Current topics in developmental biology, volume 112

Melanie Whiteand Nicolas Plachta review how adhesion cooperates with the cytoskeleton to drive the earliest cellular events in the preimplantation mouse embryo:compaction, change in cell

CURRENT TOPICS IN

DEVELOPMENTAL BIOLOGY

“A meeting-ground for critical review and discussion of developmental processes”

A.A Moscona and Alberto Monroy (Volume 1, 1966)

SERIES EDITOR

Paul M Wassarman

Department of Developmental and Regenerative Biology Icahn School of Medicine at Mount Sinai

New York, NY, USA

CURRENT ADVISORY BOARD

FOUNDING EDITORS

A.A Moscona and Alberto Monroy

FOUNDING ADVISORY BOARD

525 B Street, Suite 1800, San Diego, CA 92101-4495, USA

125 London Wall, London, EC2Y 5AS, UK

The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, UK

First edition 2015

Copyright © 2015 Elsevier Inc All rights reserved.

No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher Details on how to seek permission, further information about the Publisher’s permissions policies and our arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our website: www.elsevier.com/permissions

This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may be noted herein).

Notices

Knowledge and best practice in this field are constantly changing As new research and experience broaden our understanding, changes in research methods, professional practices,

or medical treatment may become necessary.

Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any information, methods, compounds, or experiments described herein In using such information or methods they should be mindful of their own safety and the safety of others, including parties for whom they have a professional responsibility.

To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein.

ISBN: 978-0-12-407758-4

ISSN: 0070-2153

For information on all Academic Press publications

visit our website at store.elsevier.com

Department of Physiology, Development and Neuroscience, The Gurdon Institute, University of Cambridge, Cambridge, United Kingdom

Institute for Molecular Bioscience, The University of Queensland, Brisbane, Queensland, Australia

xi

CREST, Japan Science and Technology Agency; Division of Neurophysiology, Department

of Physiology and Cell Biology, Kobe University Graduate School of Medicine, and Faculty

of Health Sciences, Kobe University Graduate School of Health Sciences, Kobe, Japan Nicolas Plachta

European Molecular Biology Laboratory (EMBL) Australia, Australian Regenerative Medicine Institute, Monash University, Clayton, Victoria, Australia

Cell adhesion is a fundamental determinant of development in metazoanorganisms For over a century—from the early observations of HV Wilson,through the seminal studies of Townes and Holtfreter, and since—we haveendeavored to understand how adhesion helps make multicellular organismsmore than just the sum of their parts We now know that physical interac-tions between cells and their environment (other cells and components ofthe extracellular matrix) influence critical parameters of development,including tissue cohesion, cellular patterning, differentiation, and populationcontrol These diverse functional effects reflect the complex ways in whichdistinct adhesion systems interact with cellular processes such as signaling,the cytoskeleton, and membrane trafficking In this volume, we aim to sur-vey recent developments in understanding how the cellular and molecularmechanisms of adhesion determine the development of organisms and theirconstituent organs.

The early chapters in this volume endeavor to define some of the keyprocesses that allow adhesion to influence development Melanie Whiteand Nicolas Plachta review how adhesion cooperates with the cytoskeleton

to drive the earliest cellular events in the preimplantation mouse embryo:compaction, change in cell shape, polarity, and cell fate Franc¸ois Fagottothen addresses one of the long-standing problems in developmental biology:understanding how boundaries are formed in the embryo Building on thelong-standing realization that boundaries reflect physical differencesbetween populations of cells, Fagotto outlines how different cell–cell adhe-sion systems may cooperate with the cytoskeleton to segregate cellpopulations at boundaries

We then have a series of chapters that focus on the mechanisms by whichcadherin cell adhesion molecules influence animal development Here, amajor advance has come from the realization that cadherins cooperate withthe contractile apparatus, that is, the actomyosin cytoskeleton Accordingly,Rashmi Priya and Alpha Yap discuss the molecular and cellular mechanismsthat allow cadherin adhesion systems to physically interact with, and alsoregulate, the actomyosin cytoskeleton Katja R€oper then addresses howcooperation between cell–cell adhesion and contractility determines mor-phogenesis in the earlyDrosophila embryo In their chapter, Pierre McCrea,Meghan Maher, and Cara Gottardi broaden the discussion to review how

xv

cadherins and their associated proteins signal to the nucleus, a paradigm thatunderlies canonical Wnt signaling and also impinges on other fundamentaldevelopmental pathways, such as the Hippo signaling pathway.

Of course, cadherins are not the only adhesion systems that influencedevelopment Another large family of cell–cell adhesion molecules arethe nectins and nectin-like proteins Kenji Mandai, Yoshimi Takai, and theircolleagues discuss the fundamental cell biology of nectins and review howthese molecules affect the development of many organs in the body AidanMaartens and Nicholas Brown then outline developments in understandinghow integrin cell–matrix adhesion molecules contribute toDrosophila devel-opment, including notable developments in how integrins influence cellfate, cell migration, and cell polarity

The two subsequent chapters focus on developmental processes thatintegrate adhesion, signaling, and the cytoskeleton Pierre Savagner discussesthe concept of epithelial-to-mesenchymal transition, providing a historicaland conceptual framework for this complex phenomenon, with its oftencontroversial mechanistic underpinnings Elias Barriga and Roberto Mayorthen take the example of neural crest migration to consider how adhesiveevents generate collective patterns of cell migration

Finally, we examine how cell adhesion influences the development ofindividual organs Anne Lagendijk and Benjamin Hogan review how cellsignaling and cell–cell adhesion cooperate during vascular development.Eliah Shamir and Andrew Ewald focus on how individual and collective cellmigration are regulated by cell–cell adhesion to drive epithelial morphogen-esis of the mammary gland Kaelyn Sumigray and Terry Lechler review howmultiple junctions (adherens, tight, and desmosomes) contribute to devel-opment of the epidermis as a fundamental biological barrier in the body.Lauren Friedman, Deanna Benson, and George Huntley consider the rolethat cadherins play in the nervous system, with a particular focus on under-standing their role in synapse formation and the generation of synapticnetworks, the bases of neural activity And in the final chapter of this vol-ume, Alexander Combes, Jamie Davies, and Melissa Little discuss how celladhesion drives self-organization in the embryonic kidney, providinginsights relevant to tissue engineering and regenerative medicine

We hope that the contributions in this volume illustrate some of the ferent perspectives that are now being used to understand how cell adhesioncontributes to development A final perspective lies in the relationshipbetween development and disease Many of the cellular mechanisms andbiological processes that we consider are also implicated in disease Thus,

we also sought, where possible, to highlight how basic biology illuminatesour understanding of disease and vice versa We hope that these reviews willthen be a useful guide to students of fundamental biology and pathology.And we will be well pleased if they prompt further research at the interfacebetween these disciplines.

ALPHAS YAP

CHAPTER ONE

How Adhesion Forms the Early Mammalian Embryo

Melanie D White, Nicolas Plachta1

European Molecular Biology Laboratory (EMBL) Australia, Australian Regenerative Medicine Institute, Monash University, Clayton, Victoria, Australia

Abstract

The early mouse embryo is an excellent system to study how a small group of initially rounded cells start to change shape and establish the first forms of adhesion-based cell –cell interactions in mammals in vivo In addition to its critical role in the structural integrity of the embryo, we discuss here how adhesion is important to regulate cell polar- ity and cell fate Recent evidence suggests that adherens junctions participate in signaling pathways by localizing key proteins to subcellular microdomains E-cadherin has been identified as the main player required for the establishment of adhesion but other mech- anisms involving additional proteins or physical forces acting in the embryo may also con- tribute Application of new technologies that enable high-resolution quantitative imaging of adhesion protein dynamics and measurements of biomechanical forces will provide a greater understanding of how adhesion patterns the early mammalian embryo.

1 THE MOUSE PREIMPLANTATION EMBRYO

AS A MODEL OF ADHESION IN MAMMALIAN

DEVELOPMENT

Most research on adhesion has been performed on cells in tissue ture due to their availability and ease of manipulation However, it is only

cul-Current Topics in Developmental Biology, Volume 112 # 2015 Elsevier Inc.

during true cellular differentiation within an embryo that the contribution ofadhesion to development can be examined directly The mouse preimplan-tation embryo provides an ideal system to study adhesion mechanisms thatare based exclusively on cell–cell interactions A glycoprotein membrane,the zona pellucida, encloses the preimplantation embryo so cell–cell adhesioncan be studied in the complete absence of extracellular matrix interactions.Preimplantation development naturally occurs within the oviduct, but itcan be recapitulated in vitro without adversely affecting the developmentalpotential of embryos (Summers & Biggers, 2003) Mouse embryos can beeasily removed from the maternal oviducts and cultured in simple mediaconditions Under these ex utero conditions, the embryos develop almost

as rapidly as they do in utero and if transferred back to the uterus they canimplant and continue developing to produce viable offspring

During the first 2 days of development, the fertilized mouse eggundergoes three cleavage divisions to produce an 8-cell embryo

The first major cell morphological changes begin as the 8-cell embryoundergoes compaction Concomitant with a rise in intercellular adhesion,the cells flatten their membranes against each other, maximizing contactand forming a highly packed mass This process of increased adhesion andembryo compaction occurs ubiquitously during preimplantation

Figure 1 Imaging preimplantation development in the mouse embryo (A) DIC images showing development of mouse embryo from 1-cell to blastocyst stage (B) Microinjection of mRNA or DNA into the pronucleus allows visualization of proteins

of interest throughout preimplantation development In the example shown, the brane is labeled with mCherry and the nucleus is labeled with H2B-GFP ICM, inner cell mass; TE, trophectoderm.

mem-development in different mammalian species and is an absolute requirementfor embryo viability This process is very similar in mouse and humans, mak-ing the preimplantation mouse embryo an ideal model to study the role ofadhesion in cell shape, position, and fate in early mammalian development.

