Targeting mitophagy for neuroprotection: mechanisms and therapeutic opportunities

#1

https://www.nature.com/articles/s41514-026-00424-3

chatGPT(5.5paid):

Overall assessment

This is a narrative review, not an experimental paper. It surveys how mitophagy contributes to neuronal maintenance and neurodegeneration, with particular emphasis on the idea that mitophagy is bidirectional: insufficient clearance permits damaged mitochondria to accumulate, whereas excessive clearance may deplete the neuronal mitochondrial pool and cause energetic failure. It then reviews potential therapeutic approaches, mainly for Alzheimer’s and Parkinson’s diseases. The uploaded document is an unedited accepted manuscript in npj Aging.

Summary

1. Why mitophagy matters in neurons

Neurons are unusually dependent on mitochondrial quality because they:

  • have high and continuous ATP requirements;
  • need mitochondria at distant axons, dendrites and synapses;
  • depend on mitochondria for calcium buffering;
  • are long-lived and generally cannot dilute damaged organelles through cell division.

Mitophagy removes dysfunctional mitochondria through the autophagosome–lysosome system. The authors argue that this supports:

  • ATP production;
  • control of mitochondrial ROS;
  • calcium homeostasis;
  • synaptic function;
  • suppression of innate immune signalling;
  • maintenance of the mitochondrial network.

A useful point made by the paper is that dysfunctional mitochondria may not merely produce less ATP; severely depolarised mitochondria can allow ATP synthase to run in reverse and consume ATP.

2. Mitophagy pathways

The review divides mitochondrial clearance into two main groups.

Ubiquitin-dependent mitophagy

The central pathway is PINK1–Parkin:

  1. A fall in mitochondrial membrane potential prevents normal PINK1 import and degradation.
  2. PINK1 accumulates on the outer mitochondrial membrane.
  3. PINK1 activates and recruits Parkin.
  4. Parkin ubiquitinates outer-membrane proteins such as mitofusins and Miro.
  5. OPTN and NDP52, assisted by TBK1 phosphorylation, connect ubiquitinated mitochondria to LC3-positive phagophores.
  6. The mitochondrion is engulfed and ultimately degraded by lysosomes.

The review also describes Parkin-independent ubiquitin systems involving:

  • MUL1;
  • RNF185;
  • AMBRA1–HUWE1;
  • ubiquitination following outer-membrane rupture and exposure of inner-membrane proteins.

Receptor-mediated mitophagy

Mitochondrial proteins containing LC3-interacting regions can recruit autophagosomes without requiring Parkin-mediated ubiquitination. These include:

  • BNIP3;
  • NIX/BNIP3L;
  • FUNDC1;
  • BCL2L13;
  • FKBP8;
  • PHB2.

NIX is highlighted because its overexpression can restore mitophagy in some PINK1- or Parkin-deficient patient-derived neurons. FUNDC1 is also linked to hypoxic responses, mitochondrial–ER contacts and calcium signalling.

The pathway table on pages 7–8 is one of the more useful parts of the review because it places canonical and compensatory pathways side by side.

3. Mitophagy as a double-edged process

The central conceptual argument is that neuronal survival requires an optimal rate and completion of mitophagic flux, rather than indiscriminate stimulation.

Too little mitophagy

Insufficient clearance causes:

  • accumulation of depolarised mitochondria;
  • ATP deficiency;
  • ROS generation;
  • release of mitochondrial DNA;
  • activation of NLRP3 and cGAS–STING inflammatory pathways;
  • synaptic dysfunction;
  • increased apoptosis and possibly ferroptosis.

Too much mitophagy

Excessive mitophagy may cause:

  • loss of still-functional mitochondria;
  • reduced mitochondrial DNA copy number;
  • ATP depletion;
  • excessive fission;
  • degradation of fusion proteins such as Mfn2;
  • collapse of axonal and synaptic energy supply.

