Mitochondria in Neurons: Roles in Energy, Calcium, and Disease

Mitochondria do more in neurons than in almost any other cell type. Because neurons fire constantly, rarely divide, and can stretch axons over a meter long, they depend on mitochondria not just for energy but for managing calcium signals, clearing damaged organelles, and even communicating metabolically with neighboring glial cells. When any of these roles falters, the consequences show up as some of the most devastating neurological diseases we know. Understanding how mitochondria serve neurons, and how they fail, is central to understanding conditions from Alzheimer’s to ALS.

Why Neurons Are Especially Hungry for Energy

The brain accounts for roughly two percent of body weight but consumes about twenty percent of the body’s oxygen. Most of that energy goes to maintaining the electrical signals neurons use to communicate. Synaptic activity alone places enormous energy demands on local sites: ATP drives action potentials, powers the assembly and remodeling of synapses, and fuels the filling and recycling of the tiny vesicles that carry neurotransmitters across synaptic gaps.¹ This ATP comes from two routes: glycolysis, which is fast but relatively inefficient, and mitochondrial oxidative phosphorylation, which extracts far more energy per glucose molecule. Neurons lean heavily on the mitochondrial route, which is one reason these organelles are so densely packed at synaptic terminals and along axons where signals travel.

Neurons also get metabolic help from their neighbors. Astrocytes, the star-shaped glial cells that surround neurons, respond to heightened neuronal activity by ramping up their own glucose uptake and glycolysis, then releasing lactate into the surrounding space. Neurons take up that lactate and feed it into their own mitochondria for oxidative ATP production.² This arrangement, known as the astrocyte-neuron lactate shuttle, has become a widely accepted framework for brain energy metabolism.³ The implication is that neuronal mitochondria are not just processing what the neuron picks up directly from the blood; they are the final processing station for fuel delivered by an entire support network.

Getting Mitochondria Where They Need to Be

A typical motor neuron in the spinal cord can have an axon stretching from the lower back to the toes. Mitochondria manufactured in the cell body need to travel that entire distance, and they need to stop and park precisely where energy demand is highest, like synaptic terminals and nodes of Ranvier. This is not passive diffusion. Neurons have an elaborate transport system built on microtubule tracks running the length of axons and dendrites.

The molecular machinery that steers mitochondria along these tracks involves adaptor proteins from the TRAK family. TRAK1 binds to both kinesin (which moves cargo toward axon tips) and dynein (which moves cargo back toward the cell body), and it is concentrated in axons, where it supports normal axon growth. TRAK2, by contrast, predominantly interacts with dynein and is more abundant in dendrites, where it supports dendritic development.⁴ This division of labor means axons and dendrites use partly different molecular strategies to position their mitochondria, which makes sense given how different these compartments are in shape, function, and metabolic need.

Getting mitochondria to their destination is only half the problem. They also need to stop and stay put. An axon-targeted protein called syntaphilin acts as a molecular anchor: when syntaphilin is present on a mitochondrion, it locks onto microtubules and the organelle stops moving. Deleting the gene for syntaphilin in mice dramatically increases the proportion of mobile mitochondria in axons and decreases overall mitochondrial density along the axon, which in turn affects synaptic function.⁵ The balance between mobile and stationary mitochondria is itself a form of regulation: more anchored mitochondria means more local energy supply, while freeing mitochondria up lets them redistribute to new sites of demand.

Calcium Buffering and Signaling

Calcium is the most versatile signaling molecule in neurons. It triggers neurotransmitter release, activates gene expression, and modulates synaptic strength. But calcium in the wrong amount, at the wrong time, is toxic. Mitochondria act as calcium sinks, absorbing surges and releasing calcium back slowly, which prevents dangerous spikes in the cytoplasm.

The main entry point for calcium into mitochondria is the mitochondrial calcium uniporter, or MCU. In pyramidal neurons, the MCU has been confirmed as the principal route for calcium uptake during action potential firing.⁶ Calcium does not just sit passively inside mitochondria once it enters. It gets pumped back out, primarily through a sodium-calcium exchanger called NCLX.⁷ The balance between MCU-mediated uptake and NCLX-mediated efflux sets the tone of mitochondrial calcium at any given moment.⁸

When this balance tips, the consequences are serious. MCU dysfunction can lead to calcium overload inside mitochondria, which triggers oxidative stress and programmed cell death.⁹ On the other side, loss of NCLX-dependent calcium efflux promotes mitochondrial calcium overload, metabolic breakdown, oxidative stress, and cognitive decline in models of Alzheimer’s disease.¹⁰ Calcium buffering is also not purely a defensive maneuver. In the hippocampus, the way mitochondria handle calcium appears to influence synaptic plasticity itself, with structures like the mitochondrial permeability transition pore and voltage-dependent anion channels playing roles in how synapses strengthen or weaken over time.¹¹

Mitochondrial calcium handling also matters in autoimmune neurological disease. In experimental models of multiple sclerosis, neuronal MCU helps support axon survival and remyelination at metabolically stressed sites by undocking mitochondria, buffering calcium, and boosting ATP production where it is most needed.¹²