In addition, the cells of the mouse embryo are relatively large, facilitatingimaging of subcellular processes Furthermore, there are many availablegenetic tools that are applicable in the mouse for manipulation of proteins

of interest Pronuclear microinjection of mRNA or DNA is a established technique for expression of exogenous proteins and mouseembryos are resilient enough to withstand this process with high efficiency

available carrying targeted endogenous genes or expressing various genic constructs

trans-1.1 Adhesion molecules in the preimplantation

mouse embryo

Early studies identified a critical role for calcium-dependent adhesion inembryo compaction, subsequent spatial segregation of the inner cell mass(ICM) and formation of the first differentiated tissue, the trophectoderm

calcium ions or using antibodies targeting a cell surface glycoproteindecompacted embryos and prevented blastocyst formation (Ducibella &

gly-coprotein was identified as uvomorulin, now more commonly known asE-cadherin (Hyafil, Babinet, & Jacob, 1981)

Although usually expressed in epithelial cell layers, E-cadherin is alsoexpressed from the very early stages of development It is initially maternallyderived in the oocyte and at the 2-cell stage, de novo E-cadherin zygotic syn-thesis starts (Vestweber, Gossler, Boller, & Kemler, 1987) Embryos lackingzygotic E-cadherin are preimplantation lethal They do undergo compac-tion due to residual maternal E-cadherin but fail to form a blastocyst

knock-down E-cadherin expression in just half of the embryo prevents those cellsfrom integrating into the compacting embryo (Fig 2) (Fierro-Gonzalez,

E-cadherin are unable to compact or form a blastocyst and they appear asloose aggregates of cells (Stephenson, Yamanaka, & Rossant, 2010) Delet-ing maternal E-cadherin alone delays compaction until the late morula stagebut embryos then develop normally due to zygotic E-cadherin expression

3 Adhesion in the Early Mouse Embryo

(De Vries et al., 2004) Adhesion does not develop until the late morula stage

in these embryos indicating that although E-cadherin and its binding ners are expressed (Ohsugi et al., 1996; Sefton, Johnson, & Clayton, 1992;

complexes at very early stages

E-cadherin is uniformly distributed in the cell membrane until the 8-cellstage when PKC-α-mediated phosphorylation of β-catenin, a key proteinregulating E-cadherin intracellular signaling, is thought to activate the com-paction process (Fig 3; Pauken & Capco, 1999) Compaction can beblocked by inhibition of PKC-α and induced early at the 2- and 4-cell stages

Figure 2 E-cadherin is required for cell –cell adhesion and embryo compaction (A) Microinjection of one cell at the 2-cell stage results in an embryo expressing a control siRNA and a membrane-Cherry marker in half of its cells The transgenic cells have nor- mal morphology and integrate into the compacting embryo mass (B) An siRNA targeting E-cadherin reduces cell –cell adhesion in the transgenic half of the embryo The nontransgenic cells compact normally but the E-cadherin knockdown cells are very spherical and do not integrate into the embryo mass (C) Treating the embryo with the DECMA-1 E-cadherin function-blocking antibody reduces adhesion and causes all cells

to become very spherical The embryo does not compact Scale bar ¼10 μm.

Figure 3 E-cadherin localization changes during preimplantation development E-cadherin is distributed throughout the membrane until the late 8-cell stage Then,

it begins to accumulate in cell –cell junctions and is predominantly localized to olateral regions by the 16-cell stage.

bas-by PKC-α activation (Ohsugi, Ohsawa, & Semba, 1993; Winkel, Ferguson,

for-ming adherens junctions between cells and connecting to the actin eton via catenin proteins (Ozawa, Ringwald, & Kemler, 1990) As adhesioninitiates, the actin cytoskeleton is reorganized to define the orientation of thefirst cellular polarity in the embryo (Stephenson et al., 2010)

cytoskel-After compaction has occurred, tight junctions begin to assemble atapicolateral cell–cell junctions in a process that requires prior activation ofE-cadherin-mediated adhesion (Fleming, McConnell, Johnson, &

of tight junction formation is tightly regulated by staggered expression ofthe constituent proteins from the 8- to 32-cell stage (Sheth et al., 1997).E-cadherin-mediated adhesion may also stabilize tight junction proteins,preventing their turnover once they are assembled at the membrane

tight junctions then forms a permeability seal between adjacent epithelialcells This allows the formation of the blastocoel cavity and the generation

of the blastocyst During blastocyst expansion, small, strongly adhesive tions called desmosomes are assembled between adjacent cells in thetrophectoderm These junctions are multiprotein complexes containingthe desmosomal cadherins, desmocollins, and desmogleins (Fleming,

function-ally synergize to maintain epithelial polarity and structure Thus, in thesequence of developmental events generating adhesion in the embryo,E-cadherin is a central player in a pathway that enables the progressive for-mation and organization of different types of anchoring junctions and theestablishment of the first forms of tissue-like organization duringdevelopment

2 ADHESION REGULATES CELL SHAPE

There are many studies demonstrating the importance of adhesion formorphogenesis in other experimental models (Lecuit, Lenne, & Munro,

2011) However, relatively little work has been performed in the tation mouse embryo to elucidate how adhesion controls cell shape The firstdiscernible changes accompanying the increased intercellular adhesion dur-ing compaction are morphological From the late 8-cell stage, cells lose theirpreviously spherical shape and flatten into a tightly packed mass with indis-tinguishable membranes (Fig 1) The formation of adherens junctions was

preimplan-5 Adhesion in the Early Mouse Embryo

widely believed to be responsible for this global change in morphology sinceperturbations of E-cadherin using antibodies or gene deletion approachesresulted in cells displaying a rounder morphology However, recent workhas demonstrated that in addition to its localization to adherens junctions,E-cadherin also accumulates in long cellular protrusions identified asfilopodia, which appear specifically during the stages of embryo compaction

junc-tions onto the apical membrane of neighboring cells and adhere to them viaE-cadherin trans interactions (Fig 4) The filopodia then maintain tension toelongate the cell’s membrane over its neighbors, bringing them into closeapposition, and facilitating compaction (Fierro-Gonzalez et al., 2013).Whereas E-cadherin localized at adherens junctions play an important role

in keeping cells together, the pool of E-cadherin in the recently discoveredfilopodia helps cells establish tight interactions at their apical membranes inorder to compact the embryo

Physically disrupting these filopodia with laser ablation causes cells torevert to a more spherical shape Moreover, interfering with the molecularcomponents present in these filopodia, which include E-cadherin,α-catenin, β-catenin, or Myo10, prevents compaction Expressing a mutantform of E-cadherin lacking the extracellular domain or treating embryoswith an E-cadherin function-blocking antibody also disrupts filopodia for-mation and embryo compaction, suggesting that the adhesive function of

Figure 4 E-cadherin-dependent filopodia control embryo compaction Two cells labeled with a mCherry protein targeted to the cell membrane extend filopodia over their nonlabeled neighbors during compaction at the late 8-cell stage Dashed box shows filopodia at higher magnification.

E-cadherin is critical Only around 60% of cells in each embryo were found

to form filopodia and there is no direct association with subsequent cell fate

so it remains to be revealed what determines whether a cell will employ thisnewly described mechanism of adhesion during preimplantationdevelopment

3 ADHESION CONTROLS CELL POLARITY

The initiation of adhesion at the 8-cell stage also directs establishment ofthe first forms of cellular polarity in the developing mouse embryo During thistime, the actin cytoskeleton is reorganized to establish an apicobasal polaritytogether with the formation of an apical microvillous pole (Handyside,

1980) and a polarized distribution of cytoplasmic organelles and cytoskeletalelements (Fleming & Pickering, 1985; Houliston, Pickering, & Maro,

Ezrin, Pard6b, and the aPKCs (PKCζ and PKCλ) are then localized to the cal domain while Par-1, Jam-1, and Na/K ATPase are found at basolateralcell–cell contacts (Barcroft, Moseley, Lingrel, & Watson, 2004; Louvet,Aghion, Santa-Maria, Mangeat, & Maro, 1996; Pauken & Capco, 2000;

Polar cells remain on the outside of the embryo and differentiate intotrophectoderm, whereas apolar cells are enclosed inside the embryo and formthe pluripotent ICM (Dyce, George, Goodall, & Fleming, 1987) Adhesiondoes not appear to be required for the initiation of polarization as cells isolatedfrom early mouse embryos can polarize in the absence of cell–cell contacts

inhibition of E-cadherin function with antibodies (Houliston, Pickering, &

dis-aggregated cells is delayed and also random in orientation Deleting bothmaternal and zygotic E-cadherin confirms that E-cadherin is required torestrict the area of the apical domain and confine basolateral proteins, ensuringcorrect segregation of apical and basolateral domains (Stephenson et al., 2010).Together, these studies reveal that E-cadherin-mediated adhesion controls thetiming and axis of polarization in the preimplantation mouse embryo

