The Inhibitory Effects of Integrin Antibodies and the RGD Tripeptide on Early Eye Development

E. Svennevik and P. J. Linser

Purpose. The authors investigated the effects of probes that disrupt integrin-extracellular matrix interactions on early eye development. Methods. Antibodies and peptides that have been shown in other studies to block the interaction of cell surface integrins with various ligands were microinjected into the preoptic regions of chick embryos. Eye morphogenesis and biochemical differentiation of ocular tissue layers were assessed by histologic and immunohistochemical analyses. Results. Antibodies that bind to the jS, subunit of integrin and block its function prevented normal eye morphogenesis but did not block expression of certain cell differentiation markers. The RGD tripeptide showed the same inhibitory capacity as did the anti-integrin antibodies. Conclusions. Integrin-based cell-cell and/or cell-extracellular matrix interactions are important in early eye morphogenesis. By contrast, certain aspects of tissue and cell differentiation, such as the expression of carbonic anhydrase II, are controlled independent of morphogenesis. Invest Ophthalmol Vis Sci. 1993; 34:1774-1784.

JL he complexity of the vertebrate eye presumably re- studies have attempted to define the mechanisms in- flects complicated developmental strategies for regula- volved. Evidence has been presented that indicates tion of morphogenesis and biochemical differentia- that both soluble and insoluble signal molecules are tion. As a model for studying "embryonic induction" involved in eye induction.1"4 Several studies have the eye has been a favored system for several de- shown that a prominent extracellular matrix (ECM) is cades.1"3 The early events of eye development, such as laid down between NE and SE as an early event in eye the thickening and folding of neural (NE) and skin morphogenesis.5"8 Furthermore, deposition of this ectoderm (SE) and the appearance of tissue-specific ECM and attachment of the apposing epithelial sheets gene products, are regulated in part by specific inter- to it seems to influence eye maturation.7 actions between the different tissue layers in the eye Numerous studies during the last several years primordium.2i3 The molecular bases of the inductive have shown that a family of surface receptor molecules tissue interactions remain obscure, although many called integrins can mediate cell-ECM interactions.9"20 To examine the role of integrin in early eye development, we have chosen to investigate the effects of in- hibiting integrin-ECM binding. The paradigm was to From the Whitney Uibondory and the. Department of Anatomy and Cell Biology, University of Florida, St. Augustine, Florida. microinject inhibitory antibodies or a synthetic tripep- Supported by grant 1-1030 from the March of Dimes Birth Defects Foundation, tide RGD,1518"20 which mimics an amino acid se- grant BNS-8819743 from the National Science Foundation (PJL), grant BBS-8804980 from the Research Experience for Undergraduates program of the quence common to several ligands of integrin, into the National Science Foundation (ES), and a grant from the Grass Foundation. region of the eye primordium. Then embryos were Submitted for publication: September 25, 1991; accepted August 25, 1992. Proprietary interest category: C7. allowed to continue development for several days. Fi- Reprint requests: P. J. Linser, The Whitney Laboratory and the Department of nally, they were analyzed for structural development Anatomy and Cell Biology, University of Florida, 9505 Ocean Shore Boulevard, St. Augustine, FL 32086. of the eye and biochemical differentiation. Two spe-

Downloaded from iovs.arvojournals.org on 09/30/2021 Integrins Affect Eye Morphogenesis 1775

