinous carcinomas by strong MUC2 expression. In all mucinous carcinomas of the ovary, breast, and pancreas, de novo expression of the MUC2 gene was observed, which differentiated mucinous and non-mucinous carcinomas of these tissues (P < 0.001). The overexpression or ectopic expression of the MUC2 gene exhibited by mucinous carcinomas of four organs indicates a common genetic lesion associated with the mucinous tumour phenotype.

Low O-acetylation of sialyl-Le(x) contributes to its overexpression in colon carcinoma metastases.

Two factors potentially determining the consistent overexpression of sialyl-Le(x) antigen in colon carcinoma and metastases were investigated: (i) the expression of the mucins MUC1 and MUC2, known to carry sialyl-Le(x), by Northern blotting; (ii) the extent of sialic acid O-acetylation, by Western blotting and HPLC. RNA and sialyl-Le(x)-positive mucins were purified from normal colonic mucosa (N), primary carcinomas (T) and their liver metastases (M). Northern blots showed that mRNA expression both of MUC1 and of MUC2 decreases during the progression of the disease, and is lowest in metastatic tissue. The expression of mucin-bound sialyl-Le(x) increased strongly from N to T and, to a lesser extent, to M. After alkali treatment of the mucins these differences disappeared, indicating that the total amount of mucin-bound sialyl-Le(x) is the same in the 3 types of tissues. The O-acetylation of mucin-bound sialyl-Le(x) gradually decreased from N to M. HPLC analysis showed that in N about 70%, in T 45% and in M only 20% of mucin-bound sialic acids are O-acetylated. Thus, the increase of sialyl-Le(x) detectable during colon-carcinoma progression is due to diminished O-acetylation and not to increased expression of mucin protein cores. The decrease of O-acetylation is therefore the primary chemical alteration contributing to colon carcinoma-associated overexpression of sialyl-Le(x).

Fucosyltransferase III and sialyl-Le(x) expression correlate in cultured colon carcinoma cells but not in colon carcinoma tissue.

The potential contribution of fucosyltransferases to the overexpression of sialyl-Le(x) antigen was investigated in the colon carcinoma cell line HT-29 and in human colon carcinoma tissue. In HT-29 cells as well as in normal or malignant colonic tissues Fuc-TIII, Fuc-TIV, Fuc-TVI but not Fuc-TV nor Fuc-TVII were detectable after RT-PCR. Sodium butyrate treatment of HT-29 cells increased (to about 200%) and DMSO treatment decreased (to about 20%) the expression of sialyl-Le(x). This modulation of sialyl-Le(x) was concomitant with the analogous increase/decrease of mRNA of Fuc-TIII but not Fuc-TIV. Fuc-TVI was not detectable by Northern blotting in HT-29 cells. In six human colon carcinomas which exhibited strong overexpression of sialyl-Le(x), the expression of Fuc-TIII-mRNA was the same or lower than in the corresponding normal colonic tissue. Thus Fuc-TIII expression may be affecting the expression of the sialyl-Le(x) moiety in HT-29 cells but not in human colon carcinoma tissue.

Low frequency of p53 gene mutation and protein expression in mucinous colorectal carcinomas.

Immunohistochemical data indicate that the frequency of p53 protein overexpression is consistently lower in the mucinous than in the non-mucinous carcinomas of the breast, ovary, pancreas and colon. This peculiar immunohistochemical behavior of the mucinous phenotype could be due to the effect of large amounts of mucus on the staining or to an actual mutation frequency difference between mucinous and non-mucinous carcinomas. This question was investigated on a group of mucinous colorectal carcinomas. DNA was extracted from paraffin sections of 16 human mucinous colorectal carcinomas and the mutation frequency was determined by sequencing of p53 exons amplified in PCR. The expression of p53 protein was determined with the avidin-biotin complex-peroxidase staining procedure and CM-1 antiserum. Twenty-five percent of the tumors, exhibited p53 protein overexpression and in 31% a mutation was detected. Concordance between the two techniques was found in 69% of tumors. Overexpression without mutation was observed in 12% and mutation without overexpression in 19%. G:C --> A:T transitions represented the most frequent lesion (80%), as previously observed in non-mucinous colorectal carcinomas. These data indicate that the mutation pattern in the p53 gene is similar in mucinous and non-mucinous colorectal carcinomas. The low frequency of p53 overexpression in the mucinous phenotype is not due to a mucus effect on the staining but is related to the low mutation frequency of p53 gene. These results lead to the hypothesis that in contrast to the nonmucinous tumors the development of the majority of colonic carcinomas with the mucinous phenotype may be independent from p53 mutations.

Characterization of the major sialyl-Lex-positive mucins present in colon, colon carcinoma, and sera of patients with colorectal cancer.

