Search

We have recently demonstrated that a high c-myc endogenous amplification level confers an apoptosis-prone phenotype to serum-deprived colon carcinoma SW613-S cells. The aim of this study was to gain new insights into the features of c-myc-dependent apoptosis, by extending our analysis to different apoptogenic stimuli. The study was carried out on clones, derived from the human colon carcinoma SW613-S cell line, which harbor different levels of endogenous c-myc amplification, and on isogenic cell lines with an enforced c-myc overexpression. Our results indicate that cells with endogenous or transfected exogenous c-myc overexpression (SW613-12A1 and -2G1mycP2Tu1 cell lines, respectively), activate the apoptotic machinery in response to the treatment with etoposide, doxorubicin and vitamin D3, which induce apoptosis through the death receptor Fas. The low levels of c-myc expression present in SW613-B3 and -B3mycC5, seem to be unable to activate Fas-mediated apoptosis, thus suggesting that only a high c-myc expression can bypass the lack of Fas receptor. Apoptosis induction mediated by DNA damage and long-term culture was independent of c-myc expression. A pathway of apoptosis characterized by the activation of the enzyme L-DNase II, was observed in both 12A1 and B3 cell lines.

We are investigating the mechanism responsible for the overexpression of the keratin 18 (K18) gene in tumorigenic clones from the SW613-S human colon carcinoma cell line, as compared with non-tumorigenic clones. We have previously shown that this mechanism affects the minimal K18 promoter (TATA box and initiation site). We report here that treatment of the cells with histone deacetylase inhibitors stimulates the activity of the promoter in non-tumorigenic cells but has no effect in tumorigenic cells, resulting in a comparable activity of the promoter in both cell types. The adenovirus E1A protein inhibits the activity of the K18 promoter specifically in tumorigenic cells. This inhibition can be reversed by an excess of CBP protein. The conserved region 1 (CR1) of E1A, which is involved in the interaction with the CBP/p300 co-activators, is necessary to the inhibitory capacity of E1A. A 79 amino acid long N-terminal fragment of E1A, encompassing the two domains of E1A necessary and sufficient for binding to CBP (N-terminus and CR1), has the same differential inhibitory capacity on the K18 promoter as wild-type E1A. Forced recruitment of GAL4-CBP fusion proteins to the K18 promoter results in a greater stimulation of its activity in non-tumorigenic than in tumorigenic cells. The histone acetyltransferase activity of CBP is essential for this differential stimulation and the presence of the CBP2 domain greatly augments the activation capacity of the fusion protein. Chromatin immunoprecipitation experiments carried out with anti-acetylated histone antibodies showed no difference in the level of histone acetylation in the region of the K18 promoter between the two cell types. The structure of chromatin in the promoter region is similar in tumorigenic and non-tumorigenic cells, as determined by mapping of DNase I hypersensitive sites and probing the accessibility of the DNA to restriction endonucleases. From all these results we conclude that alteration of an acetylation mechanism involving the CBP (or p300) protein and acting on a non-histone substrate is responsible for the higher activity of the K18 promoter in tumorigenic cells of the SW613-S cell line.

The keratin 18 (K18) gene is expressed at a normal level in cells of nontumorigenic clones derived from the SW613-S human colon carcinoma cell line, but is overexpressed in cells of tumorigenic clones. A high level of expression was also found in the cells from 10 of 15 other human colon carcinoma cell lines. The expression of the gene is downregulated in differentiating Caco-2 cells, resulting in a normal expression level. Determination of K18 mRNA half-life in growing and confluent Caco-2 cells indicated that this downregulation does not take place at a posttranscriptional level. The density of RNA polymerase molecules on the K18 gene, as measured in nuclear run-on experiments, is the same in growing and confluent Caco-2 cells, but the rate of synthesis of K18 transcripts in confluent Caco-2 cells, as determined by in vivo pulse-labeling, is 35% of that in growing cells. Nuclear run-on experiments carried out with nuclei prepared from growing or confluent Caco-2 cells treated with 5,6-dichloro-1-beta-D-ribofuranosyl benzimidazole (DRB) indicated that a reduction in both the initiation and elongation rates of RNA polymerase molecules occurs on the K18 gene in confluent Caco-2 cells. This leads to a decreased rate of K18 transcript production with no reduction in the polymerase density on the gene. Evidence is provided that the mechanisms responsible for the differential expression of the K18 gene between tumorigenic and nontumorigenic SW613-S cells and between growing and differentiating Caco-2 cells share some similarities.

Although apoptosis can be induced by the enforced expression of exogenously introduced c-myc genes, it is not clear whether overexpression resulting from the amplification of the resident c-myc gene in tumor cells is sufficient to induce apoptosis. We have investigated the relationship between c-myc gene amplification and the propensity of tumor cells to undergo apoptosis, using the SW613-12A1 and SW613-B3 cell lines, which are representatives, respectively, of tumorigenic and non-tumorigenic clones isolated from the SW613-S human colon carcinoma cell line. Tumorigenic clones are characterized by a high level of amplification and expression of the c-myc gene, whereas cells of non-tumorigenic clones have a small number of copies and a lower level of expression of this gene. Analysis of c-myc mRNA level in cells cultured under low serum conditions indicated that the expression of the gene is tightly regulated by serum growth factors in non-tumorigenic B3 cells, whereas it is poorly regulated in tumorigenic 12A1 cells, the level of mRNAs remaining relatively high in serum-starved 12A1 cells. Under these conditions, 12A1 cells showed clear evidence of apoptosis, whereas B3 cells were completely refractory to the induction of apoptosis. Moreover, the study of cell lines derived from non-tumorigenic apoptosis-resistant clones following the introduction by transfection of exogenous c-myc gene copies showed that they have acquired an apoptosisprone phenotype. Altogether, our results strongly suggest that deregulated c-myc expression due to high-level amplification confers an apoptosis-prone phenotype to tumor cells. The possible consequences of these observations for cancer therapy are discussed.

