Hypothesis

mTORC1 activity does not only regulate EZH2 translation and H3K27me3 levels; it also governs the nuclear import of the H3K27me3 demethylase Jmjd3 via phosphorylation of Importin‑7. This creates a biphasic control of CDKN2A/B expression: moderate mTOR inhibition favors Jmjd3 nuclear accumulation and transient p16^INK4a up‑regulation (a protective senescence response), whereas strong or chronic inhibition impairs Importin‑7‑mediated import, reducing nuclear Jmjd3 and paradoxically lowering CDKN2A/B transcription despite low EZH2 activity.

Mechanistic Rationale

1. mTORC1‑dependent phosphorylation of Importin‑7 – mTORC1 substrates include ribosomal protein S6 kinase (S6K) and 4E‑BP1; we propose that Importin‑7 is a direct mTORC1 target. Phosphorylation enhances its affinity for the Jmjd3‑importin complex, promoting nuclear entry.
2. Jmjd3 activity requires nuclear localization – demethylation of H3K27me3 at the CDKN2A/B promoter can only occur when Jmjd3 reaches chromatin. Cytoplasmic retention renders the enzyme inactive regardless of its expression level.
3. Dose‑dependent outcome – At basal or mildly reduced mTORC1 activity (e.g., intermittent fasting), Importin‑7 remains sufficiently phosphorylated, allowing Jmjd3 to enter the nucleus and demethylate the locus, raising p16^INK4a to a level that enforces cell‑cycle arrest without triggering chronic inflammation. At profound mTORC1 suppression (e.g., high‑dose rapamycin, genetic Raptor loss), Importin‑7 phosphorylation drops, Jmjd3 stays cytoplasmic, H3K27me3 persists, and CDKN2A/B transcription falls, mirroring the Alzheimer’s‑associated downregulation.

Testable Predictions

• Prediction 1: In primary human fibroblasts, a titration of rapamycin (0‑100 nM) will produce a bell‑shaped curve for nuclear Jmjd3 (measured by subcellular fractionation and immunofluorescence) and p16^INK4a mRNA, with peak nuclear Jmjd3 and p16 at ~10‑20 nM rapamycin.
• Prediction 2: siRNA‑mediated knockdown of Importin‑7 will mimic high‑dose rapamycin effects: decreased nuclear Jmjd3, increased promoter H3K27me3 (ChIP‑qPCR), and reduced p16^INK4a despite low mTORC1 signaling (assessed by p‑S6 levels).
• Prediction 3: Post‑mortem Alzheimer’s cortex will show (a) reduced p‑Importin‑7 (phospho‑specific antibody), (b) low nuclear Jmjd3, (c) elevated H3K27me3 at the CDKN2A/B promoter, and (d) diminished p16^INK4a mRNA compared with age‑matched controls.
• Prediction 4: Expressing a phospho‑mimetic Importin‑7 mutant (S→D) in neurons exposed to chronic rapamycin will rescue nuclear Jmjd3 localization and restore p16^INK4a expression.

Falsifiability

If any of the following are observed, the hypothesis is refuted:

• Nuclear Jmjd3 levels do not change across a range of mTORC1 activities.
• Importin‑7 knockdown does not alter nuclear Jmjd3 or p16^INK4a under low‑mTOR conditions.
• Alzheimer’s tissue shows normal or increased nuclear Jmjd3 despite low p16^INK4a.
• Phospho‑mimetic Importin‑7 fails to rescue Jmjd3 nuclear entry or p16^INK4a expression under mTORC1 inhibition.

Implications

This model reframes mTOR not merely as a regulator of the silencing arm (EZH2/H3K27me3) but as a bifunctional tuner of both repression and activation arms of the CDKN2A/B locus. It explains why longevity interventions that dampen mTOR can yield both beneficial senescence (tumor suppression, tissue homeostasis) and detrimental outcomes (impaired regeneration, neurodegeneration) depending on the depth of inhibition. Importantly, it offers a concrete molecular target—Importin‑7 phosphorylation—for fine‑tuning mTOR‑based therapies to preserve the "civilizational" growth program while engaging the "survival" senescence checkpoint when needed.