BGPT: Paper Review: Aging kidney is associated with metabolic rewiring and epigenetic reprogramming

Fuel Your Discoveries

Quick Answer Copied

Core claim: This paper links age-associated renal metabolic remodeling to changes in metabolites that act as epigenetic enzyme cofactors, with marked sex-specific patterns, and supports partial translational relevance via human urine metabolites.

Long Answer

Paper review (skeptical, evidence-based): Aging kidney is associated with metabolic rewiring and epigenetic reprogramming

DOI: 10.1007/s11357-025-02071-0

Received 1 Sep 2025; accepted 14 Dec 2025 (from the provided full text).

1) What the paper actually measured (and what it didn’t)

Organism/model: C57BL/6JN mice; kidney tissue from 4, 12, 24 months (4M, 12M, 24M), both sexes.
Metabolomics: Kidney homogenates profiled by HPLC-TripleTOF-MS/MS, including untargeted metabolomics with multivariate modeling and pathway enrichment (MetaboAnalyst; SMPDB via Mus musculus).
Epigenetic enzyme readouts: Western blot and activity assays from kidney total protein/nuclear/histone extracts: HDAC1/2/3 levels; total H3 acetylation; DNMT and TET activities; JMJD3 (JMJD3C) activity; LSD1 activity; HAT/HDAC/SIRT activity assays.
Translational angle: Human spot urine metabolite kits (normalized to creatinine) for select intermediates/cofactors (L-carnitine, pyruvate, lactate, α-KG, β-HB). The paper explicitly states limitations: no 24-h urine and no human kidney tissue access.

Skeptical limitation framing: Whole-kidney homogenates average over compartments and cell types, so metabolite/cofactor changes may reflect shifts in cell composition, microenvironment, and injury state rather than cell-autonomous epigenetic coupling. The authors acknowledge the lack of cell-type specificity as a limitation.

2) Visual: how strongly PCA/PLS-DA separated age in each sex (variance explained)

The paper reports variance captured by the first two PCA components for female and male metabolomics datasets.

Reported: female PCA first two components capture 28.15% variance (PC1=19.37%, PC2=8.78%). For males, the first two PCs account for 32.75% variance (PC1/PC2 split not separately included in the provided excerpt).

3) Core biological narrative (what’s strongest vs what’s weakest)

Metabolic remodeling with age: The paper reports a shift consistent with impaired mitochondrial lipid oxidation and altered central carbon metabolism: increased glycolysis markers (e.g., lactate/LDHA upregulation) and decreased FAO/TCA intermediates and β-HB in aged kidneys (in mice).

Sex-specific patterns: They report sex-specific signatures, including enhanced amino acid catabolism and tryptophan metabolism and reduced lysophospholipase/ammonia clearance in aged females versus males, and inflammatory oxylipin/lipid signaling elevations in aged males.

Metabolism–epigenome coupling (cofactors + enzyme activity): The paper focuses on metabolite cofactors required for epigenetic enzymes (ac-CoA, β-HB, NAD+; and α-KG, SAM, FAD). It reports altered levels/activities across age and sex (e.g., increased acetyl-CoA in aged mice but reduced total acetylated histone H3; reduced β-HB; increased HDAC1/2/3 protein/HDAC-related activity changes; stable NAD+ with sex-linked sirtuin activity differences; increased α-KG and increased JMJD3C/TET activity; changes in FAD/LSD1 activity with sex-dependent mismatches).

Mechanistic plausibility: The study’s metabolic–epigenetic premise is consistent with broader literature that epigenetic enzymes require metabolite cofactors (e.g., acetyltransferases use ac-CoA; deacetylases use NAD+; demethylases use FAD or α-KG).

Skeptical evaluation (correlation ≠ causation): The strongest evidentiary thread is the concordance between metabolite cofactors and measured enzyme activity/histone marks. The weakest part is causal direction: whole-kidney homogenates cannot establish whether metabolism drives epigenetic reprogramming, or whether both are downstream of aging/injury, altered cell composition, and changing mitochondrial function. The authors explicitly acknowledge whole-tissue homogenate and limited cell specificity.

