TAZ induces growth factor-independent proliferation through activation of EGFR ligand amphiregulin

Abstract

The Hippo signaling pathway regulates cellular proliferation and survival, thus exerting profound effects on normal cell fate and tumorigenesis. We previously showed that the pivotal effector of this pathway, YAP, is amplified in tumors and promotes epithelial-to-mesenchymal transition (EMT) and malignant transformation. Here, we report that overexpression of TAZ, a paralog of YAP, in human mammary epithelial cells promotes EMT and, in particular, some invasive structures in 3D cultures. TAZ also leads to cell migration and anchorage-independent growth in soft agar. Furthermore, we identified amphiregulin (AREG), an epidermal growth factor receptor (EGFR) ligand, as a target of TAZ. We show that AREG functions in a non-cell-autonomous manner to mediate EGF-independent growth and malignant behavior of mammary epithelial cells. In addition, ablation of TEAD binding completely abolishes the TAZ-induced phenotype. Last, analysis of breast cancer patient samples reveals a positive correlation between TAZ and AREG in vivo. In summary, TAZ-dependent secretion of AREG indicates that activation of the EGFR signaling is an important non-cell-autonomous effector of the Hippo pathway, and TAZ as well as its targets may play significant roles in breast tumorigenesis and metastasis.

Figure 1. TAZ induces malignant cell behavior. (A) Top: 3D cultures of vector control and TAZ^4SA-transduced MCF10A cells in the presence of EGF on day 4 and day 8. (Scale bar, 100 µm); bottom: 3D cultures of vector control and TAZ^4SA-transduced MCF10A cells in the absence of EGF on day 4 and day 8. (Scale bar, 100 µm) (B) Expression of wild-type (wt) and mutant TAZ results in the loss of epithelial markers and gain of mesenchymal markers. Immunoblot reveals an increase in E-cadherin (CDH1) and P-cadherin (CDH3) as well as a decrease in fibronectin-1 (FN1) and plasminogen activator inhibitor-1 (PAI1). Flag antibody detects the ectopic expression of TAZ. β-Actin was used as a loading control. (C) Overexpressing wt and mutant TAZ as well as control vector in presence of EGF in MCF10A 3D culture . (Scale bar, 100 µm) (D) TAZ mutations promote cell migration. Control and mutant TAZ-transduced MCF10A cells were plated onto 8-µm Transwell filters and allowed to migrate for 24 h. Data are the mean number of migrated cells per × 20 field of four fields from each of the triplicate wells. Error bars equal ± SD of three independent experiments. (E) Effect of TAZ on anchorage-independent growth in soft agar. Vector control and mutant TAZ-transduced MCF10A cells were plated in soft-agar assays and allowed to grow for 21 d. Data are mean number of colonies per six-well plate culture of 5 × 10⁴ cells. Error bars equal ± SD of three independent experiments.

Figure 2. Identification of AREG as a non-cell-autonomous effector of TAZ. (A) Non-cell-autonomous effect of TAZ. CherryRed-tagged vector and GFP-tagged TAZ^4SA-transduced MCF10A cells were co-cultured in 3D 8 d without EGF. Representative light and fluorescence images are shown. (Scale bar, 100 µm) (B) TAZ^4SA conditioned media induces EGF-independent growth of parental MCF10A cells in 3D cultures. Representative phase contrast images are shown. (Scale bar, 100 µm) (C) Secreted growth factor screen. Human growth factor antibody array analysis was performed using conditioned media (day 12) from vector- (top) or TAZ^4SA-transduced (bottom) MCF10A cells. The membrane was printed with antibodies for 41 growth factors and receptors, with four positive and four negative controls in the upper left corner. Five proteins were exclusively enriched in TAZ^4SA conditioned medium (rectangles). (D) Induced AREG mRNA in TAZ^4SA-transduced cells in the absence of EGF, as detected by qRT-PCR. GAPDH was used as an internal control. Data represent mean ± SD of three independent experiments. (E) Induction of AREG by wild-type and mutant TAZ in the absence of EGF, as revealed by immunoblot. β-Actin was used as a loading control. (F) Activation of the EGFR signaling induced by overexpression of wild-type and mutant TAZ.

Figure 3. AREG mediates the effect of TAZ. (A) AREG-neutralizing antibody blocks TAZ^S89A and TAZ^4SA-induced EGF-independent 3D growth. Vector control and TAZ-expressing cells were cultured in 3D assay in the absence of EGF for 12 d, together with a neutralizing antibody against AREG. Normal goat IgG is used as a control. Representative images are shown. (Scale bar, 100 µm) (B) Efficient knockdown of AREG by shRNA in vector control or TAZ^S89A and TAZ^4SA-transduced MCF10A cells, as revealed by immnoblot. β-Actin was used as a loading control. (C) Knockdown of AREG abolishes the 3D phenotype induced by TAZ^S89A and TAZ^4SA in absence of EGF (Scale bar, 100 µm). (D) Knockdown of AREG reduces the TAZ^S89A- and TAZ^4SA-induced cell migration.

Figure 4. Induction of AREG by TAZ is mediated through TEAD. (A) S51A abolishes TAZ^S89A-induced EMT of MCF10A cells. Epithelial and mesenchymal markers were detected by immunoblot. β-Actin was used as a loading control. (B) Induction of AREG by TAZ^S89A is inhibited by S51A as revealed by immnoblot. β-Actin was used as a loading control. (C) S51A abolishes the 3D phenotype induced by TAZ^S89A both in the presence (top) and absence (bottom) of EGF. Representative phase contrast images are shown. (Scale bar, 100 µm)

Figure 5. Positive correlation between TAZ and AREG in breast cancer patients. (A) Example of IHC staining of low (top) and high (bottom) TAZ expression in breast cancer TMAs. Arrows: positive TAZ staining. Low (left) and high (right) resolutions are presented. (B) Example of IHC staining of low (top) and high (bottom) AREG expression in breast cancer TMAs. Arrows: positive AREG staining. Low (left) and high (right) resolutions are presented.