https://www.nature.com/articles/s41392-026-02813-2
chatGPT(5.6maxpaid):
This is a useful, wide-ranging review whose main contribution is synthesis. Its treatment of mitochondrial mechanisms is stronger than its assessment of therapeutic evidence, and it contains several specific errors worth correcting.
The paper is Creus et al., “Mitochondrial quality control in health and disease: mechanisms and therapeutic targets”, published in Signal Transduction and Targeted Therapy in July 2026. It is a 28-page narrative review with 490 references, containing no new experiments or meta-analysis. Page numbers below refer to the uploaded PDF. Publication record
The central argument is that mitochondrial health depends on coordinated maintenance at several levels. Repairing proteins, replacing damaged components, reorganising mitochondrial networks and removing entire organelles must work together.
| Process | What the review explains | Biological significance |
|---|---|---|
| Protein quality control | Chaperones assist folding; proteases including LONP1 and ClpP remove damaged proteins | Mitochondria can repair some damage without being destroyed |
| Stress responses | Mitochondrial stress activates nuclear responses, including the OMA1–DELE1–HRI pathway and selective translation of stress-response factors | Mitochondria influence cellular gene expression and adaptation |
| Biogenesis | PGC-1α, TFAM and metabolic signalling coordinate production of mitochondrial components | Clearance must be balanced with replacement |
| Fusion and fission | MFN1/2, OPA1 and DRP1 regulate network organisation and separation of mitochondrial regions | Shape changes can support repair, redistribution or disposal |
| Mitophagy | PINK1–Parkin and receptor pathways such as BNIP3/NIX and FUNDC1 promote mitochondrial removal | Several mechanisms contribute, with different triggers and functions |
| Other maintenance mechanisms | mtDNA maintenance, cristae organisation, permeability transition, mitochondrial vesicles and transfer between cells | Quality control extends beyond conventional intracellular mitophagy |
The authors then apply this framework to metabolic disease, cardiovascular disease, neurodegeneration, cancer, autoimmune disease and ageing.
Several conclusions deserve particular attention:
- Increasing a quality-control pathway is not universally beneficial. Effects depend on tissue, disease stage, duration and compensatory responses. For example, muscle-specific deletion of Fundc1 can improve systemic metabolism through FGF21 signalling, whereas whole-body deletion worsens metabolic dysfunction.
- Fission has several functions. Division near the mitochondrial midpoint can support biogenesis; peripheral division can separate material destined for degradation. Consequently, observing fragmentation alone does not establish mitochondrial damage or successful mitophagy.
- Mitochondrial dysfunction can promote inflammation. Released mtDNA can activate cGAS–STING, inflammasomes and other immune pathways.
- Cancer creates a therapeutic complication. Quality-control mechanisms can help malignant cells survive metabolic stress and treatment, although their effects vary substantially between cancers.
- Ageing does not consistently reduce measured mitophagy. The review acknowledges studies reporting decreased, unchanged or increased activity. It also distinguishes lifespan findings in worms and flies from the much more limited evidence in humans.
The therapeutic discussion covers exercise, calorie restriction, urolithin A, NAD⁺ precursors, rapamycin, metformin, other compounds and mitochondrial transplantation. The authors acknowledge that most mechanistic evidence remains preclinical.
The novelty lies mainly in how the literature is assembled. The connection between mitochondrial quality control, disease and ageing was already well established. This paper contributes a broad reference framework rather than a new explanatory theory.
Its most useful updates are:
-
Integration across molecular, organelle and cellular scales. Including mitochondrial vesicles and transfer between cells makes the account broader than a review centred on fusion, fission and mitophagy.
-
Inclusion of more recent mechanisms of selective mitochondrial maintenance. One example is MTFP1-mediated inhibition of inner-membrane fusion, which helps isolate altered mitochondrial regions for disposal. This is a significant development discussed by the review, originating in a 2024 experimental paper. Tábara et al.
-
Recognition of distinct fission outcomes. The division between biogenesis-associated and degradation-associated fission is mechanistically useful, although the underlying discovery dates to 2021. Kleele et al.
-
A more qualified account of mitophagy during ageing. Its discussion of conflicting reporter studies is valuable. For example, a 2024 study found stable or increased mitophagy in several aged mouse tissues despite mitochondrial dysfunction and inflammation. Jiménez-Loygorri et al.
These are worthwhile inclusions, but the discoveries belong to the cited primary studies.
