Hypothesis

Aging shifts the PERK arm of the unfolded protein response from a protective regulator of satellite cell regeneration to a driver of CHOP‑dependent apoptosis in post‑mitotic tissues. We propose that this dichotomy is not merely a consequence of differential arm activation but is actively governed by a tissue‑restricted RNA‑binding protein, PERK‑modulating factor 1 (PMF1), which skews PERK downstream signaling toward ATF4‑mediated autophagy in regenerating muscle and toward CHOP‑mediated apoptosis in aged brain and heart.

Key mechanism PMF1 contains an RNA recognition motif that binds a conserved stem‑loop in the 5′‑UTR of Atf4 mRNA, enhancing its translation when PERK is active. In skeletal muscle, PMF1 is highly expressed and forms a complex with PERK and the eukaryotic initiation factor 2α phosphatase complex (PP1‑GADD34). This complex stabilizes phosphorylated eIF2α just long enough to favor selective translation of ATF4 while limiting prolonged translation repression that would favor CHOP synthesis. Consequently, ATF4 drives expression of autophagy genes (e.g., LC3, Becn1) and satellite‑cell differentiation regulators (e.g., Myod1, Myog).

In aged brain and cardiomyocytes, PMF1 expression declines, and PERK preferentially associates with the stress‑induced protein CHOP‑promoting factor (CPF). CPF recruits the transcriptional co‑activator CBP/p300 to the Chop promoter, amplifying CHOP transcription even when ATF4 levels are modest. The resulting CHOP surge suppresses autophagy via Bcl‑2 interaction and activates JNK‑mediated apoptosis.

Novel insight beyond the cited work While the provided literature notes that PERK deletion causes hyper‑active p38 MAPK in satellite cells and that SIRT1 suppresses PERK/CHOP apoptosis in heart, it does not explain why PERK’s transcriptional output diverges across tissues. Our hypothesis introduces a post‑transcriptional regulatory layer—PMF1‑dependent Atf4 translation—that can be experimentally uncoupled from PERK kinase activity, thereby resolving the paradox of PERK being simultaneously essential for regeneration and detrimental in aging.

Testable predictions

1. PMF1 knockdown in mouse satellite cells will attenuate ATF4‑dependent autophagy markers (LC3‑II, p62 degradation) after acute injury, impair regeneration, and phenocopy PERK knockout (reduced MyoD/Myog expression, rescued by p38 inhibition).
2. Overexpression of PMF1 in aged hippocampal neurons will shift PERK signaling toward ATF4, increase autophagosome formation (GFP‑LC3 puncta), reduce CHOP protein levels, and protect against tunicamycin‑induced apoptosis.
3. Disruption of the PERK‑PMF1 interaction (using a cell‑permeable peptide mimicking the PMF1 binding domain) will decrease ATF4 translation without affecting PERK autophosphorylation, leading to increased CHOP/JNK signaling in both muscle and brain contexts.
4. SIRT1 activation will increase PMF1 expression in muscle (via deacetylation of a putative transcriptional activator) but not in brain, providing a mechanistic link between the observed tissue‑specific SIRT1 effects and PERK output bias.

Experimental approach

• Generate conditional Pmf1 floxed mice; cross with Pax7‑CreERT2 for satellite‑cell deletion and CamKII‑Cre for forebrain neuron deletion.
• Subject mice to cardiotoxin injury (muscle) or tunicamycin intraperitoneal injection (brain) and assess regeneration (centrally nucleated fibers, grip strength) versus apoptosis (TUNEL, cleaved caspase‑3) and autophagy (LC3‑II/I ratio, electron microscopy).
• Perform ribosome profiling to quantify Atf4 versus Chop translation efficiency in each condition.
• Rescue experiments: administer SIRT1 activator (SRT2104) or p38 inhibitor (SB203580) to test epistasis.

Falsifiability If PMF1 manipulation fails to alter the ATF4/CHOP balance or does not affect regeneration/apoptosis as predicted, the hypothesis would be refuted, indicating that tissue‑specific PERK outcomes arise solely from kinase activity differences or alternative scaffolds.