Autophagy regulates tumor growth and metastasis

Abstract

The role of autophagy in tumorigenesis and tumor metastasis remains poorly understood. Here we show that inhibition of autophagy stabilizes the transcription factor Twist1 through Sequestosome-1 (SQSTM1, also known as p62) and thus increases cell proliferation, migration, and epithelial-mesenchymal transition (EMT) in tumor development and metastasis. Inhibition of autophagy or p62 overexpression blocks Twist1 protein degradation in the proteasomes, while p62 inhibition enhances it. SQSTM1/p62 interacts with Twist1 via the UBA domain of p62, in a Twist1-ubiquitination-dependent manner. Lysine 175 in Twist1 is critical for Twist1 ubiquitination, degradation, and SQSTM1/p62 interaction. For squamous skin cancer and melanoma cells that express Twist1, SQSTM1/p62 increases tumor growth and metastasis in mice. Together, our results identified Twist1 as a key downstream protein for autophagy and suggest a critical role of the autophagy/p62/Twist1 axis in cancer development and metastasis.

Figure 1.. Effect of autophagy inhibition on cell migration, invasion, and proliferation.

(A) Immunoblotting in cells with or without knockout (KO) of Atg3, Atg5, Atg9 or Atg12. (B) Transwell assay of cell invasion in cells as in A. (C) A radius migration assay of cells as in A. (D) Cell proliferation analysis of cells as in A. Data were obtained from three independent experiments (mean±S.D.). n=3. *, P<0.05; **, P<0.01, versus the WT group). (E) Representative images for histological and immunohistochemical analysis of p62 and E-cadherin (brown) in normal human skin (n=14), well-differentiated human squamous cell carcinoma (SCC WD) (n=25), and poorly differentiated human SCC (SCC PD) (n=10). Black squares present the region shown with higher magnification. Scale bar, 200 μm and 50 μm for left and right panels, respectively.

Figure 2.. Autophagy inhibition suppresses Twist1 degradation.

(A) Luciferase reporter assay of the E-cadherin promoter with an intact (E-cad (-108)-WT-Luc) or mutated (E-cad (-108) Ebox Mut-Luc) Ebox site in cells with or without Atg5 knockout. (B) Immunoblotting in cells with or without Atg5 knockout. (C-D) Immunofluorescence assay of p62/Twist1 (C) or LC3/Twist1 (D) in cells with or without Atg5 knockout treated with or without rapamycin (500 nM) for 6 h. Scale bar, 10 μm. DAPI is used as the nuclear counterstain. (E-F) Immunoblotting in cells with or without Atg5 knockout treated with or without CHX (100 μg/ml, E) or MG132 (10 μM, F) over a time course. (G) Immunoblotting in mouse skin with or without Atg7 knockout. (H) Immunoblotting in cells with or without Atg7 knockout in combination with Atg7 overexpression.

Figure 3.. Regulation of Twist1 by p62.

(A) Immunoblotting in cells with or without Atg5 knockout in combination with or without siRNA knockdown of p62 or Twist1. (B) Immunoblotting in cells with or without Atg5 knockout in combination with or without shRNA knockdown of p62 or Twist1. (C) Immunoblotting in cells with or without Atg5 knockout in combination with or without p62 overexpression. (D) Immunoblotting in cells with or without p62 knockout in combination with or without p62 overexpression. (E) Immunoblotting in 293T cells with or without overexpression of Myc-Twist1 and HA-p62 treated with or without CHX (100 μg/ml) over a time course.

Figure 4.. p62 interacts with Twist1.

(A) Co-immunoprecipitation (Co-IP) analysis of Twist1 interaction with p62 or Rad23B in cells with or without Atg5 deletion. (B) Immunoblotting in cells with or without Atg5 deletion in combination with overexpression of Myc-Flag-Rad23B or empty vector (EV). (C) Schematic for the constructs with or without individual domain deletions for p62 and Twist1. (D) Co-IP analysis of the interaction between p62 or p62-△UBA and Twist1 or Twist1-△WR in 293T cells. Molecular weight in kDa is marked (A and D).

Figure 5.. Role of K175 in ubiquitination, degradation, and interaction of Twist1 with p62.

(A) Immunoblotting in 293T cells overexpressed with vector (Con), WT, or Twist1 mutants. (B) Immunoblotting in 293T cells transfected with WT or Twist1 mutants followed by treatment with or without with CHX (100 μg/ml) over a time course. (C) Co-IP analysis of ubiquitination and interaction of WT and mutant Twist1 with p62 . (D) Immunofluorescence assay of p62 and Myc (Twist1) in WT MEF cells transfected with WT or mutant Twist1 following treatment with rapamycin (500 nM) for 6 h. Scale bar, 10 μm. DAPI is used as the nuclear counterstain. (E) Immunoblotting in WT MEF cells with or without overexpression of WT or mutant Twist1 in combination with p62 overexpression. (F) Schematic summary for Twist1 stabilization by p62.

Figure 6.. Role of the p62/Twist1 axis in tumor growth and metastasis.

(A) Immunoblotting in A431 cells with or without overexpression of p62 and/or Twist1. (B) Luciferase reporter assay of the E-cadherin promoter with an intact (E-cad (-108) WT-Luc) or mutated (E-cad (-108) Mut-Luc) Ebox site in cells as in A. (C) Wound healing assay of cell migration in cells as in A. (D) MTS assay of cells proliferation of cells as in A. (E) Tumor growth from cells as in A. (F) Number of lung tumor nodules per at 14 weeks following injection. (G) Immunoblotting in A375 cells with or without p62 overexpression. (H) Tumor growth from cells as in G. (I) Immunoblotting in A375 cells transfected with or without knockdown of Atg7 or p62. (J) Tumor growth from cells as in I. (K) Immunoblotting in HaCaT cells with or without p62 knockdown and treated with or without EGF/TGF-β)over a time course. (L) qRT-PCR analysis of Twist1 in HaCaT as in K treated with EGF/TGF-β for 48 h. (M) qRT-PCR analysis of p62. Data are shown from three independent experiments (mean±S.D.). n=3. *, P<0.05; compared with WT Con cells (B); #, P<0.05 compared with the WT group (B); *, P<0.05; compared with the Con, A431-p62, and A431-Twist1 groups (E); **, P<0.01; compared with the A431-Twist1 group (F).