Tyrosine kinase receptor EGFR regulates the switch in cancer cells between cell survival and cell death induced by autophagy in hypoxia

Abstract

Autophagy is an intracellular lysosomal degradation pathway where its primary function is to allow cells to survive under stressful conditions. Autophagy is, however, a double-edge sword that can either promote cell survival or cell death. In cancer, hypoxic regions contribute to poor prognosis due to the ability of cancer cells to adapt to hypoxia in part through autophagy. In contrast, autophagy could contribute to hypoxia induced cell death in cancer cells. In this study, we showed that autophagy increased during hypoxia. At 4 h of hypoxia, autophagy promoted cell survival whereas, after 48 h of hypoxia, autophagy increased cell death. Furthermore, we found that the tyrosine phosphorylation of EGFR (epidermal growth factor receptor) decreased after 16 h in hypoxia. Furthermore, EGFR binding to BECN1 in hypoxia was significantly higher at 4 h compared to 72 h. Knocking down or inhibiting EGFR resulted in an increase in autophagy contributing to increased cell death under hypoxia. In contrast, when EGFR was reactivated by the addition of EGF, the level of autophagy was reduced which led to decreased cell death. Hypoxia led to autophagic degradation of the lipid raft protein CAV1 (caveolin 1) that is known to bind and activate EGFR in a ligand-independent manner during hypoxia. By knocking down CAV1, the amount of EGFR phosphorylation was decreased in hypoxia and amount of autophagy and cell death increased. This indicates that the activation of EGFR plays a critical role in the switch between cell survival and cell death induced by autophagy in hypoxia.

Keywords: Autophagy; CAV1 (caveolin-1); autosis; cell death; epidermal growth factor receptor (EGFR); hypoxia.

Figure 1.

Autophagic flux increases with the prolonged incubation of cells in hypoxia. (A) U87 and A549 cells were incubated in hypoxia for 0, 1, 4, 16, 24, 48 and 72 h. Cells were lysed and western blotted for the autophagy protein LC3B-II in the absence and presence of the lysosomal inhibitor ammonium chloride (NH₄Cl; 30 mM). ACTB (actin, β) was used as a loading control. (B, C) Densitometry quantification of LC3B-II (n = 4) was performed as described in the Materials and methods section. (D, E) Quantification of mRFP-LC3B puncta per cell as described in the Materials and methods section. A positive autophagic flux is defined by an increase in LC3B-II protein level or in the number (#) of mRFP-LC3B puncta per cell in the presence of NH₄Cl compared to that in the absence of NH₄Cl. Error bars represent standard deviation and statistical significance calculated as *, P <0.05; **, P<0.01; ***, P<0.001.

Figure 2.

Autophagy inhibitors increase cell death at an early time of hypoxia but inhibit cell death at a later time of hypoxia. (A) U87 cells were treated with autophagy inhibitors 3-methyladenine (3-MA; 4 mM) and spautin-1 (3 µM) in hypoxia in the absence and presence of NH₄Cl. These cells were lysed and western blotted for LC3B-II. U87 cells were treated with (B) 3-MA or (C) spautin-1 in hypoxia over a 72-h time course. The amount of cell death was determined by the trypan blue exclusion assay. (D) A549 cells were treated with 3-MA in hypoxia and western blotted for LC3B-II. (E) The amount of cell death was determined following 3-MA treatment in hypoxia in A549 cells. Cell death was quantified by flow cytometry as described in the Materials and methods section. These results were representative of 3 independent experiments (n = 3). ACTB was used as a loading control. Error bars represent standard deviation and statistical significance calculated as *, P <0.05; **, P<0.01; ***, P<0.001.

Figure 3.

Knockdown of autophagy genes increases cell death at early times of hypoxia but inhibits cell death at a later time of hypoxia. (A) Knockdown of autophagy genes ATG5 and BECN1 by siRNAs is shown by a western blot of ATG5 and BECN1 in U87 cells. The protein level of ATG5 was represented by the ATG12–ATG5 complex since the binding of these 2 proteins is an essential step during the autophagy process. (B) U87 cells with knockdown of ATG5 or BECN1 were placed in hypoxia for 4, 24, 48, 72 and 144 h and amount of cell death determined by trypan blue exclusion assay. (C) Clonogenic assay of cell survival was performed in U87 cells with knockdown of ATG5 or BECN1 at 4 and 72 h as described in the Materials and methods section. Colony numbers were normalized to that in siCon cells in normoxia. (D) Knockdown of autophagy genes ATG5 and BECN1 by siRNAs in A549 cells was confirmed by western blot. (E) A549 cells with knockdown of ATG5 or BECN1 were placed in hypoxia for 4, 24 and 72 h and the amount of cell death determined as above. (F) Clonogenic assay of cell survival was performed in A549 cells with knockdown of ATG5 or BECN1 at 24 and 72 h as described in the Materials and Methods section. Colony numbers were normalized to that in siCon cells in normoxia. These results are representative of 3 independent experiments (n = 3). ACTB was used as a loading control. Error bars represent standard deviation and statistical significance calculated as *, P <0.05; **, P<0.01; ***, P<0.001.

