a

REVIEWS

Coming Into Focus: The Role of Extracellular Matrix inVertebrate Optic Cup MorphogenesisKristen M. Kwan*

Department of Human Genetics, University of Utah, Salt Lake City, Utah

The vertebrate eye acquires its basic form during the process of optic cup morphogenesis, during which the optic vesicle emergesfrom the brain neuroepithelium and, through a series of cell and tissue movements, transforms itself into the multilayered opticcup, containing neural retina (comprised of retinal progenitors), retinal pigmented epithelium, and the lens, which is derived fromthe overlying ectoderm. While great strides have been made to understand the developmental signals controlling specification, pat-terning, and differentiation of the optic cup, only in recent years have the cellular and molecular bases of optic cup morphogenesisbegun to be unraveled. One critical component of the morphogenetic process is the extracellular matrix: the complex,glycoprotein-rich layer that surrounds the optic vesicle and lens. Though the extracellular matrix has long been visualized by classi-cal histological techniques and postulated to play various roles in optic cup development, its functional role was uncertain. This isnow beginning to change, as live imaging techniques, quantitative image analyses, molecular genetics and in vitro models yieldnew insights into the process of optic cup morphogenesis and the specific influences of particular extracellular matrix componentsand their associated signaling pathways. Developmental Dynamics 00:000–000, 2014. VC 2014 Wiley Periodicals, Inc.

Key words: extracellular matrix; eye; morphogenesis; laminin; fibronectin; lens

Submitted 3 April 2014; First Decision 12 June 2014; Accepted 25 June 2014; Published 00 Month 2014

Introduction

The mature eye is a complex organ, comprised of many tissuesderived from several distinct developmental origins. Precise tissueorganization is critical for visual function: disrupting the devel-opmental processes underlying the acquisition of eye structurecan lead to a variety of human disorders of visual impairment.The basic structure of the eye is established during optic cupmorphogenesis, with elaboration of its specialized substructuresfollowing. Optic cup morphogenesis commences with the evagi-nation of the optic vesicle from either side of the brain neuroepi-thelium (Fig. 1). This small outpocketing of tissue undergoes aseries of cell and tissue rearrangements to generate the optic cup,which contains the neural retina, retinal pigmented epithelium,and lens (Chow and Lang, 2001; Yang, 2004; Martinez-Moralesand Wittbrodt, 2009; Fuhrmann, 2010). At the time when theoptic cup is fully formed (E10.5 in mouse; HH15–HH16 in chick;24 hr postfertilization in zebrafish), the neural retina is a pseu-dostratified columnar epithelium comprised of retinal progeni-tors, and the retinal pigmented epithelium a flat monolayersurrounding the entire neural retina. The lens has separated fromthe overlying ectoderm, but the anterior epithelial layer and lensfibers have not yet differentiated. The choroid fissure, a transientyet critical structure through which vasculature will enter andretinal axons will exit the eye, has formed at the ventral side of

the optic cup, but has not yet begun to fuse. Concurrent withthese dramatic morphogenetic changes, the optic vesicle and lensectoderm execute an intricate molecular program by which tis-sues are specified and patterned. Many major developmental sig-naling pathways, including FGF, Wnt, TGF-b, and Shh, arerequired for the optic cup to adopt its precise regionalized geneexpression before the onset of neurogenesis (Chow and Lang,2001; Yang, 2004; Adler and Canto-Soler, 2007; Fuhrmann,2010), though the role of each pathway in controlling specificcell movements is less clear.

While the steps of vertebrate optic cup morphogenesis weredescribed decades ago using light and electron microscopic tech-niques, little was understood about the dynamics underlying theprocess: the cell movements, single cell behaviors, and changesin cell shape underlying eye morphogenesis. In recent years, liveimaging and quantitative image analysis have revealed a greatdeal about the cellular events governing optic cup morphogene-sis. In the optically accessible zebrafish and medaka, live imagingand 4-dimensional cell tracking have allowed direct observationand analysis of the cell behaviors and movements underlyingoptic vesicle evagination and optic cup morphogenesis (Englandet al., 2006; Rembold et al., 2006; Kwan et al., 2012; Ivanovitchet al., 2013). In mouse and chick, though live embryo observationis difficult to perform, quantitative image analysis of fixedembryos has yielded critical information. Yet it has not been triv-ial to connect cellular dynamics with molecular mechanism. With

DE

VE

LO

PM

EN

TA

L D

YN

AM

ICS

*Correspondence to: Kristen M. Kwan, Department of Human Genetics,EIHG 2100, University of Utah Health Sciences Center, 15 North 2030East, Salt Lake City, UT 84112. E-mail: kmkwan@genetics.utah.edu

Article is online at: http://onlinelibrary.wiley.com/doi/10.1002/dvdy.24162/abstractVC 2014 Wiley Periodicals, Inc.