cine biochemical markers of ocular differentiation in phosphate-buffered saline (PBS) at a concentration were analyzed. One of the earliest gene products to be of 10 mg/ml. Phenol red was added to the solution to synthesized in apparent response to eye induction is ensure proper placement of the micropipette. the lens-specific protein delta crystallin.21 Thus, we used immunocytochemical analysis for delta crystallin Microinjection to evaluate successful lens induction. In addition, we Micropipettes were prepared with a tip diameter of show here that the enzyme carbonic anhydrase II (CA- approximately 20 jum. With the aid of a Nikon SMZ-10 II) is also expressed very early during eye develop- (Nikon Inc., Garden City, NJ) dissecting microscope, ment. Unlike delta crystallin, CA-II shows limited ex- the embryos were injected with approximately 4 nl of pression in both the SE (prelens) and the NE (preret- antibody solution, control hybridoma supernatant, ina) of the early eye rudiment and, thus, serves as one peptide solution, or PBS sham solution. This resulted of the earliest known markers of retina and lens bio- in the delivery of approximately 40 ng of RGD or chemical differentiation. Comparison of these two bio- RGES tripeptide or 2 ng monoclonal antibody to each chemical markers of eye induction with the morpho- embryo that was injected. The solutions were injected logic analyses allowed us to follow lens and retina dif- into the cephalic region of the embryo in the space ferentiation even in the context of disrupted between the SE and the diencephalon of the four to morphology. Our results indicate that the morpho- five somite pair chick embryo (26-29 hr of develop- logic development of the eye is strongly dependent on ment). After injection, the embryos were incubated an integrin-mediated interactions. By contrast, early bio- additional 48 hr before harvest for analysis. chemical differentiation of both lens and retina seem to occur, even when morphogenesis is severely Fixation disrupted. Thus, it appears that aspects of biochemical The embryos were harvested from the surrounding differentiation and morphogenesis are regulated by yolk and embryonic membranes and fixed in Zenker's separate mechanisms. fixative for 2 hr followed by 2 hr of a water wash. The embryos were dehydrated through an alcohol series 25 MATERIALS AND METHODS and xylene and then embedded in paraffin as before. Embryos Immunohistochemical and Histologic Analysis All the investigations presented here were conducted The embryos were sectioned at 6 fim. In most cases, in adherence to the ARVO Statement for the Use of the entire head region was serially sectioned, and all Animals in Ophthalmic and Vision Research. None of sections were analyzed. For analysis of CA-II and delta these experiments involved human subjects in any crystallin distribution, the 7C6 monoclonal antibody way. Fertilized eggs of white leghorn chickens were to CA-II was used with rabbit anti-delta crystallin in a obtained from the University of Florida Division of double-label analysis as previously described.25 In Poultry Sciences. The eggs were incubated at 38°C in a some cases, the sections were immunostained with humidified chamber. For microinjection experiments, monoclonal antibodies to fibronectin (kindly provided the egg shells were removed, and the contents were by D. Fambrough, Johns Hopkins University, Balti- 2223 placed in culture dishes, essentially as described. more, MD) again by standard techniques. After immu- The embryo cultures were maintained in a humidified nohistochemical analysis, the sections were stained sec- chamber at 38°C with an atmosphere of 1% CO2 in air. ondarily with hematoxylin and eosin, and an image of each of the serial sections was recorded on video tape Monoclonal Antibodies to Integrin for later reconstruction. For serial reconstruction, the Monoclonal antibodies to the (5X subunit of integrin images of sections were traced, and the tracings were were the JG22 antibody (generously provided by Da- digitized and reconstructed with the DANCAD3D pro- vid Gottlieb13) and the CSAT and 1-G antibodies (program of Daniel Higgins (San Francisco, CA) using an vided by Clayton Buck9). Hybridomas were grown International Business Machines AT-compatible com- under standard conditions.24 For microinjection ex- puter (CompuAdd Corp., Austin, TX). periments, hybridoma supernatants were concentrated 20-fold with an Amicon (Danvers, MA) pressure concentrator and Ultrafilter XM50 membranes. Con- RESULTS centrates were passed through a 0.2-jtm filter before Biochemical differentiation of optic tissues can be de- injection. tected as early as the 20 somite pair stage. It was shown that CA-II is immunodetectable in both the lens and Synthetic Peptide retina portions of the embryonic chick eye as early as The RGD and RGES peptides were purchased from day 3.5 of development.26 Others reported detecting Peninsula Laboratories. (Belmont, CA) and dissolved CA-II messenger RNA in the chick eye at approxi-

Downloaded from iovs.arvojournals.org on 09/30/2021 1776 Investigative Ophthalmology & Visual Science, April 1993, Vol. 34, No. 5