Figure 4.

Hypoxia increases EGFR expression but inhibits its tyrosine phosphorylation. (A) The total protein level of EGFR and the activation of its tyrosine kinase, as represented by the phosphorylation of EGFR at tyrosine1068 (P-EGFR [Y1068]), was determined by western blot in U87 and A549 cells. ACTB was used as a loading control. (B) Densitometry quantification of P-EGFR (Y1068) normalized to EGFR in U87 and A549 cells was determined. (C) EGFR binding to BECN1 was determined by immunoprecipitation (IP) of EGFR and western blotting for BECN1 in U87 cells. Total cell lysate (TCL) was western blotted for both EGFR and BECN1 as a positive control. Antibody isotype control was used as a negative control for IP. These results are representative of 3 independent experiments (n = 3). Error bars represent standard deviation and statistical significance calculated as *, P <0.05; **, P<0.01; ***, P<0.001.

Figure 5.

Activation of EGFR tyrosine kinase regulates the binding of BCL2 to BECN1 in hypoxia. (A) BCL2 protein level over a 72 h time course in hypoxia was determined by western blot in U87 and A549 cells. ACTB was used as a loading control. (B) BCL2 binding to BECN1 was determined by IP of BCL2 and western blotting for BECN1 in U87 cells. Total cell lysate (TCL) was western blotted for both EGFR and BECN1 as a positive control. Antibody isotype control was used as a negative control for IP. Gefitinib (Gef), 40 µM; EGF, 20 ng/ml. These results are representative of 3 independent experiments (n = 3).

Figure 6.

Reactivation of EGFR shifts autophagy-induced cell death to autophagy-induced cell survival in hypoxia. (A) Cell were treated with EGF (20 ng/ml) at 0, 24, and 48 h in hypoxia. They were lysed after 72 h in hypoxia and western blot of tyrosine phosphorylated EGFR (Y1068), total EGFR and ACTB. Cells cultured in normoxia (0 h) were used as a control. (B) Binding of EGFR to BECN1 after 72 h in hypoxia was determined in the presence or absence of EGF. Cells were lysed and immunoprecipitated for EGFR and western blotted for EGFR and BECN1. Total cell lysate (TCL) was also western blotted as a control. Immunoglobulin isotype (IgG) was used as negative control. (C) Effects of reactivation of EGFR with EGF on autophagic flux were determined. Cells were treated in the presence and absence of NH₄Cl and 3-MA in normoxia (0 h) and hypoxia (72 h). EGF was added as described above and cells were lysed and western blotted for LC3B and ACTB. (D) Reactivation of EGFR with EGF on autophagy-mediated cell death in the absence and presence of 3-MA was determined as described above. Cell death was determined by trypan blue exclusion assay. These results were representative of 3 independent experiments (n = 3). ACTB was used as a loading control. Error bars represent standard deviation and statistical significance calculated as *, P <0.05; **, P<0.01; ***, P<0.001.

Figure 7.

Knockdown of EGFR shifts autophagy-induced cell survival to autophagy-induced cell death at an early time in hypoxia. (A) U87 cells were knocked down for EGFR and ATG5 genes confirmed by western blotting. (B) Cells with knockdown of EGFR and ATG5 genes were placed in hypoxia and amount of autophagy was determined by western blotting for LC3B-II in the absence and presence of NH₄Cl. EGFR and ATG5 genes were knocked down in U87 cells and then placed in normoxia and hypoxia for 4 h (C) and 96 h (D) and the amount of cell death was determined by trypan blue exclusion assay. These results were representative of 3 independent experiments (n = 3). ACTB was used as a loading control. Error bars represent standard deviation and statistical significance calculated as *, P <0.05; **, P<0.01; ***, P<0.001.

Figure 8.

EGFR tyrosine kinase inhibitors (EGFR-TKIs) increase autophagy and cell death at an early time of hypoxia. (A) U87 cells were treated with EGFR-TKI gefitinib (Gef, 40 µM) and amount of EGFR tyrosine phosphorylation (P-EGFR [Y1068]) was measured by western blot. (B) Binding of EGFR to BECN1 in U87 cells following 4 h of treatment with gefitinib in hypoxia was determined by immunoprecipitation of EGFR and western blotting for BECN1. Total cell lysate (TCL) was western blotted for EGFR and BECN1 as a positive control. Isotype antibody (IgG) was immunoprecipitated as a negative control. (C) U87 cells were treated with Gef and autophagy inhibitors 3-MA or spautin-1 and the amount of autophagy was determined by western blotting for LC3B-II in the absence and presence of NH₄Cl. (D) U87 cells were then treated with Gef and 3-MA or spautin-1 and cells were placed in normoxia (N) and hypoxia (H) for 24 h. The amount of cell death was determined by trypan blue exclusion assay. (E) A549 Cells treated with Gef and 3-MA or spautin-1 were placed in normoxia (N) and hypoxia (H) and amount of cell death determined as above. (F) Clonogenic assay of cell survival was performed in A549 cells with and without Gef and 3-MA for 24 h as described in the Materials and methods section. Colony numbers were normalized to that in cells without inhibitors in normoxia. (G) Representative images of A549 cell colonies in 6-well plates for the results in (F). These results were representative of 3 independent experiments (n = 3). ACTB was used as a loading control. Error bars represent standard deviation and statistical significance calculated as *, P <0.05; **, P<0.01; ***, P<0.001.

