Report

Mechanical strain induces E-cadherin–dependent Yap1 and β-catenin activation to drive cell cycle entry

See allHide authors and affiliations

Science  29 May 2015:
Vol. 348, Issue 6238, pp. 1024-1027
DOI: 10.1126/science.aaa4559

Stretching cell sheets promotes proliferation

Mechanical strain regulates the development, organization, and function of multicellular tissues. But how? Cadherins mechanically couple neighboring epithelial cells through extracellular interactions and sequester the transcription factors β-catenin and Yap1. To find out more, Benham-Pyle et al. stretched epithelial cell sheets. This mechanical strain induced rapid cell cycle reentry, DNA synthesis by sequential nuclear accumulation, and transcriptional activation of Yap1 and β-catenin. Thus, cell-cell junctions are mechanically responsive structural scaffolds providing signaling centers that coordinate transcriptional responses to externally applied force.

Science, this issue p. 1024

Abstract

Mechanical strain regulates the development, organization, and function of multicellular tissues, but mechanisms linking mechanical strain and cell-cell junction proteins to cellular responses are poorly understood. Here, we showed that mechanical strain applied to quiescent epithelial cells induced rapid cell cycle reentry, mediated by independent nuclear accumulation and transcriptional activity of first Yap1 and then β-catenin. Inhibition of Yap1- and β-catenin–mediated transcription blocked cell cycle reentry and progression through G1 into S phase, respectively. Maintenance of quiescence, Yap1 nuclear exclusion, and β-catenin transcriptional responses to mechanical strain required E-cadherin extracellular engagement. Thus, activation of Yap1 and β-catenin may represent a master regulator of mechanical strain–induced cell proliferation, and cadherins provide signaling centers required for cellular responses to externally applied force.

Cellular responses to mechanical force are important during development and disease and involve reinforcing cell-cell and cell–extracellular matrix (ECM) adhesions, increased cytoskeletal stiffness, and regulation of cell fate (14). Increased ECM stiffness leads to cytoskeleton reorganization and cell cycle progression by activating the Hippo pathway transcription factors Yap and Taz (5) downstream of actin remodeling factors (6), indicating that Yap is a mechanotransducer. However, less is known about signaling from cadherin-mediated cell-cell junctions following applied force.

Classical cadherins couple neighboring cells through trans interactions between opposed extracellular domains and force-dependent linkage of the cytoplasmic domain to the actin cytoskeleton through β-catenin and α-catenin (711), resulting in constitutive tension on E-cadherin at the plasma membrane (10). The cadherin-catenin complex is thought to regulate growth signaling by sequestering the transcription factors β-catenin and Yap1 (1216) in the cytoplasm. However, it is unclear whether cadherin-mediated adhesion is required for the activation of β-catenin and Yap1 in response to mechanical force.

To model mechanical force in multicellular tissues, dense monolayers of quiescent kidney epithelial (Madin-Darby canine kidney, MDCK) cells were formed on compliant silicone substrates in an integrated strain array (ISA) [fig. S1; see also (17)]. The ISA was used to apply and maintain different levels of static biaxial stretch for different times (2 to 24 hours). Cells were then processed for imaging, data acquisition, and analysis (see supplementary materials). Mechanical strain induced rapid cell cycle reentry (Ki67 positive; Fig. 1, A and C, and fig. S2, A and C) and subsequent DNA replication and progression through S phase [EdU positive; Fig. 1, A and E, and fig. S2, B and D; see also (6)] into G2 (Geminin positive; fig. S3). Most cells had entered S phase after 24 hours of strain application (Fig. 1E, “24T”), and higher levels of strain resulted in higher levels of cell cycle reentry (fig. S2).

Fig. 1 Mechanical strain induces cell cycle reentry and sequential activation of Yap1 and β-catenin.

