We have compared the expression patterns of the non-allelic insulin 1 and 2 genes during prolonged in-vitro culture of the mouse islet cell line β-TC3, where transformation by the SV40 T oncoprotein is targeted to the differentiated β-cell phenotype, and the rat islet cell line NHI-6F, in which the β-cell phenotype is induced by transient in-vivo passage. The NHI-6F clone carries, in addition, a single copy of a transfected silent human insulin gene which contains 3 kb of regulatory sequences known to confer β-cell-specific expression. Insulin gene expression was measured by an assay based on a reverse transcription-polymerase chain reaction, to determine whether the ancestral rodent insulin 2 genes (and the human homologue in the NHI-6F cells) are regulated differently from the duplicated rat and mouse insulin 1 genes.
We have shown that activation of insulin gene expression in the NHI-6F cells includes transcriptional activation of all three genes, but that extended propagation of tumour cells in vitro leads to a selective and equal decline in the quantities of transcripts from the rat 2 and human genes relative to transcripts from the rat 1 gene. In the later passages, insulin transcripts were derived almost exclusively from the rat 1 gene. In early in-vitro passages of the mouse endocrine cell line β-TC3, the expression pattern of the mouse 1 and 2 insulin genes resembled that seen in isolated mouse islets. After more than 45 in-vitro passages, expression of the duplicated mouse 1 gene decreased tenfold when compared with the ancestral mouse 2 gene. As previously shown for NHI-6F cells, the differential expression of non-allelic insulin genes in the β-TC3 line was also clearly evident at the cellular level, where a subpopulation of cells selectively expressed readily detectable levels of mouse C-peptide 2 immunoreactivity while devoid of C-peptide 1. Our results suggest that the maintenance of insulin gene expression in rodent tumour cells is influenced by enhancer sequences which are not shared by the ancestral and duplicated insulin genes, and that either species-specific conditions or transformation-related differences exist between the rat and mouse cell lines that govern which gene remains active during prolonged in-vitro propagation.