Why can aging muscle become more oxidative while function declines? Cardiolipin and compensation

The supported conclusion is that cardiolipin synthesis contributes to muscle structure, metabolic adaptation and function, with evidence of modifiability. An oxidative shift alone is not a rejuvenation marker, and these studies do not establish a human healthspan benefit from increasing cardiolipin.

01Cardiolipin depletion differs from longer lipid chains

Cardiolipin (CL), predominantly in the inner mitochondrial membrane, has four fatty-acid tails. Quantity, tail composition, oxidation and membrane distribution can affect cristae, respiratory organization and transport. CRLS1 synthesizes CL; remodeling follows synthesis. Increasing CRLS1 does not establish restoration of every compositional or functional feature, nor an equivalent effect from consuming one fatty acid. 1

Four questions require separate evidence: does CL decline; can loss cause pathology; how do cells compensate; and does enhancing synthesis improve aged function? An age association does not answer the other questions. Severe depletion in young animals cannot alone establish natural-aging rescue.

The 2026 study compares young and aged male mice in oxidative soleus (SOL) and predominantly glycolytic extensor digitorum longus (EDL), reporting lower CL and Crls1 with age. It then uses constitutive muscle knockout MUCKO and adult tamoxifen-inducible iMUCKO. Main constitutive experiments span 4–16 weeks of age, with growth restriction, atrophy and premature death. This severe pathology is not simply normal aging on a shorter timescale. 2

02More mitochondrial material can accompany impaired performance

After constitutive depletion, EDL citrate synthase (CS) activity and COX IV-stained area fraction increase toward SOL levels; SOL does not show the same increase. CS is a content proxy and staining area is not an exact organelle count. Cristae are disrupted in both muscles; respiratory supercomplexes decline in pooled hindlimb preparations. 2

The table checks figures, Methods and author statistics. Already adjusted P values are retained without further adjustment; these are not independently reconstructed individual respiration measurements. 2 6

A definition example explains the denominator. If tissue respiration J divided by CS proxy C equals j, then J=C×j. Choose C=2 times baseline and j=0.3, 0.5 or 0.7 times baseline: J becomes 0.6, 1.0 or 1.4 times baseline. Identical directions of content and content-normalized respiration can accompany decreasing, unchanged or increasing tissue totals. These chosen parameters are not fitted to animals, and CS can itself be regulated. Without paired original measurements and covariance, multiplying group means does not recover observed tissue energy supply.

03How far does causality for oxidative switching extend?

MUCKO EDL loses IIb proportion and gains IIa/IIx features, with protein and transcriptional changes. Nonsignificant total fiber-count comparisons do not certify absence of selective loss. Together with switching in C2C12 myotubes, however, findings support active cellular remodeling, rather than explaining everything through fast-fiber death. Some IIx classification uses absence of other markers; this is not fiber lineage tracing. 2

EDL ERRγ (Esrrg) increases. Esrrg knockdown or its inhibitor GSK5182 blocks selected CL-depletion-associated myosin changes in cultured myotubes; downstream Myog perturbation supports a regulatory chain. This is meaningful cellular perturbation evidence, but does not directly establish whether blocking compensation improves or worsens force in aged intact animals. 2

The paper classifies 626 of 773 altered genes as ERRγ dependent, partly because they no longer meet a differential threshold with inhibition. Significance in one comparison and nonsignificance in another do not establish a significant difference between effects. A factorial interaction test is closer to that question. We checked sample dictionaries without rerunning the RNA model, so do not treat 626/773 as a certified genome-wide causal proportion. Registration contains 24 culture samples; duplicated sequencing-file rows do not create 48 biological samples. 9

A competing explanation remains: the oxidative program may mitigate part of an oxidative-stress burden without repairing the membrane or contractile apparatus. Marker direction and net functional direction can differ.

