Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
, 11 (12), 1444-50

YAP-dependent Induction of Amphiregulin Identifies a Non-Cell-Autonomous Component of the Hippo Pathway

Affiliations

YAP-dependent Induction of Amphiregulin Identifies a Non-Cell-Autonomous Component of the Hippo Pathway

Jianmin Zhang et al. Nat Cell Biol.

Abstract

The Hippo signalling pathway regulates cellular proliferation and survival, thus has profound effects on normal cell fate and tumorigenesis. The pivotal effector of this pathway is YAP (yes-associated protein), a transcriptional co-activator amplified in mouse and human cancers, where it promotes epithelial to mesenchymal transition (EMT) and malignant transformation. So far, studies of YAP target genes have focused on cell-autonomous mediators; here we show that YAP-expressing MCF10A breast epithelial cells enhance the proliferation of neighbouring untransfected cells, implicating a non-cell-autonomous mechanism. We identify the gene for the epidermal growth factor receptor (EGFR) ligand amphiregulin (AREG) as a transcriptional target of YAP, whose induction contributes to YAP-mediated cell proliferation and migration, but not EMT. Knockdown of AREG or addition of an EGFR kinase inhibitor abrogates the proliferative effects of YAP expression. Suppression of the negative YAP regulators LATS1 and 2 (large tumour suppressor 1 and 2) is sufficient to induce AREG expression, consistent with physiological regulation of AREG by the Hippo pathway. Genetic interaction between the Drosophila YAP orthologue Yorkie and Egfr signalling components supports the link between these two highly conserved signalling pathways. Thus, YAP-dependent secretion of AREG indicates that activation of EGFR signalling is an important non-cell-autonomous effector of the Hippo pathway, which has implications for the regulation of both physiological and malignant cell proliferation.