4) Visual: study design & sample sizes (as stated)

Cohort	Tissue	Ages	Sex	N per group (metabolomics)	Other assays
Mouse	Kidney homogenate; nuclear/histone extracts	4M, 12M, 24M	Male + Female (age/sex matched)	4M: n=10 (both sexes); 24M: n=10 (both sexes); 12M: n=8 (both sexes)	Western blot (LDHA/CPT1A/cPLA2/HDACs) + activity assays (HAT/HDAC/SIRT; DNMT/TET; JMJD3/LSD1) + global H3 acetylation
Human	Spot urine	Age-stratified (figures mention <45 vs >65)	Men + Women	Figure legends report small n (e.g., n=5 per group for some analyses)	Creatinine-normalized urine L-carnitine, pyruvate, lactate, α-KG, β-HB via commercial kits

Sample size statement for kidney metabolomics is explicit. Human spot urine study is described as small-n with limitations.

5) Translational concordance check (mouse kidney ↔ human urine)

The paper attempts cross-species translation by comparing directional changes in key metabolites/cofactors.

Reported directions:

L-carnitine: reduced in 24M female mice and also lower urinary in older women (>65).
Pyruvate: decreased in aged mice of both sexes; in humans decreased only in older women (no change in older men).
Lactate/β-HB: the paper reports discrepancies (mouse renal lactate/β-HB decline vs human spot urine changes not always significant).
α-KG: mouse kidney α-KG increases with age, while human urine α-KG decreases with age.

Interpretation caution: These direction mismatches could reflect compartmental differences (tissue vs excretion), renal clearance changes, or the use of spot urine rather than 24-h collections. The authors explicitly flag the missing 24-h urine limitation.

6) Mechanistic critique: the “cofactor mismatch” pattern is an opportunity, not a bug

A particularly interesting report in the paper is: acetyl-CoA increases but total acetylated histone H3 decreases with age. The authors discuss decoupling (they note HAT activity changes from earlier work).

What would strengthen the mechanistic claim? Demonstrating (1) chromatin occupancy (e.g., H3K9ac/H3K27ac/H3K4me3/H3K27me3) at specific loci, (2) cell-type-specific cofactor gradients, and (3) causality tests (perturb metabolism and observe epigenetic shifts in the same cells). The current study provides cofactor and enzyme activity readouts, but not locus-resolved chromatin maps. This is consistent with the study’s own limitation about homogenate-level inference.

7) Blind spots / alternative explanations (most important)

Cell composition vs cell-autonomous causality: Whole kidney homogenates can shift metabolite/epigenetic profiles simply due to changes in proportions of tubule vs interstitial vs immune cells during aging. (The paper partially supports functional story with nuclear enzyme assays but still cannot isolate cell-autonomous coupling.)
Spot urine as a noisy readout: Urine metabolite levels can be influenced by systemic production, filtration, and excretion rates; thus lack of 24-h urine can blur tissue-specific changes.
PLS-DA overfitting risk: Supervised methods like PLS-DA can yield separations that don’t generalize without rigorous cross-validation. The excerpt describes VIP selection but does not show how overfitting risk was mitigated. This is a general metabolomics methodological risk (not a claim of wrongdoing), and readers should seek the full methods/validation details.
Epigenetic inference is global, not locus-resolved: The study measures global H3 acetylation and activities of several enzymes, but not the downstream epigenomic targets that would link directly to transcriptional outputs. Cofactor/enzymatic changes could be compensatory or secondary to oxidative stress and mitochondrial dysfunction.

8) Suggested falsification targets (what would disprove the paper’s main mechanistic framing)

Break the metabolite–epigenome linkage: If manipulating cofactors (ac-CoA-related input, β-HB-class HDAC inhibition logic, or α-KG-driven demethylase logic) fails to move the corresponding enzyme activities and global histone marks in the same direction within renal cells, the “coupling” becomes less mechanistically meaningful (becoming purely correlative). (This is the logical falsification of their associative framework; see their own cofactor→enzyme logic.)
Cell-type causality: Cell-type-specific metabolomics and chromatin profiling should reproduce the coupling; if coupling disappears after accounting for cell composition, whole-kidney correlations may be driven by changing cell proportions.
Cross-species robustness: In larger human cohorts and with paired or more physiologically averaged urine collections (24-h), the urine biomarkers should preserve the same age/sex relationships. The paper itself notes the current translational limitations.

Author reviews (tap for more)

(Names are taken from the provided full-text metadata; full given names beyond initials weren’t explicitly extractable from the excerpt.)

Feedback:

Updated: May 01, 2026