SenNet has begun building an atlas of human senescent cells

NIH has shown that senescent cells cannot be identified by a single protein: each tissue requires its own set of markers

The NIH SenNet consortium has compiled a catalog of markers for cellular senescence, a state in which a cell stops dividing after damage but remains active. In three new studies, researchers showed that no single protein or gene can identify these cells. Identification instead requires a set of markers specific to the cell type and the cause of senescence.

Researchers often look for senescent cells using p16, p21, or inflammatory proteins. This approach is like trying to identify a person from a single word they use. Some cells that express the marker will be senescent, some will not, and many senescent cells will not express it at all. This is one reason why senolytics, drugs intended to remove senescent cells, have so far been difficult to develop into targeted therapies.

In the SenCat study, published on June 11, the team examined RNA and proteins in 14 types of primary human cells across more than 30 forms of senescence. The researchers induced this state through several mechanisms: exhaustion of the cells' capacity to divide, DNA damage, oxidative stress, and oncogene activation. No single marker was shared across all cases. Cell type and the cause of damage had the strongest effects on the resulting profile.

The researchers did identify several shared processes: damage response, metabolic remodeling, and tissue repair. Using these processes, the authors built a model that integrates multiple signals. The model distinguished senescent cells from normal cells in independent datasets and showed how different types of senescent cells accumulate in mouse organs.

A more difficult question is whether this cellular trace can be detected in the blood of a living person. In another study, researchers tested the SenCat protein signatures in 2 272 participants from two long-term studies of aging. Signatures associated with different cell types pointed to different health trajectories. The immune-cell signal was associated with subsequent mortality and disease onset, the kidney-cell signal with kidney disease, and the adipose-cell signal with body mass index.

For now, these are observational associations. Proteins in the blood do not show that senescent cells caused the disease, nor can they identify people who would benefit from a senolytic. In a review of human dietary trials, proteins from the inflammatory program of senescence changed more often than direct markers of senescent-cell accumulation. SenCat takes the next step by linking blood proteins to signatures from specific cell types. Instead of asking, “How many old cells are in the body?”, researchers can ask which cells have changed, in which tissue, and what future risk is associated with that change.

Senolytic therapy must identify a harmful cellular state and establish its role in a particular tissue before an intervention can be selected. SenCat provides a map for that work. Clinical studies must now determine whether this map can be used to select patients and alter the course of disease.