Figures

Figure 1
Figure 1. YAP-induced secreted factor enhances EGF-independent growth of MCF10A cells
(a) Non-cell-autonomous effect of YAP. Vector (tagged with Cherry marker) and either YAP-Wt or YAP-S127A (tagged with GFP) transduced MCF10A cells were cultured in Matrigel either separately or as a 1:1 mixture for 25 days without EGF. Representative light and fluorescence images are shown. (Scale bars, 100μm) (b) Both YAP and YAP-S127A promote the EGF-independent growth of MCF10A cells in 3D culture. Cells transduced with vector, YAP or YAP-S127A were cultured in Matrigel for 25 days in the absence of EGF. Representative phase contrast images are shown. (Scale bars, 100μm) Immunoblotting analysis of endogenous and exogenous YAP, using antibodies to detect Flag-tag, YAP or phosphorylated S127 residue which is mutated in YAP-S127A. β-Tubulin used as loading control. (c) YAP-Wt and YAP-S127A conditioned media induce the EGF-independent growth of parental MCF10A cells in 3D culture. Equal numbers of MCF10A cells were plated in Matrigel and fed for 25 days with medium from vector, YAP or YAP-S127A transduced 3D cultures. Representative phase contrast images are shown. Numbers represent total acini in culture +/− SD (four 200X fields). (Scale bars, 100μm)
Figure 2
Figure 2. Amphiregulin (AREG) mediates YAP-induced EGF-independent growth
(a) Secreted growth factor screen. Human Growth Factor Antibody Array analysis (RayBiotech, Norcross, GA) was performed using the 25th day conditioned medium from vector or YAP-S127A transduced cells grown in the presence (left) and absence (right) of EGF. The membrane is encoded with antibodies for 41 growth factors and receptors, with four positive and four negative controls in the upper left corner. Four proteins were predominantly enriched in YAP-S127A EGF-deprived conditioned medium (arrows). (b) Immunoblot of candidate YAP target genes from lysates of YAP-S127A and vector transduced MCF10A cells cultured on 2D monolayers withdraw EGF over night. (c) Induced AREG mRNA upregulation in YAP-S127A transduced cells in the absence of EGF, as detected by qRT-PCR. (d) In vivo binding of YAP to the AREG promoter by ChIP assay. The known YAP target promoter CTGF is shown as control. (e)AREG neutralizing antibody blocks YAP-S127A induced EGF-independent growth. YAP-S127A cells were cultured in Matrigel with 3D assay medium in the absence of EGF for 25 days, together with neutralizing antibody against AREG, IGFBP6, PDGF-AA or M-CSF-R, with normal goat IgG as control. Representative images are shown. (Scale bars, 100μm)
Figure 3
Figure 3. Regulation of AREG by the Hippo pathway
(a) Recombinant AREG has an equivalent effect on MCF10A 3D acini growth as EGF. (Scale bars, 100μm) (b) YAP-S127A induces activation of ErbB receptor family. Human EGFR Phosphorylation Antibody Array analysis (RayBiotech, Norcross, GA) was performed using lysate from either vector or YAP-S127A transduced cells cultured in the presence and absence of EGF. Each dot presents the tyrosine phosphorylation of ErbB family members at a specific site. (c) AREG expression induced by YAP-S127A is independent from EGFR activation. qRT-PCR analysis of AREG expression in vector or YAP-S127A transduced cells in the presence and absence of EGFR inhibitor (Erlotinib). (d) Effectiveness of EGFR inhibition by erlotinib, demonstrated by abrogation of EGFR phosphorylation, shown by immunoblot. (e) EGFR inhibitor treatment abolishes YAP-induced 3D culture growth in the absence of EGF, indicating the requirement for EGFR signaling. (Scale bars, 100μm) (f) qRT-PCR analysis of AREG expression in ACHN cells infected with shRNAs targeting YAP. The Hippo pathway is activated in these cells by mutation of the upstream regulator Salvdador. (g) Knockdown of LATS1/2 induces expression of AREG. Immunoblotting analysis of AREG, LATS1 and LATS2; after treatment of MCF10A cells with control or LATS1/2 siRNA. β-Tubulin used as loading control.
Figure 4
Figure 4. Requirement of AREG for YAP-mediated cell migration and 3D acini formation
(a) Efficient knockdown of AREG with two independent lentiviral constructs, sh-AREG A9 and G12, introduced into vector or YAP-S127A transduced MCF10A cells. Knockdown of AREG inhibited the activation of AKT (p-AKT) and ERK (p-ERK) in YAP-S127A transduced cells. (b) Knockdown of AREG abrogates EGF-independent 3D acini formation induced by YAP-S127A. MCF10A cells transduced by vector or YAP-S127A were infected with lentivirus containing either control or AREG-targeting shRNAs and cultured on Matrigel for 25 days in the absence of EGF. Representative phase contrast images are shown. Insets: quantified 3D acini formation numbers as mean of four 200X fields. (Scale bars, 100μm) (c) Knockdown of AREG inhibits YAP-induced cell migration. YAP and YAP-S127A transduced MCF10A cells were infected by lentivirus containing the control or AREG-targeting shRNAs, plated onto 8-μmtranswell filters and allowed to migrate for 24 hrs in the absence of EGF. (d) Knockdown of AREG has no effect on YAP-induced EMT. Immunoblot of EMT markers (E-cadherin, Fibronectin and vimentin) in vector and YAP-S127A transduced MCF10A cells co-infected with shAREG (A9 and G12). β-Tubulin used as loading control.
Figure 5
Figure 5. Evolutionary conservation of interactions between EGFR and Hippo pathways in Drosophila
(a) Scanning electron micrograph (SEM; 200X) of adult eye from wild type (GMR-Gal4/+). (Scale bars, 100μm) (b) Over-expression of wts in the eye (GMR-wts/+) displays normal morphology. (c, d) Eyes from GMR-wts/Egfrf2 and GMR-wts/+; vnC221/+ display a rough eye phenotype. (e) Over-expression of hpo in the eye (GMR-hpo/+) leads to a smaller and rough eye phenotype (f, g) GMR-hpo/Egfrf2; GMR-hpo/+; vnC221/+ lead to a glassy appearance and fused ommatidia were apparent. (h) YAPS127A induces dpERK activation. Immunoblot of dpERK was performed using wing imaginal discs from either control (w1118) or wing discs with YAPS127A overexpression (genotype: nubbin-Gal4/UAS-YAPS127A). β-Tubulin used as a loading control. (i) qRT-PCR analysis of Egfr ligands using eye discs either from control (w1118)or YkiS168A expressing flies (genotype: yw eyFlp/+; tub>w+>ykiSA/+). Ct volume normalized by ribosomal protein 49; β-Tubulin expression was used as internal negative control. (*p<0.001; **p<0.003 based on T-test)

Similar articles

See all similar articles

Cited by 175 PubMed Central articles

See all "Cited by" articles

Publication types

MeSH terms

Substances

Feedback