4 ADHESION DETERMINES CELL FATE

Polarization has been closely linked to cell fate Disruption of the apicalproteins Par3 and PKCλ preferentially directs cells in the embryo toward an

7 Adhesion in the Early Mouse Embryo

ICM fate (Plusa et al., 2005) By determining the orientation of polarity,intercellular adhesion has a critical role in specifying cell fate The maternaland zygotic deletion of E-cadherin demonstrated that E-cadherin-mediatedadhesion is important to restrict trophectoderm fate (Stephenson et al., 2010).More cells in the mutant embryos express the trophectoderm-specific markerCdx2, and the normal spatial allocation of trophectoderm cells to the outside

of the embryo and ICM cells to the inside is disrupted Recent work has nowestablished how adhesion affects cell fate indirectly through polarity andshown that it also has an additional direct role in fate determination (Fig 5).Differentiation of the outer cells of the morula into trophectoderm requiresexpression of the transcription factors Cdx2 and Gata3, which in turn is driven

Figure 5 Adhesion and polarity determine cell fate in the preimplantation mouse embryo The apical polarity complex in outer cells sequesters components of the Hippo signaling pathway preventing its activation Unphosphorylated Yap can enter the nucleus and drive expression of trophectoderm-specific genes In inner cells, Amot localizes to adherens junctions where it binds to Lats1/2 and the E-cadherin adhesion complex via Nrf2 Lats 1/2 phosphorylates Amot, activating it and this complex phos- phorylates Yap Phosphorylated Yap is excluded from the nucleus and the Hippo path- way is activated, allowing transcription of inner cell mass specific genes.

by the transcription factor Tead4 and its coactivator, Yap1 (Ralston et al.,

2010) Tead4 knockout mice can specify an ICM but do not form thetrophectoderm (Nishioka et al., 2008; Yagi et al., 2007) The Hippo signalingpathway regulates Yap1 subcellular localization by phosphorylation (Nishioka

Yap1, excluding it from the nucleus, promoting transcription of ICM-specificgenes, and repressing trophectoderm fate In the absence of sufficient Hipposignaling, Yap1 is free to enter the cell nucleus and induce transcription ofthe trophectoderm-specific genes Cdx2 and Gata3 However, Yap1, Tead4,and the members of the Hippo signaling pathway are expressed in all cells ofthe preimplantation mouse embryo So, how is Hippo signaling suppressed

in outer cells and activated in inner cells? The answer lies in sequestration ofkey members of the pathway by proteins involved in polarity and adhesion.Angiomotin (Amot) and Nf2 are required to activate Lats1/2 and switch

on the Hippo signaling pathway (Hirate et al., 2013) Nf2 is uniformly tributed through the membrane but in outer cells, Amot is sequestered bycomponents of the apical polarity complex and localized to the apicaldomain (Hirate et al., 2013) Here, it is bound to actin and held in an inactivestate Lats1/2 may also be sequestered by the polarity complex as it too has anapical localization in outer cells (Cockburn, Biechele, Garner, & Rossant,

dis-2013) The apical localization of Amot and Lats1/2 prevents activation ofthe Hippo signaling pathway, unphosphorylated Yap1 enters the nucleusand the cell reverts to a trophectoderm fate

In inner cells that lack apical polarity, Amot can interact with Lats1/2 andthe E-cadherin adhesion complex at adherens junctions via Nf2 (Hirate

junctions where it switches the Hippo signaling pathway on lated Yap1 is excluded from the nucleus and ICM genes are transcribed.When Nf2 is removed by maternal and zygotic deletion, the Hippopathway cannot be activated and the mutant embryos fail to generate anICM (Cockburn et al., 2013) Instead, inner cells express trophectoderm-specific genes demonstrating that, regardless of their position in the embryo,cells revert to a trophectoderm fate when the components of the Hippopathway are not correctly localized to adherens junctions

Phosphory-5 EMERGING TECHNOLOGIES TO STUDY ADHESION

In vivo imaging of developing mouse embryos has recently been used

to track cell progeny and fate (Bischoff, Parfitt, & Zernicka-Goetz, 2008;

9 Adhesion in the Early Mouse Embryo

Kurotaki, Hatta, Nakao, Nabeshima, & Fujimori, 2007; Morris et al., 2010;

imaging technologies that allow tracking of the levels, stability and tion of proteins controlling adhesion will enable visualization of the estab-lishment of adhesion and how this influences downstream processes such aspolarity and fate For example, photobleaching methods like fluorescencerecovery after photobleaching (FRAP) allow measurement of the overallstability of membrane and cytoplasmic proteins regulating adhesion FRAPhas been useful to elucidate how E-cadherin is positioned and maintained atadherens junctions in different cellular contexts (de Beco, Gueudry,

applied to study adhesion yet in the early mouse embryo Methods like rescence correlation spectroscopy (FCS), based on the study of the fluores-cent fluctuations in a small volume, can provide even more detailedinformation about the biophysical properties of proteins (Digman &

related to cell adhesion, such as actin (Gowrishankar et al., 2012) Morerecently, FCS has been applied to living preimplantation mouse embryos

to study the dynamics of nuclear gene-regulatory proteins (Kaur et al.,

2013) In the future, it will be important to establish its use for cytoplasmicand membrane-bound proteins involved in adhesion to reveal their dynamicbehavior throughout development in different subcellular contexts Super-resolution microscopy approaches are already starting to yield impressivelydetailed information about adhesion complexes in fixed cells (Guillaume

So far, these methods are restricted to thin specimens or sliced tissues, butthey may soon be applicable in more complex three-dimensional structuressuch as the mouse embryo and may even allow data to be gathered at dif-ferent time points, as has been achieved in living brain slices (Berning,

Finally, the study of adhesion will be greatly advanced by the development

of engineered proteins E-cadherin has been engineered to contain a able F€orster resonance energy transfer probe that can be used to directly mea-sure the tensile forces transmitted through the cytoplasmic domain ofE-cadherin (Borghi et al., 2012) So far, this approach has only been used

stretch-in cultured epithelial cells, but it will be highly stretch-informative to visualize tensionacross adherens junctions in the developing mouse embryo Optogenetic con-trol of Rho-family GTPases engineered to be light-sensitive has been used to

remodel actin and alter cell shape in vitro (Leung, Otomo, Chory, & Rosen,

inactivation can be used to selectively inactivate proteins (reviewed in

tech-nologies to express light-sensitive adhesion proteins in the preimplantationmouse embryo, and specifically alter their activity through targeting of laserstimulation to defined subcellular microdomains

6 QUESTIONS FOR THE FUTURE

Recent studies have just begun to explain the connection between celladhesion, polarity, and fate in the early mammalian embryo However,many important questions remain unanswered and the development of sev-eral new technologies will enable them to be addressed in future work.For example, it remains unclear whether, aside from its role in activatingthe Hippo pathway, adhesion has any other signaling functions in the mousepreimplantation embryo E-cadherin has been shown to negatively regulatethe receptor tyrosine kinases EGFR, IGF-1R, and c-Met and inhibit cellgrowth in vitro in an adhesion-dependent manner (Qian, Karpova,

Cdc42 are activated upon E-cadherin-mediated cell contact formation

in vitro (Betson, Lozano, Zhang, & Braga, 2002; Calautti et al., 2002;Kim, Li, & Sacks, 2000; Kovacs, Ali, McCormack, & Yap, 2002; Noren,

sig-naling functions for adhesion in the preimplantation mouse embryo ever, E-cadherin has recently been shown to be required to activate Igf1rsignaling at adherens junctions for trophectoderm formation (Bedzhov,

it also influences signaling from other receptor tyrosine kinases Given the

in vitro evidence and the recent role for adherens junctions in modulating theHippo signaling pathway, it seems likely that adherens junctions may serve asscaffolds where signaling proteins are recruited and regulated Careful imag-ing studies will be required to determine if this is the case in the embryo

In addition, it will be important to establish if adhesion is involved insignaling through the E-cadherin-dependent filopodia that control compac-tion A similar mechanism has been reported in Drosophila where filopodia-like structures known as cytonemes are involved in spatial patterning of the

11 Adhesion in the Early Mouse Embryo

embryo (Roy, Hsiung, & Kornberg, 2011) Further research will ascertainwhether adhesion contributes to spatiotemporally regulated signaling outside

of adherens junctions in the preimplantation mouse embryo

Most of our current knowledge about adhesion in the mouse embryorelates to E-cadherin However, it is likely that other proteins and mecha-nisms also play a role When E-cadherin is disrupted, cells become rounderbut still maintain some adhesive properties (Fierro-Gonzalez et al., 2013;

of this remaining adhesion may be explained by residual expression of otherless known cadherin members (for example, P-Cadherin;Stephenson et al.,

2010), other proteins, such as GalTase (Bayna, Shaper, & Shur, 1988) mayalso participate

Adhesion is one of the main physical forces acting between cells of theembryo However, it remains largely unknown how the adhesion forceinteracts with some of the other main forces acting in the embryo In par-ticular, it is now well recognized that cells behave like viscoelastic fluids andthe opposing forces of adhesion and cortical tension determine the degree ofcell–cell contact (Lecuit & Lenne, 2007) Cells have been shown to behavelike fluid objects, with a tendency to maximize their intercellular adhesion inthe same way that liquids maximize their intermolecular attraction andsimultaneously minimize their free surface energy (Foty, Forgacs,

E-cadherin is likely to initiate this adhesive process by bringing cell surfacesinto contact and providing the first anchoring point The adhesion may then

be reinforced by cells minimizing their cell–liquid interfacial tension asrecently proposed in the zebrafish embryo (Maitre et al., 2012) Researchaddressing the overriding question of how adhesion and tension interact

to control cell–cell interactions and embryo patterning has begun in othernonmammalian systems like the zebrafish (Maitre et al., 2012) and Drosophila

actomyosin cortex or indirect measurements of tensile forces Similar workwill likely be focused in the mammalian embryo in the future

In summary, future work should provide a better understanding of howadhesion is integrated with other key processes patterning the early mamma-lian embryo and controlling cell–cell interactions in vivo The application ofnew technologies permitting real time and quantitative studies will helpreveal how adhesion controls cell shape, fate, and position and vice versa

in the mammalian embryo

Barcroft, L C., Moseley, A E., Lingrel, J B., & Watson, A J (2004) Deletion of the ATPase alpha1-subunit gene (Atp1a1) does not prevent cavitation of the preimplantation mouse embryo Mechanisms of Development, 121(5), 417–426.

