Hypoxia-induced inactivation of the protein kinase mammalian target of rapamycin (mTOR) is dominant over mTOR-activating signals from growth factors and nutrients, but the functional pathway is unknown. William Kaelin and colleagues, and Jan Reiling and Ernst Hafen have now mapped this pathway using both mammalian and Drosophila models and show that the hypoxia-induced gene REDD1 is required.

mTOR is regulated by a range of upstream proteins including the TSC1–TSC2 complex, which is mutated in patients with the cancer-predisposing syndrome Tuberous sclerosis. Kaelin, Brugarolas and co-workers used Tsc2-wild-type (Tsc2+/+) or Tsc2-null (Tsc2−/−) mouse embryo fibroblasts (MEFs) to study the effects of hypoxia on mTOR, using phosphorylation of the mTOR target protein S6 kinase (S6K) as a marker of mTOR activity. Tsc2+/+ MEFs showed decreased S6K phosphorylation in response to hypoxia, as expected, but Tsc2−/− MEFs did not, indicating that a functional TSC1–TSC2 complex is required. Hypoxia-inducible factor-1α (HIF1α) expression is also regulated through mTOR and is normally downregulated in response to prolonged hypoxic conditions. However, HIF1α levels remain high in Tsc2−/− MEFs and this can be reversed by the mTOR inhibitor rapamycin, indicating that mTOR regulates HIF1α in hypoxic conditions. In addition, Tsc2−/− MEFs have a proliferative advantage over Tsc2+/+ MEFs in hypoxic conditions indicating that a failure to downregulate mTOR might contribute to tumour formation in patients with Tuberous sclerosis.

Next, Kaelin and colleagues asked if this pathway requires de novo mRNA synthesis and found that globally inhibiting transcription blocked the hypoxia-induced downregulation of S6K phosphorylation. REDD1 is a HIF target gene transcribed in response to hypoxia and DNA damage, and results from Hafen's laboratory indicated that REDD1 Drosophila orthologues might function in the TOR pathway. So Kaelin and colleagues, in collaboration with Hafen and Reiling, examined the mTOR response in Redd1−/− MEFs. These cells do not downregulate S6K phosphorylation in response to hypoxia, so mTOR remains active but, S6K phosphorylation is downregulated in response to the exogenous expression of REDD1. Small interfering RNA experiments established that REDD1 requires TSC2 to inhibit mTOR under hypoxic conditions and are consistent with REDD1 acting upstream of the TSC1–TSC2 complex, which in turn inhibits mTOR as shown in the figure.

These findings support those of Hafen and Reiling in Drosophila, where the REDD1 orthologues scylla and charybdis were identified in a screen for suppressors of AKT function, which acts upstream of TSC in the TOR pathway. Like REDD1, Scylla and Charybdis are induced by hypoxic conditions, scylla being a direct target of Drosophila HIF1. Moreover, complex genetic experiments carried out by Reiling and Hafen show that Scylla and Charybdis act downstream of AKT, but upstream of TSC to regulate TOR and S6K activity.

Overall, these results indicate that inhibition of the mTOR pathway by hypoxia is likely to be important for tumour inhibition. Whether REDD1 is mutated in human tumours is as yet unclear, but these results indicate that REDD1 might function as a tumour-suppressor gene.