Can different combinations of inputs give the same average mTORC1 activity but a different time course — and a different outcome?

Framing gap · evidence stands at: Mechanism in cells · Atlas ID F6

The gap

mTOR listens to several things at once: food (amino acids), hormones like insulin, the cell's energy level, and stress. Experiments usually change one of these at a time and read the answer once. But two different mixes of signals could give the same average reading while one produces a quick spike and the other a slow, steady rise. If the cell responds to the shape and not only the average, those two mixes would do different things, and a single reading could not tell them apart.

Technical framing: mTORC1 integrates amino acids through the Rag GTPases, growth factors and insulin through AKT and TSC2, energy through AMPK, and lysosomal and stress signals besides. Most studies vary one input at a time and read the result at steady state. Two different input combinations could therefore be reported as producing the same mTORC1 activity while producing different kinetics: a fast transient against a slow plateau, or a steady level against an oscillation. An averaged readout would call them identical.

What changed

The inputs are now well mapped at the molecular level (Goul, Peruzzo and Zoncu, Nat Rev Mol Cell Biol 2023, doi:10.1038/s41580-023-00641-8). In mouse liver, the rise of mTORC1 after a meal needs dietary protein, while insulin's AKT-TSC2 route explains the insulin response but not the feeding response (KAL2026), so two physiological inputs reach the same complex by routes that can be separated. Live reporters and single-cell imaging of mTORC1, AMPK and AKT signalling make it possible to record kinetics rather than end points.

What is still open

Whether input combinations matched for average mTORC1 activity produce distinguishable time courses in single cells, and whether those time courses lead to different outputs such as TFEB localisation, autophagic flux or protein synthesis. We have not found a study that matches the average and compares the dynamics directly.

How it could be tested

In cells carrying a live mTORC1 reporter, titrate amino acids against insulin or growth factors to find several combinations that give the same time-averaged reporter signal. Record single-cell kinetics for each, together with TFEB localisation and a translation readout. Different kinetics with different outputs at equal average supports the idea. Identical kinetics, or different kinetics with identical outputs, argue against it.

Why it matters

If inputs are encoded in the time course, then what a cell registers about its nutrient state is partly lost in a steady-state measurement, and diet composition and meal timing could matter in ways a single fasted or fed sample would not show.

Evidence behind this question

Gap not yet confirmed by a logged literature search. The text states that the comparison was not found, but the literature search behind that statement is not logged. (checked 2026-10-15)

How it relates to the other frontier questions: The input-side version of F1: F1 asks whether outputs read the pattern, F6 asks whether different input combinations with the same average produce different patterns in the first place.

Related evidence

• Goul, Peruzzo and Zoncu 2023 (GOU2023) — Review of how amino acids, growth factors, energy and stress signals reach mTORC1.
• Kalafut et al. 2026 (KAL2026) — In mouse liver the post-meal rise of mTORC1 needs dietary protein, while insulin's AKT-TSC2 route explains the insulin response but not the feeding response.

Assumption not yet verified

• Live reporters of mTORC1, AMPK and AKT — The text says single-cell imaging makes kinetics recordable. No citation is given yet.

The grouping is the curator's reading of how each paper is used here. It has not been checked by a second reader.

Bears on these open questions

• Which methionine sensor, if either, carries the benefit of methionine restriction?
• A sensor/autophagy biomarker panel as a surrogate endpoint for human trials

← All frontier questions · Open questions