Reactive Oxygen Species and the Cost of Running Hot

Mitochondrial energy production has a byproduct: reactive oxygen species, or ROS. At low levels, ROS serve as signaling molecules, fine-tuning cellular responses to changing conditions. The rapidly changing levels of ROS in mitochondria actually make them well-suited for this kind of physiological signaling.¹³ But neurons are uniquely vulnerable when ROS production tips from helpful to harmful. Unlike most cells in the body, neurons are long-lived and do not divide, so they cannot dilute accumulated damage by splitting into daughter cells. When mitochondrial ROS production overwhelms the cell’s antioxidant defenses, the result is oxidative stress that damages proteins, lipids, and DNA, eventually triggering mitochondrial dysfunction and cell death.

This is not a purely theoretical concern. Research has shown that even under normal conditions, peroxynitrite (a reactive nitrogen species produced in mitochondria) modifies proteins in mitochondrial complex I, suggesting that neurons live under a constant low-grade oxidative burden.¹⁴ Mutations, environmental toxins, or chronic oxygen deprivation can shift this baseline burden into outright pathology. The fact that neurons are essentially non-replaceable makes this chronic oxidative exposure especially dangerous over a human lifespan.

Quality Control Through Fission, Fusion, and Mitophagy

Neurons cannot afford to keep broken mitochondria around. They maintain quality through three interconnected processes. Mitochondrial fusion merges two organelles, mixing their contents and compensating for minor damage in either one. Fission splits a mitochondrion in two, often segregating a damaged portion from a healthy one. And mitophagy is the cell’s recycling program, engulfing and digesting mitochondria that are beyond repair.

Fusion depends on proteins called mitofusins (Mfn1 and Mfn2) on the outer membrane; deleting either one disrupts normal mitochondrial shape and movement. The major mitophagy pathway involves two proteins whose names have become nearly synonymous with Parkinson’s disease research: PINK1 and parkin. Together, these proteins recognize and tag damaged mitochondria for destruction.¹⁵ When this cleanup system fails, damaged mitochondria accumulate, leak ROS, and contribute to neurodegeneration. Defective PINK1-parkin mitophagy has been linked not only to Parkinson’s but also to prion disease, where deficiency in this pathway accelerates the buildup of dysfunctional organelles.

Mitochondrial Dysfunction in Neurodegenerative Disease

The threads described above, energy supply, calcium handling, ROS, and quality control, converge in neurodegenerative disease. Rather than a single defect, these diseases typically involve multiple mitochondrial systems going wrong at once.

In Parkinson’s disease, the progressive loss of dopamine-producing neurons in the brain is closely tied to mitochondrial dysfunction that appears early in the disease process. The abnormalities span the spectrum: impaired electron transport chain function, altered mitochondrial shape and dynamics, mitochondrial DNA mutations, and disrupted calcium handling. These features appear in both inherited and sporadic forms of the disease.¹⁶

In Alzheimer’s disease, amyloid-beta, one of the hallmark proteins of the condition, directly interacts with a mitochondrial structure called the permeability transition pore (mPTP). When amyloid-beta promotes mPTP formation through interaction with cyclophilin D, the result is a cascade of harm: excessive ROS generation, collapse of mitochondrial membrane potential, calcium dysregulation, reduced respiration, and release of factors that trigger cell death.¹⁷

ALS, typically framed as a disease of motor neurons, also has a strong mitochondrial dimension. The protein TDP-43, which misfolds and accumulates in the cytoplasm of motor neurons in most ALS cases, directly disrupts mitochondrial dynamics. Overexpression of TDP-43, especially its ALS-associated mutant forms, reduces mitochondrial length and density in neurites. Paradoxically, suppressing TDP-43 increases mitochondrial length and density but still impairs mitochondrial movement, suggesting TDP-43 has a specific regulatory role in mitochondrial dynamics that is lost when the protein is either too abundant or too scarce.¹⁸

SARM1 and the Pathway to Axon Destruction

When an axon is injured or metabolically stressed, it does not simply wither passively. It self-destructs through a regulated program, and a protein called SARM1 sits at the center of that process. SARM1 is an enzyme that consumes NAD+, a molecule essential for energy metabolism. Once SARM1 is activated, the resulting NAD+ depletion triggers a cascade: ATP levels collapse, mitochondrial movement halts, mitochondria depolarize, calcium floods in, and the axon breaks apart.¹⁹ The rapid ATP collapse after SARM1 activation is not fully explained by NAD+ loss alone, suggesting SARM1 may also degrade ATP directly and cause mitochondrial dysfunction through additional mechanisms.²⁰

What makes this story especially interesting is the feedback loop between SARM1 and mitochondria. In a rat model of Charcot-Marie-Tooth disease type 2A (a hereditary neuropathy caused by mutations in mitofusin 2), dysfunctional mitochondria activate SARM1, and activated SARM1 then feeds back to make the mitochondrial damage worse. When researchers deleted the Sarm1 gene in these rats, the result was dramatic: axon degeneration, neuromuscular junction abnormalities, and muscle wasting were all rescued, and even the mitochondrial defects themselves, including problems with number, size, and cristae structure, were substantially suppressed.²¹ This positions SARM1 inhibition as a potential therapeutic strategy not just for Charcot-Marie-Tooth but for other neurodegenerative diseases where mitochondrial pathology is prominent.

