An analysis of 3.8 million immune cells from nearly 2,000 people finds that immune ageing arrives in distinct waves rather than a steady slide, with T cells driving both and men and women following different trajectories.

Ageing is usually described as a gradual process, a slow accumulation of wear that eventually shows up in a blood test or a knee. For the immune system, that picture may be wrong in an important way. A study from Duke-NUS Medical School in Singapore reports that immune cells in the bloodstream do not decline at a constant rate at all. Instead, their gene activity shifts sharply in two concentrated windows: around the age of 40, and again after 60.

“We tend to think of ageing as a gradual process, but our findings show that the immune system does not simply decline at a steady rate,” said Associate Professor Jacques Behmoaras of Duke-NUS’ Centre of Biomedical Data Science, the study’s senior author. “Instead, we see distinct periods of rapid change, particularly around 40 and again after 60, with T cells playing a major role.”

The work, published in Nature Communications, pooled publicly available single-cell datasets to assemble a map of 3.8 million individual immune cells drawn from nearly 2,000 healthy people between the ages of 19 and 97. The researchers note that the pooled population included people of Asian ethnicity, a group under-represented in much of the existing ageing literature.

Two waves, two different sets of cells

Single-cell analysis lets researchers read the gene activity of individual cells rather than averaging across a whole blood sample, which makes it possible to see which specific cell types are changing and when. Across the pooled data, the genes responsible for RNA and protein synthesis — the basic machinery any cell needs to do its job — declined most steeply at the two peak points rather than drifting downward across the decades.

T cells, the immune system’s coordinators and killers, dominated both windows. They showed the most pronounced changes in gene activity during people’s 40s, and they were the main drivers of the second wave of change after 60. But the study found that the specific T cell subtypes most affected in the 40s are not the same ones most affected after 60, which suggests the two waves are not simply the same process repeating with more force.

“Our findings build on previous evidence that T cells are among the immune cells most affected by ageing,” said Professor Antonio Bertoletti, one of the study’s authors. “Understanding what drives these nonlinear changes in T cells over time will be an important next step.” He added that “changes in T cell function may help explain why older adults are more susceptible to infections and inflammatory conditions.”

Men and women do not follow the same curve

The second headline result concerns sex. How these specialised T cell populations change over a lifetime differed between men and women in the pooled data. The study does not claim to explain why, but the observation lands on a long-standing puzzle: autoimmune conditions fall disproportionately on women, and the reasons remain contested.

Behmoaras was careful about how far the finding reaches. “Understanding what drives these windows of immune ageing could eventually help us identify when, and for whom, interventions may have the greatest benefit,” he said. “It could also shed light on why some autoimmune conditions disproportionately affect women.” The operative words are could and eventually; the study maps a pattern rather than identifying a mechanism or a treatment target.

The scale of the clinical problem is not small. More than 600,000 people in Singapore, roughly 11 per cent of the population, live with an autoimmune disease.

Predicting biological age — and the limits of the data

The team also built machine-learning models to predict biological age from immune cell profiles, part of a broader push across ageing research to find measures that track how old a body is functioning rather than how many birthdays it has had. The release does not report how accurate those models were, which is the number that would matter most for any practical use.

Two features of the study design are worth keeping in view when reading the result. First, this is a meta-analysis: the data were collated from existing public datasets rather than generated by recruiting and sampling a single cohort. That approach buys enormous scale — 3.8 million cells is far beyond what one laboratory would typically generate — but it inherits whatever differences in sampling, processing and population make-up the original studies carried. Apparent transitions at particular ages can in principle reflect how the contributing studies were assembled rather than biology alone.

Second, the design is cross-sectional. It compares different people of different ages at a single point in time; it does not follow the same individuals as they age through their 40s and into their 60s. A wave that appears at 40 in a population snapshot is a strong hypothesis about what happens to individuals, not yet a demonstration of it. Longitudinal sampling would be needed to confirm that any one person’s immune system lurches on this schedule.

Why the shape of the curve matters

If immune ageing really is concentrated in windows, the practical implications run toward timing. Interventions aimed at preserving immune function — vaccination schedules, immunomodulating drugs, anything intended to slow immune decline — are currently organised around broad age bands. A curve with identifiable inflection points invites the question of whether there are better and worse moments to act, and whether those moments differ for men and women.

“The study gives us a more detailed map of how the human immune system changes across the lifespan, and importantly, how those trajectories differ between men and women,” said Professor Sheemei Lok, Duke-NUS’ Interim Vice-Dean for Research, who was not an author on the paper. “That knowledge is fundamental if we want to move away from treating ageing as a one-size-fits-all approach and move towards more precise approaches to maintaining health as people grow older.”

That is the ambition. What the study delivers for now is a reference map: a population-scale account of which immune cells change, roughly when, and how the pattern diverges by sex. Whether the waves hold up in individuals followed over time, and whether anything can usefully be done about them, are the next questions.

---

Notes

1. “Sex-specific trajectories of nonlinear immune aging at single-cell level.” Nature Communications, 21 August 2026. DOI: 10.1038/s41467-026-76737-4.
2. Duke-NUS Medical School. “Immune ageing accelerates around 40 and 60, with key differences between men and women.” EurekAlert!, 1 October 2026. https://www.eurekalert.org/news-releases/1146248.