mTORC1 signaling under hypoxic conditions is controlled by ATM-dependent phosphorylation of HIF-1α

Abstract

The mTOR complex-1 (mTORC1) coordinates cell growth and metabolism, acting as a restriction point under stress conditions such as low oxygen tension (hypoxia). Hypoxia suppresses mTORC1 signaling. However, the signals by which hypoxia suppresses mTORC1 are only partially understood, and a direct link between hypoxia-driven physiological stress and the regulation of mTORC1 signaling is unknown. Here we show that hypoxia results in ataxia telangiectasia mutated (ATM)-dependent phosphorylation of hypoxia-inducible factor 1-alpha (HIF-1α) on serine(696) and mediates downregulation of mTORC1 signaling. Deregulation of these pathways in pediatric solid tumor xenografts suggests a link between mTORC1 dysregulation and solid tumor development and points to an important role for hypoxic regulation of mTORC1 activity in tumor development.

Figure 1. ATM regulates mTORC1 signaling in response to hypoxia

A, Western blot analysis of Wild type (wt), ATM^+/+ and ATM^−/− MEFs exposed to hypoxia (0.2% O₂) for the indicated times. Cell extracts were analyzed by western blot using antibodies as shown. B, Western blot analysis of three different human ATM deficient cell lines (hATM^−/−) compared with the normal human fibroblast (NHFB). Cells were exposed to hypoxia (0.2% O₂) for the indicated periods of time. At 24 hr ATM was phosphorylated (Ser¹⁹⁸¹) under hypoxia in NHFBs. C, D Experiments were repeated under mild hypoxia (1% O₂)

Figure 2. Hypoxia-induced HIF-1α activity is attenuated in ATM^−/− MEFs

A, ATM^+/+ and ATM deficient cells exposed to hypoxia (0.2% O₂) for the indicated times and cell extracts were blotted for anti-Hif-1α and β-tubulin antibodies. B, Immunofluorescence staining for Hif-1α (red) of ATM^+/+ and ATM deficient cells. Cells were exposed to hypoxia (0.2% O₂) for 6 h and immunofluorescence was detected as described in methods. C, Quantification of fluorescence intensity for HIF-1α on section images with 150-ms exposure time (no saturation) has been shown. Relative intensity = intensity of the object (cell) – intensity of the background. For each of ATM^+/+ or ATM^−/−, n = 15 randomly selected nucleus were quantitated. ***P < 0.001. Mean ± s.e.m. D, Luciferase assay for ATM^+/+ and ATM deficient cells transfected with HRE-Luc plasmid. After transfection, cells were incubated for 24 h under normoxic conditions (21% O₂) and exposed to hypoxia (0.2% O₂) for the indicated times. Cell lysates were assayed for the dual-luciferase activity. Error bars represent the mean standard deviation (s.d.), n=2. E, wtES cells, Hif-1α^−/−ES and NHFB, transfected with control or siRNA against Hif-1α (Dharmacon), exposed to hypoxia (0.2% O₂) for the indicated times and cell extracts were blotted for antibodies as shown.

Figure 3. ATM is necessary for hypoxia-induced expression of REDD1 protein and REDD1 is sufficient for the regulation mTORC1 signaling by hypoxia in ATM deficient cells

A, Western blot analysis of extracts from NHFB and a human ATM deficient cell line (GM03395) exposed to hypoxia (0.2% O₂) for the indicated times and blotted for anti-REDD1 and β-tubulin antibodies. B, ATM^−/− MEFs and a human ATM deficient cell line (GM03395) were transduced with either a REDD1 or a GFP expressing lentiviruses and exposed to hypoxia (0.2% O₂) for the indicated times. Cell extracts were analyzed by western blot using antibodies as shown. C, ATM^−/− MEFs were electrophorated by Flag-ATM or pcDNA plasmids as a control. After 30h incubation, electroporated cells were exposed to hypoxia (0.2% O₂) for the indicated times. Cell extracts were analyzed by western blot using antibodies as shown. D, wtMEFs and p53^−/− MEFs were treated with hypoxia for the indicated times and cell extracts were analyzed by western blot using antibodies as shown, (see also supporting data in Figure S1)

Figure 4. Ser696 phosphorylation by ATM kinase is necessary for Hif-1α physiological activity in hypoxia