(A) Distribution of Yap1 1 hour after no strain or 15% strain, TBSmCherry after 6 hours, Ki67 after 8 hours, β-catenin after 16 hours, TOPdGFP after 16 hours, and EdU incorporation after a 1-hour pulse (“Pulse”) before fixation at 6 hours, or total EdU incorporation over 24 hours (“Total”). Insets show higher magnification of the region outlined by a dotted line. (B to E) Levels of TBSmCherry (B), Ki67 (C), TOPdGFP (D), and EdU (E) in MDCK monolayers 2 to 24 hours after mechanical strain; for methods used for quantification, see supplementary materials. (F) Summary of cell cycle responses to mechanical strain. Scale bars: 25 μm. All quantifications were from at least three independent experiments with two replicate monolayers per experiment (table S1). Quantifications were mean ± SEM; unpaired t test P values <0.05 (*), <0.01(**), and <0.001 (***). n.s, not significant.

We examined whether the cadherin-associated transcriptional activators Yap1 and β-catenin responded to mechanical strain. In the absence of mechanical strain, Yap1 localized in the cytoplasm and cell cortex [Fig. 1A and fig. S4A; see also (13)]. β-Catenin localized at cell-cell contacts (Fig. 1A and fig. S5A), as expected due to cadherin binding (7) and proteasome-mediated degradation of excess cytoplasmic β-catenin (18, 19). Upon mechanical strain, Yap1 and β-catenin relocalized to the nucleus, but on different time scales. Nuclear Yap1 was detected within 1 hour of strain application, peaked at 6 hours, and then declined rapidly to background levels (Fig. 1A and fig. S4, A and B). In contrast, nuclear β-catenin was not observed until 6 hours following strain and remained over 24 hours (Fig. 1A and fig. S5, A and B).

We next determined if nuclear localization of Yap1 and β-catenin corresponded to their transcriptional activities. Analysis of the TBSmCherry reporter for Yap1 transcriptional activity [fig. S4C; (13)] revealed that, like Yap1 nuclear accumulation, activation following strain was rapid and peaked at 6 hours (Fig. 1, A and B, and fig. S4, D and E), then decreased before the majority of cells entering S phase (EdU positive; Fig. 1E and fig. S2, B and D). In contrast, β-catenin transcriptional activity measured with the TOPdGFP reporter (20) increased rapidly 6 hours after strain application, at the same time that nuclear β-catenin was detected (Fig. 1, A and D, and fig. S5). β-Catenin transcriptional activity then remained high (Fig. 1D and fig. S5, D and E) as cells proceeded through S phase (Fig. 1E and fig S2, B and D). Thus, mechanical strain induced both Yap1- and β-catenin–mediated transcriptional activities, but at different times after strain application and transiently in the case of Yap1 (Fig. 1F).

Although Yap1 activation preceded β-catenin activation by several hours following strain, we tested whether their activation was coupled. Expression of the YAP1-TEAD inhibitory peptide (YTIP) disrupts interactions between Yap1 and TEA domain (TEAD) transcription factors and prevents transcription of Yap1/TEAD-targeted genes [fig. S6C; (21)]. When mechanical strain was applied to MDCK cells transiently expressing GFP- or red fluorescent protein (RFP)–tagged YTIP, YTIP-positive cells did not have increased Yap1 activity (Fig. 2A), Ki67 staining (Fig. 2, D and E, and fig. S6, E and F), or EdU incorporation (Fig. 2, F and G), in contrast to their untransfected neighbors. Similar results were obtained with Verteporfin (fig. S8), a small-molecule inhibitor of Yap1 binding to TEAD transcription factors (22). However, inhibition of Yap1 activity with YTIP or Verteporfin did not block increased nuclear β-catenin levels (Fig. 2B and fig. S8) or β-catenin transcriptional activity following mechanical strain (Fig. 2, B and C, and figs. S6, G and H, and S8). In the presence of Verteporfin, Yap1 was still detected in the nucleus after strain application even though it could not bind to TEAD transcription factors (fig. S8). Yap1 has been reported to co-regulate β-catenin transcriptional activity (23), but our results showed that in response to mechanical strain, nuclear Yap1 levels peaked before nuclear β-catenin was detected and then decreased while nuclear β-catenin and TOPdGFP levels remained high (figs. S4A and S5).

Fig. 2 Yap1-mediated transcription is required for cell cycle reentry after mechanical strain, but not for β-catenin nuclear accumulation or transcriptional activity.

