A mouse that has forgotten how to navigate a maze it once knew is, in neuroscience terms, a model of what dementia does to a human mind. The animal still has its neurons. The wiring is largely intact. But something inside those cells has gone dark. In a study published in Nature Neuroscience in early 2025, a research team showed it could flip that switch back on, restoring lost memories in dementia-model mice by directly recharging the mitochondria inside their neurons.
The tool behind the rescue is an engineered receptor called mitoDREADD-Gs. Unlike conventional chemogenetic receptors, which sit on a neuron’s outer surface and influence the whole cell, this one is designed to embed in the mitochondrial membrane, the double-layered shell around the organelle that generates most of a cell’s energy. When activated by a small-molecule drug, mitoDREADD-Gs triggers a targeted signaling chain (Gs protein to protein kinase A) that ramps up mitochondrial energy output without setting off the cascade of side effects a surface receptor would cause.
The result, across multiple mouse models, was striking. Animals that had lost the ability to form and recall memories regained performance on standard learning and memory tasks that matched healthy controls. The work represents a shift in how scientists think about neurodegeneration: not just as irreversible damage, but potentially as an energy crisis that can be reversed if the right cellular machinery is reactivated.
Why mitochondria matter in dementia
Mitochondrial dysfunction has been implicated in Alzheimer’s disease and related dementias for decades. Neurons are among the most energy-hungry cells in the body, and when their mitochondria falter, synapses weaken, signaling slows, and memories fail to consolidate. Most existing Alzheimer’s therapies, including the recently approved antibodies lecanemab and donanemab, target amyloid plaques, the protein clumps long considered the disease’s primary driver. But a growing body of research suggests that energy failure inside neurons may be an equally important, and perhaps more treatable, piece of the puzzle.
The mitoDREADD-Gs study builds on more than a decade of preclinical evidence. Earlier work using MitoQ, a mitochondria-targeted antioxidant, showed that reducing oxidative stress inside mitochondria slowed memory loss and extended lifespan in 3xTg-AD mice, a triple-transgenic Alzheimer’s model that develops amyloid plaques, tau tangles, and synaptic damage. A separate study in the same mouse line found that MitoQ preserved spatial memory and reduced early neuropathology.
Those antioxidant experiments established that protecting mitochondria could slow cognitive decline. The new chemogenetic approach asks a bolder question: once neurons are already failing, can you restart their engines and bring cognition back? In the mouse models tested, the answer was yes.
How the experiments worked
The research team tested mitoDREADD-Gs in two distinct scenarios. In one, they used pharmacological agents to impair cognition in otherwise healthy mice, creating an acute model of brain energy failure. In the other, they worked with transgenic mice engineered to develop progressive neurodegeneration over time. Both models are standard in the field, though neither perfectly replicates the full complexity of human Alzheimer’s disease.
In both cases, activating the receptor normalized mitochondrial respiration, the process by which mitochondria convert nutrients into ATP, the cell’s energy currency. Cognitive deficits measured through maze navigation, object recognition, and other behavioral assays reversed. A research highlight in Nature Biotechnology described the work as a proof of concept for using engineered receptors to control metabolism inside specific organelles, framing the key question not as whether mitochondria matter in dementia (that is well established) but whether restoring their function after damage has set in can recover lost cognition.
What could complicate the picture
Mouse results, even compelling ones, carry a long list of caveats on the road to human relevance. No published data yet show how human neurons, or neurons derived from human stem cells, respond to mitoDREADD-Gs. Mouse brains are orders of magnitude smaller than human brains, metabolize drugs differently, and do not develop the full spectrum of pathology seen in Alzheimer’s patients. Translating a chemogenetic tool that requires viral delivery of an engineered gene into a clinical therapy would demand years of safety, dosing, and manufacturing work that has not started.
There is also a technical wrinkle specific to DREADD systems. These receptors are typically activated by clozapine-N-oxide (CNO), a small molecule assumed to be pharmacologically inert. But studies have shown that CNO can convert back into clozapine, an antipsychotic with its own receptor-binding activity, inside living animals. That means behavioral improvements in a DREADD experiment could theoretically reflect off-target drug effects rather than on-target mitochondrial activation. Newer DREADD studies have increasingly adopted alternative ligands like deschloroclozapine (DCZ) to sidestep this problem, but whether the mitoDREADD-Gs team used CNO, DCZ, or another activator matters for interpreting the strength of their controls.
Long-term safety data are absent as well. Chronically boosting mitochondrial energy production through PKA signaling could, in theory, alter gene expression, increase reactive oxygen species at high activity levels, or disrupt the balance between mitochondrial fusion and fission, processes that neurons depend on to distribute energy along their axons. The study demonstrates acute cognitive rescue over the span of its experimental timeline, not durable disease modification over months or years.
One open question worth tracking: whether this kind of energy rescue works better in early-stage disease than in advanced neurodegeneration. Mitochondria in neurons experiencing early tau pathology likely retain more spare respiratory capacity, the reserve energy they can produce above baseline demand. Once widespread fragmentation and structural remodeling of the mitochondrial inner membrane (cristae) have occurred, there may be less functional machinery left to recharge. The published data do not yet parse that distinction.
Where this fits in the search for dementia treatments
More than 55 million people worldwide live with dementia, according to the World Health Organization, and that number is projected to nearly triple by 2050. The approved amyloid-targeting antibodies have shown modest effects on cognitive decline in clinical trials but come with significant side effects, including brain swelling and microbleeds, and work only in patients with confirmed amyloid pathology. The field has been searching for complementary approaches that address other aspects of the disease.
Mitochondrial-targeting strategies represent one such avenue. Beyond MitoQ and mitoDREADD-Gs, researchers are exploring NAD+ precursors (like nicotinamide riboside), mitophagy enhancers that help cells clear damaged mitochondria, and small molecules that stabilize mitochondrial membrane potential. None of these has yet succeeded in a large human trial for dementia, but the collective weight of preclinical evidence has made mitochondrial biology one of the more active frontiers in neurodegeneration research.
What makes the mitoDREADD-Gs result notable is not just that it worked, but what it implies: that neurons deep into an energy crisis may not be beyond rescue. If that principle holds in human tissue, it could reshape how clinicians think about the therapeutic window for dementia, the period during which intervention can still make a difference. For now, that remains an “if” supported by mouse data and a strong mechanistic rationale, not clinical proof. But it is a more hopeful “if” than the field has had in some time.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
