Age‑dependent FTO‑mediated demethylation of proteostasis transcripts drives neuronal decline via loss of m6A‑dependent translation

In aging neurons, upregulated activity of the RNA demethylase FTO selectively removes m6A marks from transcripts encoding core proteostasis factors (LAMP2A, proteasome subunits, ATF6α, PERK). This loss of m6A diminishes YTHDF1‑dependent translation and alters RNA export/stability, leading to reduced protein synthesis of these key machines and precipitating proteostasis collapse.

Mechanistic Rationale

• m6A generally enhances translation through YTHDF1 binding and can promote nuclear export via YTHDC1; loss of the mark therefore decreases ribosome loading and cytoplasmic mRNA abundance.
• ’s analysis shows that normal brain aging is associated with hypomethylation of ubiquitin‑mediated proteolysis genes while synaptic genes gain m6A[1]. This transcript‑specific pattern argues against a global METTL3/14 decline and points to active demethylation of select mRNAs.
• Proteostasis failure in aged neurons occurs at the protein level—LAMP2A stability drops without mRNA change[3] and UPR sensors PERK/ATF6α lose activity[4]—suggesting a post‑transcriptional block that could stem from defective translation of their transcripts.
• FTO expression rises with age in several tissues and can preferentially target m6A sites near start codons, a motif enriched in LAMP2A and ATF6α 5’UTRs (predicted from existing epitranscriptomic maps). Demethylation in these positions would specifically impair initiation complex formation.

Testable Predictions

1. m6A mapping: DART‑seq or miCLIP of young versus aged mouse hippocampal neurons will reveal significant loss of m6A peaks on LAMP2A, PSMC5, ATF6A, and EIF2AK3 (PERK) transcripts, while synaptic transcripts (e.g., SYN1, GRIN1) show gains.
2. FTO manipulation: Knockdown of FTO in aged neurons will restore m6A levels on proteostasis mRNAs, increase their ribosome occupancy (Ribo‑seq), and elevate LAMP2A and proteasome subunit proteins without altering mRNA abundance.
3. Functional rescue: Restoring m6A (via FTO knockdown or overexpression of a catalytically dead METTL3) will improve chaperone‑mediated autophagy flux (measured by KFERQ‑reporter clearance) and reduce lipofuscin accumulation in aged mouse cortex.
4. YTHDF1 dependence: The rescue effect will be abolished in neurons lacking YTHDF1, confirming that m6A’s translational role—not its impact on decay—is critical for proteostasis maintenance.

Experimental Approach

• Cohort: Primary hippocampal neurons derived from young (2 mo) and aged (24 mo) mice; parallel cultures from human iPSC‑derived neurons treated with pro‑aging stressors (e.g., low‑dose rotenone).
• m6a profiling: Perform DART‑seq followed by motif analysis focusing on 5’UTR and start‑codon regions.
• Fto perturbation: Use AAV‑shRNA or CRISPRi to knockdown FTO; include overexpressing wild‑type FTO as a gain‑of‑function control.
• Readouts:
  • RNA‑seq for transcript levels.
  • Ribo‑seq or puromycin‑based SUnSET assay for translation efficiency.
  • Western blot / quantitative proteomics for LAMP2A, PSMA5, ATF6A, PERK.
  • Functional assays: DQ‑BSA uptake for CMA, Lysotracker for lysosomal activity, thioflavin‑T for tau/Aβ aggregation.
  • Imaging: Lipofuscin autofluorescence, ubiquitin‑positive inclusions.
• Controls: Scramble shRNA, catalytically inactive FTO mutant, YTHDF1 knockout neurons.

Potential Outcomes & Interpretation

• If predictions hold: Loss of m6A on proteostasis transcripts is a causal driver of age‑related translational insufficiency, linking epitranscriptomic dysregulation directly to proteostasis collapse. This would shift focus from METTL3/14 loss to demethylase gain‑of‑function as a therapeutic target.
• If m6a levels remain unchanged: The hypothesis is falsified; alternative mechanisms (e.g., altered RNA‑binding protein composition or microRNA regulation) must be sought for the observed protein deficits.
• If fto knockdown restores m6a but not protein levels: Suggests that m6a loss is correlative or that additional layers (e.g., increased proteasomal degradation of nascent chains) dominate.

By directly testing whether age‑induced demethylation of specific mRNAs compromises their translation, this work bridges the discordant m6a patterns observed in the aging brain with the well‑characterized post‑transcriptional failure of neuronal proteostasis.