DOES A GENOTOXIC CARCINOGEN CONTRIBUTE TO HUMAN BREAST-CANCER - THE VALUE OF MUTATIONAL SPECTRA IN UNRAVELING THE ETIOLOGY OF CANCER

The p53 tumour suppressor gene is turning out to be a useful reporter for the stigmata of past genotoxic exposure. About half of all human c ancers contain p53 mutations most of which occur in those regions (exo ns 5 - 8) of the gene that are highly conserved during evolution. Muta tions are mainly of the missense type and their frequency and distribu tion vary among different kinds of cancer. The ability to detect all s ix possible base-substitution mutations in the p53 gene in human tumou rs makes it possible to construct mutational spectra for different can cers at a locus clearly implicated in carcinogenesis. Transitions at o ne particular hotspot-the CpG dinucleotide-occur frequently in many ca ncers and may reflect endogenous mutation. A reduction in the proporti on of CpG mutations at the expense, for example, of an increase in GC to TA transversions may signal the effect of an exogenous mutagen. We exploited these features of the p53 gene to examine the evidence that a previously unsuspected genotoxic exposure may contribute to the high incidence of breast cancer in women living in rich industrialized cou ntries. We compiled a mutational spectrum of p53 from 120 breast cance rs and compared it with the spectrum from 145 colorectal cancers and 2 46 lung cancers. A germline p53 spectrum was constructed using data fr om 27 patients. Two hundred germline mutations in the haemophilia B ge ne served as a 'background' spectrum. The spectrum of mutations in the p53 gene in breast cancer revealed a reduction in the proportion of G -->A and C-->T transitions at CpG dinucleotides compared with colorect al cancer (P < 0.0005) and an increase in G-->T transversions (P < 0.0 005). Other mutations showed no significant differences from colorecta l cancer or germline mutational spectra. In breast cancer, as in lung cancer, G-->T transversions were over-represented at CpG dinucleotides and there was also a G-->T hotspot at codon 157 that was not seen in colorectal cancer. Moreover, G-->T transversions were much more common on the coding strand, as in lung cancer. Thus, the mutational spectru m in the p53 gene of breast cancer differs significantly from that tho ught to be attributable to endogenous or background mutagenic processe s. It resembles more closely the lung cancer spectrum, which is probab ly caused by exogenous mutagenic chemicals. These findings pose the fo llowing questions. Is a mutagenic agent of exogenous or endogenous ori gin involved in the aetiology of breast cancer? Is the breast epitheli um and/or its neoplastic derivatives less efficient at repairing DNA d amage than are colorectal epithelial cells?