Figure 1 on page 9 illustrates this proposed continuum: basal mitophagy supports homeostasis, deficient mitophagy permits damaged mitochondria to accumulate, while excessive mitophagy removes too much mitochondrial mass.

4. Disease-specific discussion

Alzheimer’s disease

The review proposes reciprocal interactions among amyloid-β, tau and mitochondrial quality control:

  • Aβ may suppress PINK1/Parkin recruitment and reduce mitophagy receptors.
  • Damaged mitochondria produce ROS that can worsen Aβ and tau pathology.
  • Tau may impair mitochondrial transport and mitophagic completion.
  • Some tau fragments may instead cause inappropriate Parkin recruitment and excessive mitochondrial removal.
  • Defective lysosomal clearance can leave mitophagy initiated but incomplete.

The resulting model is not simply “too little mitophagy”; different disease stages or cellular compartments may show blocked, incomplete or excessive activity.

Parkinson’s disease

The PD section centres on:

  • PINK1 and PRKN mutations;
  • α-synuclein interference with TOM20-mediated mitochondrial protein import;
  • complex I dysfunction;
  • defective Miro degradation and mitochondrial arrest;
  • loss of Mfn2- and OPA1-dependent network organisation;
  • mtDNA-triggered cGAS–STING inflammation;
  • links between impaired mitophagy and lipid peroxidation or ferroptosis.

The review also emphasises compensatory non-Parkin pathways, particularly NIX-mediated clearance.

Huntington’s disease

Mutant huntingtin is presented as disrupting several stages simultaneously:

  • mitochondrial fission and transport;
  • autophagosome formation;
  • OPTN–Rab8-dependent trafficking;
  • lysosomal completion.

At the same time, VCP or Rhes–NIX signalling may cause excessive mitochondrial removal. The authors therefore argue that global autophagy activation, including rapamycin, could be ineffective or harmful when downstream transport and lysosomal fusion remain blocked.

ALS

The review discusses:

  • TDP-43 and SOD1 toxicity;
  • Drp1-dependent fragmentation;
  • OPTN and TBK1 defects;
  • impaired axonal mitochondrial transport;
  • possible harm from excessive autophagy in some TDP-43 models.

It proposes restoring mitochondrial transport and flux rather than simply increasing autophagy initiation.

5. Therapeutic approaches

The review covers a broad collection of compounds and interventions, including:

  • resveratrol;
  • urolithin A;
  • melatonin;
  • β-asarone;
  • kaempferol;
  • aloe-emodin;
  • rhapontigenin;
  • UMI-77;
  • PINK1/Parkin modulators;
  • SIRT/AMPK/PGC-1α-related interventions;
  • Drp1 inhibition;
  • exercise and dietary interventions;
  • potential gene therapies.

The therapeutic message is that interventions should ideally:

  1. identify whether the disease context involves deficient, excessive or incomplete mitophagy;
  2. preserve functional mitochondria;
  3. restore transport and lysosomal completion;
  4. coordinate clearance with mitochondrial biogenesis.

What is novel?

Because this is a review, it does not present a new experiment or discovery. Its novelty lies mainly in how it organises and interprets the literature.

1. Strong emphasis on bidirectionality

The most distinctive feature is the sustained argument that “enhancing mitophagy” is not universally beneficial. The paper gives substantial space to pathological overactivation, mitochondrial depletion and excessive fission, rather than treating mitophagy deficiency as the sole problem.

This is not an entirely new concept, but it is less commonly foregrounded in therapeutic mitophagy reviews.

2. Integration of clearance with neuronal geography

The review usefully treats mitophagy as a spatial problem. In neurons, initiation at a synapse is insufficient if damaged mitochondria or mitophagosomes cannot undergo retrograde transport to lysosome-rich regions of the soma. The discussion of SNAPIN, Miro, OPTN, Rab8 and axonal transport helps move the analysis beyond simple measurements of LC3 or Parkin.

