New Form of Brain Cell Death May Explain How Dementia Destroys Neurons

Inside a dying brain cell, the first thing to go is the nucleus. Not the whole cell at once, not some tidy programmed shutdown, but the nucleus itself: the membrane around it buckles, the smooth sphere puckers and shrivels like fruit left too long in a bowl, and then it falls apart. Researchers at King’s College London have a name for this. They call it karyoptosis, and they think it accounts for a great deal of the neuronal loss that has long been pinned, rather vaguely, on dementia.

That word, karyoptosis, has been kicking around the lab for a decade. What is new is where they have now found it.

For a long time the picture went something like this. In Alzheimer’s disease, in frontotemporal dementia (FTD), in motor neurone disease, toxic proteins pile up inside neurons until the cells give out. Fine. But how, exactly? Apoptosis, the orderly self-destruct sequence cells use when they are damaged or no longer needed, was the obvious candidate, and for years it took most of the blame. The trouble is that apoptosis never quite added up. Mature neurons are unusually resistant to it, and even when you tot up all the apoptotic deaths you can find, you are left with a stubborn surplus of dead cells that died some other way.

Karyoptosis, the King’s team argue, is at least part of that other way.

The work, published in Nature Communications, started in cells in a dish. When the researchers jammed the cell’s protein-clearing machinery (the autophagy-lysosome system, which acts as a kind of cellular recycling depot), proteins built up and the nuclear lamina began to come apart. The lamina is a mesh of proteins that lines the inside of the nucleus and holds its shape; one of its key components is a protein called LaminB1. As LaminB1 destabilised, the nucleus lost its round profile, shrank, and started shedding bits of itself, packaged into little bubbles, into the space outside the cell. DNA damage followed. But, crucially, it followed: the genetic harm came after the nuclear collapse, not before it. That ordering matters, because it separates karyoptosis from the slow decline of cellular senescence, which it superficially resembles.

“The death and loss of cells in the brain drives many symptoms experienced by people living with dementia,” says Rebecca Casterton, a senior researcher at the UK Dementia Research Institute at King’s and the paper’s first author. “We have started to lay out the road map of how karyoptosis works.”

Finding the Switch

A road map is only useful if it points somewhere, and here it points at a protein called p38 MAP kinase. Kinases are the cell’s switches, flicking other proteins on and off by tacking phosphate groups onto them, and the team found that p38 phosphorylates LaminB1 at a specific spot, destabilising it and tipping the nucleus toward ruin. When they blocked p38 with a drug, or engineered neurons so that LaminB1 could no longer be switched at that particular site, the nuclear damage eased off. They saw the rescue in rat neurons, then in fruit flies carrying the genetic fault behind one inherited form of ALS and FTD, and finally in human neurons grown from stem cells. Across all three, dialling down that one molecular interaction kept nuclei intact and cells alive longer.

“By specifically targeting the interaction between p38 MAP kinase and LaminB1 we may slow down the process of cell death, buying time for more pinpointed therapies against specific neurodegenerative diseases,” says Manolis Fanto, who led the work.

None of which would mean much for human patients if karyoptosis stayed in the dish. So the team went to the brain bank. They took post-mortem frontal cortex from people who had died with Alzheimer’s or frontotemporal lobar degeneration, plus tissue from people of similar age who had neither, and fed thousands of individual cells through a machine-learning sorting method that knew nothing about which brain each cell came from. The algorithm, blind, picked out clusters of cells bearing the shrivelled, low-circularity nuclei that mark karyoptosis. When the labels were finally revealed, those clusters were heavily overrepresented in the diseased brains. Roughly 35 per cent of cells from the frontal cortex of Alzheimer’s patients showed signs of karyoptosis, against about 15 per cent in healthy aged controls.

That 15 per cent in the healthy brains is worth dwelling on, mind you. It hints that karyoptosis is not purely a disease phenomenon; some of it seems to come along with ordinary ageing, with disease piling an extra ~18 to 20 per cent on top. And there are caveats the authors are upfront about. They have mapped the mechanism mostly in simplified models, the exact role of those expelled nuclear bubbles is still murky, and karyoptosis is certainly not the only way a neuron can die. Apoptosis and regulated necrosis are still in the mix.

Buying Time

Still, the strategic appeal is obvious, and it is a slightly different kind of appeal from most dementia research. A great deal of effort goes into the upstream causes, the toxic proteins themselves, the tangles and clumps. Karyoptosis sits downstream, at the moment of death, which means a drug aimed at it would not cure the disease so much as keep neurons breathing while other, more targeted treatments got to work on the root. Widen that window, the thinking goes, and everything else becomes more feasible.

“For decades, we’ve known that toxic proteins build up in Alzheimer’s disease and frontotemporal dementia, but exactly how they lead to the loss of brain cells has remained unclear,” says Sara Rodrigues of Alzheimer’s Research UK, which funded much of the work. She reckons the find could “help widen the window for therapies that tackle the underlying causes of disease.”

Whether a p38 blocker can be made selective and safe enough for human brains is the next question, and not a small one; kinases tend to have their fingers in many pies. But the cell biology has shifted. The shrivelling nucleus, once just a melancholy detail in a slide of dying tissue, now looks like something you might be able to stop.

DOI / Source: https://doi.org/10.1038/s41467-026-73802-w

Frequently Asked Questions

How does karyoptosis actually kill a brain cell?

It starts at the nucleus rather than the cell as a whole. Toxic protein build-up destabilises the nuclear lamina, the protein scaffold that holds the nucleus in shape, so the nucleus shrivels, sheds fragments of itself, and disintegrates, with DNA damage following afterward rather than triggering the collapse. That sequence is what sets it apart from other known forms of cell death, and it is exactly where researchers now hope to intervene.

Why does this matter for dementia treatment when it does not cure the disease?

Most dementia research targets the upstream causes, the toxic proteins themselves, whereas karyoptosis happens downstream at the moment a neuron dies. A drug that slowed it would not remove the underlying disease but could keep neurons alive longer, widening the window for other therapies aimed at the root cause. Buying that time may be what makes more targeted treatments viable in the first place.

Is it true that this kind of cell death happens in healthy brains too?

To some degree, yes. Around 15 per cent of cells in healthy aged brains showed karyoptosis markers, against roughly 35 per cent in Alzheimer’s tissue, suggesting the process is partly a feature of normal ageing that disease then amplifies. How much of ordinary age-related brain decline it accounts for is still an open question.

Could blocking p38 kinase become a real therapy?

That is the hope, since switching off the p38 interaction with LaminB1 kept neurons intact across rat, fly and human stem-cell models. The catch is that kinases are involved in many processes throughout the body, so any drug would need to be selective and safe enough for long-term use in the brain. Getting that balance right is the next major hurdle.