mately 48 hr of development.27 In Figure 1, we show pair stage.28 The use of monoclonal antibodies to CA- that CA-II antigen first begins to accumulate in ocular II has increased the sensitivity of the immunohisto- tissues at 20 somite pairs of development (48 hr). Fig- chemical technique, making it possible, for the first ure 1 also shows that no CA-II is detectable in the time, to pinpoint the earliest period of CA-II accumu- ectodermal structures of the eye rudiment (prelens lation. At this time, CA-II can be detected in a few and preretina) at the 18-somite-pair stage of develop- mesenchymal cells throughout the embryo, particu- ment, which is only a few hours before the 20-somite- larly at the borders of epithelial organs, such as the brain (Fig. 1). These may represent cells destined to give rise to choroid and blood vessels. Erythrocytes also show CA-II immunoreactivity at this early developmental stage. Most other cells and tissues of the embryo do not contain CA-II at this time. Hence, it is likely that CA-II expression in the very early eye rudiment indicates a specific role for this enzyme in eye development. During subsequent stages of eye development, CA-II continues to be expressed in the ecto- dermally derived portions of the eye.26-29"31 The concentration and cellular distribution of the enzyme pro- gressively changes in the eye, and several distinct phases of eye development have been described on the basis of the CA-II concentration and distribution.32 A second marker of biochemical differentiation in the developing eye is the lens-specific protein delta crystallin.21 Figure 1 shows that delta crystallin is also expressed at approximately the same time as the earliest appearance of CA-II in the eye. However, delta crystallin is only expressed in the lens rudiment, thus making it a useful secondary marker of lens differentiation, which can be used to discriminate NE- and SE-derived optic tissues. In microsurgical experiments, we determined that, as reported earlier,2-3 the SE and NE become adherent to each other by the ten-somite-pair stage of development. Also as reported by others,8 we found that, by ten somite pairs of development, the ECM FICURE 1. This figure shows immunohistochemical localiza- between the SE and NE already showed immunode- tion of CA-II (A-C) and the lens marker protein delta crys- tectable fibronectin, laminin, and heparan sulfate pro- tallin (D) in the early embryonic chick, eye rudiment. (A) teoglycan. Because the SE and NE are readily sepa- Shows a low power magnification view of the optic vesicle region (OV) of an 18-somite-pair stage embryo. CA-II stain- rated surgically before the ten-somite-pair stage, we ing is not evident in the OV but rather is present in scattered reasoned that blockade of the cell-ECM interactions cells (arrows) that border several epithelial structures, in- should be initiated at an earlier time. Thus, injection cluding the forming eye. A region of light staining in the of probes into embryos was performed on embryos at floor of the neural tube (arrow head) is also marked. (B) four to five somite pairs (26-29 hr) of development. Higher magnification view of the region in A outlined in Three different probes with the reported capacity white (ie, the OV). Note absence of staining in the OV but to block or reduce integrin-ECM interaction were prominently stained cells in the mesenchymal tissue between tested: (1) the JG22 monoclonal antibody,13 (2) the the neural ectoderm and skin ectoderm in the region of the 9 CSAT monoclonal antibody, and (3) the RGD syn- forming eye. (C, D) A single section of a 20-somite-pair stage 18 20 embryo at high magnification (as in B) immunostained for thetic tripeptide. - JG22 and CSAT both bind to the the simultaneous localization of CA-II (C) and delta crystal- j8] subunit of chicken integrin. Each antibody concen- lin (D). Note bright staining for CA-II in the newly folded trate was paired with control hybridoma supernatants presumptive neural retina lying between the opticoel (OC) prepared from either nonsecreting SP20 cells or the and the lens placode (LP). Light CA-II reactivity is also seen 1-G cell line, which secretes an anti-integrin antibody in the LP. Arrows again indicate CA-II-positive cells be- that does not block integrin—ECM binding.17 In ex- tween the epithelial layer of the eye rudiment and the head periments with the RGD peptide, control embryos re- mesenchyme. In D, the position of the LP is defined by im- ceived the PBS vehicle, which was used to dissolve and munostaining for delta crystallin. deliver the peptide, or the RGES peptide (also in PBS).

Downloaded from iovs.arvojournals.org on 09/30/2021 Integrins Affect Eye Morphogenesis 1777

In all cases, one side of an embryo in shell-free culture phogenesis were limited to the injected side of the em- was injected with probe solution, and the opposite side bryo. However, in one case using the JG22 antibody, was left undisturbed. In some cases, live hybridoma contralateral inhibition of eye development seemed ap- cells were included in the injected material in an at- parent. tempt to maintain levels of antibody.17 Figure 2 shows a comparison between a control Injection of JG22, CSAT, or the RGD tripeptide (injected with concentrated SP20 supernatant) and gave similar and dramatic effects on the morphogene- the right and left eyes of a JG22-treated embryo. The sis of the eye. In most cases, the effects on eye mor- planes of section in the paired figures are close to be-

FIGURE 2. This figure shows histologic and immunohistochemical analyses of the ocular regions of two embryos that were injected with either 20-fold concentrated SP20 hybridoma supernatant (A-D) or JG22 hybridoma supernatant (E-H). (A, B) Show hematoxylin and eosin-stained sections of the injected and uninjected sides, respectively, of the control, SP20- treated embryo. In both cases, normal lens vesicles (lv) are apparent, and normal eye structure in general was found. The eye cup has formed as a result of invagination, and the retina is thickened relative to earlier stages. Also labeled is the position of the opticoel (oc) as in Figure ]. (C, D) Show immunofluorescence of CA-II in parallel sections to those in A and B, respectively. (E, F) Show the injected and uninjected sides, respectively, of an embryo that received 20-fold concentrated JG22 supernatant. Note that the lens vesicle is incomplete, and the optic cup has not invaginated to form a double-walled structure. The neural ectodermal layer, which should give rise to the neural retina (nr), is small and does not exhibit normal retinal architecture for this stage of development. Also note that the structural abnormalities extend to both sides of the embryo. In (G) and (H), immunofluorescent staining for CA-II shows that, even in the grossly disrupted morphology, CA-II has accumulated in the ocular NE- and SE-derived tissues.

