Lysosomal EGFR acts as a Rheb-GEF independent of its kinase activity to activate mTORC1

Abstract

Oncogenic mutations in EGFR often result in EGF-independent constitutive activation and aberrant trafficking and are associated with several human malignancies, including non-small cell lung cancer. A major consequence of EGFR mutations is the activation of the mechanistic target of rapamycin complex 1 (mTORC1), which requires EGFR kinase activity and downstream PI3K/AKT signaling, resulting in increased cell proliferation. However, recent studies have elucidated kinase-independent roles of EGFR in cell survival and cancer progression. Here, we report a cis mTORC1 activation function of EGFR that is independent of its kinase activity. Our results reveal that lysosomal localization of EGFR is critical to mTORC1 activation, where EGFR physically binds Rheb, acting as a guanine exchange factor (GEF) for Rheb, with its Glu804 serving as a potential glutamic finger. Genetic knock-in of EGFR-E804K in cells reduces the level of GTP-bound Rheb, and significantly suppresses mTORC1 activation, cell proliferation and tumor growth. Different tyrosine kinase inhibitors exhibit distinct effects on EGFR-induced mTORC1 activation, with afatinib, which additionally blocks EGFR's GEF activity, causing a much greater suppression of mTORC1 activation and cell growth, and erlotinib, which targets only kinase activity, resulting in only a slight decrease. Moreover, a novel small molecule, BIEGi-1, was designed to target both the Rheb-GEF and kinase activities of EGFR, and shows a strong inhibitory effect on the viability of cells harboring EGFR mutants. These findings unveil a fundamental event in cell growth and suggest a promising strategy against cancers with EGFR mutations.

Fig. 1. Lysosomal EGFR is crucial for mTORC1 activation.

a, b Quantification of the p-T389-S6K1/S6K1 (a) and p-S473-AKT/AKT (b) ratios in NSCLC patients tested by IHC. Data are presented as means ± SD. Two-tailed unpaired t-test. c Afatinib causes much stronger inhibition of mTORC1 activation than erlotinib, although both are capable of suppressing EGFR tyrosine kinase activity. PC9 and HCC827 cells were treated with erlotinib or afatinib at the indicated dose for 12 h and analyzed by western blotting. d LY3000328 or Dyngo-4a impairs the activation of mTORC1 in cells harboring mutant EGFR. PC9, HCC827, or NCI-H1975 cells were treated with DMSO, 50 μM LY3000328 for 24 h, or 50 μM Dyngo-4a for 2 h, and analyzed by western blotting.

Fig. 2. EGFR-TKD directly binds Rheb.

a Co-localization of endogenous mutant EGFR with Rheb and LAMP1 (top) or mTOR (bottom). Immunofluorescence analysis of endogenous Rheb (green), EGFR (pink) and LAMP1 (top) or mTOR (bottom) (red) in EGFR-mutated cells using super-resolution SIM. Scale bar, 1 μm (enlarged view, 0.1 μm). b Afatinib inhibits the interaction between EGFR and Rheb in cells. PC9 cells treated with or without 25 nM erlotinib or afatinib for 12 h were subjected to immunoprecipitation and analyzed by western blotting. c The interaction between EGFR and Rheb is independent of EGFR kinase activity. HEK-293T cells stably expressing empty vector, HA-tagged EGFR WT or KD were subjected to immunoprecipitation and analyzed by western blotting. d Afatinib disrupts the EGFR–Rheb interaction in vitro. Purified EGFR-TKD was incubated with GST or GST-tagged Rheb_1–169 (Rheb, unless specified) with or without 100 μM erlotinib or 100 μM afatinib, precipitated with GST beads, and subjected to SDS-PAGE analysis. Coomassie blue staining is shown.

Fig. 3. EGFR is a GEF for Rheb.

