Which genes' splicing changes with age? A genome-wide analysis of 2.8 million junction tests in GTEx v8

Inside every cell of the human body, a complex assembly line works tirelessly to turn genetic instructions into the proteins that keep us alive. These instructions are written in DNA, but they are not used directly. First, a copy is made in the form of a molecule called RNA. This RNA copy is often like a rough draft, containing both useful sections and unnecessary filler. Before the cell can use this draft to build a protein, it must be edited. This editing process, known as splicing, involves cutting out the filler and stitching the useful parts together. Remarkably, the cell can often choose to stitch these parts together in different ways, creating multiple versions of the same protein from a single gene. This flexibility allows a single gene to perform different jobs in different contexts. However, as we grow older, the precision of this editing process begins to falter. The question of which specific genes lose their ability to be edited correctly, and how this breakdown happens across the body, has remained largely a mystery until now.

A recent study set out to map this landscape of aging by examining the genetic material from nearly 18,000 tissue samples collected from hundreds of donors. The researchers focused on a massive dataset known as GTEx, which contains genetic snapshots from 49 different types of human tissues, ranging from the brain and heart to the liver and skin. By analyzing the junctions where RNA segments are stitched together, the team could see exactly how the splicing process was changing as the donors aged. They looked at nearly three million specific editing events across more than 15,000 genes to see if the patterns of these edits shifted consistently with the passage of time.

The investigation revealed that splicing changes are not a rare occurrence but a widespread feature of human aging. Out of the nearly one million specific editing events they tested, about 12 percent showed a clear link to age. This means that in roughly one out of every nine places where the cell edits its genetic instructions, the pattern shifts as we get older. These changes were not scattered randomly; they were concentrated in a specific group of genes. While the majority of genes showed some sign of change in at least one tissue, the most dramatic shifts occurred in a few hundred genes that stood out as being particularly sensitive to aging.

Perhaps the most surprising discovery was that the genes most affected by aging were often the very ones responsible for controlling the splicing process itself. The study found that the machinery that edits RNA—specifically the proteins that act as the editors—was frequently changing its own instructions as people aged. Genes that regulate how other genes are spliced, such as those producing proteins like HNRNPH1 and SF1, were among the most consistently altered across the body. This suggests a feedback loop where the aging of the editing tools leads to errors in how those tools are made, potentially causing a cascade of further errors throughout the cell. This finding aligns with the idea that the loss of control over these fundamental processes is a key driver of aging.

When the researchers looked at the direction of these changes, they found a consistent trend. As people aged, the editing process tended to drift away from the "principal" or standard version of a gene's instruction. Instead of sticking to the most common and likely correct version, the cell began to produce more of the alternative, less common versions. In nearly two-thirds of the cases where a change was detected, the standard version lost its dominance. This shift suggests that the cell's ability to maintain the most efficient and accurate version of its proteins is gradually eroding, leading to a more chaotic mix of protein variants.

However, the study also uncovered a significant complication in interpreting these results. The data came from tissue samples taken after the donors had passed away, and the circumstances of death can vary greatly. The researchers found that in many tissues, the age of the donor was tightly linked to how they died, particularly whether they had been on a ventilator before passing. This connection created a confusing signal where it was difficult to tell if a change in splicing was due to aging itself or simply a result of the stress of the final days of life. When the researchers adjusted their analysis to account for these factors, such as the time between death and tissue collection and the use of life support, the number of genes showing age-related changes dropped dramatically. In tissues like the brain, where the circumstances of death were less linked to age, the changes remained more consistent. This indicates that while aging does alter splicing, the raw data from post-mortem samples often exaggerates the extent of these changes because of the confounding effects of how people die.

Despite this noise, a core group of genes remained stubbornly stable. The study identified nearly 600 genes that showed no sign of changing their splicing patterns with age, regardless of the tissue or the circumstances of death. These genes, which include essential housekeeping proteins involved in basic cellular functions like energy production and protein building, appear to be remarkably resilient. Their stability suggests that the aging process does not affect every part of the cell equally; some fundamental systems remain precise and unchanged even as others begin to drift.

The researchers also confirmed that their earlier observations about a specific set of enzymes involved in a metabolic pathway were not unique. One gene, AFMID, which had previously been noted for its dramatic shift toward shorter, truncated versions of its protein, was found to be among the top 5 percent of all genes most affected by aging across the entire genome. This reinforced the idea that while some genes are more vulnerable than others, the tendency to shift away from standard forms is a general rule of aging.

Ultimately, this work paints a picture of aging as a gradual loss of fidelity in the cell's editing room. The tools that manage the process begin to falter, leading to a shift away from the standard, most efficient versions of proteins. While the environment of the final days of life can muddy the waters of observation, the underlying signal is clear: the precise control of genetic instructions weakens with time, particularly in the very systems designed to maintain that control. At the same time, a small but vital group of genes holds firm, providing a glimpse of the cellular stability that persists even as the rest of the system begins to change.