Na/K-Bayna, E M., Shaper, J H., & Shur, B D (1988) Temporally specific involvement of cell surface beta-1,4 galactosyltransferase during mouse embryo morula compaction Cell, 53(1), 145–157.

Bedzhov, I., Liszewska, E., Kanzler, B., & Stemmler, M P (2012) Igf1r signaling is pensable for preimplantation development and is activated via a novel function of E-cadherin PLoS Genetics, 8(3), e1002609.

indis-Berning, S., Willig, K I., Steffens, H., Dibaj, P., & Hell, S W (2012) Nanoscopy in a living mouse brain Science, 335(6068), 551.

Betson, M., Lozano, E., Zhang, J., & Braga, V M (2002) Rac activation upon cell-cell tact formation is dependent on signaling from the epidermal growth factor receptor Jour- nal of Biological Chemistry, 277(40), 36962–36969.

con-Bischoff, M., Parfitt, D E., & Zernicka-Goetz, M (2008) Formation of the abembryonic axis of the mouse blastocyst: Relationships between orientation of early cleav- age divisions and pattern of symmetric/asymmetric divisions Development, 135(5), 953–962 Borghi, N., Sorokina, M., Shcherbakova, O G., Weis, W I., Pruitt, B L., Nelson, W J.,

embryonic-et al (2012) E-cadherin is under constitutive actomyosin-generated tension that is increased at cell-cell contacts upon externally applied stretch Proceedings of the National Academy of Sciences of the United States of America, 109(31), 12568–12573.

Calautti, E., Grossi, M., Mammucari, C., Aoyama, Y., Pirro, M., Ono, Y., et al (2002) Fyn tyrosine kinase is a downstream mediator of Rho/PRK2 function in keratinocyte cell- cell adhesion Journal of Cell Biology, 156(1), 137–148.

Cockburn, K., Biechele, S., Garner, J., & Rossant, J (2013) The Hippo pathway member Nf2 is required for inner cell mass specification Current Biology, 23(13), 1195–1201.

de Beco, S., Gueudry, C., Amblard, F., & Coscoy, S (2009) Endocytosis is required for E-cadherin redistribution at mature adherens junctions Proceedings of the National Acad- emy of Sciences of the United States of America, 106(17), 7010–7015.

De Vries, W N., Evsikov, A V., Haac, B E., Fancher, K S., Holbrook, A E., Kemler, R.,

et al (2004) Maternal beta-catenin and E-cadherin in mouse development Development, 131(18), 4435–4445.

Digman, M A., & Gratton, E (2011) Lessons in fluctuation correlation spectroscopy Annual Review of Physical Chemistry, 62, 645–668.

Ducibella, T., & Anderson, E (1975) Cell shape and membrane changes in the eight-cell mouse embryo: Prerequisites for morphogenesis of the blastocyst Developmental Biology, 47(1), 45–58.

Dyce, J., George, M., Goodall, H., & Fleming, T P (1987) Do trophectoderm and inner cell mass cells in the mouse blastocyst maintain discrete lineages? Development, 100(4), 685–698.

Fierro-Gonzalez, J C., White, M R., Silva, J., & Plachta, N (2013) Cadherin-dependent filopodia control preimplantation embryo compaction Nature Cell Biology, 15, 1424–1433 Fleming, T P., Garrod, D R., & Elsmore, A J (1991) Desmosome biogenesis in the mouse preimplantation embryo Development, 112(2), 527–539.

Fleming, T P., McConnell, J., Johnson, M H., & Stevenson, B R (1989) Development of tight junctions de novo in the mouse early embryo: Control of assembly of the tight junction-specific protein, ZO-1 Journal of Cell Biology, 108(4), 1407–1418.

Fleming, T P., & Pickering, S J (1985) Maturation and polarization of the endocytotic system in outside blastomeres during mouse preimplantation development Journal of Embryology and Experimental Morphology, 89, 175–208.

13 Adhesion in the Early Mouse Embryo

Fleming, T P., Sheth, B., & Fesenko, I (2001) Cell adhesion in the preimplantation malian embryo and its role in trophectoderm differentiation and blastocyst morphogen- esis Frontiers in Bioscience, 6, D1000–D1007.

mam-Foty, R A., Forgacs, G., Pfleger, C M., & Steinberg, M S (1994) Liquid properties of embryonic tissues: Measurement of interfacial tensions Physical Review Letters, 72(14), 2298–2301.

Gowrishankar, K., Ghosh, S., Saha, S., Rumamol, C., Mayor, S., & Rao, M (2012) Active remodeling of cortical actin regulates spatiotemporal organization of cell surface mole- cules Cell, 149(6), 1353–1367.

Guillaume, E., Comunale, F., Do Khoa, N., Planchon, D., Bodin, S., & Rouviere, C (2013) Flotillin microdomains stabilize cadherins at cell-cell junctions Journal of Cell Science, 126(Pt 22), 5293–5304.

Gauthier-Handyside, A H (1980) Distribution of antibody- and lectin-binding sites on dissociated blastomeres from mouse morulae: Evidence for polarization at compaction Journal of Embryology and Experimental Morphology, 60, 99–116.

Heintzelman, K F., Phillips, H M., & Davis, G S (1978) Liquid-tissue behavior and ferential cohesiveness during chick limb budding Journal of Embryology and Experimental Morphology, 47, 1–15.

dif-Hirate, Y., Hirahara, S., Inoue, K., Suzuki, A., Alarcon, V B., Akimoto, K., et al (2013) Polarity-dependent distribution of angiomotin localizes Hippo signaling in preimplan- tation embryos Current Biology, 23(13), 1181–1194.

Houliston, E., Pickering, S J., & Maro, B (1987) Redistribution of microtubules and icentriolar material during the development of polarity in mouse blastomeres Journal of Cell Biology, 104(5), 1299–1308.

per-Houliston, E., Pickering, S J., & Maro, B (1989) Alternative routes for the establishment of surface polarity during compaction of the mouse embryo Developmental Biology, 134(2), 342–350.

Huang, J., Huang, L., Chen, Y J., Austin, E., Devor, C E., Roegiers, F., et al (2011) ferential regulation of adherens junction dynamics during apical-basal polarization Jour- nal of Cell Science, 124(Pt 23), 4001–4013.

Dif-Hyafil, F., Babinet, C., & Jacob, F (1981) Cell-cell interactions in early embryogenesis:

A molecular approach to the role of calcium Cell, 26(3 Pt 1), 447–454.

Indra, I., Hong, S., Troyanovsky, R., Kormos, B., & Troyanovsky, S (2013) The adherens junction: A mosaic of cadherin and nectin clusters bundled by actin filaments Journal of Investigative Dermatology, 133(11), 2546–2554.

Javed, Q., Fleming, T P., Hay, M., & Citi, S (1993) Tight junction protein cingulin is expressed by maternal and embryonic genomes during early mouse development Development, 117(3), 1145–1151.

Johnson, M H., & Maro, B (1984) The distribution of cytoplasmic actin in mouse 8-cell blastomeres Journal of Embryology and Experimental Morphology, 82, 97–117.

Johnson, M H., Maro, B., & Takeichi, M (1986) The role of cell adhesion in the nization and orientation of polarization in 8-cell mouse blastomeres Journal of Embryology and Experimental Morphology, 93, 239–255.

synchro-Kaur, G., Costa, M W., Nefzger, C M., Silva, J., Fierro-Gonzalez, J C., Polo, J M., et al (2013) Probing transcription factor diffusion dynamics in the living mammalian embryo with photoactivatable fluorescence correlation spectroscopy Nature Communi- cations, 4, 1637.

Kim, S H., Li, Z., & Sacks, D B (2000) E-cadherin-mediated cell-cell attachment activates Cdc42 Journal of Biological Chemistry, 275(47), 36999–37005.

Kovacs, E M., Ali, R G., McCormack, A J., & Yap, A S (2002) E-cadherin homophilic ligation directly signals through Rac and phosphatidylinositol 3-kinase to regulate adhe- sive contacts Journal of Biological Chemistry, 277(8), 6708–6718.

Kurotaki, Y., Hatta, K., Nakao, K., Nabeshima, Y., & Fujimori, T (2007) Blastocyst axis is specified independently of early cell lineage but aligns with the ZP shape Science, 316(5825), 719–723.

Larue, L., Ohsugi, M., Hirchenhain, J., & Kemler, R (1994) E-cadherin null mutant embryos fail to form a trophectoderm epithelium Proceedings of the National Academy

of Sciences of the United States of America, 91(17), 8263–8267.