Downloaded from iovs.arvojournals.org on 09/30/2021 1778 Investigative Ophthalmology 8c Visual Science, April 1993, Vol. 34, No. 5

ing equivalent in their location in the embryo, al- failure of eye morphogenesis. The optic cup has not though exact matching was not possible. This explains invaginated on the injected side of the embryo, and why the lens looks filled in Figure 2A and open in the NE has also not thickened; these occur normally Figure 2B. In all cases, double-labeling and serial sec- during retinal differentiation. Serial analysis of the en- tion analyses allowed unambiguous identification of tire eye on both sides of the embryo showed that the tissues as they are labeled in the figures. In this case, uninjected side eye was small, although it possessed the right eye was injected with the concentrated JG22 several characteristics of a normal eye as follows: dou- supernatant. The control eye (SP20) shows normal eye ble-layered optic cup, closed lens vesicle, and lens vesi- structure with the double-layered optic cup and a cle detached from the SE. The eye on the injected side closed lens vesicle. Immunostaining for CA-II shows of the embryo was markedly disrupted. The lens rudi- strong labeling of the SE- and NE-derived tissue ment was still connected to the SE and had not formed layers. The JG22-injected embryo shows a generalized into a closed vesicle. The NE portion of the eye rudi-

FIGVJRE 3. This figure compares parallel sections of the injected

Downloaded from iovs.arvojournals.org on 09/30/2021 Integrins Affect Eye Morphogenesis 1779

ment had not invaginated and formed a double- fibronectin antibodies. In the CSAT treated eyes, the layered optic cup. Despite the obvious inhibition of ocular ECM appeared much thinner and did not eye morphogenesis with the JG22 antibody treatment, clearly separate the SE- and NE-derived layers of the the SE- and NE-derived cell layers showed strong reac- eye rudiment as in the controls. Hence, it is possible tivity with the CA-II monoclonal antibodies. The ap- that the CSAT antibody directly affected the form and parent staining intensity differences might be caused distribution of the ocular ECM. In the absence of by differences in the expression and, hence, regula- more definitive measurements of fibronectin content, tion of the markers, although these differences might this observation remains speculative. also be the result of histologic variability beyond our Figure 4 shows a computer-generated diagram of control. Because of the inherent imprecision of histo- the serial analysis of the CSAT-treated embryo shown logic and immunohistochemical techniques when ap- in Figure 3. This figure demonstrates the extent of the plied to the hundreds of samples analyzed in this morphologic effects of the CSAT antibody and the study, we believe that only qualitative analyses are fact that the effects were primarily limited to the in- valid. jected side of the embryo. This figure also shows that Figure 3 shows typical results as generated with the results presented are not caused by the selection of the CSAT antibody to integrin. In this experiment, the section planes that do not cut through equivalent re- left eye of each embryo was injected with 20-fold con- gions of the eyes being compared. Again, all embryos centrated hybridoma supernatant from either the analyzed were sectioned and studied through the en- CSAT cell line or the 1-G cell line. The 1-G antibody, tire extent of the head and eyes. which reacts with integrin but does not block binding to ECM,17 had no effects on eye development (data not shown). With both, injected and contralateral nonin- jected eyes looked normal. The CSAT antibody, however, caused considerable disruption of eye morphogenesis, as described with the JG22 antibody. In this case, the antibody inhibition was limited to the injected side of the embryo (Fig. 3). Immunohistochemical analyses of CA-II and delta crystallin distribution in these embryos again showed that these two proteins were expressed even in the disrupted morphology. Delta crystallin defined the lens portion of the eye. CA-II staining was present in the easily identified lens, connected SE, and the NE adjacent to the lens rudiment. CA-II staining was also present in a group of loosely organized cells internal to the distinguishable optic NE. Serial section analysis of the entire head region seemed to indicate that these unusual cells were derived from the optic NE. That is, the loose masses of cells showed a gradual compacting and continuity with NE at some distance from the site of maximum dispersion. In addition, in most embryos injected with inhibitory antibodies to integrin, the optic stalk (preoptic nerve) was disorganized or appar- ently missing. In some cases, even the NE of the brain on the injected side of the embryo showed apparent loss of epithelial organization. To examine whether or not the antibody injections also disrupted the ECM of the eye, we immunostained the specimens for fibronectin in parallel sections to those analyzed for CA-II and delta crystallin. In some cases, fibronectin labeling was performed as a FIGURE 4. This figure shows a computer-generated recon- simultaneous double-label analysis in combination struction of the injected (C) and uninjected (B) side eyes of with either delta crystallin or CA-II. Figure 3 also the embryo described in Figure 3. (A) Shows a section of the shows that the ECM that normally lies between the entire embryo head region with the neural retina (R) and lens and the preretina NE is readily visualized with lens (L) of the uninjected side eye labeled.