3. Attention to alternative and compensatory pathways

Rather than equating mitophagy with PINK1–Parkin, the authors bring together:

  • receptor-mediated mechanisms;
  • alternative E3 ubiquitin ligases;
  • inner-membrane receptors such as PHB2;
  • compensatory NIX signalling.

This is especially relevant because basal mitophagy in many mammalian tissues is not fully dependent on Parkin.

4. Disease- and stage-specific therapeutic framing

The review argues that the same intervention might be beneficial at one stage and harmful at another. For example, increasing initiation may be inappropriate where lysosomal completion or axonal transport is already blocked.

That “flux-matching” interpretation is more sophisticated than a simple activator-versus-inhibitor classification.

Critique

1. The review is not systematic

The authors do not clearly describe:

  • databases searched;
  • search terms;
  • inclusion or exclusion criteria;
  • date limits;
  • study-quality assessment;
  • procedures for resolving conflicting evidence.

Consequently, the selection of studies may be subjective. It is impossible to know whether the cited examples are representative or were chosen because they fit the proposed bidirectional model.

A formal systematic or scoping-review method would have considerably strengthened the paper.

2. The evidence hierarchy is poorly controlled

Cell culture observations, toxin-treated rodents, transgenic models, patient-derived neurons and human clinical studies are frequently discussed in close succession without enough separation.

This creates a risk that:

  • a change in LC3-II in cultured cells;
  • improved behaviour in a mouse model; and
  • clinical neuroprotection in humans

appear to be successive points on the same evidential continuum. They are not.

Most therapeutic claims remain preclinical, and several compounds described as promising have:

  • uncertain brain exposure;
  • pleiotropic mechanisms;
  • doses that may not be achievable in humans;
  • no demonstration that mitophagy is the principal mediator of benefit.

3. Mitophagy initiation is often conflated with completed flux

This is perhaps the most important technical weakness.

Changes in PINK1, Parkin, LC3-II, p62 or mitochondrial colocalisation do not by themselves prove increased mitochondrial degradation. They may indicate:

  • greater initiation;
  • blocked autophagosome turnover;
  • impaired lysosomal degradation;
  • accumulation of intermediates.

The paper recognises transport and lysosomal failure conceptually, but it does not consistently apply that caution when evaluating individual studies or compounds.

A stronger review would distinguish:

  • damage recognition;
  • ubiquitination or receptor activation;
  • sequestration;
  • transport;
  • lysosomal fusion;
  • actual mitochondrial degradation;
  • compensatory mitochondrial biogenesis.

4. “Excessive mitophagy” is insufficiently defined

The paper correctly identifies the lack of a damage threshold, but it sometimes labels mitochondrial loss as excessive mitophagy without fully excluding other causes, such as:

  • inhibited biogenesis;
  • direct mitochondrial toxicity;
  • apoptosis-associated mitochondrial destruction;
  • proteasomal degradation of dynamics proteins;
  • nonspecific bulk autophagy;
  • failure of fusion or replication.

To establish pathological excessive mitophagy, studies ideally need to demonstrate increased completed flux together with inappropriate removal of functionally competent mitochondria. That level of evidence is uncommon.

5. Membrane-potential selection deserves more precise treatment

The review presents ΔΨm loss as the key trigger for PINK1 accumulation, which is broadly correct for experimentally induced depolarisation. However, it does not explore several important qualifications:

  • basal neuronal mitophagy can occur without dramatic global depolarisation;
  • local mitochondrial damage may be heterogeneous;
  • PINK1–Parkin behaviour in neurons differs from that in commonly used immortalised cells;
  • mild loss of ΔΨm may be reversible and not necessarily justify organelle destruction;
  • mitochondrial fission can segregate damaged subdomains rather than condemn an entire mitochondrion.

The paper identifies the unresolved “damage threshold” but does not develop a quantitative or mechanistic framework for it.

