PI3K/AKT/mTOR Signalling in T Cells: Pathway, Function and Assays

PI3K does not switch on a protein — it rewrites a membrane. Everything downstream follows from one chemical event: the 3-hydroxyl of a phosphoinositide is phosphorylated, PIP₂ becomes PIP₃, and every PH-domain protein in the cytosol suddenly has somewhere to dock. In a T cell that single change converts antigen recognition into growth, glycolysis and a choice between a short-lived effector and a memory cell. Reading the pathway well means knowing which phosphorylations are causal, which are merely convenient readouts, and where the brakes sit.

From the TCR to PIP₃: building the lipid signal

Peptide–MHC engagement of the TCR/CD3 complex brings Lck into the synapse to phosphorylate the CD3 ITAMs and dock ZAP-70, which phosphorylates the adaptor LAT and recruits Grb2. Lck is itself restrained by Csk, which phosphorylates its inhibitory C-terminal tyrosine, so the cascade is gated by a kinase–kinase competition before any lipid is made. A weak agonist peptide can therefore give apparently normal ZAP-70 phosphorylation and still fail to sustain PI3K output.

The PI3K step changes the chemistry of the signal. p85 docks on phosphotyrosine motifs — the YMNM motif of CD28 and its equivalent on ICOS — relieving its inhibition of the p110δ catalytic subunit, and active PI3K then phosphorylates the 3-position of PIP₂ to give PIP₃. Nothing is switched on by proteolysis; the cell rewrites its inner leaflet and every PH-domain protein responds. ICOS couples to p110δ more strongly than CD28, which is why it dominates in follicular helper cells.

Three activities erase that signal, and they are not equivalent. PTEN removes the 3-phosphate to regenerate PIP₂; SHIP-1 removes the 5-phosphate, producing a different lipid with its own effectors. PD-1 acts higher up: its ITSM recruits SHP-2, which dephosphorylates CD28 in preference to the TCR itself, so checkpoint blockade largely restores the CD28–PI3K arm rather than TCR signalling. CTLA-4 reaches the same end differently, by stripping B7 ligands away from CD28 altogether.

AKT: two phosphorylation sites, two kinases

PIP₃ recruits PDK1 and AKT through their PH domains, and PDK1 phosphorylates AKT at T308 in the activation loop. Full activity needs a second phosphorylation at S473, delivered by mTOR in its mTORC2 configuration with mLST8. The sites are not redundant: T308 alone supports TSC2 and GSK3β phosphorylation, but nuclear exclusion of FOXO also requires S473. PHLPP1 strips S473 specifically while PP2A acts on T308, so the two phosphatases decide which substrates lose signal first.

AKT then works mostly by inhibition. It switches off GSK3β, phosphorylates BAD to release it from Bcl-2 family partners, and drives FOXO1 and FOXO3 out of the nucleus. Losing FOXO1 output is much of what an activated T cell is, because FOXO1 maintains IL-7R, CD62L and TCF-1 — the quiescence and homing programme. Sustained AKT signalling does not merely add an effector programme; it dismantles the naive and memory one.

TSC1/TSC2, RHEB and the nutrient checkpoint

mTORC1 — mTOR with RAPTOR, mLST8 and the inhibitory subunit DEPTOR — is not switched on by AKT directly; AKT releases two brakes instead. It phosphorylates TSC2, displacing the TSC1–TSC2 complex from the lysosomal surface, and it phosphorylates PRAS40 off RAPTOR. Because TSC1/TSC2 is the GAP for RHEB, its removal lets RHEB-GTP accumulate and allosterically activate the kinase. RHEB-GTP is the only direct activator here; everything above it is release of inhibition, not positive input.

The nutrient input arrives separately, and in a different place. Leucine imported by LAT1 and glutamine handled by glutaminase are sensed in part through Sestrin 2 and relayed to the RagA/B GTPases, which deliver mTORC1 to the lysosomal surface where RHEB sits. mTORC1 therefore demands coincidence of two independent signals, and leucine starvation cannot be rescued by costimulation — one concrete reason amino-acid depletion in a tumour microenvironment is functionally immunosuppressive.