a Afatinib decreases the level of GTP-bound Rheb in PC9 and HCC827 cells. PC9 and HCC827 cells treated with or without 25 nM erlotinib or 25 nM afatinib for 12 h, were subjected to immunoprecipitation using Rheb-GTP agarose and analyzed by western blotting. b The Rheb-D60V mutant preferentially interacts with endogenous WT EGFR. HEK-293T cells were transfected with empty vector, 3× FLAG-Rheb-WT, -D60V, or -Q64L as indicated, subjected to immunoprecipitation, and analyzed by western blotting. c EDTA increases the interaction between EGFR and Rheb at endogenous levels. HeLa cells were lysed in the absence or presence of EDTA, subjected to immunoprecipitation, and analyzed by western blotting. d EGFR preferentially interacts with nucleotide-free or GDP-bound Rheb in vitro. Purified EGFR-TKD was incubated with GST, nucleotide-free GST-Rheb, GST-Rheb with GDP, or GST-Rheb with GTP as indicated, and then precipitated with GST beads and subjected to SDS-PAGE analysis. Coomassie blue staining is shown. e Top, the reconstitution peaks in the size-exclusion chromatography analysis of the purified EGFR-TKD-WT and Rheb-D60V complex. The reconstitution peak for the complex is shown by a dotted line and observed at 10.22 mL. Bottom, SDS-PAGE analysis of each reconstitution sample. The peak for the complex of EGFR-TKD and Rheb is indicated. f EGFR induces Rheb nucleotide exchange from the GDP-bound to the GTP-bound state. A guanine nucleotide exchange assay was performed in vitro using purified EGFR-TKD and Rheb. The relative fluorescence reflects the guanine nucleotide exchange activity. The initial fluorescence intensity was set to 1. Human RhoGEF Dbs (hDbs) and RhoA were used as positive controls. Curves are representative of three independent experiments. Data are presented as means ± SEM. g EGFR induces Rheb nucleotide exchange in a dose-dependent manner. An in vitro guanine nucleotide exchange assay was performed using purified Rheb and a concentration gradient of EGFR-TKD-WT as indicated. h LAMP2-V5-HER2-ICD, LAMP2-V5-IGF1R-ICD, or LAMP2-V5-c-MET-ICD failed to activate mTORC1. HEK-293T cells stably expressing LAMP2-V5-EGFR-TKD-WT, LAMP2-V5-HER2-ICD, LAMP2-V5-IGF1R-ICD, or LAMP2-V5-c-MET-ICD were serum-starved for 24 h and analyzed by western blotting.

Fig. 4. EGFR-Glu804 is a potential glutamic finger indispensable for GEF activity.

a AlphaFold2-Multimer prediction of EGFR-TKD complexed with Rheb (amino acids 1–184). Glu804 of EGFR is shown as a stick-and-ball model and is in close proximity to the nucleotide-binding pocket of Rheb (left), which is similar to the cytohesin-2–Arf1 complex structure (PDB: 1R8Q). GDP Guanosine diphosphate, G3P Guanosine-3’-monophosphate-5’-diphosphate. b EGFR-E804K does not bind Rheb in cells. HEK-293T cells stably expressing SFB-Rheb were transfected with HA-tagged EGFR-WT, -KD, or -E804K, then subjected to immunoprecipitation using anti-HA antibody, and analyzed by western blotting. c EGFR-E804K does not directly bind Rheb in vitro. Purified EGFR-TKD-WT or -E804K was incubated with GST or GST-tagged Rheb as indicated, precipitated with GST beads, and subjected to SDS-PAGE. Coomassie blue staining is shown. d EGFR-E804K failed to induce Rheb nucleotide exchange. An in vitro guanine nucleotide exchange assay was performed using purified EGFR-TKD-WT or -E804K and Rheb as indicated and analyzed as described in Fig. 3f. e LAMP2-V5-EGFR-TKD-WT or -KD, but not LAMP2-V5-EGFR-TKD-E804K, triggers the activation of mTORC1. HEK-293T cells stably expressing the indicated plasmids were serum-starved for 24 h and analyzed by western blotting. f Afatinib, but not erlotinib, impairs the EGFR-mediated Rheb nucleotide exchange. An in vitro guanine nucleotide exchange assay was performed using purified Rheb and EGFR-TKD-WT with the addition of erlotinib or afatinib as indicated and analyzed as described in Fig. 3f. g Afatinib inhibits mTORC1 activation in TSC2-deficient MEFs. WT and TSC2-deficient MEFs were serum-starved, treated with MK-2206 (10 μM), erlotinib (10 μM), or afatinib (10 μM) for 24 h, stimulated with 100 ng/mL EGF for 30 min and analyzed by western blotting. h Afatinib decreases the level of GTP-bound Rheb. TSC2-deficient MEFs treated with or without 10 μM erlotinib or 10 μM afatinib as indicated for 24 h were subjected to immunoprecipitation using Rheb-GTP agarose and analyzed by western blotting. i MK-2206 does not inhibit mTORC1 activation in PC9 and HCC827 cells. PC9 and HCC827 cells were treated with MK-2206 (1 μM), erlotinib (25 nM), or afatinib (25 nM) for 12 h as indicated, and analyzed by western blotting.