DEVELOPMENTAL DYNAMICS 00:000–000, 2014DOI: 10.1002/DVDY.24162

1

such a complicated process, intrinsic and extrinsic signals mustcollaborate to yield the precise tissue-level morphogeneticchanges responsible for shaping the optic cup. Embryologicalapproaches in chick have uncovered many of these signals andtheir effects on optic cup formation. Additionally, improvedmolecular and genetic tools, in particular conditional knockoutsin mice, have made it possible to begin to assess the roles of avariety of pathways during optic cup morphogenesis. The extrac-ellular matrix, a complex glycoprotein network known to playimportant roles throughout development, is a critical regulator ofoptic cup morphogenesis and, as recent work is revealing, theunderlying cellular dynamics.

Many different extracellular matrix molecules are present dur-ing optic cup morphogenesis. These include proteoglycans con-taining glycosaminoglycan (GAG) chains such as heparan sulfateand chondroitin sulfate, polysaccharides such as hyaluronic acid,and mesh-forming glycoproteins such as laminin, fibronectin,and collagen (Hendrix and Zwaan, 1975; Wakely, 1977; Kurkinenet al., 1979; McAvoy, 1981; Webster et al., 1983; Parmigiani andMcAvoy, 1984; Webster et al., 1984; Tuckett and Morriss-Kay,1986; Svoboda and O’Shea, 1987; Hilfer and Randolph, 1993;Peterson et al., 1995). The roles of GAG chains and other polysac-charides during optic cup morphogenesis are very poorly under-stood; this problem is compounded, in the case of proteoglycans,by our lack of knowledge regarding the specific roles of theirdiverse core proteins. Relative to the mesh-forming elements,there is less evidence for defined and specific roles of these fac-tors in optic cup morphogenesis, therefore, this review focusesprimarily on the mesh-forming elements laminin, fibronectin,and collagen. Laminins function as heterotrimers of a, b, and g

chains. There are multiple forms of each chain in all vertebrates,with laminin-111 (the heterotrimer of lama1, lamb1, and lamc1)believed to be the predominant isoform through early develop-mental stages (Colognato and Yurchenco, 2000; Miner andYurchenco, 2004). Depending on the specific heterotrimer, lami-nins can bind several different receptors, including integrins anddystroglycan. In addition, laminins are associated with othermatrix components such as fibronectin, collagen, and entactin.Fibronectin functions as a homodimer (linked by disulfide bonds)that is assembled into the matrix and can interact with collagens.

Integrins commonly serve as fibronectin receptors, binding bymeans of a well-characterized RGD (Arginine-Glycine-Aspartate)motif on the fibronectin protein. Collagen IV has been longknown to be present surrounding the early eye (Svoboda andO’Shea, 1987; Hilfer and Randolph, 1993; Matsuda and Keino,2001): unlike more common collagens that wind into a tight tri-ple helix and comprise connective tissue, collagen IV containsstructural differences that result in its assembly into the basementmembrane meshwork that associates with many other matrixmolecules (Hudson et al., 1993; Kalluri, 2003).

All of these extracellular matrix molecules and their receptorscan, in principle, influence optic cup morphogenesis in multipleways. Contact with the matrix can be critical for cell survival:loss of attachment to the matrix can induce a specific form ofprogrammed cell death, anoikis (Frisch and Francis, 1994; Julianoet al., 2004). The matrix can also act as a scaffold for cell migra-tion, influencing direction, duration, and type of movement orother protrusive activity (Adams and Watt, 1993; Daley andYamada, 2013). Furthermore, the matrix surrounds polarized epi-thelial tissues, and the presence or absence of particular matrixcomponents can affect the establishment or maintenance of tis-sue polarity (Martin-Belmonte and Mostov, 2008). Finally, thematrix can influence cell–cell signaling: many secreted signalingmolecules bind to the matrix, which affects diffusion and altersthe range of signaling. Furthermore, in some cases, extracellularmatrix components may be required for signaling: heparan sul-fate in the matrix serves as an obligate cofactor for FGF bindingto its receptor, and may play a role in Wnt signaling. In this way,matrix may influence tissue specification and patterning whilealso controlling cell behavior.

Previous reviews have addressed eye development, and theroles of extracellular matrix components and integrin receptorsthroughout eye development into postembryonic stages (Chowand Lang, 2001; Yang, 2004; Hausman, 2007; Martinez-Moralesand Wittbrodt, 2009; Fuhrmann, 2010). The goal of this review isto focus specifically on the early step of optic cup morphogenesis:recent cell biological insights, gained by means of quantitativeimage analysis and live imaging, are reshaping our view of theextracellular matrix and its role in this important, dynamicprocess.

DE

VE

LO

PM

EN

TA

L D

YN

AM

ICS

Fig. 1. Overview of optic cup morphogenesis and extracellular matrix components. In this schematic of a generic vertebrate, the optic vesicleundergoes a series of dramatic shape changes to form the optic cup: it first evaginates from the brain neuroepithelium, makes close contact withpre-lens ectoderm as the lens placode forms, then invaginates with the lens to form the optic cup. Laminin, fibronectin, and collagen IV are allpresent throughout optic cup morphogenesis; when listed in red, it has been demonstrated to be functionally required during that stage of opticcup formation. Inset: Closer view of retina-lens interface at the end of optic cup morphogenesis. Note that ECM is present not just at the retina-lens interface, but also surrounding the eye and lens structures throughout optic cup morphogenesis. blue, laminin; orange, fibronectin; purple,collagen.

