Caveolin-1 Modulates Mechanotransduction Responses to Substrate Stiffness through Actin-Dependent Control of YAP

Abstract

The transcriptional regulator YAP orchestrates many cellular functions, including tissue homeostasis, organ growth control, and tumorigenesis. Mechanical stimuli are a key input to YAP activity, but the mechanisms controlling this regulation remain largely uncharacterized. We show that CAV1 positively modulates the YAP mechanoresponse to substrate stiffness through actin-cytoskeleton-dependent and Hippo-kinase-independent mechanisms. RHO activity is necessary, but not sufficient, for CAV1-dependent mechanoregulation of YAP activity. Systematic quantitative interactomic studies and image-based small interfering RNA (siRNA) screens provide evidence that this actin-dependent regulation is determined by YAP interaction with the 14-3-3 protein YWHAH. Constitutive YAP activation rescued phenotypes associated with CAV1 loss, including defective extracellular matrix (ECM) remodeling. CAV1-mediated control of YAP activity was validated in vivo in a model of pancreatitis-driven acinar-to-ductal metaplasia. We propose that this CAV1-YAP mechanotransduction system controls a significant share of cell programs linked to these two pivotal regulators, with potentially broad physiological and pathological implications.

Graphical abstract

Figure 1

CAV1 Modulates YAP Activity (A) Variations in gene expression in the RNA-seq analysis between cells grown on stiff and soft substrates, showing all identified genes (left bar) and YAP-target genes alone (right bar). Genes significantly upregulated by ECM stiffness are boxed, and the enrichment for YAP target genes was analyzed using the Fisher exact test. Genes highlighted green are those that were upregulated on the stiff substrate, whereas those highlighted red were downregulated. (B) qRT-PCR analysis of Ctgf and Ankrd1 expression in WT and Cav1KO MEFs grown on stiff and soft substrates for 24 hr. Data are normalized to WT cells grown on a stiff substrate. n = 3. (C) TEAD transcriptional activity in WT and Cav1KO MEFs expressing the 8xGTIIC-luciferase reporter and grown on stiff or soft substrates for 24 hr. Luciferase activity was measured and normalized as described in STAR Methods. Data are normalized to WT MEFs grown on stiff substrate. n = 4. (D) Confocal immunofluorescence images of YAP expression in WT and Cav1KO MEFs grown on stiff or soft substrate. F-actin was stained with fluorophore-conjugated phalloidin (red; left column), and nuclei were stained with Hoechst (blue in merged images; third column). The right column shows zoomed views of the YAP ROI (boxed in white in the YAP images). (E) Percentage of cells from analysis as in (D) with predominantly nuclear YAP (N), predominantly cytosolic YAP (C), or an even nuclear-to-cytosolic distribution (N/C). Randomly selected images from 3 independent experiments were analyzed (60–200 interphase cells per condition). (F) SEEK computational gene co-expression analysis, showing expression correlation (Z score) between YAP target genes and the rest of the genome. (G) Confocal immunofluorescence images of YAP in WT and Cav1KO MEFs plated on different micropatterns with a fibronectin-coated grid that allows cells to spread to a predefined size of 2,025 μm² or 300 μm². Nuclear contours are outlined with dotted gray lines. (H) ImageJ quantification of YAP subcellular distribution in cells plated on micropatterns of 3 grid sizes (300 μm², 1,024 μm², and 2,025 μm²). Data are presented as the nuclear to total cell staining intensities; 10–20 cells were analyzed from 2 biological replicates per condition. The boxplots show the median, 1^st and 3^rd quartiles, and 90^th and 10^th percentiles (whiskers). (I) qRT-PCR analysis of the YAP targets Ctgf and Ankrd1 in cells subjected to cyclic mechanical stretching (CS; see STAR Methods) and unstretched cells. n = 4. Data in (B), (C), (E), and (I) are presented as means ± SEM; ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.005, and ^∗∗∗∗p < 0.0005. See also Figure S1 and Table S1.

Figure 2

Hippo-Kinase-Independent YAP Serine Phosphorylation Determines Impaired Nuclear Translocation and Blunted YAP-Dependent Transcriptional Activity in Cav1KO Cells (A) Western blot of Ser112-phopshorylated YAP and total YAP in WT and Cav1KO MEFs and Cav1KO MEFs reconstituted with CAV1 (o_CAV1) or IRES-GFP. (B) Confocal immunofluorescence of cells transfected with YAP-FLAG or YAP(S5A)-FLAG and stained with anti-FLAG antibody (green), fluorophore-conjugated phalloidin (red), and Hoechst (blue). (C) FLAG distribution from analysis as in (B), represented as the ratio of nuclear-to-cytosolic intensities. n = 6–11. (D) Western blot analysis of FLAG subcellular distribution in MEFs transfected with YAP-FLAG or YAP(S5A)-FLAG followed by biochemical fractionation. RHO-GDI and Tef-1 were used as cytosolic and nuclear markers, respectively. The percentage of total YAP located in the nuclear fractions was quantified (right graph). (E) TEAD transcriptional activity in MEFs transfected with YAP-FLAG and YAP(S5A)-FLAG, measured by 8xGTICC-luciferase reporter assay. Data are normalized to growth on the soft substrate in each experiment. n = 3 (E). The fold-change with respect to untransfected cells is indicated above the bars. (F) qRT-PCR for Ctgf and Ankrd1 expression in WT and CAV1 KO MEFs transfected with control or Lats1 and 2 siRNAs. Data are normalized to WT control. n = 10. (G) qRT-PCR for Ctgf and Ankrd1 expression in WT and CAV1 KO MEFs transfected with control or NF2 siRNAs. Data are normalized to WT control. n = 4. Data are presented as mean ± SEM; ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.005, and ^∗∗∗∗p < 0.0005. See also Figure S2.

