Regulation of epithelial-mesenchymal transition and metastasis by TGF-β, P-bodies, and autophagy

Abstract

Processing bodies (P-bodies) are ribonucleoprotein complexes involved in post-transcriptional mRNA metabolism that accumulate in cells exposed to various stress stimuli. The treatment of mammary epithelial cells with transforming growth factor-beta (TGF-β), triggers epithelial-mesenchymal transition (EMT), and induces the formation of P-bodies. Ectopic expression of the transcription factor TWIST, which stimulates EMT downstream of the TGF-β receptor, also promotes P-body formation. Removal of TGF-β from treated cells results in the clearance of P-bodies by a process that is blocked by inhibitors of autophagy. Activators of autophagy enhance P-body clearance and block EMT. Blockage of P-body formation by disruption of the gene for DDX6, a protein essential for P-body assembly, blocks EMT and prevents tumor cell metastasis in vivo. These studies suggest critical roles for P-body formation and autophagy in transitions of cancer cells between epithelial and mesenchymal phenotypes and help explain how autophagy functions to promote or suppress tumor cell growth during different stages of tumorigenesis.

Keywords: P-body; autophagy; epithelial-mesenchymal transition (EMT); metastasis; transforming growth factor beta (TGF-β).

Figure 1. TGF-β induces the formation of P-bodies

(A) NMuMG cells were treated without (Control) or with TGF-β for 24 h or sodium arsenite for 1 h, fixed and stained with anti-DCP1A (green) or anti-G3BP (red) to detect P-bodies or stress granules, respectively. Nuclei were stained with DAPI (blue). Bar, 10 μm. (B) The number of granules (P-bodies or stress granules) per cell was quantified (n > 100 cells per treatment). Data represents means ± SEM for triplicate experiments.

Figure 2. TGF-β receptor signaling is required for the induction of P-bodies

(A) NMuMG cells were treated with TGF-β for the indicated times, fixed and stained with anti-DCP1A to visualize P bodies (green). Nuclei were stained with DAPI (blue). Bar, 10 μm. (B) The number of P-bodies/cell was quantified. Data represent average number of P-bodies per cell (n > 100 cells per treatment) ± SEM for triplicate experiments. (C) NMuMG cells were treated for the indicated times with TGF-β. Cell lysates were examined by Western blotting using antibodies against E-cadherin or glyceraldehyde 3-phosphate dehydrogenase (GAPDH). (D) NMuMG cells were not treated (Untreated), treated with TGF-β alone or with a combination of TGF-β and SB-431542 or TGF-β and Actinomycin D (ActD) for 24 h. Cells were fixed and stained with anti-DCP1A to visualize P bodies (green). Nuclei were stained with DAPI (blue). Bar, 10 μm. (E and F) Average number of P-bodies per cell ± SEM (n > 150 cells per treatment) from panel D were quantified from triplicate experiments. Data were analyzed by ANOVA, ^*P < 0.001. (G) Lysates from NMuMG cells treated for 24 h without (-) or with (+) TGF-β in the absence (-) or presence (+) of the indicated concentrations of SB-431542 were examined by Western blotting using antibodies against phospho-SMAD2 (pSmad2) or Smad2/3.

Figure 3. NMuMG cells expressing TWIST have high levels of P-bodies

(A) NMuMG cells lacking or expressing TWIST were treated without (Control) or with TGF-β (10 ng/ml) for 24 h, fixed and stained with anti-DCP1A to visualize P bodies (green). Nuclei were stained with DAPI (blue). (B) Average number of P-bodies per cell ± SEM (n > 150 cells per treatment) was quantified from triplicate experiments. ^*P < 0.001. (C) NMuMG cells expressing TWIST were treated without (Control) or with TGF-β or SB-431542 for 24 h, fixed and stained with anti-DCP1A to visualize P bodies (green). Nuclei were stained with DAPI (blue). Bar, 10 μm.

Figure 4. TGF-β promotes P-body accumulation HMLE and DG75 cells

(A) HMLE cells were treated with TGF-β for the indicated times in days. Cells were fixed and stained with antibodies against DCP1A (green). Nuclei were stained with DAPI (blue). (B) HMLE cells were treated without (-) or with (+) TGF-β for 120 h. Cell lysates were examined by Western blotting using antibodies against E-cadherin, SYK, and GAPDH. (C) HMLE cells and HMLE cells expressing TWIST (TWIST) were fixed and stained with antibodies against DCP1A (green). Nuclei were stained with DAPI (blue). Bar, 10 μm.

Figure 5. Inhibitors of autophagy block P-body clearance

(A) NMuMG cells were treated without (Control) or with TGF-β for 24 h and then allowed to recover for 15 h (Clearance) in the presence of DMSO carrier or the autophagy inhibitor DBeQ at the concentrations indicated. Cells were fixed and stained using antibodies against DCP1A (green). Nuclei were stained with DAPI (blue). (B) Average number of P-bodies per cell ± SEM (n > 100 cells per treatment) was quantified from triplicate experiments. Data were analyzed by ANOVA, ^*P < 0.001 (C) NMuMG cells were untreated (Control) or treated with TGF-β for 24 h. TGF-β was removed and cells were cultured in fresh media in the absence (Clearance) or presence of Chloroquine (10 µM) for 24 h. Cells were fixed and stained with anti-DCP1A to visualize P bodies (green). DAPI (blue) was used to visualize nuclei. Bar, 10 μm. (D) NMuMG cells were treated without (Control) or with TGF-β for 24 h and then allowed to recover for 16 or 24 h in the absence of TGF-β (Clearance). Cells were fixed and stained using antibodies against DCP1A (green) and LC3 (red). Nuclei were stained with DAPI (blue). Bar, 10 μm.

