Ratio of two types of cells predicts health for years to come

Researchers analyzed approximately 12.4 million individual immune cells from blood samples of 2,609 predominantly healthy adults aged 20 to 90+ years — data were collected in eight cohorts from North America, the UK, Asia, and Australia. They noticed a pattern: the blood of young people was generally rich in immune cells that had not yet encountered specific pathogens and retained flexibility for new infections or vaccines. In older individuals, the picture became more complex and diverse; in some participants, the number of cells associated with inflammation increased noticeably.

To visualize this shift, experts placed each participant on a условном spectrum — their position was determined by the balance of two types of CD8 memory T cells that secrete different granzymes. Granzymes are molecules that such cells use to destroy affected cells or coordinate immune responses. Cells secreting granzyme B act as direct killers of diseased cells. Cells with granzyme K have been less studied; it is believed that they send alarm signals and recruit additional immune system forces.

The ratio of these cells can predict future health. To test the hypothesis, the team turned to the UK Biobank database — it contains medical data from 500,000 initially healthy participants aged 40–69, who were followed for up to 15 years, as well as information about hundreds of proteins in their blood. Direct counts of CD8‑T cells were not present in the UK Biobank, so the researchers first built a computer model using a small sample where such counts were available: it learned to determine which type of cell (with granzyme B or K) predominates based on the characteristic fingerprint of proteins in the blood. The model was then applied to samples from 50,000 UK Biobank participants, and their position on the immune aging spectrum was calculated.

Fifteen years of observation revealed a clear pattern: among people who were healthy at the time of inclusion in the study and whose immune system initially tended to favor granzyme‑B cells, the risk of death was higher a decade later. This same group was more likely to experience chronic conditions, including type 2 diabetes, hypertension, liver diseases, and kidney failure. In contrast, those who had more granzyme‑K cells showed a more favorable aging trajectory.

According to Marina Terekhova, a co‑author of the study, a higher level of granzyme‑B cells in an apparently healthy person does not mean a formal diagnosis, but it may signal that the immune system is actively responding to something. It turns out that healthy aging of the immune system gradually shifts towards granzyme‑K cells, while a shift towards granzyme‑B cells is a marker of unhealthy aging. Moreover, this imbalance can arise long before symptoms appear, which is why people born in the same year face different risks decades later.

Currently, mapping a person onto such a spectrum requires complex and expensive specialized equipment. However, the Artemov Laboratory is working to transform this approach into a simple blood test that can be performed using standard equipment. The goal is to create an accessible tool for early diagnosis: it should detect subclinical immune stress long before signs of chronic diseases appear. As Artemov explains, this is similar to checking the oil level in a car before the warning light comes on; noticing a deviation in time will allow for an in‑depth examination and, possibly, intervention long before the disease develops.