Lecuit, T., & Lenne, P F (2007) Cell surface mechanics and the control of cell shape, tissue patterns and morphogenesis Nature Reviews Molecular Cell Biology, 8(8), 633–644 Lecuit, T., Lenne, P F., & Munro, E (2011) Force generation, transmission, and integration during cell and tissue morphogenesis Annual Review of Cell and Developmental Biology, 27, 157–184.

Leung, D W., Otomo, C., Chory, J., & Rosen, M K (2008) Genetically encoded photoswitching of actin assembly through the Cdc42-WASP-Arp2/3 complex pathway Proceedings of the National Academy of Sciences of the United States of America, 105(35), 12797–12802.

Levskaya, A., Weiner, O D., Lim, W A., & Voigt, C A (2009) Spatiotemporal control of cell signalling using a light-switchable protein interaction Nature, 461(7266), 997–1001 Louvet, S., Aghion, J., Santa-Maria, A., Mangeat, P., & Maro, B (1996) Ezrin becomes restricted to outer cells following asymmetrical division in the preimplantation mouse embryo Developmental Biology, 177(2), 568–579.

Maitre, J L., Berthoumieux, H., Krens, S F., Salbreux, G., Julicher, F., Paluch, E., et al (2012) Adhesion functions in cell sorting by mechanically coupling the cortices of adhering cells Science, 338(6104), 253–256.

Morris, S A., Teo, R T., Li, H., Robson, P., Glover, D M., & Zernicka-Goetz, M (2010) Origin and formation of the first two distinct cell types of the inner cell mass in the mouse embryo Proceedings of the National Academy of Sciences of the United States of America, 107(14), 6364–6369.

Nishioka, N., Inoue, K., Adachi, K., Kiyonari, H., Ota, M., Ralston, A., et al (2009) The Hippo signaling pathway components Lats and Yap pattern Tead4 activity to distinguish mouse trophectoderm from inner cell mass Developmental Cell, 16(3), 398–410 Nishioka, N., Yamamoto, S., Kiyonari, H., Sato, H., Sawada, A., Ota, M., et al (2008) Tead4 is required for specification of trophectoderm in pre-implantation mouse embryos Mechanisms of Development, 125(3–4), 270–283.

Noren, N K., Niessen, C M., Gumbiner, B M., & Burridge, K (2001) Cadherin ment regulates Rho family GTPases Journal of Biological Chemistry, 276(36), 33305–33308.

engage-Ohsugi, M., Hwang, S Y., Butz, S., Knowles, B B., Solter, D., & Kemler, R (1996) Expression and cell membrane localization of catenins during mouse preimplantation development Developmental Dynamics, 206(4), 391–402.

Ohsugi, M., Larue, L., Schwarz, H., & Kemler, R (1997) Cell-junctional and cytoskeletal organization in mouse blastocysts lacking E-cadherin Developmental Biology, 185(2), 261–271.

Ohsugi, M., Ohsawa, T., & Semba, R (1993) Similar responses to pharmacological agents of 1,2-OAG-induced compaction-like adhesion of two-cell mouse embryo to physiolog- ical compaction Journal of Experimental Zoology, 265(5), 604–608.

Ozawa, M., Ringwald, M., & Kemler, R (1990) Uvomorulin-catenin complex formation is regulated by a specific domain in the cytoplasmic region of the cell adhesion molecule Proceedings of the National Academy of Sciences of the United States of America, 87(11), 4246–4250.

Pauken, C M., & Capco, D G (1999) Regulation of cell adhesion during embryonic paction of mammalian embryos: Roles for PKC and beta-catenin Molecular Reproduction and Development, 54(2), 135–144.

com-15 Adhesion in the Early Mouse Embryo

Pauken, C M., & Capco, D G (2000) The expression and stage-specific localization of protein kinase C isotypes during mouse preimplantation development Developmental Biology, 223(2), 411–421.

Plachta, N., Bollenbach, T., Pease, S., Fraser, S E., & Pantazis, P (2011) Oct4 kinetics dict cell lineage patterning in the early mammalian embryo Nature Cell Biology, 13(2), 117–123.

pre-Plusa, B., Frankenberg, S., Chalmers, A., Hadjantonakis, A K., Moore, C A., Papalopulu, N., et al (2005) Downregulation of Par3 and aPKC function directs cells towards the ICM in the preimplantation mouse embryo Journal of Cell Science, 118(Pt 3), 505–515.

Qian, X., Karpova, T., Sheppard, A M., McNally, J., & Lowy, D R (2004) mediated adhesion inhibits ligand-dependent activation of diverse receptor tyrosine kinases EMBO Journal, 23(8), 1739–1748.

E-cadherin-Ralston, A., Cox, B J., Nishioka, N., Sasaki, H., Chea, E., Rugg-Gunn, P., et al (2010) Gata3 regulates trophoblast development downstream of Tead4 and in parallel to Cdx2 Development, 137(3), 395–403.

Rauzi, M., Verant, P., Lecuit, T., & Lenne, P F (2008) Nature and anisotropy of cortical forces orienting Drosophila tissue morphogenesis Nature Cell Biology, 10(12), 1401–1410.

Roy, S., Hsiung, F., & Kornberg, T B (2011) Specificity of Drosophila cytonemes for tinct signaling pathways Science, 332(6027), 354–358.

dis-Sano, Y., Watanabe, W., & Matsunaga, S (2014) Chromophore-assisted laser inactivation— Towards a spatiotemporal-functional analysis of proteins, and the ablation of chromatin, organelle and cell function Journal of Cell Science, 127(Pt 8), 1621–1629.

Sefton, M., Johnson, M H., & Clayton, L (1992) Synthesis and phosphorylation of uvomorulin during mouse early development Development, 115(1), 313–318 Sheth, B., Fesenko, I., Collins, J E., Moran, B., Wild, A E., Anderson, J M., et al (1997) Tight junction assembly during mouse blastocyst formation is regulated by late expres- sion of ZO-1 alpha + isoform Development, 124(10), 2027–2037.

Shirayoshi, Y., Okada, T S., & Takeichi, M (1983) The calcium-dependent cell-cell sion system regulates inner cell mass formation and cell surface polarization in early mouse development Cell, 35(3 Pt 2), 631–638.

adhe-Stephenson, R O., Yamanaka, Y., & Rossant, J (2010) Disorganized epithelial polarity and excess trophectoderm cell fate in preimplantation embryos lacking E-cadherin Development, 137(20), 3383–3391.

Summers, M C., & Biggers, J D (2003) Chemically defined media and the culture of malian preimplantation embryos: Historical perspective and current issues Human Repro- duction Update, 9(6), 557–582.

mam-Thomas, F C., Sheth, B., Eckert, J J., Bazzoni, G., Dejana, E., & Fleming, T P (2004) Contribution of JAM-1 to epithelial differentiation and tight-junction biogenesis in the mouse preimplantation embryo Journal of Cell Science, 117(Pt 23), 5599–5608 Vestweber, D., Gossler, A., Boller, K., & Kemler, R (1987) Expression and distribution of cell adhesion molecule uvomorulin in mouse preimplantation embryos Developmental Biology, 124(2), 451–456.

Vinot, S., Le, T., Ohno, S., Pawson, T., Maro, B., & Louvet-Vallee, S (2005) Asymmetric distribution of PAR proteins in the mouse embryo begins at the 8-cell stage during com- paction Developmental Biology, 282(2), 307–319.

Wales, R G (1970) Effects of ions on the development of the preimplantation mouse embryo in vitro Australian Journal of Biological Sciences, 23, 421–429.

Wang, A Z., Ojakian, G K., & Nelson, W J (1990) Steps in the morphogenesis of a ized epithelium I Uncoupling the roles of cell-cell and cell-substratum contact in

polar-establishing plasma membrane polarity in multicellular epithelial (MDCK) cysts Journal

of Cell Science, 95(Pt 1), 137–151.

Whitten, W K (1971) Nutritional requirements for the culture of preimplantation embryos

in vitro Advances in Bioscience, 6, 129–139.

Winkel, G K., Ferguson, J E., Takeichi, M., & Nuccitelli, R (1990) Activation of protein kinase C triggers premature compaction in the four-cell stage mouse embryo Develop- mental Biology, 138(1), 1–15.

Yagi, R., Kohn, M J., Karavanova, I., Kaneko, K J., Vullhorst, D., DePamphilis, M L.,

et al (2007) Transcription factor TEAD4 specifies the trophectoderm lineage at the beginning of mammalian development Development, 134(21), 3827–3836.

Ziomek, C A., & Johnson, M H (1980) Cell surface interaction induces polarization of mouse 8-cell blastomeres at compaction Cell, 21(3), 935–942.

17 Adhesion in the Early Mouse Embryo

Regulation of Cell Adhesion

and Cell Sorting at Embryonic

2.7 Direct investigation of basic properties: Embryonic boundaries are not stable physical structures, but the dynamic product of cell –cell interactions 26

3.6 Boundaries reflect abrupt discontinuities in tissue properties 33

4 Molecular Base of Separation in Vertebrates: Ephrins –Eph Signaling 37

Current Topics in Developmental Biology, Volume 112 # 2015 Elsevier Inc.