Figure 6. Autophagy promotes P-body clearance

(A) NMuMG cells expressing TWIST were cultured in the absence (Control) or presence of rapamycin at the indicated concentrations for 24 h and stained with anti-DCP1A to visualize P bodies (green). Nuclei were stained with DAPI (blue). Bar, 10 μm. (B) NMuMG cells expressing TWIST were cultured in the presence of rapamycin at the indicated concentrations and in the absence (-) or presence (+) of SBI-0206965 for 24 h. The average number of P-bodies per cell ± SEM (n > 150 cells per treatment) was quantified from triplicate experiments. Data were analyzed by one-way ANOVA ^*P = 0.002. (C) Lysates from NMuMG and NMuMG-TWIST cells were examined by Western blotting using antibodies against p62 and GAPDH. (D) NMuMG-TWIST cells were treated without (-) or with (+) TGF-β for 24 h. Cells were allowed to recover in fresh media lacking (-) or containing rapamycin (20 nM) (+). Cell lysates were examined by Western blotting using antibodies against p62 and GAPDH.

Figure 7. Functional P-bodies are important for EMT

(A) NMuMG cells were left untreated (Control) or were treated with TGF-β, rapamycin (20 nM), or a combination of TGF-β and rapamycin (TGF-β + Rap) for 24 h. Cells were fixed and stained with antibodies against DCP1A (green). Nuclei were stained with DAPI (blue). Bar, 10 μm. (B) Average number of P-bodies per cell ± SEM (n > 150 cells per treatment) from the experiment in panel A was quantified from triplicate experiments. ^*P < 0.003 compared to control cells. ns, not significant. (C) NMuMG cells were treated without (-) or with (+) TGF-β in the absence (-) or presence (+) of rapamycin (20 nM) for 24 h. Cell lysates were examined by Western blotting using antibodies against E-cadherin or GAPDH. (D) NMuMG cells were transfected with siRNA against DDX6 (DDX6 siRNA) or a scrambled siRNA (scrRNA) and cultured for 48 h. Transfected cells were compared to untransfected cells (NMuMG) by Western blotting with antibodies against DDX6 and E-cadherin. (E) NMuMG cells transfected with siRNA against DDX6 (DDX6 siRNA) or a scrambled siRNA (scrRNA) and cultured for 48 h were then treated without (-) or with (+) TGF-β. Cells were fixed and stained with antibodies against DCP1A (green). Nuclei were stained with DAPI (blue). Bar, 10 μm. Cells also were imaged by light microscopy (lower panel). (F) NMuMG cells were transfected with an empty vector or a vector expressing EGFP-DCP1A (green). Control and transfected cells were treated with TGF-β for 24 h and stained with antibodies against E-cadherin (red). Bar, 10 μm. The arrows indicate the location of E-cadherin at the plasma membrane of EGFP-DCP1A-expressing cells.

Figure 8. Lack of TGF-β-induced EMT in DDX6-deficient 4T1 cells

(A) 4T1 mammary carcinoma cells and 4T1 cells lacking DDX6 (DDX6KO) were treated without (-) or with (+) TGF-β for 48 h. Cells were fixed and stained with antibodies against DCP1A (green). Nuclei were stained with DAPI (blue). Bar, 10 μm. (B) Lysates from 4T1 cells and 4T1 cells lacking DDX6 (DDX6KO) were analyzed by Western blotting using antibodies against DDX6 and GAPDH. (C) Average number of P-bodies per cell ± SEM (n > 100 cells per treatment) from the experiment in panel A was quantified from triplicate experiments. ^*P < 0.001 compared to control, untreated cells. ns, not significant. (D) 4T1 cells (WT) or 4T1 cells lacking DDX6 (DDX6KO) were examined by light microscopy before (-TGF-β) and after (+TGF-β) treatment for 48 h with TGF-β.

Figure 9. DDX6 is important for metastasis

(A) 4T1 (WT) and DDX6 deleted 4T1 cells (DDX6KO) expressing firefly luciferase were engrafted onto the mammary fat (2.5x10⁴ cells/mouse). Deletion of DDX6 resulted in reduced primary tumor growth as assessed by bioluminescence and caliper measurements. (B) Upon necropsy, the mammary fatpad tumors were removed, fixed and stained via immunohistochemistry for expression of E-cadherin (Ecad) and the proliferive marker Ki67. Images are representative of three separate tumors for each group. (C) Pulmonary metastasis was quantified in WT and DDX6K0 tumor-bearing animals by bioluminescence readings taken at the indicated time points. (D) Representative bioluminescent images from control and DDX6KO tumor bearing mice. (E) Numbers of pulmonary metastases were confirmed upon necropsy and fixation of lung tissues. Arrows indicate pulmonary metastases. (F) Quantification of the number of pulmonary metastases in WT and DDX6KO tumor-bearing animals. (G) H&E stained histological sections from lungs of three different WT and DDX6KO tumor-bearing animals. Arrows indicate pulmonary metastases. Graphical data in panels (A), (C) and (F) are the mean ± SD of n = 4 mice per group resulting in the indicated P values.