Energy status enters through LKB1 and AMPK, which raises TSC2 GAP activity and also phosphorylates RAPTOR directly. AMPK then activates ULK1 at S555 while mTORC1 inhibits the same protein at S757 — two opposing kinases competing on a single substrate to set autophagy on or off. mTORC1 likewise holds TFEB phosphorylated in the cytosol; when mTORC1 activity falls, TFEB translocates to the nucleus and switches on the lysosomal gene set.

Output: translation, glycolysis and lineage

By far the best-characterised mTORC1 outputs are translational ones. S6K1 phosphorylates ribosomal protein S6 at S235/S236, and mTORC1 phosphorylates 4E-BP1 to release eIF4E for cap binding. Phospho-RPS6 is the convenient readout, but it is 4E-BP1 release that actually gates translation of the growth machinery — and rapamycin separates the two, blocking S6K1 almost completely and 4E-BP1 only partially. That partial, substrate-selective inhibition is precisely why rapalogues are immunomodulatory rather than broadly cytotoxic.

Metabolically, the pathway then hands off to two transcription factors. Myc is induced early and drives glucose and glutamine uptake; HIF-1α, stabilised by mTORC1 even under normoxia, sustains GLUT1, hexokinase 2 and LDHA, while SREBP1 supplies lipogenesis for membrane expansion. Aerobic glycolysis here is not about ATP yield at all; it is about making biosynthetic intermediates quickly, and about the moonlighting effect that glycolytic enzymes have on effector cytokine translation.

Fate follows amplitude. IRF4 scales with signal strength and, with NFAT, drives the effector programme, while T-bet and Eomes specify cytotoxic and memory-associated features. Restrain the pathway and FOXO1 and TCF-1 stay active, with CPT1A-dependent fatty-acid oxidation and PGC-1α-driven mitochondrial biogenesis replacing glycolysis. This is why low-dose mTOR inhibition during priming improves memory formation, and why IL-2 through CD25 and 4-1BB costimulation give convergent metabolic phenotypes: both simply raise amplitude. The human genetics agree from both directions, since PTEN loss causes lymphoproliferation while p110δ gain-of-function produces activated PI3Kδ syndrome with senescent effectors and poor memory.

Key targets and matching reagents

Studying PI3K–AKT–mTOR signalling in vivo

The pathway is intracellular, so in vivo work manipulates the receptors that feed it and reads the output ex vivo. Experiments fall into three groups.

1. Setting the input: TCR and costimulation

Anti-CD3 (low endotoxin) supplies the primary stimulus, while anti-CD28, anti-ICOS (CD278) and anti-4-1BB add or remove the costimulatory input that recruits p85–p110δ. Because CD28 and ICOS engage PI3K through the same class of motif, comparing them isolates receptor dwell time from signalling chemistry. Low endotoxin is not optional: LPS activates mTORC1 in myeloid cells and will contaminate any whole-tissue metabolic readout. Use matched isotype controls from the In Vivo range.

2. Removing the brakes: checkpoints and the IL-2 amplifier

Anti-PD-1 (RMP1-14) and anti-CTLA-4 (9D9) both raise CD28-dependent PI3K output, making them the cleanest tools for increasing flux through this pathway without transfecting anything. Anti-IL-2 (JES6-1A12) and anti-CD25 act on the cytokine amplifier instead, and JES6-1A12 is useful precisely because it redirects IL-2 rather than simply removing it. Pairing a checkpoint antibody with CD25 blockade separates the priming contribution from the expansion contribution.

3. Reading the output quantitatively

Convert phenotype into numbers using phospho-specific kits for phospho-RPS6 (S235/S236), phospho-4E-BP1 (T36), phospho-Rictor (T1135) and phospho-ULK1 (S556), covering mTORC1 output, mTORC2 feedback and the autophagy switch. Add total-protein kits for AKT, mTOR, PTEN, FOXO1 and TSC2, plus metabolic readouts for GLUT1, HK2, LDHA and HIF-1α. Lysate handling is the limiting step: the mTORC1 signal decays within minutes of harvest.

Featured products for this pathway

Frequently asked questions

For research use only. Not for use in diagnostic or therapeutic procedures.