Hypothesis

Selective activation of mTORC1 downstream effectors that trigger autophagy and stress resistance—while sparing mTORC2‑dependent Akt signaling—will prolong lifespan in mice without compromising acute immune responses to viral challenge.

Rationale

• Rapamycin and caloric restriction extend life by mimicking scarcity, but they blunt mTORC2 activity, which impairs T‑cell development and increases mortality after infection 2.
• mTORC1 inhibition alone improves protein turnover and reduces senescence markers, yet does not reproduce the full immunosuppressive profile of pan‑mTOR blockers 4.
• Age‑dependent changes in mTOR signaling mean that young animals show little phospho‑S6K response to CR, whereas middle‑aged rats display a robust decrease 3, indicating that timing and tissue specificity matter.
• Graph neural networks such as TxGNN can prioritize compounds that engage specific pathways without prior disease‑drug examples 5, but current pipelines omit immune competence as a filter 7.

Novel mechanistic insight

We propose that a subset of mTORC1‑biased allosteric modulators can preferentially activate the Rag‑GTPase–lysosome axis, leading to TFEB‑driven lysosomal biogenesis and autophagy, while leaving the mTORC2‑ribosome‑S6K4EBP1 arm untouched. This selective bias would preserve Akt‑mediated survival signals in lymphocytes, maintaining their proliferative capacity and cytokine production during infection. In contrast, pan‑mTOR inhibitors dampen both complexes, causing a hormetic stress state that looks like longevity only in pathogen‑free labs.

Testable predictions

1. In vivo lifespan – Mice treated with a mTORC1‑biased compound (e.g., RapaLink‑1 analogs that favor FKBP12‑rapastin binding) will show median survival increase comparable to rapamycin 1, but only when administered after 12 months of age to match the age‑dependent window observed in CR studies.
2. Immune competence – Upon intranasal West Nile Virus infection, treated mice will exhibit viral clearance and survival rates indistinguishable from vehicle controls, whereas rapamycin‑treated cohorts will suffer heightened mortality 2.
3. Transcriptomic signature – RNA‑seq of splenocytes will reveal unchanged expression of Akt‑phosphorylation targets (FoxO1, GSK3β) and preserved IL‑2 production, while liver tissue will show elevated LC3‑II and lysosomal genes (Lamp1, Ctsb) 4.
4. GNN validation – A TxGNN model retrained with an explicit edge label for 'immunotoxicity' (derived from public assay data) will rank the mTORC1‑biased set higher than pan‑mTOR inhibitors for longevity‑relevant phenotypes 57.

Falsifiability

If the mTORC1‑biased agent fails to extend lifespan, or if it reproduces the immunosuppression seen with rapamycin, the hypothesis is refuted. Likewise, a lack of differential pathway activation (e.g., persistent mTORC2 inhibition measured by phospho‑Akt‑S473) would invalidate the mechanistic claim.