Figure 9.

CAV1 regulates EGFR tyrosine phosphorylation, autophagy and cell death in hypoxia. (A) The change in CAV1 protein levels when A549 cells were incubated in hypoxia for 0, 24, 48 and 72 h was determined by western blot. (B) A549 cells with CAV1 knockdown were placed in hypoxia for 24 h and the amount of EGFR tyrosine phosphorylation, EGFR and ACTB was determined. (C) A549 cells were knocked down for ATG5 or BECN1 and placed in hypoxia for 24 h. The amount of ATG5, BECN1, tyrosine phosphorylated EGFR, EGFR and CAV1 was determined by western blotting. (D) CAV1 was overexpressed in HEK293 cells and the amount of CAV1 and ACTB in normoxia and CAV1, P-EGFR (Y1068), EGFR and ACTB in hypoxia after 24 and 48 h were determined by western blotting. These results were representative of 3 independent experiments (n = 3). ACTB was used as a loading control.

Figure 10.

Hypoxia-treated cells show morphological features of autosis. A549 cells treated in hypoxia for 0 h (normoxia) and 72 h were analyzed by electron microscopy (EM) as described in Materials and methods section. The features of autosis were identified based on the criteria defined in the literature. Phase 1 of autosis exhibits autophagic bodies (autophagosomes and autolysosomes), and dilated and fragmented endoplasmic reticulum (ER). A transition stage between Phase 1 and Phase 2 of autosis (Phases 1 to 2) exhibits swollen perinuclear spaces that contains cytoplasmic materials, electron-dense mitochondria, an empty ballooning space with membrane starting to merge with outer nuclear membrane, and focal nuclear concavity. Phase 2 of autosis exhibits empty focal ballooning perinuclear spaces and a marked decrease in cytoplasmic organelles. White arrows show dilated and fragmented ER. Black arrows show the swollen perinuclear space that contains cytoplasmic materials. Triangles show an empty ballooning space with membrane starting to merge with outer nuclear membrane. A, autophagic body (autophagosome or autolysosome); E-PNS, empty focal ballooning perinuclear space; ER, endoplasmic reticulum; INM, inner nuclear membrane; M, mitochondria; N, nucleus; NM, nuclear membrane; ONM, outer nuclear membrane.

Figure 11.

Hypoxia-induced autophagic cell death is autosis. (A) A549 cells were treated with and without gefitinib (Gef, 40 μM), the autosis inhibitors digoxin (Digo, 5 μM) and digitoxigenin (Di, 5 μM), and the lysosomal inhibitor ammonium chloride (NH₄Cl) (30 mM) in hypoxia for 4 h. Cells were lysed and western blotted for the autophagy protein LC3B-II. ACTB was used as a loading control. (B) A549 cells were treated with and without the autosis inhibitor Digo and Di at different concentrations (Digo/Di-#, Digo/Di-# μM), the apoptosis inhibitor zVAD (10 μM), and the necroptosis inhibitor necrostatin-1 (Nec-1, 10 μM), in normoxia and hypoxia for 144 h, and cell death was determined by trypan blue exclusion assay. (C) A549 cells were treated with and without Gef and Digo at different concentrations in normoxia and hypoxia for 24 h and cell death was determined as above. These results were representative of 3 independent experiments (n = 3). Error bars represent standard deviation and statistical significance calculated as *, P <0.05; **, P<0.01; ***, P<0.001.

Figure 12.

A graphic model demonstrating the roles of hypoxia-induced autophagy in cell death of cancer cells. In an in vitro cell culture system, at an early time of hypoxia the activity of EGFR tyrosine kinase is high which inhibits autophagy and maintains the autophagic flux at a low level; at these time points, autophagy is a pro-cell survival mechanism. Autophagy degrades the lipid raft protein CAV1 that binds to EGFR and positively regulates its tyrosine kinase activity. Prolonged incubation of cells in hypoxia leads to a decrease of CAV1 protein levels and therefore inhibition of EGFR activation. The consequence is an elevation of autophagic flux that promotes cell death. Taken together, our data suggests that the differential roles of autophagy on hypoxia-induced cell death can be switched by inhibition or reactivation of EGFR.