Hypothesis

Aging-related epigenetic noise first accumulates as heterogeneous H3K27me3 marks at super‑enhancers, increasing transcription factor binding entropy and destabilizing homeostatic network motifs. Targeted erasure of these marks restores TF binding specificity before global GRN topology collapses, thereby lowering inflammaging without inducing ectopic pluripotency.

Mechanistic Rationale

Single‑cell multi‑omics data show that aged fibroblasts and muscle cells display increased variance in H3K27me3 at loci bound by master regulators such as EZH2 and CTCF [1][2]. This variance correlates with higher TF binding entropy measured by motif‑scoring across accessible regions, predicting loss of negative feedback loops and mutual inhibition structures that preserve cell fate [4]. We propose that the epigenetic scar at super‑enhancers acts upstream of GRN topology drift: when H3K27me3 becomes noisy, master TFs bind promiscuously, weakening circuit stability and allowing inflammatory NF‑κB programs to leak. Conversely, precise removal of the scar—using a dCas9‑HDAC3 fusion guided to H3K27me3‑rich super‑enhancers—should sharpen TF occupancy, reinstate feedback loops, and suppress NF‑κB target expression.

Experimental Design

1. Identify scar loci – Perform scATAC‑seq and scCUT&Tag for H3K27me3 on young (3 mo) and aged (20 mo) mouse liver hepatocytes; call super‑enhancers with MACS2 and rank by H3K27me3 variance (entropy > 0.6).
2. Perturbation – Deliver AAV‑dCas9‑HDAC3 with sgRNAs targeting the top 10 scar loci to aged mice; control groups receive AAV‑dCas9‑dead or non‑targeting sgRNA.
3. Readouts – After 4 weeks, collect liver for:
   • scRNA‑seq to compute GRN information transfer (using the info‑theory metric from [1])
   • scCUT&Tag for H3K27me3 and HDAC3 occupancy
   • ATAC‑seq footprinting to quantify TF binding entropy at perturbed sites
   • Serum cytokines (IL‑6, TNF‑α) and histology for fibrosis.
4. Scaling test – Repeat in human liver organoids derived from donors >65 y, measuring same endpoints.

Predictions and Falsifiability

• Prediction 1: Scar‑targeted HDAC3 recruitment will reduce H3K27me3 variance at super‑enhancers by ≥30 % compared with controls (p < 0.01).
• Prediction 2: TF binding entropy at those loci will drop correspondingly, restoring ≥15 % of lost GRN information transfer measured in young tissue.
• Prediction 3: Serum IL‑6 and TNF‑α will decrease by ≥25 % and fibrosis scores improve, without ectopic expression of pluripotency markers (OCT4, NANOG).
• Falsification: If HDAC3 recruitment fails to lower H3K27me3 variance or TF binding entropy, and GRN information transfer and inflammatory markers remain unchanged, the hypothesis that epigenetic scar precedes GRN collapse is falsified. Likewise, induction of pluripotency markers would indicate off‑target reprogramming, contradicting the claimed specificity. By linking epigenetic scar clearance to upstream restoration of network motifs, this hypothesis offers a testable, mechanistic bridge between the observed reversibility of histone‑modification drift [3] and the limited rescue achieved by single‑TF perturbations [1][2]. Success would prioritize enhancer‑focused epigenome editing as a safer route to organismal rejuvenation.