I thought I would run the original paper through chatGPT: (5.6maxpaid)
Overall assessment
Donega et al., “Oxygenaging: A Physiological Framework for Geroscience” is a thought-provoking narrative review rather than an experimental study. Its central physiological synthesis is valuable, but it does not yet establish deterioration of oxygen homeostasis as a primary cause of ageing. The strongest evidence concerns age-related loss of cardiopulmonary and microvascular reserve; the weakest concerns mitochondrial oxygen limitation, RNA-splicing failure and the purported gerotherapeutic effects of intermittent hypoxia and hyperbaric oxygen.
I would cite it as a hypothesis-generating framework, not as proof of a new hallmark of ageing or of oxygen therapy extending healthspan.
Summary
The authors define “Oxygenaging” as the progressive loss of the capacity to match oxygen supply to metabolic demand across the entire oxygen cascade:
$$
F_{IO_2}\rightarrow P_{AO_2}\rightarrow P_{aO_2}/S_{aO_2}/C_{aO_2}
\rightarrow D_{O_2}\rightarrow P_{mvO_2}\rightarrow P_{isO_2}\rightarrow P_{mitoO_2}
$$
Their important distinction is that Oxygenaging is not simply systemic hypoxaemia. It is declining resilience: oxygen delivery may appear adequate at rest but become spatially or temporally mismatched to demand during exercise, illness or other stress.
The proposed sequence is:
-
Pulmonary decline
Loss of elastic recoil, premature small-airway closure and greater ventilation–perfusion heterogeneity produce a modest age-related fall in arterial (P_{O_2}) and reduced ventilatory reserve.
-
Central cardiovascular decline
Lower maximal heart rate and stroke-volume reserve constrain maximal cardiac output and therefore peak oxygen delivery:
$$
D_{O_2}=C_{aO_2}\times Q
$$
-
Microvascular decline
Endothelial dysfunction, reduced nitric oxide, pericyte loss, capillary rarefaction, fibrosis and increased diffusion distance make local perfusion less responsive to metabolic demand. The authors illustrate this in brain, kidney, skin and skeletal muscle.
-
Failure of cellular adaptation
Acute hypoxia normally stabilises HIF-1α/HIF-2α, which:
- stimulates VEGF and erythropoietin;
- reduces mitochondrial oxygen demand through PDK1-mediated PDH inhibition;
- suppresses mTORC1 and protein synthesis;
- promotes glycolysis, mitophagy and redox protection.
In ageing, the authors propose that this becomes maladaptive. NAD⁺ depletion, succinate accumulation, endothelial senescence, impaired erythropoietin production and persistent mTOR activity produce chronic “pseudohypoxia” without an effective reparative response.
-
Mitochondrial and iron pathology
Suppressed mitochondrial biogenesis combined with impaired mitophagy supposedly leaves fewer, older and ROS-leaking mitochondria. Labile iron then amplifies reoxygenation injury and ferroptosis—the proposed intersection with “Ferroaging.”
-
Epigenetic and RNA-processing effects
Because TET DNA demethylases and JmjC histone demethylases require oxygen and 2-oxoglutarate, hypoxia may alter DNA and histone methylation. The paper also links hypoxia to PKM isoform switching, SRSF6 poison-exon regulation and increased intron retention.
-
Interventions
Exercise, calorie restriction, metformin and rapamycin are interpreted as indirectly improving oxygen supply–demand matching. Intermittent hypoxia, hyperbaric oxygen therapy and intermittent hypoxic–hyperoxic training are presented as more direct attempts to retrain oxygen sensing, angiogenesis and mitochondrial quality control. The authors appropriately stress that ageing narrows the hormetic therapeutic window.
What is genuinely novel?
| Component |
Novelty |
| “Oxygenaging” terminology |
Apparently a new label for this integrated concept |
| Whole-cascade perspective |
The most useful novelty: it joins lungs, circulation, microvasculature, mitochondria and molecular ageing |
| Supply–demand mismatch rather than simple hypoxia |
Conceptually valuable and more physiologically realistic |
| Integration with splicing, epigenetics and ferroptosis |
Relatively unusual, although largely speculative |
| Individual mechanisms |
Mostly established oxygen physiology, HIF biology, vascular ageing and mitochondrial biology |
| Experimental evidence |
None generated by this paper |
| Quantitative theory or diagnostic index |
None provided |
Thus the novelty is principally conceptual organisation and branding, not discovery of a new pathway.
