Somatic mutations and ageing: where longevity may hit a limit

Somatic mutations and ageing are connected by a basic biological problem: cells accumulate random DNA changes throughout life, and that process cannot be made perfectly error-free. A mathematical model now asks what would happen if every other age-dependent mechanism were removed and somatic mutation were left as the only driver of ageing. Under its central assumptions, median lifespan was about 156 years.

That number is not a prediction that humans can live to 156, nor an estimate of what a future rejuvenation treatment might achieve. This was a computational experiment designed to isolate one component of ageing while deliberately suppressing the age-dependent effects of the others.

What happens when ageing is switched off

The researchers began with demographic data from 16 populations. For the main calculations, they used Swiss life-table data and permanently fixed mortality at the level seen at age 30. In other words, the risk of dying was allowed to continue, but it was no longer allowed to rise with age.

In this artificial population, median lifespan reached 1,759 years. The model also produced a much larger “maximum” of 29,221 years. However, maximum lifespan had a specific statistical definition here: the point at which only one person in 100,000 of the original population would remain alive.

Those millennia therefore should not be interpreted as the biological capacity of the human body. They are a mathematical consequence of maintaining a relatively low young-adult mortality rate indefinitely.

The authors then reintroduced one age-related process: somatic mutation. They drew on previously published single-cell sequencing data from human neurons, cardiomyocytes, liver cells and airway basal cells. Healthy-control samples were selected, and the airway analyses were restricted to non-smokers.

A somatic mutation is a DNA change acquired by an ordinary body cell after conception. Most such changes do not kill a cell, so mutation counts alone are not enough. The researchers also estimated the probability that a mutation would disrupt a gene essential for a particular cell and cause lethal damage.

Why the brain and heart became the bottlenecks

The outcome depended strongly on whether a tissue could replace damaged cells.

Mature neurons are essentially non-dividing, and cardiomyocyte renewal is also very limited. When one of these cells is lost, replacement is difficult. In the brain model, median lifespan was 194 years and the statistical maximum was 557 years. For the heart, the corresponding estimates were 208 and 868 years.

The liver behaved very differently. Hepatocytes can proliferate and replace lost cells. Even without support from liver progenitor cells, the median time to the modelled liver-failure threshold was 37,664 years. When progenitor cells were included, none of the simulated trajectories reached liver failure during 100,000 years of simulation.

This does not mean that a real human liver could survive for tens of thousands of years. It means that, within this model, cell replacement was sufficient to compensate for mutation-driven cell loss.

Airway basal cells were similarly resilient. They can divide, but the model assigned them a finite replicative capacity, so eventual failure still occurred. The median time to the critical threshold was about 4,359 years.

When the brain, heart, liver and airway epithelium were combined and organ failures were assumed to be independent, median lifespan was 156 years and the statistical maximum was 470 years. Because real organs do not age independently, the authors also calculated mathematical bounds for different dependence structures. These gave a median range of 146–194 years and a maximum range of 210–557 years.

What the model cannot establish

One of the most uncertain inputs was the probability that a single mutation actually kills a cell. This quantity cannot be measured directly from the available datasets. It was estimated using sets of genes considered essential for cellular survival together with computational predictions of variant damage. Some of the evidence used to define essential genes came from mouse experiments.

The model also covers only a limited set of cell types and focuses mainly on mutations that cause cell death. It does not capture gradual functional decline in surviving mutant cells, expansion of mutant clones, the age-dependent rise in cancer risk or the full network of signalling between organs. The cell-loss thresholds used to define organ failure are also indirect estimates.

There is an additional complication. The mutation accumulation rates came from tissues in real ageing humans, where oxidative stress and other age-related processes already affect DNA. In the hypothetical organism imagined by the model, those influences are supposed to be removed, so the underlying mutation rate could be different.

The study therefore does not establish 156 years as a new human lifespan limit. Its more interesting conclusion is mechanistic: even in an organism freed from most other forms of ageing, random DNA changes could still place a substantial constraint on longevity, particularly in tissues that cannot readily replace lost cells. Somatic mutation alone, however, is not sufficient to explain the lifespan humans actually experience.