04A higher ATP/ADP ratio is not greater ATP production

Our exploratory analysis uses author-normalized lipid and metabolite tables, six controls and four constitutive knockout samples per muscle. All 23 canonical CL annotations are retained without selection by outcome. KO/control ratios of their summed signals are 0.247 in SOL and 0.306 in EDL. Welch difference intervals exclude zero; Holm-adjusted P across two muscles is approximately 0.000048 and 0.005889. 4

This supports depletion under that definition but does not reproduce every value in the authors' total-CL figure. The 23-species sum does not match bar magnitudes, and some species increase. Species selection and normalization are insufficiently mapped. Methods specify dry-weight normalization while an axis uses protein normalization. We do not guess deletions to match the figure or treat signal sums as calibrated molar totals. Repeated values of 0.0001 lack a fully specified imputation/detection-limit rule; all-constant comparisons receive no meaningful P or interval.

Descriptive point estimates; intervals and tests are in the analysis tables. CL uses 23 annotated signals. ATP and ADP compare group means; ATP/ADP is calculated within each column before comparing means. None measures natural-aging rescue or metabolic flux.

EDL ATP mean signal is 0.826 times control, ADP 0.285, and mean within-column ATP/ADP signal ratio 1.911. SOL values are 1.117, 0.851 and 1.346. EDL's ratio has raw Welch P≈0.0086; Holm adjustment across four endpoints and two muscles gives 0.0691. Other endpoints also fail 0.05 in that family. Descriptive numerator/denominator directions therefore do not establish robust energy-state improvement across eight comparisons. 5

An ATP pool is not ATP produced per second. Unvalidated inter-metabolite MS response prevents interpreting these signals as molar adenylate energy charge. Sheet names specify protein correction while Methods specify dry weight. A shared sample denominator cancels in ATP/ADP; distinct response factors do not automatically cancel. Glucose uptake, metabolite accumulation and protein enrichment likewise do not identify complete carbon flux. Reported glucose-to-serine tracing supports EDL rerouting, but we did not locate a sufficiently detailed tracer/recovery protocol and do not infer an exact whole-body flux proportion. 2

05Functional evidence includes unchanged readouts

Five weeks after adult induction, body-weight-normalized four-limb grip falls, author unadjusted t-test P=0.0037. Isolated SOL/EDL tetanic-contraction AUC and force normalized to estimated muscle cross-sectional area show no significant difference, corresponding adjusted P>0.999999. Grip also involves behavior, recruitment and whole-animal condition; it does not establish isolated contractile failure. Nonsignificance does not establish equivalence. 2 6

Specific-force normalization uses muscle mass, density, optimal length and muscle-specific fiber-length coefficients. That denominator is not individual histological fiber area.

With tamoxifen withdrawal, iMUCKO Crls1 partly recovers. Authors report gastrocnemius (GAS) CL rising from 27% to 67% of control, with mitigation of wasting, weight loss and premature death. This does not establish identical recovery across muscles or normalization of every endpoint. Existing Cre excision is irreversible. Proposed incorporation of unrecombined satellite-cell-derived nuclei fits regeneration markers but is not direct lineage quantification checked here. Nonsignificant survival versus control within the displayed observation window is not a complete normal-lifespan equivalence test. The survival figure labels **** while the source table gives depletion/control P=0.0152 and recovery/depletion P=0.0231; we do not report the stronger star-implied significance. Original death/censoring data are unavailable for adjudication. Figure 2 recovery panels do not supply grip or isolated-force endpoints for that same recovery cohort; we cannot add force restoration. 2 6

06Retain naturally aged enhancement, with its actual endpoints

The independent 2024 study injects AAV9-mCrls1-GFP into tibialis anterior (TA) of approximately 22-month-old or older C57BL/6J mice. Five weeks later, versus AAV9-GFP, CL, muscle wet weight and fiber area increase. Selected-frequency tetanic force is higher, labeled n=5 per group for function and n=13 for mass; single-twitch force does not show the same significant improvement. This is positive enhancement evidence in naturally aged muscle without additional CTX injury. 10