Strengths
- It restores whole-organism physiology to a field often dominated by isolated molecular hallmarks.
- The focus on dynamic perfusion and microregional heterogeneity is more persuasive than a claim of uniform age-related hypoxia.
- It recognises bidirectional feedback: vascular and mitochondrial ageing impair oxygen homeostasis, while oxygen instability may amplify those same processes.
- It correctly rejects the simplistic idea that more oxygen is necessarily better.
- It acknowledges tissue differences, sex, frailty, comorbidity and the shrinking hormetic window in older people.
- Figures 1 and 2 provide an unusually clear overview of the proposed multiscale relationships.
Major criticisms
1. The central causal step is not demonstrated
The review sometimes describes oxygen-cascade dysfunction as a “main cause” or “driver” of ageing, but most evidence is cross-sectional and equally consistent with oxygen dysregulation being a consequence of ageing, inactivity, fibrosis or mitochondrial loss.
No study cited follows the complete sequence in the same organism:
declining oxygen delivery → reduced mitochondrial (P_{O_2}) → HIF/epigenetic/splicing changes → functional ageing or shortened lifespan.
The more defensible claim is the paper’s softer one: oxygen homeostasis is an interface and amplifier of ageing processes.
2. Lower tissue oxygen does not necessarily mean oxygen-limited mitochondria
The paper never establishes the oxygen threshold at which normal age-related reductions actually constrain oxidative phosphorylation.
A recent direct measurement in human skeletal muscle estimated an apparent mitochondrial (P_{50}) of approximately 0.5 mmHg, compared with resting intracellular oxygen pressures around 34 mmHg. This implies a broad oxygen-independent range at rest, with oxygen sensitivity becoming important mainly during severe hypoxia or intense exercise. The study involved young adults, so it does not exclude severely hypoxic microregions in older tissue, but it sets an important burden of proof that the review does not meet. Erol et al., 2026
Similarly, slow phosphocreatine recovery can reflect reduced mitochondrial content or intrinsic mitochondrial dysfunction, rather than inadequate oxygen delivery. The direction of causality cannot be inferred from the imaging measurements presented.
Table 1 also gives an illustrative fall in arterial saturation from approximately 92% to 85%. That would be clinically important hypoxaemia, not the normal consequence of the cited 15–20 mmHg age-related fall in arterial (P_{O_2}), which generally remains on the flatter part of the oxyhaemoglobin curve. It conflicts with the paper’s own statement that resting saturation is usually preserved.
3. HIF regulation is oversimplified and contains a directional error
The paper states that SIRT1 “deacetylates and activates HIF-1α.” The cited Lim study reported the opposite direct effect: declining NAD⁺/SIRT1 permits HIF-1α acetylation and activation, whereas SIRT1 deacetylates and represses HIF-1α transcriptional activity. SIRT1 can activate HIF-2α, which may be part of the confusion. Lim et al., 2010
More generally, the framework invokes both:
- insufficient inducible HIF responses in aged tissues; and
- chronically stabilised HIF-1α from pseudohypoxia.
Both can occur in different tissues and time frames, but the review does not operationally distinguish them. Without that distinction, almost any HIF result can be incorporated into the model.
4. Acute hypoxia and cancer biology are repeatedly extrapolated to normal ageing
Several molecular examples come from tumour models:
- The cited PKM study demonstrated a c-Myc–hnRNP–PKM2 pathway in glioma and cancer metabolism, not HIF-driven splicing failure in normal ageing. David et al., 2010
- The SRSF6 poison-exon work used acute hypoxia in HeLa and other cancer contexts. It demonstrated an organised adaptive response, not age-related loss of splicing fidelity. Machado et al., 2023
- The “lipid lock,” MYC/TFAM suppression and several ferroptosis links are biologically plausible but not shown to form a unified pathway during ordinary human ageing.
The paper therefore conflates three different phenomena: regulated alternative splicing, stress-induced transcript suppression and erroneous age-related splicing. They are not interchangeable.
