As discussed previously, although ras mutations are rare in breast cancer, over-expression of ras p21 is frequently detected. Mutation of ras is just one of a number of ways that ras signal transduction pathways are activated. The relative infrequency with which ras mutations are seen in human breast cancer may reflect a very large number of active growth factor receptors in breast epithelium which, whether mutant or amplified, result in overexpression of ras. The mechanism(s) by which ras is overexpressed in human breast cancer is not known.

Normal c-Ha-ras-transfected cells, MCFlOAneoN (Basolo et al., 1991), have an extra copy of c-Ha-ras but the protein is not overexpressed (unpublished) and the cells do not form xenograft lesions (Miller et al., 1993). Insertion of the mutant ras resulted in increased p21 protein and transformation of MCFIOA cells. Although the preneoplastic phenotype in xenografts was observed following ras transfection, we have found that not all clones of MCFlOAneoT are able to form lesions in vivo (Miller et al., 1996).

Restriction size fragment analysis demonstrated, however, that clones unable to form preneoplastic lesions retained the activated c-Ha-ras and confirmed that the insertion site of the activated c-Ha-ras was the same for these clones as for the variants selected for ability to form lesions in vivo. Furthermore, Western blotting with antibodies specific for the codon 12 valine c-Ha-ras demonstrated that mutant p21 protein was comparable in nonlesion-forming clones and lesion-forming variants as well (Miller et al., 1996).

Thus, activated c-Ha-ras is not sufficient to produce the preneoplastic phenotype of the MCFl OAT human breast stem cells. This was an important finding because it offers the opportunity to identify the genetic alteration(s)  involved in this critical early stage in proliferative breast disease by comparing the nonlesion-forming clones with the preneoplastic variants selected by their ability to form preneoplastic lesions.

The karyotype of the MCFIOAT cells in vivo is remarkably stable despite longterm passage in xenografts. The only alterations have occurred during extended in v/rro maintenance. An additional marker chromosome, t(3;17)(pl3;pl2), appeared in MCFIOA cells in vitro between the time of immortalization and the time of transfection with T24 ras.

Likewise, MCFlOAneoT cells were initially identical to the untransfected MCFIOA cells but acquired an additional, apparently normal copy of chromosome 9 prior to the first xenograft experiment. Subsequent MCF10AT# variants isolated from xenograft lesions have been unchanged regardless of the histology of the lesion from which the cells were derived.

However, we do not suggest that progression to carcinoma occurs without karyotypic alterations in the MCFIOAT model. Breast cancer cells are notoriously difficult to establish in culture from solid tumors; therefore we hypothesize that the preneoplastic stem cell with the depicted karyotype has a growth advantage in vitro.