2 KWAN ET AL.

Optic Vesicle Evagination

The optic vesicle initially emerges from the brain neuroepitheliumas a outpocketing of cells on either side of the midline, and thecellular mechanisms driving this movement have begun to beelucidated. Using in vivo imaging techniques, it was reported thatsingle cell migration drives evagination in medaka: in a shorttimelapse movie, cells could be visualized moving individuallyand exchanging neighbors (Rembold et al., 2006). A more recentreport from zebrafish describes slightly different behaviors,including cell elongation and intercalation (Ivanovitch et al.,2013). Two distinct populations of cells are initially observed,with “marginal” cells contacting the extracellular matrix andhaving an elongated morphology, and “core” cells populating themidline region and having a rounded morphology. Over time, themarginal cells elongate further, and the core cells commenceelongation and intercalate amongst the marginal cells. This com-bination of cell elongation and intercalation, potentially com-bined with some active cell migration, may drive fish opticvesicle evagination. Live imaging has not been used in amniotes,so whether these single cell behaviors drive evagination in mouseand chick remains a mystery.

Even at this early stage of eye development, extracellularmatrix components including laminin, fibronectin, and collagenIV surround the nascent optic vesicle of all vertebrate speciesexamined to date (Parmigiani and McAvoy, 1984; Svoboda andO’Shea, 1987; Hilfer and Randolph, 1993; Peterson et al., 1995;Matsuda and Keino, 2001). Evidence that these proteins playfunctional roles in morphogenesis has only recently begun toemerge. For example, a role for laminin in evagination cellbehaviors has been identified in zebrafish, through characteriza-tion of laminin-c1 (lamc1) mutant (sleepym86) and morpholinoknockdown embryos (Ivanovitch et al., 2013). Evagination stilloccurred in the absence of laminin, though cell elongation wasdisorganized, with cells sometimes extending beyond the edge ofthe optic vesicle. Elongated cell morphologies were unstable:some marginal cells were observed with a rounded morphology.Therefore, in zebrafish, laminin appears to coordinate, organize,and delimit the polarized elongation of optic vesicle cells duringevagination. Several other laminin chains (lama1, lama2, lama4,and lamb2) are also expressed through these early developmentalstages in zebrafish (Sztal et al., 2011), though an optic cup phe-notype has only been described for lamb1 and lamc1 mutants(Parsons et al., 2002).

With respect to amniotes, although live imaging is complicatedby embryo accessibility, many features of optic cup morphogene-sis have been recapitulated in vitro, using mouse embryonic stemcells and a defined culture protocol (Eiraku et al., 2011). It wasinitially discovered that stem cell aggregates, when treated transi-ently with activin, could be induced to express retinal markers,yet no epithelial structures resembling optic vesicles or optic cupsformed (Ikeda et al., 2005). When the aggregates were cultured inMatrigel, a complex, incompletely defined extracellular matrixmixture particularly rich in basement membrane components,they exhibited morphogenetic movements resembling optic vesi-cle evagination and optic cup invagination (Eiraku et al., 2011).The distinct results observed in the absence and presence ofMatrigel suggest that the extracellular matrix plays an essentialrole in the morphogenetic process, separate from induction of ret-inal identity. Notably, purified laminin and entactin were foundto be as effective as Matrigel in supporting these morphogenetic

events. Therefore, in this in vitro system, laminin, and the lesswell-understood entactin, appear to be critical, though the exactcellular basis of their contribution is unknown.

In contrast to laminin, the functional role of fibronectin or col-lagen IV in evagination remains unclear; functional experimentstesting a role for collagen IV have not been performed. Withrespect to fibronectin, injection of RGD peptides (which blockfibronectin recognition by its integrin receptors) and antibodiesthat block integrin-b1 binding to extracellular matrix ligands(fibronectin and also possibly laminin) result in severely disor-ganized optic cups in chick (Svennevik and Linser, 1993), butevagination does occur. Optic vesicles were not analyzed forother defects, for example, in cell or tissue morphology, directlyafter evagination. In the future, it will be interesting to determinewhether impaired cell behavior and morphology are observedwhen adhesion to fibronectin is compromised: given that opticvesicles did evaginate in the zebrafish laminin-c1 mutant but cellbehaviors were perturbed, it is possible that more subtle defectsin dynamic behaviors may be observed in this case as well. Inzebrafish, neither the fibronectin mutant natter (fn1a) nor theintegrin-a5 mutant appears to have any early optic cup morpho-genesis defects (Hayes et al., 2012), arguing against a requirementfor fibronectin and its downstream signaling in evagination.However, there is a second fibronectin gene in zebrafish, whichmay act redundantly (fn1b), and integrin-a5 is maternally depos-ited (Julich et al., 2005), meaning that mutants may retain somewild-type message and protein, perdurance of which may be suf-ficient to support evagination. Further experiments will berequired to definitively determine a role for fibronectin and colla-gen IV in evagination.

