Contact inhibition controls cell survival and proliferation via YAP/TAZ-autophagy axis

Contact inhibition enables noncancerous cells to cease proliferation and growth when they contact each other. This characteristic is lost when cells undergo malignant transformation, leading to uncontrolled proliferation and solid tumor formation. Here we report that autophagy is compromised in contact-inhibited cells in 2D or 3D-soft extracellular matrix cultures. In such cells, YAP/TAZ fail to co-transcriptionally regulate the expression of myosin-II genes, resulting in the loss of F-actin stress fibers, which impairs autophagosome formation. The decreased proliferation resulting from contact inhibition is partly autophagy-dependent, as is their increased sensitivity to hypoxia and glucose starvation. These findings define how mechanically repressed YAP/TAZ activity impacts autophagy to contribute to core phenotypes resulting from high cell confluence that are lost in various cancers.

(i) Number of GFP (autophagosomes) and mRFP dots (autophagosomes and autolysosomes) in HeLa cells. HeLa cells stably expressing mRFP-GFP-LC3 were seeded at LC and HC and subjected to automated Cellomics visualization and analysis. Bars represent the mean±s.d. (n=3; **P<0.01; two-tailed one sample t-test). More than 600 cells were counted per condition, per experiment.
Note: Throughout the figures, n = number of independent biological replicates unless otherwise stated.

Supplementary Figure 2: High confluency cells have increased levels of aggregate-prone proteins/ autophagy substrates when compared to low confluency cells, despite mTORC1 inhibition.
(a) Representative western-blot of p62 in MCF10A cells. GAPDH was used as loading control.
Quantification of p62 densitometry relative to GAPDH is shown on the right. Bars represent the mean±s.d. (n=3; *P<0.05; two-tailed one sample t-test).

(b)
Representative confocal images of HeLa cells transfected with N17-97QP-GFP and then reseeded at low (LC) and high (HC) confluencies. Bars represent the relative ratio of percentages of cells with aggregates between HC and LC conditions (n=3; *P<0.05; two-tailed t-test). Scale bar is 10 µm.
(c) Representative western-blot of A53T-synuclein tagged with GFP in MCF10A cells. GAPDH was used as loading control. Quantification of A53T-synGFP densitometry relative to empty GFP is shown on the bottom panel: S4 mean±s.d. (one representative experiment in triplicates: ***P<0.001; two-tailed t-test). The experiment was repeated with similar results.
(d) Representative western-blot of A53T-synuclein tagged with GFP in HeLa cells. Quantification of A53T-synGFP densitometry relative to empty GFP is shown on the bottom panel: mean±s.d. (one representative experiment in triplicates: **P<0.01; two-tailed t-test). The experiment was repeated with similar results.
(f) Representative western-blot of the mTORC2 substrate, pAkt(S473) in primary MEFs. Gapdh was used as loading control.
(g) Representative western-blot of pULK1(S555) (a well-characterised AMPK substrate) in MCF10A cells. Quantification of pULK1(S555) densitometry relative total ULK1 levels is shown on the bottom panel: mean±s.d. (one representative experiment in triplicates: **P<0.01; two-tailed t-test). The experiment was repeated with similar results. (c) Representative images of YAP/TAZ and immunostaining in HeLa cells seeded at low and high confluencies. The graphs show the percentages of cells with nuclear or cytoplasmic localization (more than 400 cells were counted per condition) and the percentages of BrdU-positive cells (more than 600 cells were counted per condition). Nc = nucleus, Cyt = cytosol.  (a) LC3-II levels assessed by immunoblotting in MCF10A cells exposed to either control (Ctrl) or YAP siRNAs. The cells were treated with vehicle (DMSO) or bafilomycin A1 (BafA1). GAPDH was used as loading control. The graphs show the LC3-II/GAPDH densitometry: mean±s.e.m. (n=5; NS -not significant; two-tailed one sample ttest).
(c) LC3-II levels assessed by immunoblotting in MCF10A cells exposed to either control (Ctrl) or YAP/TAZ siRNAs (YAP and TAZ double knockdown). The cells were treated with vehicle (DMSO) or bafilomycin A1 (BafA1) at 200 nM for 12 hours. GAPDH was used as loading control. The graphs show the LC3-II/GAPDH densitometry: mean±s.d. (n=3 independent experiments performed in triplicates; *P<0.05; two-tailed one sample ttest).
(e) Representative images of GFP-LC3 dots in HeLa cells stably expressing GFP-LC3. HeLa cells were exposed to either Ctrl or YAP/TAZ siRNAs. Quantification of GFP-LC3 dots numbers is shown on the right panel (n=6 wells of at least 100 cells counted per well; *P<0.05; two-tailed t-test). Scale bar throughout the panel is 10 µm.
(h) Representative LC3-II immunoblot in HeLa cells exposed to increasing concentrations of verteporfin for 16 hours. LC3-II levels were quantified in the absence or presence of bafilomycin A1 (400 nM) for the last 4 hours. The experiment was performed in triplicates and repeated at least twice with similar results: mean±s.d. (***P<0.001, **P<0.01; two-tailed one sample t-test).
(c) Representative images of YAP/TAZ immunostaining in highly confluent MCF10A cells exposed to LATS1/2 double knockdown. Scale bar is 10 µm.
(d) Representative images of endogenous immunostaining of LC3 in MCF10A cells treated as in (c). Quantification of number of LC3 dots per cell is shown in the bottom panel, presented either as the mean±s.d. (***P<0.001; twotailed one sample t-test) or as percentages of cells with more than 20 dots per cell. More than 35 cells were counted per LC condition and more than 70 cells were counted per HC condition. The experiment was repeated with similar results at least another two times. YAP /TAZ cytoplasmic localization at high cell density is rescued by LATS1/2 KD (Nc = nucleus; Cyt = cytosol). See Fig. 1k.