(A) Percentage of TBSmCherry-positive cells in MDCK monolayers 4 hours after no strain or 15% strain in control (wild type, WT) and YTIP-GFP–expressing cells; (B) β-catenin and TOPdGFP distributions in YTIP-RFP–expressing cells and TOPdGFP quantification (C) after 8 hours; (D) Ki67 distribution and quantification (E) after 8 hours; and EdU incorporation (F) and quantification (G) after 24 hours in WT or YTIP-expressing cells after no strain or 15% strain. YTIP-expressing cells outlined with white dashed lines; insets show higher magnification of the region outlined by a dotted line. Scale bars: 25 μm. All quantifications (see supplementary materials) were from at least three independent experiments with two replicate monolayers per experiment (table S1). Quantifications were mean ± SEM; unpaired t test P values <0.05 (*), <0.01(**), and <0.001 (***). n.s, not significant.

Thus, transient Yap1 activation was required for strain-induced cell cycle reentry, but neither Yap1/TEAD-mediated gene transcription nor cell cycle reentry was required for β-catenin nuclear accumulation or transcriptional activity. Additionally, β-catenin transcriptional activity was not sufficient for strain-induced cell cycle reentry or progression in the absence of Yap1 activation.

We next determined if β-catenin transcriptional activity was required for strain-induced cell cycle reentry and progression. The β-catenin–Engrailed chimera (βEng) selectively inhibits β-catenin–mediated transcription without affecting cadherin-mediated cell-cell adhesion (24) or density-dependent inhibition of proliferation (fig. S7). Application of mechanical strain in βEng-expressing cells induced nuclear accumulation of β-catenin (Fig. 3A) but did not result in β-catenin transcriptional activity (Fig. 3D) or progression of cells into S phase (Fig. 3, A and C). However, cell cycle reentry (Fig. 3, A and B) and Yap1 nuclear localization and transcriptional activity (Fig. 3, A and E) were still induced by mechanical strain of βEng cells, similar to normal MDCK cells. Similar results were obtained with iCRT3 (fig. S8), a small-molecule inhibitor of β-catenin binding to TCF (25). Thus, β-catenin transcriptional activity was not required for Yap1 nuclear accumulation and transcriptional activity or cell cycle reentry following strain but was required for cell cycle progression into S phase.

Fig. 3 β-catenin transcriptional activity was required for cell cycle progression through G1 into S phase after mechanical strain.

(A) Distributions of Yap1 4 hours after no strain or 15% strain, Ki67 and β-catenin after 8 hours, and EdU incorporation after 24 hours in control (WT) and βEng-expressing MDCK monolayers. Insets show higher magnification of the region outlined by a dotted line. Quantification of Ki67 (B) and TOPdGFP (D) 8 hours after no strain or 15% strain, EdU (C) after 24 hours, and TBSmCherry (E) after 4 hours in control (WT) and βEng MDCK cells. Scale bars: 25 μm. All quantifications (see supplementary materials) were from at least three independent experiments with two replicate monolayers per experiment (table S1). Quantifications were mean ± SEM; unpaired t test P values <0.05 (*), <0.01(**), and <0.001 (***). n.s, not significant.

These results are consistent with a model in which mechanical strain in a quiescent epithelial cell monolayer causes the transient nuclear localization and transcriptional activation of Yap1, which is required for cell cycle reentry. Independently, strain also induces the nuclear localization and transcriptional activation of β-catenin, which is required for progression into S phase. In mammalian tissues, Yap1/TEAD-targeted genes promote proliferation [connective tissue growth factor (CTGF), fibroblast growth factor (FGF), Amphiregulin (AREG), Ki67], anti-apoptosis (Birc5), and adhesion (Dsc3) (2628), whereas β-catenin/TCF/LEF–targeted genes include additional cell cycle regulators (c-Myc, Cyclin D1, AuroraA, cdc25) (29). Activation of Yap1 and β-catenin gene targets, therefore, is congruent with our model of cell cycle reentry following mechanical strain. Mechanisms of Yap1 and β-catenin nuclear localization are complex and involve many pathways (12, 15, 16, 3032), and many cell surface receptors regulate responses to mechanical strain, including integrin-based adhesions to the ECM (33, 34). However, a specific role for E-cadherin extracellular domain binding between cells has not been tested.