The review’s principal strengths are its breadth, useful diagrams and willingness to describe contradictory findings. Its discussion of compensation is particularly valuable: improved glucose tolerance following disruption of a mitochondrial protein does not necessarily mean that the affected tissue has healthier mitochondria. It also includes negative human findings, including a urolithin A heart-failure trial.
My main criticisms concern how evidence is selected, measured and interpreted.
-
The review does not systematically establish the strength of evidence.
There is no reproducible search strategy, formal assessment of study bias or consistent grading of evidence. That is acceptable for a narrative overview, but it limits confidence about which mechanisms and interventions have the strongest support.
Tables identifying species, tissue, sample size, intervention duration, measured outcomes and study limitations would make its therapeutic conclusions more useful.
-
Protein abundance is too easily treated as evidence of pathway activity.
Several human findings concern the expression of Parkin, PINK1, BNIP3 or other regulatory proteins. These measurements cannot by themselves establish mitophagic flux: the rate at which mitochondrial material completes the degradation process.
Increased markers could reflect stronger initiation, greater damage requiring disposal, or accumulation because clearance is impaired. Reduced markers could reflect either reduced capacity or reduced demand. Although the authors recognise this problem in places, they do not apply the distinction consistently.
A related problem is that proteins such as MFN2 have several functions. A phenotype caused by deleting MFN2 cannot automatically be attributed exclusively to fusion or mitophagy.
-
The clinical interpretation occasionally exceeds what the cited trials establish.
On page 19, the review highlights improved muscle performance with urolithin A. However, the cited 2022 trial did not show a significant improvement in its primary endpoint, peak power output. Positive strength and biomarker findings remain relevant, but should be presented alongside that primary result. Singh et al.
The same page describes clinical effectiveness of mitochondrial transplantation. The cited study was a single-centre retrospective comparison involving 24 children: 10 treated and 14 controls. It provides preliminary evidence, with substantial scope for selection and treatment differences to influence outcomes. It cannot establish efficacy with the confidence of a randomised trial. Guariento et al.
Conversely, the negative urolithin A heart-failure study included only 10 patients, with four-week treatment periods. Its failure to detect cardiac improvement is informative but does not establish that longer treatment is ineffective. Jamialahmadi et al.
-
Its description of Mdivi-1 omits an important pharmacological limitation.
The paper repeatedly presents Mdivi-1 as a DRP1 inhibitor and uses it to discuss benefits of inhibiting mitochondrial fission.
Experimental work showed that Mdivi-1 can inhibit respiratory complex I and alter ROS production independently of DRP1. Consequently, benefits observed with this compound cannot, by themselves, demonstrate that inhibiting DRP1-mediated fission caused the benefit. This caveat was already documented in 2017 and deserves explicit treatment. Bordt et al.
-
There are identifiable factual and editorial errors.
On page 13, the sentence using “respectively” reverses the following established associations:
Gene Correct association OPA1 Autosomal dominant optic atrophy MFN2 Charcot–Marie–Tooth neuropathy type 2A The paper’s own references 263–265 establish the correct mapping. Alexander et al., Züchner et al.
Page 15 also groups HSP60 with mitochondrial proteases, although it is a chaperonin. These errors do not invalidate the overall framework, but they show why specific statements should be checked against the primary literature.
-
The mtDNA–ageing discussion needs a more discriminating causal framework.
Page 18 contrasts prematurely ageing POLG-mutator mice with models that accumulate mtDNA abnormalities without comparable lifespan shortening. That is a useful comparison.
However, interpretation requires attention to the identity of mutations, their abundance within individual cells, tissue distribution, timing and clonal expansion. “More mtDNA mutations” is not a uniform biological exposure. These differences limit what can be concluded about ordinary human ageing from comparisons between genetically altered mouse models.
For your mitochondria → citrate → acetyl-CoA → chromatin hypothesis, the most relevant material concerns cristae organisation, mitochondrial stress signalling, mtDNA-triggered inflammation and interactions between mitochondrial maintenance and nuclear responses.
The review does not directly examine citrate export, ACLY-dependent acetyl-CoA supply, histone-acetylation thresholds or impaired splicing fidelity as a connected mechanism. Its chromatin discussion mainly concerns histone methylation/demethylation during mitochondrial stress responses. It therefore supplies useful surrounding biology, while leaving your specific proposed causal sequence untested.
If helpful, I can set up “Watch for corrections to this review” to flag published fixes to the issues identified above.
I must admit 5.6max is quite good in reviewing these things.