Hypothesis

Age‑related drift of master regulator transcription factors (TFs) such as FOXO3 and NF‑κB alters the biophysical state of Polycomb repressive complex 2 (PRC2/EZH2) condensates, leading to aberrant phase separation that erodes heterochromatin at lamina‑associated domains and endogenous retroviral elements. This biophysical change precedes and drives the observed increase in epigenetic noise and GRN rewiring, rather than being a mere downstream consequence.

Mechanistic Model

1. TF flux modulates PRC2 concentration and post‑translational state – In young cells, high FOXO3 activity promotes EZH2 phosphorylation that keeps PRC2 in a dynamic, liquid‑like state permissive for targeted H3K9me3/H3K20me3 deposition at specific loci. Aging reduces FOXO3 and increases NF‑κB, shifting EZH2 toward a hypophosphorylated, oligomer‑prone conformation.
2. Aberrant phase separation creates stiff PRC2 aggregates – The altered EZH2 recruits additional PRC2 subunits and chromatin, forming solid‑like foci that sequester H3K9 methyltransferases and HP1, thereby depleting heterochromatin marks globally while causing focal hyper‑methylation at inflammatory promoters.
3. Feedback to TF accessibility – Loss of heterochromatin lifts repression of endogenous retroviral sequences, triggering innate immune signaling that further activates NF‑κB, reinforcing the TF drift and stabilizing the aberrant condensate state.
4. Noise emergence – Stochastic variations in condensate size and composition generate cell‑to‑cell heterogeneity in chromatin accessibility, measurable as increased epigenetic noise in scATAC‑seq data, particularly in postmitotic tissues where dilution through cell division is absent.

Testable Predictions

• Prediction 1: Acute optogenetic increase of FOXO3 nuclear activity in aged fibroblasts will dissolve PRC2 solid foci, restore H3K9me3/H3K20me3 levels at lamina‑associated domains, and reduce transcriptional noise within 24 h.
• Prediction 2: Pharmacological inhibition of EZH2 oligomerization (using a small‑molecule that blocks its SAM‑domain interaction) will mimic the effect of FOXO3 activation, even without altering TF levels.
• Prediction 3: In vivo expression of a phosphomimetic EZH2 mutant in aged mice will prevent age‑dependent heterochromatin loss, lower inflammatory gene signatures, and improve tissue‑specific functional readouts (e.g., grip strength, cardiac ejection fraction) compared with wild‑type controls.
• Prediction 4: Single‑cell multi‑omics (scATAC‑seq + scRNA‑seq) from sorted cardiomyocytes will reveal a bimodal distribution of PRC2 condensate markers (e.g., EZH2 intensity, H3K27me3 foci) that correlates with noise scores; shifting the distribution toward the low‑noise mode via TF flux rescue will decrease the proportion of high‑noise cells.

Experimental Approach

• Optogenetic TF control: Fuse FOXO3 to a light‑inducible nuclear localization signal (LOVpep) in primary human fibroblasts from elderly donors; apply blue light pulses and measure PRC2 dynamics via live‑cell EZH2‑HaloTag imaging.
• Condensate pharmacology: Treat cells with EZH2 oligomerization inhibitors (e.g., PGC‑744) and assess heterochromatin marks by CUT&RUN for H3K9me3/H3K20me3.
• In vivo validation: Generate AAV9‑mediated cardiac‑specific expression of EZH2‑S21E (phosphomimetic) in 24‑month‑old mice; perform longitudinal echocardiography, fibrosis staining, and snRNA‑seq + snATAC‑seq at 3‑month intervals.
• Data analysis: Use hidden Markov models to infer condensate states from imaging data; link state transitions to changes in gene expression variance (noise) and chromatin accessibility across cell types.

If these experiments confirm that restoring youthful TF‑dependent PRC2 phase behavior reverses heterochromatin loss and noise, the hypothesis establishes TF flux as a causal upstream regulator of epigenetic noise and GRN rewiring, providing a mechanistic bridge between the correlative observations in the cited literature and actionable intervention strategies.