chatGPT(5.5)paid.
Overall assessment
This is a narrative review and hypothesis paper, not a primary experimental study. Its central argument is that the relationship between energy stress and autophagy in Parkinson’s disease is biphasic:
- Mild ATP depletion activates or permits protective autophagy through pathways such as AMPK–ULK1.
- More severe or prolonged ATP depletion prevents autophagy from being completed because autophagosome formation, vesicle trafficking, lysosomal acidification, fusion and degradation all require energy.
- Autophagy failure then worsens mitochondrial quality, ATP production, oxidative stress and α-synuclein accumulation, producing a self-reinforcing pathological cycle.
The paper is conceptually strong and biologically plausible, but the specifically Parkinson’s-related evidence for ATP depletion as a direct cause of autophagy failure remains largely indirect.
1. Summary
Background
Parkinson’s disease is associated with several interacting abnormalities:
- mitochondrial complex I dysfunction;
- reduced mitochondrial membrane potential and ATP production;
- oxidative stress;
- α-synuclein accumulation;
- defective macroautophagy, mitophagy and chaperone-mediated autophagy;
- lysosomal dysfunction.
Dopaminergic neurons of the substantia nigra may be especially susceptible to energy failure because of their extensive axonal arborisation, autonomous pacemaking, synaptic activity and long-distance transport requirements.
The authors note an apparent paradox. Energy deprivation is normally regarded as an autophagy stimulus, principally through AMPK activation and mTORC1 inhibition. Yet autophagy itself consumes ATP.
Energy requirements of autophagy
The review identifies several energy-dependent components of autophagy:
- transcription and translation of autophagy-related proteins during sustained responses;
- ATP-dependent ATG conjugation reactions and LC3 lipidation;
- phagophore expansion and closure;
- membrane remodelling;
- vesicular transport;
- autophagosome–lysosome fusion;
- vacuolar H⁺-ATPase-dependent lysosomal acidification;
- lysosomal enzyme function and cargo degradation.
The authors therefore distinguish autophagy initiation from productive autophagic flux. A cell may display signalling or marker changes consistent with autophagy induction while being unable to complete cargo degradation.
Evidence in Parkinson’s disease
The paper reviews evidence that PD brains and experimental models show defective autophagic clearance, including:
- accumulation of autophagic vacuoles and lysosomal depletion in post-mortem tissue;
- impaired autophagosome–lysosome fusion in patient-derived neurons;
- reduced LAMP2A and HSC70, suggesting impaired chaperone-mediated autophagy;
- defects linked to genes including SNCA, GBA1, ATP13A2, LRRK2, PINK1, PRKN, PARK7 and VPS35;
- autophagy impairment following rotenone, MPP⁺, MPTP and 6-hydroxydopamine exposure.
The authors correctly emphasise that PD autophagy failure is multifactorial. ATP depletion operates alongside α-synuclein toxicity, oxidative and nitrosative stress, lysosomal abnormalities and disease-associated mutations.
Proposed biphasic model
The model shown in the figures is:
- Mild mitochondrial dysfunction causes a modest ATP decline.
- Energy sensing induces autophagy and mitophagy.
- Damaged mitochondria and α-synuclein can still be cleared.
- With progressive mitochondrial dysfunction, ATP falls below a context-dependent threshold.
- Autophagy becomes incomplete or fails at progressively earlier stages.
- Damaged mitochondria accumulate, increasing ROS and reducing ATP further.
- α-Synuclein clearance deteriorates, creating a vicious cycle.
The authors mention a roughly 50% ATP reduction as a previously proposed threshold for autophagy inhibition, but appropriately conclude that there is unlikely to be a universal numerical threshold.
Therapeutic implications
The paper argues that simply inducing autophagy may fail in cells that lack the energy to execute it. It proposes combining:
- bioenergetic interventions such as creatine, nicotinamide riboside, NMN, pyruvate, ketones, acetyl-L-carnitine or terazosin;
with:
- autophagy-promoting interventions such as rapamycin, trehalose, lithium, spermidine, resveratrol or ambroxol.
The authors suggest that these approaches may be most effective early or during prodromal disease, before extensive neuronal loss.
2. Novelty
Main conceptual novelty
The principal novelty is not the observation that autophagy requires ATP, nor that mitochondrial dysfunction and autophagy failure occur in PD. The novel contribution is the integration of these observations into a disease-stage-dependent framework:
Energy depletion is initially an autophagy signal but can eventually become an autophagy constraint.
This reframes apparently contradictory findings in which complex I inhibitors or metabolic stress sometimes increase autophagy markers and sometimes suppress them.
Distinction between induction and completion
A particularly useful contribution is the emphasis that:
- increased LC3-II or autophagosome numbers do not necessarily indicate successful autophagy;
- ATP depletion may permit initiation while preventing lysosomal completion;
- with greater energy loss, even autophagosome formation may fail.