Lens Placode Formation

Following evagination, the developing optic vesicle establishes aregion of contact with the overlying ectoderm, and lens placodedevelopment commences. It is known that the region between theprospective retina and lens placode is rich in extracellular matrix,separating the tissues: in chick, histological analysis demon-strated strong accumulation of glycoproteins, and electronmicroscopy revealed a dense fibrous matrix between the two tis-sues (Hendrix and Zwaan, 1975; Wakely, 1977; Kurkinen et al.,1979; McAvoy, 1981; Webster et al., 1983; Parmigiani and McA-voy, 1984; Hilfer and Randolph, 1993). Antibody staining forspecific extracellular matrix components demonstrated that lami-nin, fibronectin, and collagen IV are found at this interface inchick, mouse, and rat (Svoboda and O’Shea, 1987; Hilfer andRandolph, 1993; Matsuda and Keino, 2001). Furthermore, focaladhesion markers—indicators of integrin adhesion to the extracel-lular matrix—accumulate at this tissue–tissue interface in frog (Liand Sakaguchi, 2002). Matrix deposition has been demonstratedin mouse to be dependent upon Pax6 expression in the prospec-tive lens (Huang et al., 2011): a conditional Pax6 KO (using theLe-Cre driver in which expression begins at E9 [Ashery-Padanet al., 2000]) failed to show the characteristic accumulation ofPAS and Alcian Blue staining at the retina-lens interface. PASdetects polysaccharides and glycoproteins, while Alcian Bluemarks acidic polysaccharides such as glycosaminoglycans—there-fore, decreased staining indicated deficits in extracellular matrixproduction, and that the prospective lens is a significant sourceof the accumulating ECM. The result of the Pax6 conditionalknockout was failure of placode formation and initiation of lens

DE

VE

LO

PM

EN

TA

L D

YN

AM

ICS

ECM IN OPTIC CUP MORPHOGENESIS 3

invagination. These results suggest that loss of extracellularmatrix could be a factor significantly impairing lens developmentin a lens-specific Pax6 knockout.

The preplacode matrix has been proposed to serve as an adhe-sive that would allow signaling to occur between the retina andlens. Certainly, FGF signaling is a critical part of the reciprocalsignaling between the prospective retina and lens placode, andheparan sulfate, a necessary cofactor for FGF presentation, is acomponent of the interface extracellular matrix. Disruption ofheparan sulfate biosynthesis in the mouse Ndst (N-acetylglucos-amine N-deacetylase-N-sulfotransferase) knockout abrogatesFGF response in the early lens and leads to disruption of lensdevelopment (Pan et al., 2006). But extracellular matrix appearsto have other, more active roles besides facilitating growth fac-tor signaling. Before lens placode formation, fibronectin wasconditionally knocked out in a mouse head explant system,using a ubiquitously expressed tamoxifen-regulated Cre (Huanget al., 2011). Knockout embryos displayed a dramatic reductionin fibronectin staining, and based on quantitative image analy-sis, lens placode thickening was significantly decreased, and thearea of close apposition (a measure of ectodermal spreading)between the prospective retina and lens placode was greaterthan control. Under control circumstances, placode formation isaccompanied by a restriction in ectodermal spreading: the ECMmay play a critical role in restricting ectodermal spreading topromote placode formation. Surprisingly, BMP signaling, whichis required for lens formation, was unaffected by fibronectinknockout. These data suggest that the fibronectin component ofthe interface extracellular matrix plays a crucial role in restrict-ing the spread of ectoderm cells and facilitating placode forma-tion and thickening. Furthermore, this role appears to beindependent of at least one of the developmental signalingpathways required for lens induction; notably, FGF signalingwas not assayed.

Beyond these results with fibronectin, little is known of the func-tional role of other extracellular matrix components during placodeformation: a role for laminin or type IV collagen has not been tested.

Retina and Lens Invagination

Subsequent to placode formation, the retina and lens both inva-ginate, with the retina and RPE enwrapping the lens to yield theoptic cup. The invagination process can be divided into retinaand lens behaviors, which have been inferred through histologi-cal studies or live imaging in fish. The retina thickens and eachcell appears to elongate and decrease the surface area in contactwith the basal extracellular matrix. In amniotes, lens cells arefairly cuboidal, but as the lens invaginates and increases its sur-face area contact with the ECM, each individual cell mayaccordingly change its shape to increase its surface area contactwith the matrix. In fish, the lens invaginates as a solid masswithout a lumen, but the outer cells appear to undergo a similarincrease in surface area contact with the matrix (Greiling andClark, 2009). At some point, the retina and lens lose the closeapposition apparent during placode formation, and little isknown of this process: it is possible that either active de-adhesion or partial extracellular matrix degradation may play arole. Invagination proceeds, and there is some evidence thatmovements may be coordinated between the retina and lens.This coordination may be important only for proper lens mor-phogenesis, as embryological experiments demonstrated that

