Mutation

With the development of transgenic mouse models harboring chromosomally integrated reporter genes, it became possible to directly test the hypothesis that somatic mutations in a neutral (with no function) gene accumulate with age in various organs and tissues (Gossen and Vijg, 1993). Using one of these models, harboring the lacI gene as a target, Lee et al. (1994) were the first to demonstrate an age-related increase in mutation frequency in spleen from about 3 × 10^-5 in mice of a few weeks old to 1-2 × 10^-4 in 24-month-old animals. Subsequent results from other laboratories indicated age-related increases in mutation frequency in some, but not all, organs. Dollé et al. (1997), for example, demonstrated that mutation frequencies at a lacZ transgene increase with age in the liver, while such an increase was virtually absent in the brain. The increased susceptibility to spontaneous mutagenesis of liver versus brain corresponds to observed higher frequency of focal pathological lesions in the mouse liver as compared with the brain (Bronson, 1990). More recently, the pattern of organ specificity in age-related mutation accumulation was expanded with the observation of age-related increases in mutation frequencies in spleen, heart, and small intestine, but not in testes (Dollé et al., 2000).

Essentially the same results obtained by Dollé et al. (1997) were found by Ono et al. (2000). These investigators used the original mouse model of Gossen et al. (1989), harboring the same transgene. They also observed an age-related increase of mutation frequency in the liver, heart, and spleen, but virtually no increase in the brain and testes. Dollé et al. (1997; 2000) also observed striking organ specificity with respect to the mutational spectra of the old animals. While in the small intestine and the brain virtually only point mutations accumulated (the small increase in the brain was almost totally due to point mutations), in the liver and especially in the heart large deletion mutations were a prominent part of the spectrum (Dollé et al., 1997).

In interpreting these data it should be noted that the observed increases were modest (varying from less than twofold, to more than fourfold) and appeared to level off at middle age (Lee et al., 1994; Dollé et al., 1997). The relatively small age-related increase of mutant frequencies can be interpreted as evidence against a major role for somatic mutations in aging (Warner and Johnson, 1997). However, although transgenic reporter genes do not suffer from a selection bias (as is the case with most selectable endogenous targets), it still provides an underestimate of the real mutation load and its adverse effects. Homologous (mitotic) recombination, for example, leading to deletion of entire reporter gene copies is a frequent mutational event and goes undetected in the transgenic assays. Most of the transgenic models also do not account for mutational hot spots, and such important functional end points as cell death are missed. Indeed, to put the results on mutant frequencies of different organs and tissues at various age levels into context, it will be necessary also to assess cell proliferation and cell death. Most important, it will be necessary to determine at some point the critical level of cellular mutation loads in terms of physiological consequences. The question is whether the mutation loads observed in a tissue at old age have physiological consequences. A glimpse of an answer can possibly be obtained from another type of model system.