Alterations in gene expression by carcinogens were analyzed on three unstable alleles of the white (w+) locus of Drosophila melanogaster: white-crimson (wc); white-ivory 16 (wi16); and white-unstable 11 (wu11). Two of these alleles (wi16 and wu11) were spontaneous mutant derivatives of wc, which is known to harbor a transposable element. The compounds studied were dimethylnitrosamine, 7,12-dimethylbenz(a)anthracene, and aflatoxin B1. These carcinogens were topically applied on the early larval stages, and the genetic effects assayed were the alterations in eye color either to wild-type (w+) or to other w mutants, initiated both somatically and germinally, as well as the simultaneously induced X-chromosome recessive mutations.

The tested compounds influenced the different unstable w alleles in a highly selective manner, both as a function of the inducing agent and the organization of the genome in the target cells. The same treatments raised the somatic reversions to w+ above the corresponding controls for wc and wi16, but not for wu11, whereas the simultaneous induction of other w mutant phenotypes occurred appreciably only with wc. Furthermore, these treatments gave high and variable somatic reversions to w+ with wi16, whereas the simultaneously induced germinal events were uniformly very low.

The frequencies of altered expression at the unstable test loci, whether in the soma or germ line, were quantitatively uncorrelated with the mutagenic effects of the treatments in terms of the yield of X-chromosome recessive mutations assayed in the progeny of males emerging from the same treated larvae. There was also an association between the time of the induction of these alterations by the tested carcinogens in the soma and the cellular stage in genomic differentiation. Reversions to w+ were induced preferentially after the onset of genetic determination, whereas changes to the w mutant phenotypes occurred predominantly during the predetermination phases.

The genetic properties of transposable elements and the manner of their response to carcinogens supported the hypothesis that nonviral cancer might arise from molecular processes similar to those involved in the evolution of retroviruses.


This work was supported by grants to the Institute of Cancer Research (Chester Beatty Research Institute, Royal Cancer Hospital) from the Medical Research Council and the Cancer Research Campaign.

This content is only available via PDF.