removing the pre-lens ectoderm (before lens placode formation)did not inhibit retina invagination (Hyer et al., 2003). In addi-tion, the in vitro mouse ES cell system supports retina invagina-tion entirely in the absence of overlying ectoderm (Eiraku et al.,2011). Therefore, retina invagination does not appear to requirethe physical presence of lens tissue. On the other hand, filopo-dia, first described many years ago (Mann, 1928), extend intothe space between the retina and lens. When these filopodiawere ablated specifically in the lens ectoderm (using a lens spe-cific knockout of Cdc42, IRSp53, or focal adhesion kinase[FAK]), the lens invaginated, but exhibited defective, collapsedmorphology, with decreased lens pit depth (Chauhan et al.,2009). Retina invagination, however, proceeded normally. Fromthese quantitative analyses, it was concluded that the filopodiaare required to generate force to aid lens invagination and itscoordination with retinal invagination. Finally, the lens sepa-rates from the overlying ectoderm: the connection between thetissues is constricted, and cell death is likely involved in remov-ing the final remnant in amniotes, although not in zebrafish(Ozeki et al., 2001; Greiling et al., 2010).

Throughout these stages, the same extracellular matrix com-ponents—laminin, fibronectin, and type IV collagens—are pres-ent in the basal lamina surrounding the eye and lens (Kurkinenet al., 1979; McAvoy, 1981; Webster et al., 1983; Parmigianiand McAvoy, 1984; Svoboda and O’Shea, 1987; Hilfer andRandolph, 1993; Peterson et al., 1995), yet the functional roleof these factors is still being determined. In zebrafish, bothlaminin-b1 (lamb1) and laminin-c1 (lamc1) mutants exhibit a“protruding lens” phenotype at optic cup stages (Parsons et al.,2002); although the basis of this phenotype has not yet beencharacterized in detail, it suggests that some aspect of invagi-nation and the ability of the retina to enwrap the lens dependson laminin function. These mutants (as well as laminin-a1)later display severe lens defects, so it is also possible that earlylens invagination or lens capsule formation is affected (Seminaet al., 2006; Lee and Gross, 2007). In mouse, the filopodiaextending into the retina-lens interspace are in contact withand surrounded by laminin (Chauhan et al., 2009). In thefuture, it will be interesting to determine whether filopodial for-mation and dynamics are dependent on laminin (or any othermatrix component) and, by extension, whether the matrix isrequired to generate the force necessary for proper lens mor-phology during invagination.

How other matrix components contribute to invagination isless clear. In chick, injection of RGD peptides (which blockfibronectin-integrin interaction) or integrin-b1 blocking antibod-ies results in severely disorganized optic cups (Svennevik andLinser, 1993). It is difficult to determine to what extent thedefects are due to specific disruption of invagination movements,or loss of the tissue’s structural integrity. In mouse, fibronectinhas been conditionally knocked out only before lens placode for-mation: this blocked placode formation, and subsequent invagi-nation was also inhibited (Huang et al., 2011). To test a role forfibronectin specifically during invagination, it will need to beknocked out after placode formation; it will be interesting todetermine if fibronectin is actually required for multiple, distinctsteps of optic cup morphogenesis. Zebrafish fibronectin mutantsdo not have any apparent defect in optic cup morphogenesis,with eye defects arising only later, during lens cell differentiationand morphogenesis (Hayes et al., 2012). This could possibly bedue to redundancy: as noted above, the fibronectin gene has

DE

VE

LO

PM

EN

TA

L D

YN

AM

ICS

4 KWAN ET AL.

undergone a duplication in fishes. Similar to the evagination andplacode formation processes, no specific role for collagen IV hasbeen identified in invagination. Much more work is requiredhere: mutations in collagen IV genes are responsible for humansyndromes with ocular presentation, including Alport syndrome(Gregory et al., 1996; Savige and Colville, 2009). Patients withAlport syndrome can present with variable retinopathy and alsoanterior lenticonus, an anterior bulging of the lens capsule, sug-gesting a critical role for collagen IV in lens capsule formation:these defects may be present at the time of lens separation, butexamination of a disease model will be required to determine justhow early lens defects can be observed.

In contrast to specific extracellular matrix components,slightly more is known of the requirements for extracellularmatrix receptors and their downstream signaling molecules dur-ing invagination. Integrins serve as receptors for diverse extrac-ellular matrix components, and many are expressed throughoutinvagination stages. Downstream of integrins, integrin-linkedkinase and other focal adhesion components are present andlocalized to the basal surface of the retina and lens duringinvagination. Precise regulation of integrin trafficking has beendemonstrated to be important for optic cup invagination inmedaka: the ojoplano mutant displays increased integrin-b1internalization and failure of invagination (Martinez-Moraleset al., 2009). The ojoplano protein functions by antagonizingNumb pathway-mediated integrin endocytosis (Bogdanovicet al., 2012), suggesting that dysregulation of extracellularmatrix attachment may underlie the failure of invagination inmutants. Integrin-linked kinase (ILK) has been conditionallyknocked out in the mouse lens: early deletion (using the Le-Creline, which drives recombination starting at E9) results in disor-ganization of the lens anterior epithelial layer, leading toincomplete remodeling of the extracellular matrix at the ante-rior margin of the lens (Cammas et al., 2012). Subsequently,the lens capsule extracellular matrix ruptures and the lensdegenerates. This suggests that extracellular matrix remodelingand signaling are critical for lens separation at the conclusionof invagination. Looking ahead, it will be valuable to link theeffects of loss of integrin signaling pathways with specificextracellular matrix components.