The human colon carcinoma cell line, SW613-S, is composed of cells with a high-level amplification of the MYC proto-oncogene that are tumorigenic in nude mice and of cells with a low-level amplification of MYC that are not tumorigenic. Transcripts from FGF-3, a member of the fibroblast growth factor gene family, accumulate in cells from tumorigenic clones, but are undetectable in those from non-tumorigenic clones. Nuclear run-on analyses indicate that this differential FGF-3 expression is regulated at the level of transcription initiation. Determination of the structure of the FGF-3 transcripts indicates that they are generated by splicing of the three exons and termination at the single polyadenylation site predicted from the genomic sequence. Their size heterogeneity is due to multiple initiation sites spanning a 700 base-pair long promoter region. FGF-3 is activated in tumors induced in nude mice by MYC-transfected cells from non-tumorigenic clones. However, in most of the cell lines established from these tumors, FGF-3 expression tends to be lost upon in vitro propagation. Thus, in these transfectant cell lines, the presence of exogenous MYC gene copies is not sufficient to activate FGF-3 expression and in vivo growth is also required.

Clonal analysis has shown that the SW613-S human colon-carcinoma cell line is heterogeneous: some cell clones display a high level of amplification of the c-myc gene and are tumorigenic in nude mice, whereas others have a small number of copies of this gene and are non-tumorigenic. Tumorigenic clones can proliferate in a chemically defined serum-free medium, whereas non-tumorigenic clones cannot. Suramin, like anti-insulin-like growth factor (IGF) or anti-IGF-I receptor antibodies, efficiently inhibits the growth of tumorigenic clones in defined medium. Inhibition by suramin or by anti-IGF antibodies can be reversed by pure IGF-I or IGF-2. Pure IGF-1 or IGF-2 or culture medium conditioned by tumorigenic clones can stimulate DNA synthesis in cells of non-tumorigenic clones. Co-culture with cells of tumorigenic clones sustains the growth of non-tumorigenic clones in defined medium. Cells of both tumorigenic and non-tumorigenic clones express high-affinity IGF-1 receptors at their surface but tumorigenic clones produce on average 5 times more IGF-1 and 25 times more IGF-2 than non-tumorigenic ones. These results indicate that autocrine growth stimulation of tumorigenic clones by IGFs through the IGF-1 receptor is essential for their ability to grow in defined medium. Since cells of tumorigenic clones produce IGF-2 at levels 80 times higher than IGF-1 and since an antibody strictly specific for IGF-1 has no effect on DNA synthesis in cells of tumorigenic clones grown in defined medium, IGF-2 is very likely the main effector in the autocrine loop.

Two types of clones have been isolated from the SW613-S human colon carcinoma cell line. Clones with a high level of amplification of the c-myc gene are tumorigenic in nude mice and can proliferate in chemically defined, serum-free medium, whereas clones with a low level of amplification are nontumorigenic and cannot multiply in defined medium. The expression level of the insulin-like growth factor type 1 (IGF-1) gene is low in tumorigenic clones and undetectable in nontumorigenic clones. Tumorigenic clones produce high levels of IGF-2 (and IGF-binding proteins), compared to nontumorigenic clones. This is the consequence of a differential transcriptional regulation of the IGF-2 gene between the two types of clones. This regulation consists of a modulation of the activity of promoters P3 and P4. Overexpression of the IGF-2 gene is constitutive in tumorigenic clones: it is stably maintained during in vitro propagation of the cells. Tumorigenic cell lines obtained after transfer of c-myc gene copies into the cells of nontumorigenic clones exhibit a high level of expression of the IGF-2 gene when they are grown in vivo, as subcutaneous tumors in nude mice. This high level of expression is lost in most of these cell lines when they are returned to in vitro culture conditions indicating that, in these cells, IGF-2 overexpression is not constitutive but inducible by in vivo growth conditions. We had previously shown that tumorigenic clones use the overproduced IGF-2 as an autocrine growth factor. The results reported here suggest than IGF-2 overexpression has an important role in the tumorigenic phenotype of these cells.

The c-myc gene is amplified and the c-Ki-ras gene is mutated in the SW613-S human colon carcinoma cell line. Two cell types with different levels of c-myc amplification are present in the SW613-S cell population and representative cell clones can be isolated. The clones with a high level of amplification and expression of the c-myc gene are tumorigenic in nude mice whereas those with a low level are not. The tumorigenic clones secrete transforming growth factor alpha (TGF-alpha) in the culture medium whereas the non-tumorigenic clones do not produce any detectable amount. Accordingly the level of TGF-alpha mRNA is higher and the transcription rate of the gene is increased in the tumorigenic clones. The acquisition of the tumorigenic phenotype by cells of non-tumorigenic clones, following introduction of c-myc gene copies by transfection, is accompanied by an increase in the steady-state level of TGF-alpha mRNA. These findings suggest a role for an elevated level of TGF-alpha production in the tumorigenic phenotype of SW613-S cells. The possibility that this role is indirect is discussed.