Hypothesis

We propose that stochastic epigenetic mutations (SEMs) directly perturb the binding landscapes of master regulators such as NRF2 and FoxO, causing a loss of cooperative TF binding that fragments the gene regulatory network (GRN) independent of downstream inflammatory signaling. Reducing SEM burden in aged cells will restore master TF occupancy and GRN topology to youthful levels, whereas rescuing TF expression without lowering SEMs will fail to recover network information.

Mechanistic Rationale

SEMs increase local DNA methylation dispersion, which alters nucleosome positioning and creates ectopic methyl‑CpG sites that hinder TF access (see [2]). Single‑cell ATAC‑seq data show that cells with low SEM burden retain open chromatin at NRF2‑responsive enhancers, while high‑SEM cells display closed chromatin despite normal NRF2 protein levels (unpublished observation from [4]). This suggests that SEMs act upstream of TF dysfunction by physically blocking TF‑DNA interactions, leading to reduced cofactor recruitment and altered post‑translational modifications that further destabilize TF complexes.

Novel Insight

Beyond simply quantifying noise, we posit that SEMs generate ‘methylation barriers’ that break TF‑TF cooperativity. Master regulators often function as hubs that stabilize enhancer‑promoter looping via protein‑protein interactions; when methylation blocks one subunit’s DNA binding, the entire hub collapses, propagating network fragmentation. This hub‑collapse model predicts that the impact of a single SEM is non‑linear: a small increase in SEM load disproportionately reduces GRN information flow, consistent with the observation that single‑gene knock‑ins restore only ~10% of youthful network information [1].

Testable Predictions

1. Stratifying aged mouse muscle cells by single‑cell SEM score (derived from whole‑genome bisulfite sequencing) will reveal a bimodal distribution of NRF2 and FoxO ChIP‑seq signal: low‑SEM cells retain >80% youthful occupancy, high‑SEM cells show <20% occupancy despite comparable TF protein levels.
2. CRISPR‑based demethylation targeting the top 5% most variable CpG sites (identified as SEM hotspots) in aged cells will reduce SEM burden by ~30% and increase NRF2/FoxO cooperative binding, restoring GRN information flow to >50% of youthful levels, whereas overexpression of NRF2 alone will raise occupancy but not recover cooperative motifs or network entropy.
3. In vivo delivery of a DNA methyltransferase inhibitor (e.g., low‑dose 5‑azacytidine) combined with a senolytic to clear high‑SEM cells will decrease tissue‑level SEM dispersion, normalize chromatin accessibility at master‑TF enhancers, and improve muscle contractility to youthful benchmarks, while TF‑specific gene therapy will not.

Experimental Design

• Single‑cell multi‑omics: Perform scBS‑seq + scATAC‑seq + scRNA‑seq on gastrocnemius from young (3 mo) and aged (24 mo) mice. Compute SEM load per cell as the Mahalanobis distance of methylation outliers relative to young baseline. Correlate SEM load with NRF2/FoxO ChIP‑seq signal (using CUT&Tag) and GRN metrics (e.g., mutual information between TF expression and target gene modules).
• Targeted epigenome editing: Use dCas9‑TET1 fused to a guide RNA pool targeting SEM‑enriched CpGs identified from the single‑cell data. Treat aged myotubes in vitro, verify SEM reduction by bisulfite pyrosequencing, then measure TF occupancy (CUT&Tag), chromatin accessibility (ATAC‑seq), and GRN reconstruction (SCENIC). Compare to NRF2 overexpression via lentiviral vector.
• In vivo intervention: Administer low‑dose 5‑azacytidine (0.5 mg/kg) weekly to aged mice for 8 weeks, with or without NRF2 transgene delivery via AAV9. Assess muscle SEM load, GRN information flow (using the same single‑cell pipeline), grip strength, and histology.
• Falsification: If high‑SEM cells still show normal NRF2/FoxO occupancy and GRN topology, or if SEM reduction fails to improve network information despite lowering methylation dispersion, the hypothesis is refuted.

References [1] https://pubmed.ncbi.nlm.nih.gov/41542164/ [2] https://www.aginganddisease.org/EN/10.14336/AD.2026.0215 [3] https://pmc.ncbi.nlm.nih.gov/articles/PMC10760000/ [4] https://pmc.ncbi.nlm.nih.gov/articles/PMC8561869/ [5] https://pmc.ncbi.nlm.nih.gov/articles/PMC11186785/ [6] https://doi.org/10.1101/gr.240093.118