Hypothesis

Progressive epigenetic noise creates a bistable regime in transcription factor (TF) networks where aged cells reside in a stable, noisy attractor that resists full reversion by single TF overexpression due to hysteresis driven by retrotransposon‑derived RNA sequestration of chromatin regulators.

Mechanistic Basis

Epigenetic drift—loss of precise DNA methylation and histone patterns—precedes genomic instability and increases TF targeting density across immune and oncogenic pathways [1]. This drift weakens NuRD complex activity, making chromatin more permissive to spurious transcription [2]. Consequently, endogenous retrotransposons (SINE/Alu, ERVs) become derepressed, producing RNAs that can bind and titrate away TFs and co‑factors such as EZH2 [5], [6].

We propose that these retrotransposon RNAs form a negative feedback loop: they sequester TFs attempting to restore youthful chromatin states, thereby raising the effective concentration of TF needed to shift the network back to the young attractor. The system exhibits hysteresis—once the noisy attractor is entered, a larger TF perturbation is required to exit than to enter it. This explains why overexpression of individual TFs like EZH2 restores only ~10% of network information [4]; the applied TF dose lies below the hysteresis threshold.

Testable Predictions

1. Threshold Shift: In aged tissues, the dose‑response curve for TF‑induced network reversion will be right‑shifted relative to young cells, measurable by single‑cell multi‑omics after graded TF induction.
2. RNA Sequestration: Retrotransposon RNAs will physically associate with TFs (e.g., EZH2, FOXM1) in aged cells; RNase H treatment or antisense knockdown will lower the TF dose needed for network reset.
3. NuRD Overexpression Bypasses Hysteresis: Restoring NuRD subunit levels (e.g., CHD4, MTA2) in aged cells will reduce retrotransposon RNA accumulation and decrease the hysteresis width, allowing lower TF doses to achieve full rejuvenation.
4. Information‑Loss Ceiling Correlates with RNA Load: The fraction of GRN information recoverable by TF knock‑in will inversely correlate with endogenous retrotransposon RNA abundance across tissues and species.

Experimental Design

• Graded TF Induction: Use doxycycline‑titratable EZH2 or FOXM1 constructs in primary human fibroblasts and mouse liver; collect scRNA‑seq, scATAC‑seq, and CUT&Tag for H3K27ac at 0, 12, 24, 48 h.
• Quantify Bistability: Fit TF dose vs. proportion of cells exhibiting youthful GRN signature (PANDA/LIONESS‑derived) to a sigmoid; compare EC50 between young and aged samples.
• Retrotransposon Perturbation: Transfect LNA‑gapmers targeting abundant SINE/Alu RNAs; repeat TF dose‑response to assess leftward shift in EC50.
• NuRD Rescue: CRISPRa‑mediated upregulation of CHD4 in aged cells; measure changes in retrotransposon RNA levels, chromatin accessibility at ERV loci, and TF occupancy.
• Information Metric: Apply the information‑theoretic framework from [4] to compute recovered GRN information after each perturbation.

Falsifiability

If aged cells show no right‑shifted TF dose‑response, or if retrotransposon RNA knockdown does not lower the TF threshold for network reversion, the hysteresis model is invalid. Likewise, if NuRD overexpression fails to reduce retrotransposon RNA or improve TF efficacy, the proposed mechanistic link between chromatin regulator loss and RNA‑mediated sequestration must be rejected.

Implications

This hypothesis reframes the 10% restoration ceiling not as a hard limit on TF potency but as a systems‑property barrier arising from epigenetic noise‑driven RNA sequestration. It suggests combinatorial strategies—simultaneous TF activation and retrotransposon silencing or NuRD reinforcement—are required to overcome hysteresis and achieve comprehensive GRN rejuvenation.