Force axes use mN/g, normalized to muscle mass: neither absolute force nor physiological-area-specific force. Fatigue curves start higher but converge later, without directly establishing sustained fatigue-resistance improvement. Bilateral injection diagrams and group n do not fully resolve animal independence/pairing; we do not reconstruct paired tests or dots as individual data. Methods claim complete randomization with insufficient blinding details. Individual source tables for fresh force-effect/interval estimation are unavailable. 10 11

Separate CTX-injured aged-muscle experiments support mass and fiber-size regeneration and must be distinguished from uninjured aged contractility. Neither establishes human mobility, fall reduction or longevity. The April 2024 correction changes funding identifiers in acknowledgments, not these checked functional results. 12

The 2026 Discussion still describes aged Crls1 enhancement as future work. This identifies validation absent from that study without erasing the conditional positive 2024 experiment. Replication across laboratories, muscles, sexes, longer durations and whole-animal activity remains needed.

07More antioxidant is not invariably better

Eight weeks of NAC drinking-water treatment in depleted mice worsens wasting indices. Glucose-uptake comparisons include four untreated and three NAC-treated knockout mice. Reduction is supported in EDL, adjusted P<0.0001, while SOL has P=0.2647. Both muscles cannot be declared confirmed, nor can this small sample estimate a stable population effect. High NAC can interfere with insulin signaling, acknowledged by authors, so ROS removal alone is not a uniquely identified mediator. 2 6

The 2013 primary study supplies an opposite condition: NAC improves myotube formation and reduces selected atrophy pathways in primary cells from muscle ERRγ-deficient mice. These are mainly cultured differentiation/morphology results, not aged-animal force verification. Damage context, dose, timing and endpoint matter. Conditional adverse results cannot establish universal NAC harm; cellular rescue does not establish an anti-aging supplement indication. 13

08What is missing for human conclusions?

The 2026 main Methods describe pre-exercise samples from a male endurance-cycling cohort, four younger and five older participants. The cited 2022 study sampled vastus lateralis in long-trained cyclists, not a general older population. 14

However, the Reporting Summary cites a six-week high-intensity-training study, PMID34273538, describes men and women, and gives a different ethics identifier. Main text/legend lower age bounds differ (22/23), and a human legend incorrectly names SOL/EDL. The other PMID metadata identifies a different study; complete provenance of these nine specimens remains uncertified. The summary claims animal randomization and preliminary sample-size estimates, whereas main Statistics states no randomization or predetermination. Both acknowledge no blinding. Conflicts are retained without choosing a favorable version. 2 3 15

Even if the human age association holds, training, fiber composition, nutrition and disease must be considered. More decisive evidence would test whether enhancing synthesis consistently improves clearly normalized force, daily activity and long-term health during natural aging, whether compensation must be preserved, and safety across tissues. Current evidence supports further target testing without supplying a human cardiolipin-enhancement or antioxidant regimen.

09Research and reproducibility

Search/version cutoff: 2026-10-08. A CL review established synthesis, remodeling and measurement concepts, followed by central results, Methods, figures, relevant supplements, sample dictionaries and correction. This is question-directed research, not a systematic review. Abundance definitions were frozen after reading paper directions, making analyses exploratory. Aging, depletion and recovery remain distinct without assumed cross-muscle/cross-assay pairing. 7 8 9

Independent OOXML checks cover 12,631 populated fields/axes in three workbooks. R recalculates 550 numerical comparisons from independently extracted values, maximum difference ≈2.27×10⁻¹³. Agreement checks transcription/calculation, not peak identity, experimental provenance or causality. The reproducibility package supplies code, complete group summaries, author-statistics inventory and source hashes; originals and sample work matrices remain private. One Codex agent researched, self-reviewed and translated without independent human/professional review.