Concurrent with the dramatic movements characterizingretina and lens invagination, the choroid fissure forms: acritical structure at the ventral side of the eye through whichvasculature will enter the eye and retinal axons will exit.After the process of optic cup morphogenesis is complete, thechoroid fissure subsequently undergoes a fusion event; failureof either formation or fusion results in the birth defect uvealcoloboma (Gregory-Evans et al., 2004). Many studies haveimplicated matrix (and specifically, matrix destruction) inchoroid fissure fusion: electron microscopy and immunostain-ing for laminin demonstrate that matrix is destroyed duringfusion of the choroid fissure margins (Hero, 1989, 1990;Torres et al., 1996). Furthermore, some coloboma models dis-play prolonged presence of matrix in the fissure, suggestingthat this could be a key step in promoting fusion (Torreset al., 1996; Sehgal et al., 2008).

Yet far less is known about choroid fissure formation, or anyrole for matrix in this process. In zebrafish, laminin mutants dis-play coloboma, indicating some failure of choroid fissure devel-opment (Lee and Gross, 2007), but it is not known whetherchoroid fissure formation or fusion is the disrupted step. Further

work will be required to determine the specific roles of ECM com-ponents in different steps of choroid fissure development.

Perspectives

It is clear that live imaging and quantitative image analysis haveopened up a new era in understanding the cellular events under-lying optic cup morphogenesis. In conjunction, improved genetictools have allowed for more precise examination of the role ofextracellular matrix and downstream signaling in this importantprocess. Some of what we have learned thus far has been surpris-ing: rather than serving a passive role as a simple scaffold for theoptic epithelium, the extracellular matrix may play an active role,directing tissue–tissue signaling and guiding specific changes incell morphology and tissue structure. Much clearly remains to bedone. We still lack a clear understanding of the functions, duringspecific steps of optic cup morphogenesis, of the basic extracellu-lar matrix components addressed here. Many more componentsremain largely uncharacterized, such as entactin and a largenumber of proteoglycans and their core proteins: even the role ofcollagen IV, which is associated with human visual impairmentconditions and has long been known to be present during opticcup morphogenesis, is unclear.

Moving forward, multiple strategies can be exploited to furtherestablish the direct influences of extracellular matrix on opticcup morphogenesis: (1) in vitro models (in which extracellularmatrix components can be easily manipulated); (2) in vivo imag-ing systems, in which cellular dynamics can be directly observedand quantitatively assayed in wild-type and mutant contexts; (3)conditional genetic approaches to more precisely disrupt particu-lar matrix networks; and (4) pharmacological approaches (inaccessible systems) using small molecules to disrupt signalingpathways with precise temporal control. Combined use of thesediverse systems and approaches will uncover the precise func-tions of this critical, complex player in the dynamic process ofvertebrate eye formation.

AcknowledgmentsThanks to members of the Kwan Lab for useful discussions, and toL.C. Murtaugh for reading the manuscript. K.M.K. is supported by aCareer-Starter Research Grant from the Knights Templar EyeFoundation and a Basil O’Connor Starter Scholar Award from theMarch of Dimes.

ReferencesAdams JC, Watt FM. 1993. Regulation of development and differ-

entiation by the extracellular matrix. Development 117:1183–1198.

Adler R, Canto-Soler MV. 2007. Molecular mechanisms of opticvesicle development: complexities, ambiguities and controver-sies. Dev Biol 305:1–13.

Ashery-Padan R, Marquardt T, Zhou X, Gruss P. 2000. Pax6 activityin the lens primordium is required for lens formation and for cor-rect placement of a single retina in the eye. Genes Dev 14:2701–2711.

Bogdanovic O, Delfino-Machin M, Nicolas-Perez M, Gavilan MP,Gago-Rodrigues I, Fernandez-Minan A, Lillo C, Rios RM,Wittbrodt J, Martinez-Morales JR. 2012. Numb/Numbl-Opoantagonism controls retinal epithelium morphogenesis by regu-lating integrin endocytosis. Dev Cell 23:782–795.

Cammas L, Wolfe J, Choi SY, Dedhar S, Beggs HE. 2012. Integrin-linked kinase deletion in the developing lens leads to capsule

DE

VE

LO

PM

EN

TA

L D

YN

AM

ICS

ECM IN OPTIC CUP MORPHOGENESIS 5

rupture, impaired fiber migration and non-apoptotic epithelial celldeath. Invest Ophthalmol Vis Sci 53:3067–3081.

Chauhan BK, Disanza A, Choi SY, Faber SC, Lou M, Beggs HE,Scita G, Zheng Y, Lang RA. 2009. Cdc42- and IRSp53-dependent contractile filopodia tether presumptive lens and ret-ina to coordinate epithelial invagination. Development 136:3657–3667.

Chow RL, Lang RA. 2001. Early eye development in vertebrates.Annu Rev Cell Dev Biol 17:255–296.

Colognato H, Yurchenco PD. 2000. Form and function: the lamininfamily of heterotrimers. Dev Dyn 218:213–234.

