Genetic characterization of a human skin carcinoma progression model: from primary tumor to metastasis.

The type and number of genetic aberrations required for a fully malignant tumor are still unclear. This study describes the genetic analysis of a series of skin squamous cell carcinomas, representing the primary tumor, two recurrences, and a metastatic lesion from a single patient and cell lines established therefrom (MET-1 to MET-4). Comparative genomic hybridization demonstrated that: (i) most of the gains and losses were common for tumors and cell lines and affected chromosomes 3 (3p loss, 3q gain), 5 (5p gain, 5q loss), 7 (7p gain), 8 (8p loss, 8q gain), 11 (11q gain), and 17 (17p loss), and (ii) only one aberration was present in a tumor but not in the cell line (10 loss in tumor 4); and only few aberrations were cell line specific. From these, 10p loss and 17q gain were shared by all lines and tumor 4, suggesting that they were already present in all tumors, although in only a subpopulation of cells, whereas 20q gain (shared by all lines), 4q loss (MET-2), and 18p gain/18q loss (MET-3) seem to be culture derived. In agreement, multiplex fluorescence in situ hybridization demonstrated a set of common translocations for all lines thereby further confirming their common origin. In addition, each cell line, exhibited one or more individual translocation chromosomes, which suggested that MET-1 was a precursor of MET-4, whereas MET-2 and MET-3 developed in parallel. Whereas MET-1 to MET-3 were hypodiploid or hyperdiploid, MET-4 was characterized by polyploidization, a set of specific aberrations (t(3;7), t(X;2), i(10q)), and increased heterogeneity (varying translocations in individual metaphases). Using sequencing and expression studies, cells from all lines were wild type for p53, did not exhibit mutations in any of the ras genes (Harvey, Kirsten, or N-ras), and expressed wild-type fragile histidine triad gene (FHIT; mapped to 3p14.2, a locus underrepresented in all cells) transcripts. Thus, with the MET cell lines we present an in vivo skin carcinoma progression model that was genetically well defined, and which, despite originating from a sun-exposed site, is wild type for p53.