TGF-β induces mTORC2 activation. (A) NMuMG cells were treated with TGF-β for the indicated times before lysis and immunoblotting. (B) NMuMG cells were treated or not with TGF-β for the indicated times, in the presence or absence of SB431542 or LY294002. Cell lysates were subjected to anti-Rictor immunoprecipitation, and the immunoprecipitates were subjected to in vitro kinase assays using kinase-inactive Akt1 as a substrate, before immunoblotting of the kinase reactions, immunoprecipitates and cell lysates. The left panels are from the same gel, without differential exposure. (C) NMuMG cells were treated or not with TGF-β or insulin for the indicated times, in the presence or absence of SB431542 or GDC-0941. The kinase activity of mTORC2 was then assessed as in B. The top panels are from the same gel, without differential exposure. (D) NMuMG cells were treated or not with TGF-β for the indicated times, in the presence or absence of SB431542 or LY294002. Cell lysates were subjected to immunoprecipitation using Rictor antibody, and/or immunoblotted.

Inhibition of mTORC2 activity blocks TGF-β-induced EMT of NMuMG cells. (A,B) Control or Rictor-knockdown NMuMG cells were treated with TGF-β for the indicated times and observed by phase-contrast microscopy (A), or subjected to staining for F-actin, immunostaining for E-cadherin, and DAPI staining to visualize nuclei (B). Scale bars: 50 μm (A) and 10 μm (B).

Inhibition of mTORC2 activity inhibits the actin reorganization and migration of NMuMG cells. (A) Control or Rictor-knockdown NMuMG cells expressing GFP-tagged actin, were treated with TGF-β for 48 hours, trypsinized, and seeded sparsely before live-cell real-time epifluorescence imaging. Scale bar: 20 μm. (B) Actin stress-fiber density was quantified from linear fluorescence intensity profiles of lines drawn through the cell body (left panel), and stress fiber angles were determined with circular statistical analysis performed to generate mean resultant length (R) as a quantitative readout of stress fiber orientation in individual cells (Ctrl, n=8; shRic, n=7; 1=total alignment, 0=random in graph of alignment). Error bars represent mean ± s.e.m. ^*P<0.05; ^***P<0.001. (C) Control or Rictor-knockdown NMuMG cells were stimulated with TGF-β for the indicated times. Active GTP-bound RhoA was isolated using GST–Rhotekin-RBD–Sepharose, and detected by immunoblotting using antibody against RhoA. Total lysates were immunoblotted with antibodies against Rictor, RhoA or α-tubulin. The two panels are from the same gel, without differential exposure. (D) Control or Rictor-knockdown NMuMG cells were treated with TGF-β for 48 hours. Cell migration was observed at 0 and 24 hours after wounding, in the absence of TGF-β. Scale bar: 200 μm. The percentage of wound closure was calculated at 24 hours by the mean of eight wound widths per condition performed in triplicates. Error bars show mean ± s.e.m. ^***P<0.001. (E) Control or Rictor-knockdown NMuMG cells expressing GFP-tagged actin were observed by live-cell real-time epifluorescence imaging of GFP-tagged actin (pseudocolor) 3 hours after wounding, in absence of TGF-β. Panels are 200 msecond acquisitions at 10 minute time intervals for a total of 60 minutes. White outlines show the cell cortex at t=0. Scale bar: 10 μm. (F) Control or Rictor-knockdown NMuMG cells were immunostained for paxillin 6 hours after wounding, in absence of TGF-β. The bottom two panels represent an enlargement of the white outlined rectangles. Scale bar: 10 μm. (G) Control or Rictor-knockdown NMuMG cells expressing GFP-tagged paxillin were observed by live-cell real-time TIRFm imaging of GFP-tagged paxillin (pseudocolor) 3 hours after wounding, in the absence of TGF-β. The top six panels are 200 msecond acquisitions at 15 minute time intervals. White outlines show the cell cortex at t=0. The bottom two panels are pseudocolored z projections of the entire time course. Scale bar: 10 μm.

Inhibition of mTORC2 blocks TGF-β-induced EMT and invasive behavior of E4 squamous carcinoma cells. (A) Control or Rictor-knockdown E4 cells were treated with TGF-β for 1 hour before lysis and immunoblotting. (B–D) Control or Rictor-knockdown E4 cells were treated or not with TGF-β for 48 hours and observed by phase-contrast microscopy (B), stained for F-actin and immunostained for E-cadherin, with nuclei visualized by DAPI staining (C), or lysed and immunoblotted (D) (the two top panels are from the same gel, without differential exposure). Scale bars: 50 μm (B) and 10 μm (C). (E,F) Control or Rictor-knockdown E4 cells were treated or not with TGF-β for 1 hour (E) or 48 hours (F). RNA was extracted and Snai1 (E) or Mmp9 (F) mRNA levels were quantified by qRT-PCR. Error bars show mean ± s.e.m. (G) Control or Rictor-knockdown E4 cells were treated with TGF-β for 48 hours, and placed in Matrigel-coated Transwell chambers. After 24 hours, the DAPI-stained nuclei on the lower surface were counted in seven fields per filter in duplicate. Error bars show mean ± s.e.m. ^*P<0.05. (H) Histological evaluation of subcutaneous xenografts of control or Rictor-knockdown E4 cells 21 days after injection into nude mice. Scale bar: 100 μm. The percentage of tumor cells positive for E-cadherin staining is shown in the histogram. Error bars represent mean ± s.e.m. ^**P<0.01. (I) IVIS Luminas-quantified bioluminescence, 29 days after tail vein injection of luciferase- and GFP-labeled control or Rictor-knockdown E4 cells into nude mice. The bottom graph shows a quantification of bioluminescence over time. Error bars represent mean ± s.e.m. ^*P<0.05; ^**P<0.01. (J) Ex-vivo fluorescence stereo microscope imaging of lungs from nude mice, 29 days after tail vein injection of luciferase- and GFP-labeled control or Rictor-knockdown E4 cells. Scale bar: 5 mm. Lower panels are representative H&E-stained images of sections through the lungs with enlargement pictures in the top left corners. Arrowhead indicates a micrometastasis around a blood vessel. Scale bar: 200 μm. The bottom graph shows quantification of GFP-positive foci in the lungs. Error bars represent mean ± s.e.m. ^***P<0.001.