Daley WP, Yamada KM. 2013. ECM-modulated cellular dynamicsas a driving force for tissue morphogenesis. Curr Opin GenetDev 23:408–414.

Eiraku M, Takata N, Ishibashi H, Kawada M, Sakakura E, Okuda S,Sekiguchi K, Adachi T, Sasai Y. 2011. Self-organizing optic-cup morphogenesis in three-dimensional culture. Nature 472:51–56.

England SJ, Blanchard GB, Mahadevan L, Adams RJ. 2006. Adynamic fate map of the forebrain shows how vertebrate eyesform and explains two causes of cyclopia. Development 133:4613–4617.

Frisch SM, Francis H. 1994. Disruption of epithelial cell-matrixinteractions induces apoptosis. J Cell Biol 124:619–626.

Fuhrmann S. 2010. Eye morphogenesis and patterning of the opticvesicle. Curr Top Dev Biol 93:61–84.

Gregory MC, Terreros DA, Barker DF, Fain PN, Denison JC, AtkinCL. 1996. Alport syndrome–clinical phenotypes, incidence, andpathology. Contrib Nephrol 117:1–28.

Gregory-Evans CY, Williams MJ, Halford S, Gregory-Evans K.2004. Ocular coloboma: a reassessment in the age of molecularneuroscience. J Med Genet 41:881–891.

Greiling TM, Aose M, Clark JI. 2010. Cell fate and differentiation ofthe developing ocular lens. Invest Ophthalmol Vis Sci 51:1540–1546.

Greiling TM, Clark JI. 2009. Early lens development in the zebra-fish: a three-dimensional time-lapse analysis. Dev Dyn 238:2254–2265.

Hausman RE. 2007. Ocular extracellular matrices in development.Prog Retin Eye Res 26:162–188.

Hayes JM, Hartsock A, Clark BS, Napier HR, Link BA, Gross JM.2012. Integrin alpha5/fibronectin1 and focal adhesion kinase arerequired for lens fiber morphogenesis in zebrafish. Mol Biol Cell23:4725–4738.

Hendrix RW, Zwaan J. 1975. The matrix of the optic vesicle-presumptive lens interface during induction of the lens in thechicken embryo. J Embryol Exp Morphol 33:1023–1049.

Hero I. 1989. The optic fissure in the normal and microphthalmicmouse. Exp Eye Res 49:229–239.

Hero I. 1990. Optic fissure closure in the normal cinnamonmouse. An ultrastructural study. Invest Ophthalmol Vis Sci 31:197–216.

Hilfer SR, Randolph GJ. 1993. Immunolocalization of basal laminacomponents during development of chick otic and optic primor-dia. Anat Rec 235:443–452.

Huang J, Rajagopal R, Liu Y, Dattilo LK, Shaham O, Ashery-Padan R, Beebe DC. 2011. The mechanism of lens placode for-mation: a case of matrix-mediated morphogenesis. Dev Biol355:32–42.

Hudson BG, Reeders ST, Tryggvason K. 1993. Type IV collagen:structure, gene organization, and role in human diseases. Molec-ular basis of Goodpasture and Alport syndromes and diffuseleiomyomatosis. J Biol Chem 268:26033–26036.

Hyer J, Kuhlman J, Afif E, Mikawa T. 2003. Optic cup morphogene-sis requires pre-lens ectoderm but not lens differentiation. DevBiol 259:351–363.

Ikeda H, Osakada F, Watanabe K, Mizuseki K, Haraguchi T,Miyoshi H, Kamiya D, Honda Y, Sasai N, Yoshimura N, TakahashiM, Sasai Y. 2005. Generation of Rxþ/Pax6þ neural retinal pre-cursors from embryonic stem cells. Proc Natl Acad Sci U S A102:11331–11336.

Ivanovitch K, Cavodeassi F, Wilson SW. 2013. Precocious acquisi-tion of neuroepithelial character in the eye field underlies theonset of eye morphogenesis. Dev Cell 27:293–305.

Juliano RL, Reddig P, Alahari S, Edin M, Howe A, Aplin A. 2004.Integrin regulation of cell signalling and motility. Biochem SocTrans 32:443–446.

Julich D, Geisler R, Holley SA. 2005. Integrinalpha5 and delta/notch signaling have complementary spatiotemporal require-ments during zebrafish somitogenesis. Dev Cell 8:575–586.

Kalluri R. 2003. Basement membranes: structure, assembly androle in tumour angiogenesis. Nat Rev Cancer 3:422–433.

Kurkinen M, Alitalo K, Vaheri A, Stenman S, Saxen L. 1979. Fibro-nectin in the development of embryonic chick eye. Dev Biol 69:589–600.

Kwan KM, Otsuna H, Kidokoro H, Carney KR, Saijoh Y, Chien CB.2012. A complex choreography of cell movements shapes thevertebrate eye. Development 139:359–372.

Lee J, Gross JM. 2007. Laminin beta1 and gamma1 containinglaminins are essential for basement membrane integrity in thezebrafish eye. Invest Ophthalmol Vis Sci 48:2483–2490.

