Establishment and characterization of a spontaneously immortalized myofibroblast cell line derived from a human liver angiosarcoma.

BACKGROUND/AIM: Fibrosis and/or cirrhosis are present in the precursor stages of most liver cancers. However, little is known about the reciprocal interactions of fibroblasts, mainly responsible for fibrosis, and the other liver cells. We report here the isolation of a new liver myofibroblast cell line from a human liver angiosarcoma and its characterization. METHODS: The cells were isolated by the explant technique and characterization was performed, on one hand, using immunohistochemical and ultrastructural analysis and, in the other hand, by determining their karyotype, ras and p53 status and their tumorigenic properties. RESULTS: To date, the cells have undergone approximately 170 population doublings and are still proliferating. Immunohistochemically, they were negative for desmin, smooth muscle myosin, cytokeratin 19 and von Willebrand factor, positive for vimentin and alpha-smooth muscle actin, with an important deposition of fibronectin around the cells. Ultrastructure showed particularly cytoplasmic microfilament bundles. Their chromosome number ranged from 38 to 168 with a bimodal population, near diploid and hypotetraploid. No mutations were found in codons 12, 13 or 61 of Ha-, Ki- and N-ras genes but a homozygous missense mutation in codon 179 (CAT-->CTT) was detected in the p53 gene. They were unable to form foci in soft agar or tumors in nude mice. CONCLUSIONS: Taken together, these results show that these cells, called BM 2.2.1, exhibited typical myofibroblast-like features. Although they contained a karyotype suggestive of tumoral cells and a homozygous mutated p53 gene, they were not tumorigenic. The nature of these cells and the abnormalities of the p53 gene and the karyotype, suggest that: i) they were a component of the tumor stroma, and ii) they could have been involved in angiosarcoma development. Thus, this cell line may be valuable for the study of cellular interactions in liver carcinogenesis.

Ras gene mutations in vinyl chloride-induced liver tumours are carcinogen-specific but vary with cell type and species.

Previous studies have shown that a high proportion (5/6) of human liver angiosarcomas (ASL) associated with exposure to vinyl chloride (VC) contains a GC-->AT mutation at the Ki-ras codon 13. This mutation, however, has not been found in 5 ASL or 2 hepatocellular carcinomas (HCC) induced in rats by VC. These 2 HCC did contain a mutation at codon 61 of the Ha-ras gene. In order to extend this study and further explore the mechanisms of tumour induction, an additional 6 ASL and 6 HCC induced in rats by VC were analysed for ras gene point mutations, as well as 10 rat and 10 murine ASL induced by vinyl fluoride (VF), and 5 ASL, 6 Kupffer cell sarcomas, 4 HCC and 2 cholangiocellular carcinomas induced by Thorotrast in rats. Tumour DNA was analysed by PCR-SSCP and direct sequencing. None of the rodent ASL contained a mutation at codon 13 of the Ki-ras gene showing that the ras gene mutational pattern is species-specific. The CAA-->CTA mutation, previously found at codon 61 of the Ha-ras gene in rat HCC, was observed in 5 further VC-induced HCC but was not detected in the Thorotrast-induced HCC, suggesting carcinogen-specificity. This mutation was also absent in VC-induced ASL, which supports the cell-specificity of the ras mutational pattern in chemically induced tumours. No predominant mutation was detected in VF- and Thorotrast-induced tumours. Thus, a given mutation in a tumour may be carcinogen-specific but also depend on the species and the cell type.

Vinyl chloride-specific mutations in humans and animals.

Vinyl chloride is a potent hepatocarcinogen which reacts with DNA to generate etheno bases. In order to determine whether mutational patterns in target genes in vivo are characteristic of vinyl chloride and could be explained by the mutagenic properties of the etheno bases, human and rat liver tumours associated with exposure to vinyl chloride were analysed for point mutations in the ras and p53 genes. In this paper, we review these data and report our latest results on animal tumours. Two alterations were found which could be attributed to a direct effect of vinyl chloride: a GC-->AT transition which leads to a GGC-->GAC mutation at codon 13 of the Ki-ras gene in human liver angiosarcomas, and lesions at AT base pairs, mostly AT-->TA transversions, which lead to mutations in the p53 gene in human and rat angiosarcomas and to a CAA-->CTA mutation at codon 61 of the Ha-ras gene in rat hepatocellular carcinomas.