The expression of the mucin-bound sialyl-Lewisx epitope is increased in the tissue of most colorectal carcinomas and in the sera of about 30% of tumor patients. In colon cancer, a portion of the sialyl-Lex groups detectable with the monoclonal antibody AM-3 is located on MUC1 (C. Hanski et al., Cancer Res., 53: 4082-4088, 1993). In order to characterize the major colon carcinoma-associated sialyl-Lex-positive glycoprotein components, the tissue- and serum-derived antigens were investigated. The buoyant densities of the sialyl-Lewisx-positive antigens from tumor and normal colonic tissues and from sera of patients with colon carcinoma and healthy donors correspond to that of mucins (1.40 g/ml). The sialyl-Lex-positive mucins purified from both tissues elute under nonreducing conditions in the void volume of a Sepharose CL-2B column, indicating a molecular mass more than 2 x 10(7) daltons. They yield in immunoblot after SDS gel electrophoresis under reducing conditions a main band at an apparent M(r) 880,000. Radioactive labeling revealed that the band at M(r) 880,000 is the major protein component in sialyl-Lewisx-positive mucins both from tumor and normal colonic tissue. In sera of colon carcinoma patients, the sialyl-Lex moiety is also detectable mainly on a M(r) 880,000 glycoprotein band and, additionally, on a M(r) 140,000 molecule as well as on alpha 1-acid glycoprotein. Sera from healthy donors exhibited only a sialyl-Lex-positive glycoprotein with the apparent M(r) 140,000. Sandwich ELISA as well as immunoblots of mucins purified from the colon carcinoma cell line LS174T indicated that the sialyl-Lex moiety migrating in the M(r) 880,000 band is located on MUC2 protein core. Together, these data suggest that sialyl-Lex antigen in colon, colon carcinoma, and the sera of patients with this tumor is located on the MUC2 molecule, consisting of several subunits with an apparent M(r) 880,000, linked via disulfide bridges. The increase of sialyl-Lex expression in colon carcinomas appears to be mainly due to a more frequent transfer of sialyl-Lex moieties onto the mucin core in tumor tissue.

Altered glycosylation of the MUC-1 protein core contributes to the colon carcinoma-associated increase of mucin-bound sialyl-Lewis(x) expression.

The mucin carbohydrate epitope sialyl-Le(x), detected with the monoclonal antibody AM-3, is strongly overexpressed in > 90% of human colon carcinomas. We show here that in colon carcinoma one of the mucin cores bearing the sialyl-Le(x) group is MUC-1, whereas sialyl-Le(x) present in normal colon is not detectable on MUC-1. The amounts of MUC-1 core detectable with the monoclonal antibody BC3 in extracts of tumor tissue are 60-180% of those in normal tissue. Two other carbohydrate epitopes located on MUC-1 in mucins from normal and tumor tissue have also been characterized. In contrast to sialyl-Le(x), their expression on MUC-1 is variable and does not correlate with the malignant transformation of colonic mucosa. The transfer of the sialyl-Le(x) group onto the MUC-1 core contributes to the colon carcinoma-associated overexpression of the sialyl-Le(x) epitope.

Expression of p53 protein in invasive colorectal carcinomas of different histologic types.

BACKGROUND: This study was performed to determine whether morphologic differences of colonic cancer types can be related to different genotypes of these tumors. METHODS: Paraffin sections of 76 human invasive colorectal carcinomas were examined for the overexpression of p53 oncoprotein with the avidin-biotin complex-peroxidase staining procedure and CM-1 antiserum, which detects p53 protein in paraffin-embedded material. The tumors were categorized as mucinous (22 cases), most of which originated from adenomas, and nonmucinous, which were subdivided into carcinomas originating from adenoma-carcinoma sequence (ACS) (29 cases) and de novo (DN) carcinomas (25 cases). RESULTS: Nineteen DN carcinomas (76%), 21 ACS carcinomas (72%), and 8 mucinous carcinomas (36%) exhibited detectable amounts of p53 protein in the tumor cell nuclei. Strong overexpression of p53 protein coincided with a high percentage (> 40%) of stained nuclei in 40% of ACS and 48% of DN carcinomas versus 9% of mucinous tumors. The percentage of stained nuclei, intensity of staining, and distribution of the stained areas did not correlate with the grade of differentiation or the invasive edge of the tumors. Along with nuclear staining of the tumor area, a distinct perinuclear staining of normal epithelial cells adjacent to the tumor was observed in 48% of DN, 7% of ACS, and 9% of mucinous carcinomas. CONCLUSIONS: The current results, in combination with the recently published data on Ki-ras and c-myc alterations, indicate that mucinous carcinomas differ from nonmucinous colorectal carcinomas in their genetic lesions.