Embryonic boundaries are sharp delimitations that prevent intermingling between ferent cell populations They are essential for the development of well-organized struc- tures and ultimately a functional organism It has been long believed that this process was driven by global differences in cell adhesion strength, or expression of different types of adhesion molecules The actual picture turns out to be quite different: Bound- aries should be viewed as abrupt discontinuities, where cortical contractility is acutely upregulated in response to specific cell surface contact receptors which act as repulsive cues Cell adhesion is also modulated along the interface, in different ways depending

dif-on the type of boundary, but in all cases the process is subordinated to the functidif-on of the cortical actomyosin cytoskeleton.

1 INTRODUCTION

Development proceeds by subdivision of a single mass of cells intoprogressively smaller regions, which will eventually give rise to the tissuesand organs of the adult organism The position and size of these regionsare determined by the interplay between patterning signals and gene regu-latory networks, which have been characterized in detail and show amazingdegrees of precision and sophistication It is perhaps less widely appreciatedthat the newly determined regions become rapidly physically separated byembryonic boundaries, which impede any future exchange of cells We willsee that the property to separate from an adjacent population is acquired as aninheritable, cell-autonomous property Without this physical separation,embryonic cells, which divide frequently and are generally highly motile,would be constantly at risk of ending up in the wrong territory Short term,patterning signals are capable of reprogramming misplaced cells To con-stantly maintain a developing structure based on patterning information,however, would be a challenging task, in particular for regions undergoingintense proliferation, such as the insect imaginal discs It would certainly

be close to impossible during large scale movements such as gastrulation.Boundaries largely relieve development from these constraints andallow each separate region to further evolve into complex structures Con-sistently, experimental interference with boundary formation causes distor-tion of the insect wings (e.g., Janody, Martirosyan, Benlali, & Treisman,

2003) and catastrophic defects in the general body plan when it targetsthe initial vertebrate ectoderm–mesoderm separation (e.g., Rohani,Canty, Luu, Fagotto, & Winklbauer, 2011; Winklbauer, Medina, Swain,

& Steinbeisser, 2001) The capacity of cells to sort into different populations

is thus a fundamental property indissociable with the multicellularity ofmetazoans, which can be traced to the deepest roots of animal evolution

as the contractility of the cell cortex, or regulated repulsive reactions.One of the goals of this review is to provide an updated compilation ofthe scattered information gathered over the past years on tissue and bound-ary properties and to confront them to the theoretical models This exer-cise will highlight the paucity of evidence for global differencesbetween separating tissues, opposed to a strong case for local high tension

as hallmark of all boundaries I will then focus on the role of ephrin–Ephsignaling in building this tension at vertebrate boundaries, discussingthe different possible mechanisms through which these repulsive cuesmay control the tension and adhesive properties of the boundaries

A significant part of this essay will be devoted to the many remaining openissues in the field One should note in particular that the prototypic com-partment boundaries in Drosophila are still in want of upstream cues, which

so far have remained inexplicably elusive As for vertebrates, tissue tion clearly depends on an ephrin–Eph-mediated reaction that resemblesclassical contact inhibition, but a description of the process only based

separa-on this mechanism is likely to be a coarse oversimplificatisepara-on We willsee in particular that ephrin–Eph signaling may have multiple effectsbeyond simple repulsion We will also see that even its main target, i.e.,stimulation of actomyosin contractility, can have different effects on celladhesion and boundary properties Furthermore, several other moleculeshave been implicated in separation, which bear no obvious direct connec-tion to ephrins and Eph receptors, except for having myosin as commontarget Although far less understood, these other components must betaken into account, and I will propose some ideas for their integration

in a general model The most exciting next challenge in my opinion is

to move from a coarse description of a generic boundary to the more subtleregulations that likely provide each boundary with the right propertiesrequired for each different morphogenetic process

2 A SHORT HISTORY OF TISSUE SEPARATION

I summarize here the original discoveries of “tissue affinities” andcompartment boundaries, and briefly review the major hypotheses that wereproposed to explain these phenomena

2.1 Cell sorting and “affinities”

The field was founded by the discovery of the phenomenon of cell sorting:When cells dissociated from different embryonic regions are mixed and left

to reaggregate, they initially form a mixed aggregate, but then gradually sortinto distinct populations Remarkably, these cell clusters develop into orga-nized structures that bear the histological signatures of the tissues that wouldnormally derive from the regions from which the cells originated Firstobserved in sponges (Wilson, 1907), the phenomenon was systematicallyanalyzed in frog embryos (Holtfreter, 1939; Townes & Holtfreter, 1955),and its generality was confirmed in chicken embryos (Moscona &

con-stitute fundamental principles of metazoan organization and continue tobear deep implications for our understanding of the process: First, the factthat mixed aggregates can be produced implies that all cells of an embryosshare a common adhesive mechanism We now know that the main actorsare cadherin adhesion molecules Second, each cell, once determined,acquires an autonomous tissue identity, which can be maintained after celldissociation and isolation, and even when the cell finds itself surrounded bycells of another type This identity translates into the capacity to discriminatebetween neighbors and react by adopting a specific cell behavior, i.e., togroup with cells of the same type Holtfreter named this property “tissueaffinity” (Holtfreter, 1939) The nature of the mechanisms that mediatethe recognition of self (homotypic contacts) and nonself (heterotypic con-tacts) and drive the appropriate response remains the central question inthe field

2.2 Compartments

A second key discovery was made by Drosophila geneticists, who observedthat the expansion of proliferating clones in the embryo blastoderm and inthe larval imaginal discs was restricted by invisible yet sharp and fully imper-meable partitions Thus, these epithelial sheets were subdivided into

“compartments,” which were delimited by stably inherited “boundaries”

(Garcia-Bellido, Ripoll, & Morata, 1973; Lawrence & Green, 1975;Lawrence, Green, & Johnston, 1978; Morata & Lawrence, 1978;

1999) Similar compartment boundaries were subsequently found in thevertebrate embryo, e.g., in the brain and limb buds (Altabef, Clarke, &Tickle, 1997; Dahmann, Oates, & Brand, 2011; Fraser, Keynes, &

2.3 The differential adhesion hypothesis

Steinberg had the revolutionary idea to consider cell–cell adhesion from aphysical point of view He noticed that tissue explants behaved very muchlike liquids, from which he conceived the following analogy: Individual cellswould correspond to the molecules of a liquid and cell adhesion to the cohe-sive bonds between these molecules The principle of liquid surface tensionpredicted with astonishing accuracy many of the configurations adopted bycells and tissues For instance, single cells and pieces of tissues in isolationinvariably round up, thus minimizing the surface exposed to the medium,just as a drop of oil in water When placed against an adhesive surface,whether matrix or cells, they spread, or in biophysical terms they “wet”the surface When two groups of cells are put into contact, they either coa-lesce or, on the contrary, they remain fully separated, again similar to thebehavior of immiscible liquids Based on this analogy, Steinberg proposedthat quantitative differences in cell adhesion were sufficient to explain cellsorting and thus tissue separation, a model that was named the differentialadhesion hypothesis (DAH) (Davis, Phillips, & Steinberg, 1997;

demonstrated in vitro, using cells expressing different cadherin levels

does not seem to explain tissue separation in the embryo However, the cept of representing morphogenesis based on a simple combination of adhe-sive and tensile forces exerted on the cell surface (i.e., cell membrane and itsactin cortex) arguably constituted the most influential ideas in the field ofmorphogenesis The analogy to liquid surface tension remains a very suc-cessful way to simulate the behavior of cells and tissues (Foty & Steinberg,

2.4 Differential CAM expression

As one entered the cloning era, and cellular functions could finally start to beassigned to particular gene products, biophysical considerations on cell

sorting were temporarily left aside, and attention shifted naturally to a olutely molecular perspective The two first cell adhesion molecules(CAMs), N-CAM and E-cadherin (Edelman, 1986), soon joined by manyother CAMs, were found to preferentially bind to themselves (Inuzuka,Miyatani, & Takeichi, 1991; Matsuzaki et al., 1990; Nose, Nagafuchi, &

property and the striking tissue-specific expression of most CAMs seemed

to provide a perfect explanation for tissue segregation (Takeichi, 1995).According to this model, individualization of each tissue would rely onthe expression of a particular kit of CAMs (reviewed inOda & Takeichi,

2011) This hypothesis received a broad acceptance among developmentalbiologists, entered the textbooks, and reigned almost undisputed untilrecently

Evidence supporting this hypothesis however remained scarce (Inoue

accu-rate methods led to a reevaluation of the concept of homophilic binding.Notably, two studies showed that cells which expressed two differentcadherins—at similar levels—failed to sort (Duguay, Foty, & Steinberg,

binding is not strictly homophilic (Katsamba et al., 2009; Ounkomol,Yamada, & Heinrich, 2010; Prakasam, Maruthamuthu, & Leckband,

inter-actions were detected in vivo (Straub et al., 2011) A role for differentialexpression of homophilic CAMs in separation remains uncertain Note thatseveral studies hint at specific functions for different cadherin cytoplasmictails in regulating signaling pathways (e.g., Schafer, Narasimha,Vogelsang, & Leptin, 2014; Seidel, Braeg, Adler, Wedlich, & Menke,

could provide an alternative explanation for the multiplicity of cadherinsand for their mosaic expression

2.5 Contact inhibition

Contemporarily to the discovery of CAMs, the identification of ephrins andEph receptors as cell surface repulsive cues (Flanagan & Vanderhaeghen,