Li M, Sakaguchi DS. 2002. Expression patterns of focal adhesionassociated proteins in the developing retina. Dev Dyn 225:544–553.

Mann I. 1928. The development of the human eye. New York:Grune & Stratton.

Martin-Belmonte F, Mostov K. 2008. Regulation of cell polarityduring epithelial morphogenesis. Curr Opin Cell Biol 20:227–234.

Martinez-Morales JR, Rembold M, Greger K, Simpson JC, BrownKE, Quiring R, Pepperkok R, Martin-Bermudo MD, HimmelbauerH, Wittbrodt J. 2009. ojoplano-mediated basal constriction isessential for optic cup morphogenesis. Development 136:2165–2175.

Martinez-Morales JR, Wittbrodt J. 2009. Shaping the vertebrateeye. Curr Opin Genet Dev 19:511–517.

Matsuda M, Keino H. 2001. Possible roles of beta-catenin in evagi-nation of the optic primordium in rat embryos. Dev Growth Differ43:391–400.

McAvoy JW. 1981. The spatial relationship between presumptivelens and optic vesicle/cup during early eye morphogenesis in therat. Exp Eye Res 33:447–458.

Miner JH, Yurchenco PD. 2004. Laminin functions in tissue mor-phogenesis. Annu Rev Cell Dev Biol 20:255–284.

Ozeki H, Ogura Y, Hirabayashi Y, Shimada S. 2001. Suppression oflens stalk cell apoptosis by hyaluronic acid leads to faulty sepa-ration of the lens vesicle. Exp Eye Res 72:63–70.

Pan Y, Woodbury A, Esko JD, Grobe K, Zhang X. 2006. Heparansulfate biosynthetic gene Ndst1 is required for FGF signaling inearly lens development. Development 133:4933–4944.

Parmigiani C, McAvoy J. 1984. Localisation of laminin andfibronectin during rat lens morphogenesis. Differentiation 28:53–61.

Parsons MJ, Pollard SM, Saude L, Feldman B, Coutinho P, HirstEM, Stemple DL. 2002. Zebrafish mutants identify an essentialrole for laminins in notochord formation. Development 129:3137–3146.

Peterson PE, Pow CS, Wilson DB, Hendrickx AG. 1995. Localisa-tion of glycoproteins and glycosaminoglycans during early eyedevelopment in the macaque. J Anat 186(Pt 1):31–42.

Rembold M, Loosli F, Adams RJ, Wittbrodt J. 2006. Individual cellmigration serves as the driving force for optic vesicle evagina-tion. Science 313:1130–1134.

Savige J, Colville D. 2009. Opinion: ocular features aid the diagno-sis of Alport syndrome. Nat Rev Nephrol 5:356–360.

Sehgal R, Karcavich R, Carlson S, Belecky-Adams TL. 2008.Ectopic Pax2 expression in chick ventral optic cup phenocopiesloss of Pax2 expression. Dev Biol 319:23–33.

Semina EV, Bosenko DV, Zinkevich NC, Soules KA, Hyde DR,Vihtelic TS, Willer GB, Gregg RG, Link BA. 2006. Mutations inlaminin alpha 1 result in complex, lens-independent ocular phe-notypes in zebrafish. Dev Biol 299:63–77.

Svennevik E, Linser PJ. 1993. The inhibitory effects of integrin anti-bodies and the RGD tripeptide on early eye development. InvestOphthalmol Vis Sci 34:1774–1784.

Svoboda KK, O’Shea KS. 1987. An analysis of cell shape and theneuroepithelial basal lamina during optic vesicle formation in themouse embryo. Development 100:185–200.

DE

VE

LO

PM

EN

TA

L D

YN

AM

ICS

6 KWAN ET AL.

Sztal T, Berger S, Currie PD, Hall TE. 2011. Characterization of the lami-nin gene family and evolution in zebrafish. Dev Dyn 240:422–431.

Torres M, Gomez-Pardo E, Gruss P. 1996. Pax2 contributes to inner earpatterning and optic nerve trajectory. Development 122:3381–3391.

Tuckett F, Morriss-Kay GM. 1986. The distribution of fibronectin,laminin and entactin in the neurulating rat embryo studied by indi-rect immunofluorescence. J Embryol Exp Morphol 94:95–112.

Wakely J. 1977. Scanning electron microscope study of the extrac-ellular matrix between presumptive lens and presumptive retinaof the chick embryo. Anat Embryol (Berl) 150:163–170.

Webster EH, Jr., Silver AF, Gonsalves NI. 1983. Histochemical anal-ysis of extracellular matrix material in embryonic mouse lensmorphogenesis. Dev Biol 100:147–157.

Webster EH Jr, Silver AF, Gonsalves NI. 1984. The extracellularmatrix between the optic vesicle and presumptive lens duringlens morphogenesis in an anophthalmic strain of mice. Dev Biol103:142–150.

Yang XJ. 2004. Roles of cell-extrinsic growth factors in vertebrateeye pattern formation and retinogenesis. Semin Cell Dev Biol 15:91–103.

DE

VE

LO

PM

EN

TA

L D

YN

AM

ICS

ECM IN OPTIC CUP MORPHOGENESIS 7