This gives the proposed relationship more resolution than a simple “ATP promotes autophagy” or “ATP depletion activates autophagy” model.
Resource-allocation interpretation
The authors describe autophagy as an energetically costly investment:
- ATP is spent on degradation and recycling;
- the eventual return is improved substrate availability and mitochondrial quality;
- when ATP becomes critically scarce, the immediate energetic cost may exceed the delayed return.
This resource-allocation framing is conceptually interesting, although it remains qualitative.
Therapeutic stratification
The proposal that autophagy activators should be matched to the energetic state of the neuron is also a useful translational idea. It implies that treatment response may depend on:
- disease stage;
- neuronal ATP availability;
- residual glycolytic compensation;
- the particular autophagic step that has failed.
This could help explain why interventions that appear beneficial in experimental models have not translated consistently into clinical benefit.
Limits to novelty
The novelty should not be overstated. The paper builds heavily on existing knowledge that:
- ATP is required for vesicle trafficking and lysosomal acidification;
- severe energy depletion inhibits cellular processes;
- autophagy can become abortive;
- mitochondrial dysfunction and defective autophagy reinforce one another.
Thus, the paper’s novelty is primarily synthetic and hypothesis-generating, rather than the discovery of a new molecular mechanism.
3. Critique
Strengths
1. Clear and biologically plausible thesis
The central paradox is well identified and clearly explained. The figures are effective in distinguishing early adaptive autophagy from later energy-constrained autophagy.
2. Appropriate caution
The authors frequently acknowledge uncertainty. They note that:
- the ATP threshold is unknown;
- ATP was not measured in many cited PD studies;
- glycolysis may compensate for OXPHOS inhibition;
- pyruvate and lactate can act through mechanisms other than ATP restoration;
- autophagy dysfunction in PD has many causes.
This substantially improves the credibility of the review.
3. Focus on flux rather than static markers
The paper correctly warns against interpreting LC3-II accumulation alone as evidence of enhanced autophagy. This is a major methodological issue in the autophagy literature.
4. Broad mechanistic coverage
The review covers autophagosome formation, cargo sequestration, fusion, lysosomal acidification and degradation, rather than treating autophagy as a single reaction.
5. Testable predictions
The framework generates experiments that could falsify or support it, particularly experiments measuring ATP and autophagic flux simultaneously across graded energy depletion.
Major limitations
1. The central PD-specific causal claim is not yet demonstrated
The paper establishes three propositions reasonably well:
- ATP production can be impaired in PD.
- Autophagy can be impaired in PD.
- Severe ATP depletion can impair autophagy in several experimental systems.
What is much less firmly established is the crucial fourth proposition:
- ATP depletion is a quantitatively important cause of autophagy failure in human dopaminergic neurons during PD progression.
Much of the supporting evidence comes from:
- toxin-treated tumour-derived cell lines;
- hepatocytes and kidney tubules;
- cancer cells;
- fibroblasts with mitochondrial disorders;
- non-PD models of respiratory-chain inhibition.
These studies demonstrate plausibility, but not the size of the effect in human PD neurons.
2. Narrative rather than systematic evidence synthesis
The article is labelled as a review, but it does not describe:
- a literature-search strategy;
- inclusion or exclusion criteria;
- study-quality assessment;
- risk-of-bias evaluation;
- systematic treatment of contradictory studies.
This leaves open the possibility of selective citation. A systematic table showing model, intervention, ATP change, flux assay and affected autophagic stage would have made the argument much stronger.
3. ATP concentration is treated too globally
“Cellular ATP” is not necessarily a single homogeneous variable. Autophagy may depend on:
- ATP/ADP ratio;
- AMP/ATP ratio;
- local ATP near lysosomes or membrane-remodelling sites;
- cytosolic versus mitochondrial ATP;
- phosphocreatine buffering;
- substrate supply;
- NADH/NAD⁺ balance;
- mitochondrial membrane potential independently of ATP.
Bulk cellular ATP measurements could therefore conceal the bioenergetic state relevant to a particular autophagic compartment.
4. Mitochondrial membrane potential deserves more separation from ATP
Complex I inhibition affects more than ATP. It also alters:
- ΔΨm;
- ROS generation;
- NADH oxidation;
- calcium handling;
- mitochondrial dynamics;
- PINK1 stabilisation;
- metabolite transport.
Some effects attributed to ATP depletion might actually result from one or more of these parallel changes. For example, mitophagy signalling is directly sensitive to mitochondrial depolarisation, while lysosomal and trafficking defects may be affected by redox or cytoskeletal disruption.
The proposed model would be stronger if ATP were manipulated independently of ΔΨm and ROS.
5. Glycolytic compensation is acknowledged but insufficiently developed
Neurons, astrocytes and cultured neuroblastoma cells differ substantially in glycolytic capacity. SH-SY5Y cells may compensate for respiratory inhibition in ways that mature substantia nigra neurons cannot, while astrocyte-derived lactate may partially support neurons in vivo.