1998) pointed to a completely different model for tissue separation, based

on contact inhibition of migration, a phenomenon that had been originally

proposed based on observations on cell lines (Abercrombie, 1967) Althoughephrins and Ephs have mostly been studied for their role in development ofneuronal networks and in angiogenesis, they are also widely expressed inearly embryos (see Fagotto, Winklbauer, & Rohani, 2014 for review).The striking complementarity of ephrin/Eph expression patterns in thedeveloping hindbrain suggested a role in segmentation, which wassupported by functional data (Xu, Alldus, Holder, & Wilkinson, 1995) Sim-ilar observations were then made for segmentation of the somites (Durbin

popular than the cadherin-based models, perhaps because it was untilrecently limited to the two vertebrate segmentation processes However,ephrins and Ephs are now known to control other types of boundaries,including between germ layers (Fagotto, Rohani, Touret, & Li, 2013;

as major regulators of separation in vertebrates

2.6 Differential interfacial tension

The turn of the century witnessed strong revival of biophysical approaches

to developmental processes Myosin II-mediated contractility was known to

be involved in most aspects of cell adhesion and motility, and it becameapparent that this parameter could explain quite a few aspects of morpho-genesis The importance of cell cortex contractility had been alreadyhighlighted in a seminal critical analysis of DAH byHarris (1976) Adhesionand contractility were formally integrated into a broader theory, named dif-ferential interfacial tension hypothesis (DITH) (Brodland, 2002; Brodland,

germ layers, and it was concluded that the system was dominated by ences in cortical tension (Krieg et al., 2008; Maitre et al., 2012) Note that anold puzzle remains: When tested in reaggregation experiments, ectodermcells always sorted toward the center, endoderm to the periphery, and meso-derm in between This configuration fitted perfectly with DITH, but wasopposite to the normal organization of embryos This discrepancy may atleast partly be due to the fact that the dissected explants are artificiallyexposed to the medium, which causes a strong surface tension, whereas

differ-in the embryo, the tissues are wrapped differ-in an outer polarized epithelial layer(outer ectoderm in Xenopus, enveloping layer in zebrafish), which alleviatesinternal tensions (Ninomiya & Winklbauer, 2008)

2.7 Direct investigation of basic properties: Embryonic

boundaries are not stable physical structures, but thedynamic product of cell–cell interactions

What about the extracellular matrix (ECM)? It is a robust isolator of adultorgans Would not it be an obvious candidate for the separation of embry-onic tissues? ECM is indeed deposited soon after separation and contributes

to consolidate the boundaries (Julich, Mould, Koper, & Holley, 2009;

boundaries, which are clearly not physical fences A boundary is in fact onlyimpermeable insofar as it keeps cells of the two populations within theirrespective territories Missorted cells, on the contrary, can freely cross it

to reintegrate the proper tissue This has been unambiguously demonstrated

by following migration of single cells in mosaic notochords (Fagotto et al.,

rhombomeres (Calzolari, Terriente, & Pujades, 2014)

Further support came from in vitro reconstitution of the ectoderm–mesoderm boundary This assay, based on the simple juxtaposition of tissueexplants (Wacker, Grimm, Joos, & Winklbauer, 2000), played a significantrole in the recent progress in uncovering the mechanism of tissue separation

2004; Rohani et al., 2011; Rohani, Parmeggiani, Winklbauer, & Fagotto,

that cells react almost instantaneously to contacts with a tissue of the same

or of the other cell type, “melting” selectively in the former within minutes,while remaining stably separated from the latter This separation behavior

an explant (Wacker et al., 2000) and even between two single dissociatedcells (Rohani et al., 2014)

These observations are important because they firmly validateHoltfreter’s original interpretation that tissue separation relies on cell-autonomous properties A key experimental consequence of this property

is the possibility to study the mechanisms underlying tissue separation

in vitro using isolated cells (Rohani et al., 2014)

What is then the mechanism that allows single cells to find their way andeventually gather with cells of the same type? Classical DAH and DITH statethat boundaries result from the juxtaposition of two cell populations dis-playing global differences in cell–cell adhesion, cortical contractility, orboth Alternatively, cell populations may express different types of adhesionmolecules Contact inhibition does not presume of such differences, but

predicts local effects at heterotypic contacts One would thus expect that theanalysis of the tissue properties would provide support for one or theother model.

3 ADHESION AND CONTRACTILITY OF EMBRYONICTISSUES

There are four main models of boundaries in vertebrates: The derm is first separated from the ectoderm (Fig 1A) Its most axial portionthen splits from the paraxial (or presomitic) mesoderm (PSM) to form thenotochord (Fig 1B) The PSM eventually segments into somites

the segmentation of the hindbrain into seven rhombomeres (Fig 1C) InDrosophila, the two main models are the parasegment boundaries and thecompartment boundaries of imaginal discs (Fig 1F) I have also includedthe particular case of the egg appendages, which has been studied for thefunction of the Echinoid protein (Fig 1E, see below)

I have compiled in this section the available information on adhesion andcontractility, which is fragmentary and heteroclite, making comparisonbetween models still rather rash A brief overview of these methods will

be useful to define what has been actually measured

3.1 Methodology

Cell adhesion can be estimated in vitro by determining the degree of gation of dissociated cells, resistance of tissues to mechanical dissociation oradhesion of single cells plated on immobilized recombinant CAMs (e.g.,

as the force necessary to pull apart a pair of cells using AFM or dual aspirationpipette (Krieg et al., 2008; Maitre et al., 2012) These two methods can also beused to measure cortical stiffness of single cells (Krieg et al., 2008; Maitre et al.,

2012) The advantage of in vitro measurements is a better controlled ment and reduced parameter complexity A caveat of these single cell mea-surements is the fact that cells are bound to actively react to the artificialenvironment, including the large free surface exposed to the medium, and

environ-in the case of AFM, the type of substrate used to hold the cells (environ-inert oradherent, chemical or biological) Thus, it is important to validate thesemeasurements in vivo Cortical tension can be estimated by laser ablation(e.g.,Landsberg et al., 2009) Biosensors are being developed for direct force

Actomyosin enrichment Higher tension

Vertebrate neuroderm

Actomyosin enrichment N-cad N-cad N-cad

Germ layers

Dorsal ectoderm Dorsal mesoderm Endoderm

actomyosin High medium low

C-cad (high)

dorsal endoderm

Dorsal ectoderm (Neuroderm)

Somite Notochord

Actomyosin enrichment N-cad N-cad

Somites

Vertebrate dorsal mesoderm

PSM

Actomyosin enrichment

DE-cad

DE-cad DE-cad X

Ephrins Ephs PAPC

Ephrins Ephrins

Ephs

Ephs Ephrins Ephs

Eph PAPC Ephrin

Echinoid

Ephrins Ephs Ephs PAPC

Figure 1 Embryonic boundary models The enlarged view of each boundary provides information about cadherin distribution (C-cad, N-cad, and DE-Cad), actomyosin struc- tures the cell surface, and the identity and localization of cell surface cues functionally implicated in separation (Ephrins, Eph receptors, PAPC, and Echinoid) The boundary is represented as a dashed line (A) Separation of the dorsal ectoderm and mesoderm in the early Xenopus gastrula The process depends on a complex network of partially selective ephrin –Eph pairs setting bidirectional signals across the boundary PAPC also participate in a parallel and less understood pathway Differences in C-cadherin levels and actomyosin activity are indicated Gaps between the two tissues represent the occurrence of dynamic cycles of detachments and reattachments (B) Separation of the most axial mesoderm (notochord) from the paraxial or presomitic mesoderm (PSM) A similar ephrin –Eph network controls separation PAPC is restricted to the

measurement in live tissues (Borghi et al., 2012; Kuriyama et al., 2014) butthey have not yet been used in the context of tissue separation.

Indirect information can also be obtained from the myosin distribution(e.g.,Calzolari et al., 2014; Landsberg et al., 2009; Rohani et al., 2014) andfrom the relative cadherin levels and their degree clustering (Fagotto et al.,

2013) Another simple but quite informative criterion is the cell geometry,which is predicted to directly reflect the strength and direction of forcesexerted on the cell: At the two extremes, cells with high adhesion tend

to maximize their contacts and adopt a hexagonal shape, while highly tractile/low adhering cells are close to round In principle, a rather precisemap of local tensions may be drawn using refined morphological criteria,including membrane curvature, angles at cell vertices, or cell elongation(e.g.,Brodland et al., 2014, 2009; Lynch, Veldhuis, Brodland, & Hutson,

integrating adhesion and tension, which can be measured for whole tissueexplants (David et al., 2014; David, Ninomiya, Winklbauer, &Neumann, 2009; Kalantarian et al., 2009; Luu, David, Ninomiya, &

PSM, but a role in formation of this boundary has not yet been demonstrated Myosin activation and C-cadherin levels are similar on both sides Myosin is hyperactivated along the boundary and cadherin adhesion is inhibited, which is indicated by the space separating the two tissues Eventually, the gap is permanently stabilized by secretion of a thick layer of extracellular matrix (C) Hindbrain segmentation The hind- brain becomes segmented in seven rhombomeres r1 –7 The process depends on sev- eral ephrins and Eph receptors The central segments r3 –5 presented here express complementary sets of multiple ligands and receptors Expression in the other seg- ments is more complex All segments express N-cadherin homogenously Actin and myosin are enriched along the boundaries (D) Somitogenesis The PSM becomes pro- gressively segmented, starting anteriorly, under the control of ephrinB2, expressed in the posterior half of the newly forming somite, and EphA4, complementary expressed

in the most anterior portion of the unsegmented PSM PAPC, also expressed anteriorly, contributes to the process Eventually, somites become completely isolated from each other (empty space), through deposition of extracellular matrix (E) Drosophila dorsal appendages These two cuticular structures of the Drosophila egg are secreted

by extensions of the follicular epithelium Their primordia are delimited as two regions devoid of Echinoid The juxtaposition of Echinoid-positive and -negative cells produces

a smooth interface with high actomyosin and low DE-cadherin levels (F and F0) Drosophila compartment boundaries The embryonic blastoderm (F) and the wing imaginal disc epithelium (F0) are two examples of epithelia partitioned by sharp compartment boundaries Increased cortical tension is consistent with accumulation

of actomyosin fibers along the interface, but upstream cues have not yet been identified (?).

