The first studies of aging’s influence on gene expression looked for coherent, directional changes: gene X goes up by 5-fold in midlife; gene Y goes down 2-fold in late life; &c. One of the more fascinating recent developments in this field has been the realization that in addition to directional changes in populations of cells, the cell-to-cell variability of gene expression increases with age. This increase in transcriptional “noise” is concomitant with an increase in point mutations, and it happens both in response to advanced age in vivo and experimentally induced oxidative stress in vitro.

Progress on this front has been spearheaded by the Vijg lab (formerly at UT Health Science Center at San Antonio, now at the Buck Institute for Age Research). In a recent paper, they expand their findings to show that the increase in transcriptional variability between individual cells is also concomitant with an increase in large-scale genome rearrangements. Busuttil et al.:

Genome instability has been implicated as a major cause of both cancer and aging. Using a lacZ-plasmid transgenic mouse model we have shown that mutations accumulate with age in a tissue-specific manner. Genome rearrangements, including translocations and large deletions, are a major component of the mutation spectrum in some tissues at old age such as heart. Such large mutations were also induced by hydrogen peroxide (H2O2) in lacZ-plasmid mouse embryonic fibroblasts (MEFs) and demonstrated to be replication-independent. … To test if large rearrangements could adversely affect patterns of gene expression we PCR-amplified global mRNA content of single MEFs treated with H2O2. Such treatment resulted in a significant increase in cell-to-cell variation in gene expression, which was found to parallel the induction and persistence of genome rearrangement mutations at the lacZ reporter locus. Increased transcriptional noise was also found among single cardiomyocytes from old mice as compared with similar cells from young mice. While these results do not directly indicate a cause and effect relationship between genome rearrangement mutations and transcriptional deregulation, they do underscore the stochastic nature of genotoxic effects on cells and tissues and could provide a mechanism for age-related cellular degeneration in postmitotic tissue, such as heart or brain.

It certainly makes sense that mutations in individual genes could result in broader alterations in global transcription, and diminish the robustness of the transcriptome as these mutations accumulated — but that connection is far from proven. As the authors themselves point out at the end of the abstract, the missing piece remains the determination of a causal link between genomic changes (point mutations or larger-scale rearrangements) and transcriptional noise.

This will be tough to tease apart. For a very strong demonstration, one would need to somehow age cells without allowing the normal buildup of mutations. I suspect that the work will instead proceed via the logical converse, i.e., by asking whether cells with increased mutational rates (e.g., DNA repair mutants, or those with impaired tolerance to oxidative stress) show a more rapid increase in transcriptional noise as they age.