Therefore, the relevant variable may not be complex I dysfunction alone, but the balance among:
- OXPHOS failure;
- glucose availability;
- glycolytic reserve;
- astrocyte-neuron metabolic coupling;
- lactate and ketone utilisation.
This complicates translation from toxin-treated cell cultures to human disease.
6. Disease stage may not map simply onto ATP severity
The figures imply a relatively orderly progression from mild energy loss to severe energy loss. Human PD is heterogeneous. Different neurons within the same brain may have:
- different mitochondrial defects;
- different α-synuclein burdens;
- different lysosomal capacity;
- different calcium loads;
- different compensatory metabolic responses.
Early disease could contain severely energy-deficient neurons, while surviving neurons in advanced disease may retain relatively robust bioenergetics. The model is therefore more likely to operate at the individual-cell level than as a uniform early-versus-advanced disease distinction.
7. The proposed 50% threshold is weakly grounded
The paper appropriately backs away from a universal threshold, but repeatedly discussing a roughly 50% ATP decline may give the value more significance than the evidence warrants. The cited threshold derives from particular experimental systems and conditions, not from dopaminergic neurons in PD.
It would be better to frame the threshold in terms of energy charge and pathway-specific failure points, rather than a percentage of untreated bulk ATP.
8. Autophagy can itself initially support ATP, creating an identification problem
If autophagy supplies metabolic substrates, then reduced ATP and impaired autophagy can each cause the other. Cross-sectional correlations will not determine directionality.
Experiments need temporal resolution sufficient to establish whether:
- ATP decline occurs before flux failure;
- flux failure occurs before secondary ATP decline;
- both arise from a third process such as lysosomal damage or oxidative stress.
9. Mitophagy may have distinct energetics from general macroautophagy
The paper sometimes moves between macroautophagy, mitophagy and chaperone-mediated autophagy as though the same energetic model applies uniformly. However:
- PINK1/Parkin mitophagy has specific dependence on ΔΨm and ubiquitin signalling;
- receptor-mediated mitophagy may behave differently;
- CMA requires substrate recognition and LAMP2A-mediated translocation;
- lysosomal acidification affects all pathways but upstream energetic requirements differ.
The review would benefit from separate pathway-specific models.
10. Therapeutic recommendations are highly preliminary
Combining energy enhancers with autophagy activators is plausible, but potentially complicated.
For example:
- rapamycin can alter protein synthesis, immunity, glucose metabolism and mitochondrial function;
- AMPK activation can either stimulate or suppress autophagy depending on energy severity;
- nicotinamide precursors do not necessarily produce sustained neuronal ATP increases;
- terazosin has haemodynamic effects;
- increased mitophagy could be harmful if residual mitochondria are still essential for ATP production;
- stimulating autophagosome formation without restoring lysosomal function could worsen vacuole accumulation.
The therapeutic section is best regarded as a research agenda, not evidence supporting clinical use.
4. Experiments needed to test the hypothesis
A convincing study would use human induced pluripotent stem-cell-derived midbrain dopaminergic neurons and impose graded, reversible energy deficits.
It should simultaneously measure:
- ATP, ADP, AMP and cellular energy charge;
- mitochondrial and cytosolic ATP using compartment-targeted sensors;
- mitochondrial membrane potential;
- oxygen consumption and glycolytic flux;
- ROS;
- lysosomal pH;
- autophagic flux using tandem fluorescent LC3 reporters;
- mitophagic flux using reporters such as mt-Keima;
- α-synuclein clearance;
- neuronal survival and electrophysiological function.
Key experimental comparisons should include:
- ATP depletion with minimal mitochondrial depolarisation;
- mitochondrial depolarisation while maintaining ATP through glycolysis;
- ROS modulation without restoring ATP;
- restoration of ATP without directly activating autophagy;
- lysosomal acidification rescue;
- autophagy activation before and after ATP restoration.
The strongest evidence would be demonstration of a reproducible dose-response curve in which increasing energy loss first enhances productive flux, then causes abortive flux, and finally prevents initiation.
5. Bottom-line judgment
This is a valuable and timely conceptual review. Its strongest contribution is the distinction between energy stress as a signal to initiate autophagy and energy availability as a requirement to complete autophagy. The proposed biphasic model provides a coherent explanation for conflicting experimental findings and has potentially important implications for disease staging and combination therapy.
However, the central thesis remains a well-supported hypothesis rather than an established mechanism of Parkinson’s disease. The available literature shows biological plausibility, but much of it relies on non-neuronal systems, toxin models and studies that did not measure ATP and autophagic flux together. The paper would have been substantially stronger with a systematic evidence table, more explicit treatment of compartmental bioenergetics, and clearer separation of ATP effects from mitochondrial membrane potential, ROS and lysosomal pathology.