The Mitochondrial Theory of Aging postulates that mutations in mitochondrial DNA accumulate with age and cause various intracellular adverse effects that ultimately contribute to some age-related degenerative changes. It is usually assumed that the mutations in question are actually generated at old age. Analysis of mutational spectra (i.e. frequency distributions of mutations by kind), however, suggests that at least in some cases mutations that are observed at old age may have been originally generated early in the development.
In particular, one observes large variation of mutant frequencies from sample to sample and from individual to individual. If mutations were generated in old tissues, the large number of mutational events should be sufficient to create a stable mutational spectrum, where mutant frequencies are proportional to mutational rates. Mutational rates however, depend on biochemical reactions that should not change from one location in a tissue to the other. If however, mutations happened earlier in development in a few copies and were disseminated across growing tissue, the founder effect might have resulted in enormous variation that is observed. Of course, mutations continue to arise in the old tissue. Such de novo mutations however, may be less abundant then the "old" ones that enjoyed the advantages of dissemination and years of expansion that are not available to the "new" mutations.
The observed accumulation of mutations with age may thus reflect merely clonal expansion pre-existing "old" mutations rather than de novo generation of mutations in aged tissue. If confirmed, this hypothesis will have profound consequences on our understanding of the sources of mitochondrial mutations and the measures that should be taken to limit their abundance.
Reference: K. Khrapko et al., Mutation Res. 522:13-19