(g) LC3-II in cells overexpressing MST2-GFP.
For all immunoblots, GAPDH was used as loading control. S11 S12 Supplementary Fig. 6: Expression of actin related proteins is reduced in high confluency cells.
(a) Gene ontology analysis of genes previously identified as potential targets of YAP through ChIP-Seq. See Methods.

(b) Actin cytoskeleton gene ontology analysis for the selected potential targets (see (a)). Myosin complex and myosin-II form distinct groups.
(c) mRNA levels of CYR61 and CTGF relative to GAPDH in MCF10A cells plated at LC and HC. See also Fig. 2a. Bars -mean±s.e.m. (n=4; ***P<0.001; two-tailed one sample t-test).
(h) mRNA levels of the indicated proteins relative to GAPDH in HeLa cells exposed to YAP/TAZ siRNAs using the autophagy qPCR array (Qiagen). (j) Representative western-blots for the indicated proteins in MCF10A cells plated at various densities and exposed to YAP/TAZ siRNAs.
(k) Representative MLC2 immunoblot in HeLa cells exposed to increasing concentrations of verteporfin for 16 hours. MLC2 levels were quantified in the absence or presence of bafilomycin A1 (400 nM, 4 hours) relative to GAPDH, used as a loading control. MLC2 was revealed by ECL, while GAPDH was stained with LICOR secondary antibodies.  (a) Representative confocal images of HeLa cells exposed to either control or two individual pairs of YAP + TAZ siRNA oligos (using two distinct pairs: either YAP/TAZ_p1 or YAP/TAZ_p2). HeLa cells were immunostained for endogenous YAP/TAZ and F-actin cytoskeleton (Phalloidin 488). Scale bars are 10 µm.

(b)
Representative confocal images of MCF10A cells exposed to either control or individual pairs of LATS1 + LATS2 (LATS1/2_p1) and YAP + TAZ (Y/T_p1) siRNA oligos. MCF10A cells were immunostained for endogenous YAP/TAZ and F-actin. Scale bars are 10 µm. Distinct oligos were used in parallel experiment in (c).
(c) Representative confocal images of MCF10A cells exposed to either control or individual pairs of LATS1 + LATS2 (LATS1/2_p2) and YAP + TAZ (Y/T_p2) siRNA oligos. MCF10A cells were immunostained for endogenous YAP/TAZ and F-actin. Scale bars are 10 µm. Distinct oligos were used in parallel experiment in (b).

(b)
Percentage of MCF10A cells with either nuclear (Nc) or cytoplasmic (Cyt) YAP/TAZ localization. The cells were treated as in (a). Supplementary Fig. 12: The decrease in ATG9A -LC3 colocalisation is myosin-dependent in high confluency cells.
(a) Representative images of LC3 and ATG9A double immunostaining in MCF10A cells plates at LC and HC confluencies. MCF10A cells were transfected with either control or LATS1/2 siRNAs, plated at low and high confluencies and exposed to either DMSO (vehicle control) or blebbistatin (20 µM). Scale bar is 10 µm.  (a-d) The cells were treated with vehicle (DMSO) or bafilomycin A1 (BafA1) at 400 nM for the last 4 hours. GAPDH was used as loading control.
(e) Representative images of YAP/TAZ and F-actin (Phalloidin 488) in HeLa cells plated at low confluency and grown under either normoxia or hypoxia and glucose starvation conditions. Percentages of cells with either nuclear or cytoplasmic localization are shown in the right panel: more than 400 cells were counted per condition. Scale bars throughout the panel are 10 µm.

(f) Quantification of YAP/TAZ localization for the HeLa cells treated as in (e).
(c) Quantification of YOYO-1 negative and positive MCF10A cells plated on stiff or soft ECM and exposed or not to hypoxia and glucose starvation (more than 200 cells were counted per each condition). The experiment was repeated with similar results.
(d) Quantification of YOYO-1 negative and positive MCF10A cells plated on stiff or soft ECM and exposed or not to hypoxia and glucose starvation (more than 200 cells were counted per each condition The cells were initially exposed to control, LATS1/2, ATG7/ATG10 or ATG16L1 siRNAs. The experiment was repeated with similar results. (e) Representative images of YOYO-1 staining in MCF10A cells treated as in (d). Scale bars throughout the panel are 50 µm.
(a) Representative BrdU images of MCF10A cells exposed to LATS1/2 siRNAs, plated at low and high densities. The cells were initially exposed to control, ATG7/ATG10 or ATG16L1 siRNAs. Scale bars are 10 µm. (c) Schematic representation of the roles of the YAP/TAZ-actomyosin-autophagy axis in the clearance of aggregateprone proteins (such as polyQ-htt), survival to metabolic stress and cell proliferation.