Mitochondria During Brain Development

Mitochondria are not passive occupants that show up after a neuron is built. They play active roles during the construction phase. As neurons grow and extend their processes, mitochondrial mass increases primarily through biogenesis, the creation of new organelles, and this increase correlates with periods of active neurite outgrowth and synapse formation.²²

Dendritic spines, the tiny protrusions where most excitatory synapses form, show a particularly telling relationship with mitochondria. During the active period of spine and synapse development in cultured neurons, roughly eight to nine percent of dendritic mitochondria are found at least partially inside protrusions. As development matures and synaptogenesis slows, this fraction drops to about four percent.²³ The timing strongly suggests that mitochondria are actively involved in building and stabilizing new synaptic connections, not just powering them after the fact.

When Astrocytes Donate Their Own Mitochondria

One of the more surprising discoveries in recent neuroscience is that astrocytes can donate functional mitochondria directly to neurons. In mouse models of stroke, astrocytes released mitochondria that entered adjacent neurons, amplifying cell survival signals.²⁴ This is not a one-way street. Bidirectional mitochondrial transfer between astrocytes and neurons, including at axon terminals, has been observed through specialized extracellular vesicles containing mitochondria. In an Alzheimer’s disease mouse model, this transfer was significantly reduced. When healthy mitochondria-containing vesicles were supplied to affected neurons, they helped mobilize abnormal lipid droplet accumulation, restored fatty acid oxidation, and increased neuronal energy output, including at synapses.²⁵

The idea that mitochondria travel between cells to rescue metabolic function adds a new dimension to how we think about brain resilience. It also suggests that neurodegenerative diseases may partly involve a breakdown in this intercellular support system, not just internal mitochondrial failure within individual neurons.

SIRT3 and Exercise-Driven Protection

Inside mitochondria, a protein called SIRT3 acts as a kind of maintenance enzyme, removing chemical tags (acetyl groups) from other mitochondrial proteins to keep them functioning properly. Neurons lacking SIRT3 are markedly more sensitive to calcium overload, oxidative stress, and excitotoxicity, the kind of damage caused by excessive stimulation with the neurotransmitter glutamate. In mouse models of Huntington’s disease and epilepsy, SIRT3 deficiency increased the vulnerability of specific neuron populations.²⁶

The practical angle here is exercise. Running-wheel exercise in mice increased SIRT3 expression in hippocampal neurons, and this increase was essential for maintaining proper mitochondrial protein function and the neuroprotective effects of the exercise. In other words, one reason physical activity protects the brain may be that it boosts this specific mitochondrial maintenance pathway. SIRT3 has been identified as a potential therapeutic target for several types of brain injury, including stroke and traumatic brain injury.²⁷

Therapeutic Directions and Their Limits

Given how central mitochondrial dysfunction is to neurological disease, researchers have developed antioxidant compounds designed to accumulate specifically inside mitochondria. These mitochondria-targeted antioxidants aim to reduce oxidative stress at its source without flooding the entire cell with antioxidants, which can interfere with useful ROS signaling.²⁸ One such compound, MitoQ, prevented loss of spatial memory and reduced early neuropathology in a transgenic mouse model of Alzheimer’s disease, providing early evidence that targeting mitochondria therapeutically could slow neurodegeneration.²⁹

Translating these results to human patients remains a challenge. The blood-brain barrier limits drug delivery, neurons are notoriously difficult to reach with systemic treatments, and the sheer complexity of mitochondrial dysfunction in neurological disease means that targeting one pathway (say, oxidative stress) may not help much if the core problem is in calcium handling or transport. The SARM1 story offers a complementary approach: rather than fixing mitochondria directly, blocking the destruction pathway that dysfunctional mitochondria trigger. But SARM1 inhibitors are still in early development, and it remains unclear whether blocking axon degeneration in a chronic disease context might have unintended consequences.

New Tools for Watching Mitochondria in Real Time

Part of what has accelerated this field is the development of genetically encoded sensors that can report on mitochondrial function in living neurons. Researchers have engineered biosensors that simultaneously track extracellular and cytosolic ATP alongside calcium changes in real time. Using these tools, they demonstrated that when axons are cut, ATP is released from the injured neuron and diffuses to surrounding cells, while calcium surges propagate not only in the damaged neuron but in its intact neighbors.³⁰ Techniques like these are giving researchers a far more detailed view of how mitochondrial dysfunction unfolds in real time, moving beyond the snapshots that traditional microscopy provides. Detailed 3D reconstructions of mitochondrial shape in brain tissue have also revealed that the geometry of a mitochondrion, its curvature, surface area, and volume, can alter its metabolic output, suggesting that shape is not just a passive consequence of the cellular environment but a functional variable in its own right.³¹