Downregulation of Rictor expression suppresses TGF-β-induced mTORC2 activation. (A) NMuMG cells expressing shRNA against Rictor (shRic, sh#1 or sh#2), mSIN1 (shmSIN1) or control (Ctrl), were treated or not with TGF-β for 1 hour before lysis and immunoblotting. (B) Control or Rictor-knockdown NMuMG cells were treated with TGF-β for 1 hour before lysis and immunoblotting. The panels are from the same gel, without differential exposure. (C) NMuMG control (siCtrl) cells or cells expressing siRNA to downregulate ILK (siILK) were treated or not with TGF-β for 1 hour before lysis and immunoblotting. (D) Control or Rictor-knockdown NMuMG cells were treated with TGF-β for the indicated times before lysis, immunoprecipitation with Akt1 or Akt2 antibodies, and immunoblotting. (E,F) Control or Rictor-knockdown NMuMG cells were stimulated or not with TGF-β for the indicated times before lysis and immunoblotting.

Inhibition of mTORC2 activity attenuates TGF-β-induced changes in EMT marker expression. (A) Control or Rictor-knockdown NMuMG cells, treated with TGF-β for the indicated times, were lysed and analyzed by immunoblotting. (B,C) Control or Rictor-knockdown NMuMG cells were treated or not with TGF-β for 1 hour or 24 hours, RNA was extracted and levels of mRNA encoding E-cadherin (B) or Snail (C) were quantified by qRT-PCR. Error bars show mean ± s.e.m. (D) NMuMG cells expressing siRNA against Akt1 (siAkt1), Akt2 (siAkt2) or control (siCtrl), were treated or not with TGF-β for 1 hour before lysis and immunoblotting. (E) NMuMG cells expressing siAkt1, siAkt2, or siCtrl were treated or not with TGF-β for 1 hour or 24 hours. RNA was extracted and Snai1 mRNA levels were quantified by qRT-PCR. Error bars show mean ± s.e.m. (F) NMuMG cells expressing shRNA against Snail (shSnail) or GFP (Ctrl), were treated or not with TGF-β for 48 hours before lysis and immunoblotting. (G) Control and Snail-knockdown NMuMG cells were treated or not with TGF-β for 1 hour or 24 hours, RNA was extracted and Snai1 mRNA levels were quantified by qRT-PCR. Error bars show mean ± s.e.m. (H) Control and Snail-knockdown NMuMG cells were treated or not with TGF-β for 48 hours, and observed by phase-contrast microscopy. Scale bar: 50 μm.

Inhibition of mTORC2 activity decreases NMuMG cell invasion. (A) Control or Rictor-knockdown NMuMG cells treated or not with TGF-β for 48 hours, were placed in Matrigel-coated Transwell chambers. After 24 hours, the DAPI-stained nuclei on the lower surface were counted in four fields per filter in triplicate. Error bars show mean ± s.e.m. ^**P<0.01. (B) Control or Rictor-knockdown NMuMG cells were treated with or without SB431542 for 1 hour, and TGF-β was added or not for 24 hours. The supernatants were analyzed by gelatin zymography, and cell lysates were immunoblotted. The two top panels are from the same gel, without differential exposure. (C) NMuMG cells expressing shRNA to Rictor, or control were treated or not with TGF-β for 24 or 48 hours. RNA was extracted and Mmp9 mRNA levels were quantified by qRT-PCR. Error bars show mean ± s.e.m. (D) NMuMG cells expressing shRNA to Akt1 (shAkt1), Akt2 (shAkt2) or GFP (Ctrl), were subjected to lysis and immunoblotting. (E–G) NMuMG cells expressing shRNA to Akt1, Akt2 or control (E), shRNA to Snail or control (F), or siRNA against MMP9 (siMMP9), or siCtrl (G), were treated or not with TGF-β for 24 or 48 hours. RNA was extracted and Mmp9 mRNA levels were quantified by qRT-PCR. Error bars show mean ± s.e.m. (H, I) NMuMG cells were transfected with siMMP9 or siCtrl. In (H), the cells were treated or not with TGF-β for 24 hours. The supernatants were analyzed by gelatin zymography (Zymo), and cell lysates (Lysate) were immunoblotted with antibody against α-tubulin. The two panels are from the same gel, without differential exposure. (I) Cells were treated or not with TGF-β for 48 hours, and invasion was assessed as in A. The DAPI-stained nuclei were counted in eight fields per filter in duplicate. Error bars show mean ± s.e.m. ^*P<0.05.