3.2 Germ layers

The best studied systems are the germ layers of lower vertebrates, fish andamphibians There are discrepancies between studies, which partly reflectintrinsic specificities of the two models, but mostly differences in the type

of approach and in the interpretation of the data Thus, there is the need

to discuss the two cases in some detail

In zebrafish, adhesion, measured by AFM, was found to be lowest forectoderm, intermediate for endoderm, and highest for mesoderm cells

higher in the mesoderm (Montero et al., 2005; Ulrich et al., 2005) For tical stiffness, AFM measurements gave the highest value for ectoderm,intermediate for mesoderm, and lowest for endoderm (Krieg et al., 2008).Inference of the adhesive and tensile components of interfacial tensionhighlighted the predominant influence of cortical tension (Maitre et al.,

cor-2012) Results from cell sorting experiments also yielded configurationsconsistent with cortical tension serving as the main driving force (Krieg

of heterotypic contacts were estimated based on DITH, which assumes thatthese properties should be somewhat intermediate between those of the twotissues Yet heterotypic adhesion had been in fact measured in the originalreport, although not commented, and values were lower than for homotypicadhesion (Krieg et al., 2008), a result that did not fit with DAH/DITH.Other parameters of the boundary interface have not yet been explicitlystudied, but some information can be extracted from published images:Cadherin staining showed no particularity at the boundary (Krieg et al.,

2008), but there seems to be some myosin enrichment (Maitre et al.,

2012) The angles formed by the cell edges appear close to 90° along theboundary (Fig 1F inKrieg et al., 2008), which is a typical configuration thatreflects high interfacial tension Based on these various criteria, interfacialtension at heterotypic contacts must be significantly higher than in each

of the two tissues Although this conclusion awaits confirmation, it would

be fully consistent with the properties of the other boundaries, including theXenopus ectoderm–mesoderm boundary

Cortical properties seem to be shared between zebrafish and Xenopusembryos: Xenopus ectoderm cells are also stiffer than mesoderm cells(AFM unpublished data, Canty and Fagotto), consistent with significantlyhigher levels of activated myosin (Rohani et al., 2014) The parallel withzebrafish does not hold for adhesion, though cadherin levels were found

to be higher in the ectoderm (Angres, M€uller, Kellermann, & Hausen, 1991;

adhesion assays, reaggregation (Brieher & Gumbiner, 1994), adhesion toimmobilized recombinant cadherin extracellular domain (Zhong,

unpublished data) However, a more detailed analysis of different mesodermsubregions showed that this tissue was not homogenous, and that the pre-sumed lower adhesion of this layer only applied to the anterior region, whilethe posterior chordomesoderm appeared similar to the ectoderm in all theseaspects (Chen, Koh, Yoder, & Gumbiner, 2009; Fagotto, unpublishedobservations;Winklbauer, 2009) This was corroborated by measurements

of surface tension of ectoderm, anterior and posterior mesoderm explants

2009) The lack of correlation between tissue properties and separationbehavior is incompatible with a role of DAH/DITH The distinct charac-teristics of each region of the embryo probably reflect specific requirementfor other aspects of gastrulation, which also explains differences betweenzebrafish and Xenopus As for the boundary, it showed unique properties:heterotypic contacts across the interface were constantly disrupted by tran-sient but dramatic repulsive reactions followed by a phase of relaxation andreattachment (Rohani et al., 2011) This behavior, which was neverobserved within the tissues, suggested a mechanism of contact inhibition

We demonstrated that it was indeed directly controlled by ephrins andEph receptors interacting across the boundary (Rohani et al., 2011) Detach-ments were found to correlate with bursts of Rho activation at heterotypiccontacts (Rohani et al., 2011), consistent with myosin accumulation alongthe boundary (Rohani et al., 2014)

Both tissues showed identical cadherin and myosin staining and similar cellshapes (Fagotto et al., 2013) The boundary, however, was characterized bystrong accumulation of actomyosin structures and intense membrane bleb-bing, indicative of extreme cortical tension, and most interestingly, almostcomplete lack of cadherin clusters (see below) All these particularitiesdepended on myosin activity, which, similar to the ectoderm–mesodermcase, was activated downstream of ephrin–Eph signaling (Fagotto

et al., 2013)

3.4 Somite and hindbrain segmentation

Somitogenesis is a particularly complex process, where segmentation isaccompanied by other morphogenetic movements, including compactionand epithelization, a 90° rotation of cell alignment in the lower vertebrates,and differentiation of somite subregions This complexity makes it difficult

to distinguish those parameters that are directly involved in separation sensusstricto from those reflecting other events N-cadherin and N-CAM have dis-tinct distributions in the forming somites (Duband et al., 1987), but thesepatterns appear to relate to epithelization Another classical cadherin,cadherin 11 is specifically expressed in somites, starting first in the posteriorhalf of the newly formed somite (Kimura et al., 1995), but its appearance isalso probably more relevant for subsequent somite cohesion, not for the ini-tial segmentation Loss of N-cadherin in mice did not impair segmentation,but on the contrary led to somite fragmentation, a phenotype that wasenhanced in double N-cadherin/cadherin 11 knockouts (loss of cadherin

11 alone had no effect) (Horikawa, Radice, Takeichi, & Chisaka, 1999).Somite segmentation is actually resistant to general interference with type

I cad (Giacomello et al., 2002) Actin and myosin distribution was examined

in zebrafish Levels were higher along the somatic boundaries, but enous within the forming somites (Julich et al., 2009)

homog-Less is known about the properties of the rhombomeres The hindbrainexpresses N-cadherin homogenously Cadherin 6, however, is expressedspecifically in rhombomeres 6 and 7, but its function is not known

1997) Similar to the somites, actin and myosin are enriched at the aries but homogeneous within the rhombomeres (Calzolari et al., 2014)

bound-As already mentioned, both segmentation processes are known todepend on ephrin–Eph signaling, which readily explains the accumulation

of actomyosin along the boundary (Calzolari et al., 2014)

3.5 Drosophila tissues

In Drosophila, all criteria examined so far argue that tissue properties are ilar on both sides of compartment boundaries, whereas cortical tension ishigher along the boundaries: Cells have identical polygonal shapes onboth sides, but vertex angles approach 90° along the smooth boundary inter-face (Aliee et al., 2012) Myosin levels are indistinguishable between the tis-sues, but significantly higher at the boundary (Aliee et al., 2012; Landsberg

sim-et al., 2009; Laplante & Nilson, 2006; Major & Irvine, 2006; Monier,

Pelissier-Monier, Brand, & Sanson, 2010) Laser ablation in the wingimaginal disc confirmed that cortical tension is identical in anterior andposterior compartments, but higher along the boundary (Landsberg et al.,

2009) Cell adhesion has not yet been examined

3.6 Boundaries reflect abrupt discontinuities in tissue

properties

We have seen that, in most cases, the cell populations show no obvious ferences in terms of adhesive or tensile properties The only exception arethe ectoderm and the mesoderm, but even this case hardly supportsDAH/DITH, since there is no consistency between fish and frog tissuesand high heterogeneity between mesoderm subregions According toDAH/DITH, tissue interfaces should represent some kind of middle point(average) between tissue properties, which certainly does not predict forthe remarkable characteristics of the actual boundaries In fact, simply con-sidering the stereotypical cell alignment and smooth interface of all bound-aries from a biophysical point of view leads to the inescapable conclusionthat boundaries must be sharp discontinuities

dif-Although direct evidence for high interfacial tension so far has only beenobtained on the wing compartment boundary (Landsberg et al., 2009), wehave seen numerous indications that this is a general property of embryonicboundaries, including actomyosin enrichment (all boundaries), repulsivebehavior between ectoderm and mesoderm cells in Xenopus (Rohani

Functional evidence for the importance of contractility has been onstrated in all cases, either by biochemical inhibition of myosin function(Drosophila imaginal discs, Landsberg et al., 2009; vertebrate ectoderm–mesoderm,Rohani et al., 2011, 2014; notochord,Fagotto et al., 2013; hind-brain, Calzolari et al., 2014) or by myosin optical inactivation (Drosophilaparasegments,Monier et al., 2010)

dem-Retrospectively, it makes sense that boundary formation must be driven

by a local and robust mechanism In a DAH/DITH situation, sorting wouldrely on probing differences between different neighbors, which, in order toproduce fast and sharp separation, would need to be extreme In otherwords, such a mechanism would work only if one of the tissues is extremelycompact and the other one extremely loose These conditions would becripplingly limiting for embryonic development They would be incompat-ible, for instance, with morphogenesis of the axial mesoderm, since both the