Across 48 DNA sites, shifts in immune cell composition explained most of the change in blood methylation over a decade
Across 48 DNA sites, shifts in immune cell composition explained most of the change in blood methylation over a decade
A preprint published on August 19 analyzed paired blood samples collected ten years apart from 86 older participants. In the primary cohort, after accounting for estimated changes in immune cell composition, 32.2% of the original rate of change in the 48-site metric remained. This residual was not statistically different from zero.
Methylation refers to chemical marks on DNA. Their patterns are used to construct epigenetic measures of age, but a blood sample always contains a mixture of immune cells. Each cell type has its own methylation pattern. If the proportions of these cells change over time, the average signal measured in blood will also change, even if the same change does not occur within each cell type.
In the preprint, the author examined a metric based on 48 preselected DNA sites and tracked it in two longitudinal cohorts. In the Danish cohort of 86 participants, the author used a separate set of DNA marks to estimate changes in the proportions of seven immune cell types, then included those changes in the analysis. After this adjustment, the rate of change in the metric fell to 32.2% of its original value, while the model that included changes in cell composition explained 36.4% of the variation between participants in how much the metric changed.
Comparing each person with their own earlier measurement removes stable individual characteristics, including their usual blood cell composition. However, that composition can change over ten years.
“A variable that changes in the same direction within participants does not disappear through subtraction. It becomes part of the result itself.”
A blood methylation assay combines signals from two sources: changes in the relative proportions of cell types and changes in DNA methylation within those cells. In a 2024 study, the IntrinClock model was designed to produce the same reading across ten tested immune cell types. Its reading still increased with age in purified naïve CD8+ T-cells, which are cells that have not yet encountered an antigen.
Repeated blood analysis should report the shift in immune cell composition separately from the signal that remains after adjusting for that shift. This prevents a single value from combining changes in the cell mixture with changes in methylation within the cells.