Alzheimer’s disease is defined by familiar landmarks: amyloid plaques outside neurons and tangles of tau protein within them. A study published July 29 in Nature Neuroscience adds a different structure to that map. Researchers report dense accumulations of mitochondria inside damaged neuronal processes, which they call “mitochondrial plaques.”
The name is striking, but it should not be mistaken for a new clinical diagnosis. The work is primarily mechanistic pathology in genetically engineered mice, supported by observations in a small set of postmortem human brains. It does not show that mitochondrial plaques cause Alzheimer’s disease, predict symptoms in living people or provide a treatment target ready for patients.
What the researchers found
Mitochondria produce much of a cell’s usable energy, while mitophagy is the quality-control process that helps dismantle damaged mitochondria. The team used an APP/PSEN1 Alzheimer’s mouse model carrying a fluorescent reporter that changes signal when mitochondria enter acidic cellular compartments. This allowed the researchers to follow both mitochondrial buildup and attempted disposal across age.
They found large, rounded clusters containing both acidic and neutral mitochondria within neuronal projections. Lysosomes—the compartments that break down cellular material—were recruited to the clusters, but degradation remained incomplete. Some mitochondrial plaques developed alongside amyloid and formed mixed structures; others appeared separately, including at earlier disease stages in the model.
The researchers also identified similar structures in a second Alzheimer’s mouse model, 5xFAD. In postmortem hippocampal tissue, mitochondrial plaques were reported in four people with Alzheimer’s disease and compared with four controls. The paper’s extended data describe 222 plaques measured across the four Alzheimer’s cases. That human evidence confirms that the structures are not limited to one engineered mouse line, but the sample is far too small to establish prevalence, timing or clinical meaning.
TENS analysis: three questions, not one answer
The finding is best read through three separate questions: Is the structure real, what process produces it, and does changing that process improve human health? The study offers evidence for the first question in mouse and postmortem human tissue, a detailed working model for the second in mice, and no answer yet to the third.
Keeping those questions separate matters because a pathological structure can be a cause, a consequence or a containment response. Lysosomes gathering around damaged mitochondria could mark a failed cleanup system, but they could also reflect cells attempting to limit damage already produced by other Alzheimer’s processes. The presence of some mitochondrial plaques without nearby amyloid is suggestive about timing; it does not establish that mitochondrial failure starts the disease.
This distinction creates a two-clock problem. Researchers need to determine when mitochondrial plaques emerge in the biological course of Alzheimer’s and when they become detectable relative to memory loss in people. Postmortem tissue freezes one late moment. Longitudinal human biomarkers would be required to connect those clocks.
How this changes the mitophagy story
The new work extends an older line of research rather than replacing the amyloid and tau framework. A 2019 Nature Neuroscience study led by several of the same investigators reported impaired mitophagy in Alzheimer’s patient hippocampal samples, human neurons grown from induced pluripotent stem cells and multiple animal models. Stimulating mitophagy reduced amyloid or tau pathology and improved memory measures in worms and mice.
The National Institute on Aging summarized that earlier result as evidence from animal models and laboratory specimens of human neurons, not as proof of a therapy in people. The 2026 study adds spatial anatomy: it proposes where failed mitochondrial disposal can accumulate and how lysosomal dysfunction may turn a quality-control response into a persistent structure.
That synthesis sharpens the next experiments. A useful biomarker would need to distinguish mitochondrial plaques from ordinary mitochondrial turnover, appear before irreversible tissue loss, and track with disease progression independently of established amyloid and tau measures. A useful intervention would need to restore disposal without damaging healthy mitochondria or disrupting other lysosomal work. Neither standard has been met.
Evidence level and limitations
This is peer-reviewed, multi-model laboratory research with limited postmortem human confirmation. It is not a clinical trial, a study of living patients or evidence that enhancing mitophagy prevents dementia. Mouse models reproduce selected features of inherited or amyloid-driven disease and do not capture the full complexity of late-onset Alzheimer’s in humans.
The human comparison included only four Alzheimer’s cases and four controls, leaving major questions about age, disease stage, genetic background and other brain conditions. Postmortem measurements also cannot show whether the plaques grew before symptoms, followed degeneration or fluctuated over time. The study authors declared no competing interests, and the work was supported in part by the National Institute on Aging and other NIH grants.
The durable contribution is therefore a new object to investigate, not a new longevity promise. Mitochondrial plaques may help researchers connect energy failure, impaired cellular recycling and established Alzheimer’s pathology. Their importance will depend on whether larger human studies can turn a vivid structure into a reproducible marker—and whether experiments can show that altering it changes disease rather than merely changing the picture.
Sources: Nature Neuroscience; National Institute on Aging.
TENS Magazine conceptual illustration