A–B, Hif-1α−/− ES cells were electroporated by wtHif-1α, mutS696AHif-1α or mutS696RHif-1α plasmids according to the manufacturer’s instructions (Amaxa Biosystems). Cells were incubated for 24h in normoxia before being exposed to hypoxia (0.2% O₂) for the indicated times and finally cells were harvested for RNA analysis. Expression of Redd1 was determined by RT-PCR. Each value was normalized to GAPDH and error bars represent the mean standard deviation (s.d.). C, After Hif-1α−/− ES cells were electroporated by wtHif-1α, mutS696AHif-1α or mutS696RHif-1α plasmids, cells were incubated for 24 h in normoxia. Cells were then exposed to hypoxia (0.2% O₂) for the indicated times and cell lysates were analyzed by western blotting with using antibodies as shown. D, Ser696 phosphorylation by ATM kinase effects on HIF-1α protein stability under hypoxic conditions. After HCT116-Hif-1α−/−cells were electroporated by wtHif-1α or mutS696AHif-1α plasmids, cells were first incubated for 24h in normoxia. Cells were then exposed to hypoxia (0.2% O₂) for the indicated times in the presence of 100µM Cycloheximide (Sigma) and cell lysates were analyzed by low exposure western blotting with using antibodies as shown. At 120 min, mutS696AHif-1α showed 2.6 Fold decreased protein level compared to wtHif-1α protein level. Relative levels of wtHIF-1α and mutS696AHif-1α were quantitated and normalized to tubulin protein. E, NHFB were exposed to hypoxia (0.2% O₂) or topotecan (4µg/ml) for the indicated times and cell lysates were analyzed by western blotting with using antibodies as shown. F, NHFB and NBS-1LBI cells were exposed to hypoxia (0.2% O₂) for the indicated times and cell lysed were analyzed by western blotting with using antibodies as shown, (see also supporting data in Figure S2). G, NHFB were exposed to hypoxia (0.2% O₂) or irradiated (0.2, 0.4 and 0.6 Gray). After 3 hours, cell lysates were analyzed by western blotting with using antibodies as shown.

Figure 5

A, Tumor bearing animals were chosen randomly and treated with or without EF5 drug, given as a tail-vein injection of 10mM in 5% glucose. The injection volume (in ml) was equal to 1% of animal mass (in g). Mice were euthanized at 3hr and the tumors removed and frozen sections prepared. Sections of tumors were incubated with a Cy3-conjugated antibody to EF5 adducts (hypoxia, red). (ES: Ewing tumor, BT: brain tumor, OS: osteosarcoma, Rh: rhabdomyosarcoma, WT: Wilms’ tumor) B, ATM expression profiles of Acute Lymphocytic Leukemia (ALL) and childhood solid tumor xenografts (BT: brain tumor, ES: Ewing tumor, KT: kidney tumor, NB: neuroblastoma, OS: osteosarcoma, Rh: rhabdomyosarcoma). C-H, Immunoblot analysis of xenografts with using antibodies as shown.

Figure 6. Elevated mTORC1 signaling activity in childhood solid tumor xenografts tissues

A, Immunohistochemistry (IHC) staining of solid human tumor xenografts and normal tissues were performed using phospho-S6 ribosomal protein (Ser^235/236) [Cell Signaling Technology (91B2)] antibody. Paraffin embedded normal human tissue slides were obtained from ProSci Incorporated and paraffin embedded solid human tumor xenografts slides were prepared. IHC were developed according to the Cell Signaling Technology’s protocol (IHC-Paraffin). B, Sections of control tissues and tumors were parallel stained with ATM antibody (Genetex, Clon 5C2, 1:25 dilution) and phospho-S6 ribosomal protein (Ser^235/236) [Cell Signaling Technology (91B2)] antibody (1:75 dilution). Slides were developed as above and covered with Vectashield Mounting Medium with DAPI (Vector Labs). Secondary antibodies used were: to visualize ATM, Alexa Fluor 546 goat anti mouse (red), to visualize phospho-S6 ribosomal protein, Alexa Fluor 488 goat anti-rabbit (green) from Invitrogen. Only merged images are shown. C, EF5 treated and Cy3-conjugated (hypoxia, red) sections (from Figure 5) were parallel stained with ATM antibody (Genetex, Clone 5C2, 1:25 dilution) and phospho-S6 ribosomal protein (Ser^235/236) [Cell Signaling Technology (91B2)] antibody (1:75 dilution). Secondary antibodies used were: to visualize ATM, Alexa Fluor 350 donkey anti mouse (blue), to visualize phospho-S6 ribosomal protein, Alexa Fluor 488 goat anti-rabbit (green) from Invitrogen.

Figure 7. Loss of hypoxia signaling to TORC1 complex strongly induces p53-dependent apoptosis in human ATM deficient cells

A, Equal numbers of cells were plated and exposed to hypoxia (0.2% O₂) with or without rapamycin (100 nM) for the indicated time. Viable cells were counted by trypan blue exclusion assay (Coulter Vicell) and error bars show standard deviation for triplicate plates for a representative experiment. B, Annexin V-FITC staining for NHFB and a human ATM deficient cell line (GM03395). Cells were exposed to hypoxia (0.2% O₂) for indicated times with or without rapamycin (100 nM). The numbers refer late or early percentage of apoptosis cells. C, NHFB and a human ATM deficient cell line (GM03395) were exposed to hypoxia (0.2% O₂) for the indicated times in the presence or absence of rapamycin (100nM). Cell extracts were analyzed by western blot using antibodies as shown. D, Annexin V-FITC staining for NHFB and a human ATM deficient cell line (GM03395) transduced with or without dominant-negative p53 expressing lentiviruses and exposed to hypoxia (0.2% O₂) for the indicated times in the presence or absence of rapamycin (100 nM). Error bars represent standard deviation (s.d.), n=2. (See supporting data in Figure S3)