Design Principles of Life, Disease, and Aging

# Systems aging: why does risk rise so sharply with age? *Evidence guide: Population patterns, primary experimental results, and a systems-aging model. A fit is not unique proof of mechanism.* Start with a pattern. In adults, the risk of dying in the coming year rises roughly exponentially with age. In many modern populations it doubles about every eight years, from early adulthood until around age ninety, where the rise slows. This is the Gompertz law, named after Benjamin Gompertz, who described it in 1825. It is an approximate law. It does not describe childhood, it bends at the oldest ages, and its level differs between populations and eras. Alon's group proposed a quantitative theory, the saturating removal model. Picture a village. Houses produce garbage. Trucks collect it. Houses are damage-producing units, which accumulate slowly over decades. Garbage is the damage itself. In mammals, the model's leading candidate is senescent cells: damaged cells that stop dividing and secrete inflammatory signals. Trucks are removal processes, for senescent cells mainly immune cells. As an equation: the rate of change of damage X equals production, which rises linearly with age as eta t, minus removal, which saturates at a maximal rate beta as X grows, plus noise. When the village is young, the trucks easily keep up. As houses accumulate, removal approaches capacity. Each extra bit of garbage is cleared more slowly, so damage rises faster and faster, and fluctuations last longer. This was built from data. In mouse lungs, Karin and colleagues found in 2019 that senescent cells were cleared with a half-life of about five days in young mice, and much more slowly in old mice, consistent with saturating removal. Now add a threshold. Death, or disease onset, is modeled as damage crossing a threshold. With production rising and removal saturated, the chance of crossing rises exponentially with age. The model recovers the Gompertz law and its slowdown at very old ages. The model also explains why the rise in mortality slows at very old ages. Around then, damage production approaches the maximal removal rate, damage sits close to the threshold, and chance fluctuations dominate, so the yearly risk climbs only slowly. Noise matters. Individuals with similar parameters cross the threshold at different ages, because random fluctuations add up differently. Measurements in single E. coli cells under starvation showed the same signature: variation in lifespan driven largely by stochastic damage dynamics, rather than by initial differences. For disease, each condition gets its own threshold, with only a susceptible fraction of people at risk. Fit to a nationwide dataset on nearly a thousand diseases over fifty million life-years, this two-parameter picture captured the exponential rise and the late drop. The saturating removal model is a theory developed by Alon's group. It explains a broad set of aging patterns and makes testable predictions, but it is not a proven or complete theory of aging. Aging involves many kinds of damage, and senescent cells are one leading candidate, not the only one. Sources: [gompertz1825](https://doi.org/10.1098/rstl.1825.0026), [aging-notes](https://www.weizmann.ac.il/mcb/alon/courses/system-biology-aging-and-longevity-2026), [karin2019](https://doi.org/10.1038/s41467-019-13192-4), [yang2023](https://doi.org/10.1038/s41467-023-37930-x), [katzir2021](https://doi.org/10.1111/acel.13314).