Fig. 5. Aberrant Rheb-GEF activity of EGFR impairs mTORC1 activation and cell growth.

a mTORC1 activation was inhibited in EGFR-E804K knock-in PC9 cells. EGFR-E804K knock-in PC9 cells were analyzed by western blotting. b The EGFR–Rheb interaction was disrupted in EGFR-E804K knock-in cells. EGFR-E804K knock-in PC9 cell lysates were subjected to immunoprecipitation using IgG or anti-EGFR antibody and analyzed by western blotting. c The level of GTP-bound Rheb was decreased in EGFR-E804K knock-in cells. EGFR-E804K knock-in PC9 cell lysates were subjected to immunoprecipitation using Rheb-GTP agarose and analyzed by western blotting. d Cell vitality was decreased in EGFR-E804K knock-in cells. CCK-8 assays were performed to examine the proliferation in parental and EGFR-E804K knock-in PC9 cells (n = 4). Two-way ANOVA was used. e, f Cell proliferation was diminished in EGFR-E804K knock-in cells. Colony formation (e) and EdU (f) assays were used to evaluate the proliferation ability of parental and EGFR-E804K knock-in PC9 cells (n = 3). Two-tailed unpaired t-test. g EGFR-E804K knock-in PC9 cells exhibit reduced sensitivity to afatinib compared to parental cells. Parental and EGFR-E804K knock-in PC9 cells were treated with indicated concentrations of erlotinib or afatinib for 72 h (n = 3). The IC₅₀ values were as follows: 20.59 nM and 0.97 nM for parental cells treated with erlotinib and afatinib, respectively; 25.89 nM and 17.44 nM for EGFR-E804K knock-in cells treated with erlotinib and afatinib, respectively. h Representative parental and EGFR-E804K knock-in PC9 tumors surgically removed. i EGFR-E804K knock-in effectively inhibited tumor growth in vivo. Tumor volumes in mice were measured and calculated at specified time intervals following implantation. Two-tailed unpaired t-test. j EGFR-E804K knock-in tumors weighed less than those derived from parental cells. Five weeks after implantation, the mice were euthanized, and the tumors were excised and weighed. Two-tailed unpaired t-test was used.

Fig. 6. BIEGi-1 suppresses cancer cell growth.

a Chemical structure of BIEGi-1. b Structural model of EGFR-TKD complexed with BIEGi-1. EGFR-TKD is shown in surface representation and BIEGi-1 in ball-and-stick. The ATP-binding site (orange) and predicted Rheb-GEF interface (blue) of EGFR are indicated. c BIEGi-1 impairs the EGFR-mediated Rheb nucleotide exchange. An in vitro guanine nucleotide exchange assay was performed using purified Rheb and EGFR-TKD-WT with the addition of erlotinib, afatinib or BIEGi-1 and analyzed as described in Fig. 3f. d BIEGi-1 abolishes the interaction between EGFR and Rheb in cells. PC9 cells treated with or without 25 nM BIEGi-1 for 12 h were subjected to immunoprecipitation and analyzed by western blotting. e BIEGi-1 inhibits mTORC1 activation in EGFR-mutated cells. PC9 and HCC827 cells were treated with 25 nM erlotinib, afatinib, or BIEGi-1 for 12 h as indicated, and analyzed by western blotting. f CCK8 assays of PC-9 and HCC827 cells exposed to BIEGi-1. The IC₅₀ values were 17 nM and 20 nM for PC9 and HCC827 cells, respectively. Data points represent the means ± SD (n = 3). g BIEGi-1 and afatinib, but not erlotinib, inhibit the activation of mTORC1 induced by LAMP2-V5-EGFR-TKD. HEK-293T cells stably expressing the indicated plasmids were serum-starved, treated with 200 nM BIEGi-1, erlotinib or afatinib for 24 h, and analyzed by western blotting. WT and KD refer to LAMP2-V5-EGFR-TKD-WT and LAMP2-V5-EGFR-TKD-KD, respectively. h Model illustrating that lysosomal EGFR acts as a GEF for Rheb, with Glu804 serving as a glutamic finger to activate mTORC1. Inhibition of both GEF and kinase activities of EGFR by BIEGi-1 restricts mTORC1 activity, leading to impaired cell growth.