5. The epigenetic interpretation is too strong
The statement that oxygen directly dictates the “ticking” of the epigenetic clock overinterprets the authors’ own experiment. Severe intermittent hypoxia—cycling to 5% oxygen—temporarily increased clock estimates in old mice, with substantial reversal after normoxic recovery. No functional ageing, pathology, mortality or lifespan outcome was measured. Donega et al., 2026
The safest interpretation is that epigenetic clocks are sensitive to a reversible oxygen-stress programme, not that the animals aged five biological months and were then rejuvenated.
There is also an epigenetic imprecision: the review groups increased H3K4me3 and H3K27me3 together as chromatin-condensing changes. H3K27me3 is generally repressive, whereas promoter H3K4me3 is generally associated with active transcription. Hypoxic inhibition of different demethylases therefore produces locus- and mark-specific effects rather than uniform chromatin closure. Batie et al., 2019, Chakraborty et al., 2019
6. The intervention section overstates very preliminary evidence
This is the weakest portion of the review.
- The reported HBOT telomere study had only 30 completers and no control group; several cell-type repeated-measures results were non-significant. Crucially, the primary paper explicitly states that telomerase activity was not measured, yet Table 2 presents telomerase activation as a mechanism. Hachmo et al., 2020
- The skin study included 13 men in a fixed control-then-treatment sequence. “Senescent-cell clearance” was inferred from reduced lipofuscin staining. It did not measure circulating SASP factors, NK-cell oxidative phosphorylation or immune-mediated clearance, although Table 2 attributes these outcomes and mechanisms to it. Hachmo et al., 2021
- The cognitive HBOT trial randomised 63 participants but used a no-contact control rather than sham pressurisation. Participants therefore knew whether they received an intensive 60-session intervention; investigators also disclosed employment and equity links to an HBOT company. These factors do not invalidate the findings, but substantially weaken causal certainty. Hadanny et al., 2020
- The principal IHHT cognitive trial analysed only 34 geriatric patients receiving concurrent multimodal rehabilitation, and its authors explicitly described the mechanisms as largely speculative. Bayer et al., 2017
None of these studies demonstrates slower ageing, extended healthspan or reduced mortality.
7. The framework is not yet falsifiable
Because both too little and too much oxygen, both reduced and excessive HIF activity, and both hypoxia and hyperoxia can be classified as Oxygenaging, the framework risks becoming an explanation applied after the event.
It needs:
- direct longitudinal measurements across several cascade levels in the same subjects;
- quantitative tissue-specific oxygen thresholds;
- an operational “Oxygenaging index”;
- prospective prediction of functional decline independent of physical activity and disease;
- interventions that selectively improve oxygen delivery and show mediation through the proposed molecular pathway;
- sham-controlled trials with functional and long-term clinical outcomes.
Relevance to the citrate–acetyl-CoA–histone-acetylation hypothesis
The paper is complementary but does not directly support that pathway. It contains no substantive discussion of citrate export, SLC25A1, ACLY, nuclear acetyl-CoA or histone acetylation.
Its HIF-1α → PDK1 → PDH inhibition pathway could, in principle, reduce glucose-derived mitochondrial citrate and ACLY-derived nucleocytosolic acetyl-CoA. That would provide a possible bridge from oxygen dysregulation to histone acetylation and transcription. However, this connection is not examined or measured, and the paper instead concentrates on oxygen-dependent DNA and histone demethylation.
Likewise, its splicing section demonstrates that hypoxia can deliberately reprogramme particular splice choices; it does not establish that oxygen deficiency causes the broader decline in splicing fidelity or long-gene expression of interest.
Bottom line
The paper succeeds as a broad physiological manifesto: ageing reduces the ability to distribute and use oxygen precisely under changing demand, and this deserves more attention in geroscience. It does not yet show that oxygen-cascade failure is a primary ageing mechanism, and several molecular and therapeutic claims are more confident than their cited evidence permits. Its most defensible conclusion is:
Loss of oxygen homeostatic resilience is probably an important modifier and amplifier of ageing, but whether it is an upstream driver—and whether oxygen modulation is genuinely geroprotective—remains unproven.