Rather than simply removing E-cadherin from MDCK cells, which would result in the loss of binding sites for Yap1 and β-catenin cytoplasmic sequestration, we used MDCK cells stably expressing a mutant E-cadherin (T151) under control of a doxycycline-repressible promoter (35). T151 comprises a truncated, nonfunctional extracellular domain, but a normal plasma membrane–tethered cytoplasmic domain that binds catenins. Expression of T151 caused the down-regulation of endogenous E-cadherin (fig. S9A), resulting in the complete loss of E-cadherin–mediated cell-cell adhesion, but does not prevent the formation of tight junctions and desmosomes (35), growth to confluence, or contact inhibition (fig. S9, B to D). Unlike normal MDCK cells at high cell densities, T151 monolayers without externally applied strain were Ki67-positive (Fig. 4, A and B) and had nuclear Yap1 (Fig. 4A) and increased TBSmCherry signal (Fig. 4D), consistent with cells being in G1. Whereas T151 monolayers appeared “primed” for cell cycle progression, levels of nuclear β-catenin (Fig. 4A), TOPdGFP (Fig. 4E), and EdU incorporation were all low (Fig. 4C), indicating inhibition of G1- to S-phase transitions. Application of mechanical strain to T151 monolayers did not increase the level of EdU incorporation (Fig. 4, A and C), nuclear β-catenin (Fig. 4A), or β-catenin transcriptional activity (Fig. 4E), indicating that cells had not progressed into S phase. Thus, in multicellular monolayers, coupling between E-cadherin extracellular domains is required to block cell cycle entry and sequester Yap1 in the cytoplasm and for strain-induced nuclear accumulation and transcriptional activity of β-catenin and subsequent cell cycle progression into S phase.

Fig. 4 E-cadherin extracellular domain interactions were required for quiescence at high density, Yap1 nuclear exclusion, and β-catenin activation following strain.

(A) Distribution of Yap1 after 4 hours, Ki67 and β-catenin after 8 hours, and EdU incorporation after 24 hours in control (WT) and T151 MDCK monolayers after no strain or 15% strain. Insets show higher magnification of the region outlined by a dotted line. Quantification of Ki67 (B) and TOPdGFP (E) after 8 hours, EdU (C) after 24 hours, and TBSmCherry (D) after 4 hours in control (WT) and T151 MDCK cell monolayers after no strain or 15% strain. Scale bars: 25 μm. All quantifications (see supplementary materials) were from at least three independent experiments with two replicate monolayers per experiment (table S1). Quantifications were mean ± SEM; unpaired t test P values <0.05 (*), <0.001 (***). n.s, not significant.

Mechanical strain in epithelial monolayers results in cell cycle reentry and progression through S phase by the nuclear accumulation and transcriptional activity of first Yap1 and then β-catenin. Activation of Yap1 is required for cell cycle reentry, whereas β-catenin is required for progression from G1 to S phase. Specific inhibition of Yap1 and β-catenin transcription through use of two independent methods blocked cell cycle reentry and progression, respectively, indicating that other transcription factors were not sufficient for these critical responses to mechanical strain. Thus, activation of Yap1 and β-catenin may be a master regulator for cell cycle reentry and progression through S phase following mechanical strain and an underlying mechanism for regulation of homeostasis in adult tissues. Finally, extracellular E-cadherin engagement and β-catenin represent critical regulators of quiescence and strain-induced proliferation in multicellular assemblies. Thus, cell-cell junctions are not only mechanically responsive structural scaffolds but also signaling centers that coordinate transcriptional responses to externally applied force.

Supplementary Materials

www.sciencemag.org/content/348/6238/1024/suppl/DC1

Materials and Methods

Figs. S1 to S9

Table S1

References and Notes

  1. Acknowledgments: This work was supported by an NSF Predoctoral Fellowship to B.W.B.-P. (DGE-114747) and by grants from the NSF (EFRI-1136790) to B.L.P. and W.J.N. and the NIH to B.L.P. (EB006745) and W.J.N. (GM 35527). C. S. Simmons, B. L. Pruitt, J. Y. Sim, and P. R. Bächtold filed U.S. Patent Application Serial No. 14/376,739 on “Cell Culture Strain Array Systems and Methods for Using the Same.” Data reported in this paper are further detailed in the supplementary materials.
View Abstract

Navigate This Article