Hypothesis

Age-associated decline in NAD+ reduces SIRT1 activity, leading to hyperacetylation of KDM5A/B demethylases and diminished H3K4me3 removal in hematopoietic stem cells (HSCs). This epigenetic shift preserves active marks at tumor‑suppressor promoters while concurrent loss of KDM6B‑mediated H3K27me3 demethylation allows re‑accumulation of the repressive mark, generating expanded bivalent domains at loci such as RUNX1 and HIF1A. The resulting chromatin state poises HSCs for aberrant differentiation and clonal expansion, providing a mechanistic link between metabolic aging, demethylase dysfunction, and pre‑leukemic transformation.

Mechanistic Rationale

1. NAD+/SIRT1 axis – NAD+ levels fall with age in multiple tissues, decreasing SIRT1 deacetylase activity (3). SIRT1 directly deacetylates lysine residues on KDM5A/B, a modification required for optimal catalytic turnover (1).
2. Effect on KDM5 activity – Hyperacetylated KDM5 exhibits reduced affinity for H3K4me3 substrates, lowering demethylase flux. Consequently, H3K4me3 persists at promoters of tumor‑suppressor genes that are normally kept in a poised state.
3. Parallel KDM6B attenuation – Age‑related oxidative stress and altered α‑ketoglutarate/succinate ratios impair JmjC domain‑containing KDM6B enzymes, decreasing H3K27me3 removal. The combined loss of KDM5 and KDM6B activity favors accumulation of both marks, creating bivalent domains.
4. Stem‑cell outcome – Bivalent promoters maintain transcriptional plasticity; in aging HSCs this manifests as inappropriate low‑level expression of lineage‑specifying factors and heightened sensitivity to inflammatory cytokines, driving clonal hematopoiesis and leukemia initiation (4).

Testable Predictions

• Pharmacological rescue – Supplementing aged mice with NAD+ precursors (e.g., NR) or SIRT1 activators should restore KDM5 deacetylation, reduce H3K4me3 levels at RUNX1/HIF1A promoters, and decrease bivalent domain frequency in HSCs.
• Genetic dissection – HSC‑specific expression of an acetylation‑deficient KDM5A mutant (K→R) in aged mice will prevent bivalent domain expansion and reduce pre‑leukemic clones, whereas an acetylation‑mimic mutant (K→Q) will phenocopy aging.
• Metabolite profiling – Aged HSCs will show lowered α‑ketoglutarate/succinate ratios correlating with KDM6B activity loss; rescuing α‑ketoglutarate levels should restore H3K27me3 demethylation without affecting KDM5.
• ChIP‑seq for KDM5/6 occupancy – Conditional KDM5A/B deletion in young HSCs will mimic the aging bivalent pattern, confirming that loss of demethylase activity, not altered recruitment, drives the epigenetic shift.

Potential Confounding Factors

• Inflammatory cytokines can independently modulate KDM expression; experiments should control for serum IL‑6/TNFα levels.
• Cell‑cycle status influences histone turnover; assays should normalize for HSC proliferation rates (e.g., Ki‑67 staining).

By linking metabolic aging to the enzymatic state of KDM5/6 demethylases, this hypothesis translates correlative epigenomic observations into a causal, experimentally tractable framework.