Hypothesis

Senolytic clearance ignores a metabolic checkpoint that flips senescent cells between a KDM5‑dominated, immune‑quiet state and a KDM6‑driven, inflammatory SASP. We propose that the NAD⁺‑dependent deacetylase SIRT1 directly deacetylates KDM5B and KDM6A, altering their chromatin affinity and thereby biasing the H3K4me3/H3K27me3 balance at SASP promoters. High NAD⁺/SIRT1 activity favors KDM5B‑mediated H3K4me3 removal, keeping innate immune genes repressed and preserving a reversible, hostage‑like senescent phenotype. Low NAD⁺/SIRT1 shifts the equilibrium toward KDM6A‑dependent H3K27me3 demethylation, unlocking pro‑inflammatory SASP programs that recruit immune effectors but also promote tissue remodeling.

Mechanistic Rationale

• SIRT1 substrate preference: Both KDM5B and KDM6A contain conserved lysine residues in their catalytic domains that are acetylated in vitro (see structural models). Acetylation reduces their demethylase turnover; SIRT1 removes these acetyl groups, restoring activity.
• NAD⁺ flux as a sensor: Cellular NAD⁺ rises during oxidative stress and falls in chronic inflammation or aged tissue. This creates a bistable switch where transient NAD⁺ spikes lock senescent cells in a KDM5‑high, low‑SASP mode, while sustained NAD⁺ depletion drives a KDM6‑high, high‑SASP mode.
• Outcome dependence on tissue context: In pre‑malignant lesions, a KDM5‑high state limits immune surveillance, allowing damaged cells to persist without triggering inflammation‑mediated tumorigenesis. In wound healing, a KDM6‑high state attracts macrophages that clear debris and stimulate fibroblast proliferation.

Testable Predictions

1. Pharmacological manipulation – Treating irradiated human fibroblasts with the NAD⁺ booster NR (nicotinamide riboside) will increase SIRT1 activity, decrease H3K27me3 loss at CCL5 and TNFα promoters, and reduce SASP secretion by >40 % (measured by ELISA)[5]. Conversely, SIRT1 inhibition with EX‑527 will increase H3K27me3 loss and elevate SASP cytokines.
2. Genetic perturbation – CRISPRi‑mediated knockdown of KDM5B in senescent cells will sensitize them to SIRT1 activation, causing a premature shift to KDM6‑dependent H3K27me3 demethylation and a switch from an IL‑10‑rich to an IL‑6‑rich secretory profile[1][2]. Overexpression of a catalytically dead KDM6A (H1388A) will block SASP induction even when NAD⁺ is depleted.[4]
3. Chromatin profiling – CUT&RUN for H3K4me3 and H3K27me3 at the promoters of STING, CXCL10, and CCL5 will show reciprocal changes that correlate with intracellular NAD⁺ levels (r > 0.7, p < 0.01) across a titration of NR concentrations.[1][2]
4. In vivo relevance – In a murine liver fibrosis model, NR administration will reduce the proportion of KDM6A‑high, α‑SAS⁺ senescent hepatic stellate cells, lower collagen deposition, and improve survival without increasing hepatocellular carcinoma incidence.[3][5] Conversely, lymphoid‑specific SIRT1 knockout will exacerbate fibrosis via heightened KDM6A activity.

Falsifiability

If NAD⁺ modulation fails to alter the H3K4me3/H3K27me3 balance at SASP loci, or if SIRT1 activity does not change the demethylase occupancy of KDM5B/KDM6A (measured by ChIP‑qPCR), the hypothesis is refuted. Likewise, if forced KDM5B loss does not redirect the SASP toward an inflammatory phenotype regardless of NAD⁺ status, the proposed epigenetic gatekeeper role is invalid.

Implications

Reframing senescent cells as metabolically tunable negotiators suggests that senolytic timing should consider the NAD⁺/SIRT1 axis: boosting NAD⁺ early may lock cells into a protective, non‑inflammatory state, preserving tissue homeostasis, while later NAD⁺ depletion could be harnessed to provoke immune clearance of dangerous senescent populations.