Figure 3

YAP Activity Defects in Cav1KO MEFs Are a Consequence of Defective Actin Polymerization (A) Confocal immunofluorescence images showing actin distribution in cells plated on large squared micropatterns (2,025 μm²). Actin was detected by staining with fluorophore-conjugated phalloidin. (B and C) Quantification of actin-fiber anisotropy (fiber order/organization) in WT and Cav1KO MEFs growing on squared micropatterns (n = 7–10; B) or on a stiff ECM (n = 3; C). (D) ImageJ quantification of YAP subcellular distribution in cells treated for 24 hr with 1 μM CytD, 0.05 μM jasplakinolide (Jasplak), or DMSO. Data were obtained from 3 to 8 independent experiments and are presented as means ± SD; ^∗p < 0.05, ^∗∗∗p < 0.005, and ^∗∗∗∗p < 0.0005. (E) Confocal immunofluorescence of YAP in cells grown for 24 hr in the presence of 0.05 μM jasplakinolide or DMSO. Nuclei were detected with Hoechst (blue). (F) qRT-PCR of YAP target genes in MEFs treated for 24 hr with 1 μM CytD, 0.05 μM jasplakinolide, or DMSO (control). Data from jasplakinolide experiments (n = 5) and CytD experiments (n = 3) were normalized to WT controls. Data represent means ± SEM; ^∗p < 0.05. See also Figure S3.

Figure 4

The YAP Interactome Is Altered in Cav1KO MEFs (A and B) Mass spectrometry analysis of YWHA proteins (A) and nuclear pore components and nucelocytosolic transporters (B) that co-immunoprecipitate with YAP in WT and Cav1KO MEFs treated with or without 1 μM CytD for 24 hr. The heatmap represents the relative number of counts per protein and condition. Negative controls (first and third column) were performed in parallel by omitting the primary anti-YAP antibody. n = 5. (C) siRNA library screening scheme for identifying YAP activity regulators. (D) Plot of mean Z scores of the YAP nuclear-to-cytosolic ratio for each individual siRNA in WT and Cav1KO MEFs. (E) Mean Z score of the YAP nuclear-to-cytosolic ratio in WT and Cav1KO cells after siRNA transfection for those genes whose Z scores are above 2.5 or below −2.5. n = 3. (F) qRT-PCR of Ctgf and Ankrd1 in cells transfected with control or YWHAH siRNAs: WT and Cav1KO MEFs, WT cells treated for 24 hr with 1 μM CytD, and WT cells grown on soft substrate. n = 5. Data are presented as means ± SEM. ^∗p < 0.05 and ^∗∗p < 0.01. (G) Scheme of CAV1-YAP regulation. See also Figures S4 and S5, Table S2, and Data S1.

Figure 5

YAP Mediates the Effect of CAV1 in ECM Remodeling (A and B) Collagen gel retraction induced by MEFs transfected with YAP-Flag or YAP(S5A)-FLAG. (A) Cells were embedded in 3D collagen gels, and images were acquired 72 hr later to monitor gel retraction. Collagen organization was determined by second harmonic generation (SHG) microscopy (black and white images). Mock represents mock transfection. (B) Corresponding ImageJ quantification of gel contraction, measured as the fold change with respect to the contraction observed in mock-transfected WT cells. n = 3 experiments. (C) Confocal immunofluorescence images of CAV1 and YAP in cancer-associated fibroblasts (CAFs) extracted from the stroma of human pancreatic tumors and cultured in vitro. Nuclei were detected with Hoechst (blue). Symbols mark cells with low (^∗) or high (⁺) CAV1 levels. (D) Plot of YAP nuclear-to-cytosolic ratio against CAV1 intensity for individual CAFs as in (C). N = 97. (E) Quantification of YAP subcellular distribution (left) and CAV1 levels (right) in CAFs transfected with CAV1 siRNAs or controls. CAV1 and YAP were detected by confocal immunofluorescence microscopy. A total of 4 independent experiments (n = 4) were analyzed (∼500 cells per experiment and condition) using Columbus. See STAR Methods for details. Data are presented as means ± SD in (B) and means ± SEM in (E); ^∗p < 0.05, ^∗∗∗p < 0.005, and ^∗∗∗∗p < 0.0005.

Figure 6

CAV1 Determines YAP Activation and ADM in Caerulein-Induced Acute Pancreatitis (A) Immunohistochemistry analysis of YAP expression in pancreatic tissue of WT and Cav1KO mice 4 days after caerulein treatment. (B) Quantification of the percentage of the nuclear area covered by YAP staining 2 hr and 4 days after caerulein treatment. n = 8. (C) Quantification of the percentage pancreatic area showing extensive ADM after 4 days of chronic pancreatitis in WT mice (n = 4) and Cav1KO mice (n = 3). Data in (B) and (C) are presented as means ± SEM; ^∗p < 0.05, ^∗∗p < 0.01. See also Figure S6.