6. Some pathway descriptions are simplified or internally inconsistent

Early passages and figure legends emphasise p62 and NBR1 as major adaptors, whereas the main text later correctly gives greater prominence to OPTN and NDP52 in PINK1–Parkin mitophagy. This could leave readers with an outdated or oversimplified hierarchy.

Similarly, proteins such as BNIP3 and NIX have functions beyond selective mitophagy, including effects on membrane permeabilisation and cell death. In some cited injury models, neuronal death attributed to “excessive mitophagy” might partly reflect direct BNIP3/NIX toxicity.

7. Correlation and causality are sometimes blurred

The review frequently presents disease-associated alterations in mitophagy proteins as pathogenic drivers. In many cases they may instead be:

  • compensatory responses;
  • downstream consequences of neuronal injury;
  • markers of altered mitochondrial abundance;
  • secondary effects of inflammation or protein aggregation.

For example, lower receptor abundance in late-stage diseased tissue does not establish that receptor loss initiated the disease. Longitudinal and intervention studies are needed.

8. Disease heterogeneity is underdeveloped

AD, PD, HD and ALS are each highly heterogeneous. The paper sometimes discusses each disease as though it has a single mitophagy phenotype.

In reality, phenotypes are likely to vary with:

  • causal mutation;
  • sporadic versus familial disease;
  • neuronal subtype;
  • glial involvement;
  • disease stage;
  • brain region;
  • metabolic state;
  • age;
  • treatment history.

This matters because the paper’s proposed precision approach requires exactly this stratification.

9. Glial and intercellular mitochondrial biology receive too little attention

The review is neuron-centred, but neurodegeneration involves astrocytes, microglia, oligodendrocytes and vascular cells. Mitophagy in these populations can influence:

  • neuroinflammation;
  • lactate and metabolic support;
  • myelination;
  • removal of extracellular mitochondria;
  • intercellular mitochondrial transfer.

A therapeutic intervention that improves neuronal mitophagy but activates harmful microglial responses may not be neuroprotective overall.

10. The therapeutic section is too catalogue-like

The long list of natural products gives breadth but not enough discrimination. A more useful framework would rank candidates by:

  • molecular target;
  • evidence of direct mitophagy engagement;
  • proof of completed flux;
  • brain pharmacokinetics;
  • therapeutic window;
  • animal-model reproducibility;
  • human safety;
  • clinical-stage evidence.

Without this, weakly supported phytochemicals can appear comparable to more target-specific genetic or pharmacological interventions.

11. Human biomarkers are not adequately developed

A central translational problem is determining, in a living patient, whether mitophagy is:

  • deficient;
  • excessive;
  • initiated but blocked;
  • normal but overwhelmed by damage.

The paper mentions potential disease biomarkers but does not offer a convincing biomarker strategy for mitophagic flux. Peripheral measures, static protein levels and mitochondrial DNA copy number are unlikely to be sufficient on their own.

12. Presentation and editing issues

As an article-in-press version, it contains awkward phrasing, typographical problems and some repetition. Certain quantitative claims are presented with considerable precision but without enough discussion of experimental context, sample size or reproducibility. Final copy-editing may improve this, but some issues are substantive rather than stylistic.

Bottom line

The review’s strongest contribution is its rejection of the simplistic proposition that more mitophagy is always better. It convincingly argues that neuronal health depends on a coordinated system involving damage recognition, fission, transport, lysosomal degradation and replacement through mitochondrial biogenesis.

Its main limitation is that it does not convert that good conceptual insight into a rigorous evidential framework. The field still lacks reliable ways to determine:

  • which mitochondria are being selected;
  • whether mitophagic flux is actually completed;
  • when removal becomes excessive;
  • whether mitochondrial replacement keeps pace;
  • which patients would benefit from activation versus inhibition.

Thus, the most defensible therapeutic conclusion is not “activate mitophagy,” but restore appropriately selective, spatially completed mitophagic flux while maintaining mitochondrial biogenesis and neuronal energy reserve.