Downloaded from iovs.arvojournals.org on 09/30/2021 1780 Investigative Ophthalmology & Visual Science, April 1993, Vol. 34, No. 5

To test the role of the cell-ECM interactions on periments. Also, fibronectin distribution in the eye ru- eye morphogenesis further, we injected embryos with diment seemed disrupted by the RGD peptide. In re- either the RGD or the RGES synthetic peptides. Fig- gions where epithelial integrity was lost, fibronectin ure 5 shows that the RGD tripeptide affected eye mor- staining was also greatly reduced (Figures 5C com- phogenesis in much the same manner and degree as pared with Figure 5G). The negative control RGES10 did the inhibitory antibodies of integrin. Similar to the peptide showed no inhibition of eye morphogenesis antibody treatments, the RGD peptide had general ef- (Table 1). fects on the formation of a double-layered optic cup Table 1 presents a summary of our results. A total and caused the disruption of epithelial integrity. The of 61 embryos were treated with either JG22 or CSAT contralateral eye showed slight effects with some antibody; 53 survived to harvest. Of these, 52 showed breakdown of the NE integrity. Both CA-II and delta marked inhibition of ocular morphogenesis. All seven crystallin expression occurred, as in the previous ex- embryos survived treatment with the noninhibitory

FIGURE 5. This figure compares parallel sections of the injected (A-D) and uninjected (E-H) sides of an embryo that received the RGD tripeptide. Similar to antibodies that block the binding of integrin to ECM, this peptide (which can also block, some of these interactions) resulted in abnormal eye morphogenesis. Again, note the reduced size of the injected-side eye, failure to form an optic cup, and apparent disruption of the continuity of the NE. CA-II was, as before, expressed in the optic tissues that did develop (B). Fibronectin immunostaining (C, G) and distribution again seemed reduced (compare C with G) and delta crystallin staining (D, H) defines the lens tissue. By contrast, the uninjected-side eye appears nearly normal (E-H) with some indications of retinal disruption. Abbreviations in (A) and (E) are Iv for lens vesicle and oc for opticoel.

Downloaded from iovs.arvojournals.org on 09/30/2021 Integrins Affect Eye Morphogenesis 1781

TABLE l. Effects of Probe Injection the choroid fissure closes, and the eye becomes a truly on Eye Development chambered organ. Intraocular pressure begins to build, and the lens and optic cup separate to form the Injected No. Percent vitreous chamber of the eye. Cellular differentiation Probe Embryos Survival Normal Abnormal* Effect and specialization in all layers of the embryonic eye SP20 10 8 8 0 0 proceed during an ensuing period of rapid spheric eye JG22 25 23 1 22 95 growth. Hence, from approximately the ten-somite- CSAT 36 30 0 30 100 pair stage of development through the closure of the 1-G 7 7 7 0 0 choroid fissure, the NE and SE structures of the devel- 4 PBS 3 3 0 0 oping eye are juxtaposed to one another and held in RGD 11 8 0 8 100 RGES 10 7 7 0 0 such proximity by a prominent ECM. Embryonic induction and subsequent morpho- "No. embryos" refers to total number of embryos treated. "Sur- vival" indicates number of treated embryos that survived 48 hr genesis of the vertebrate eye is regulated in part by after treatment. "Normal" indicates number of embryos showing microenvironmental cues to the specific cells of the no detectable morphologic effects on eye development. "Abnor- preoptic NE and SE. The concept that the ECM is a mal" indicates number of embryos with defects in eye development. "Percent effect" is the percentage of surviving embryos that framework of developmentally important signal mole- showed abnormal eye development. cules has gained wide acceptance in recent years. With * Abnormal eye development ranged in severity from a minimal the isolation and characterization of specific cell sur- effect, in which the optic vesicle failed to invaginate and form the face molecules that serve as receptors for extrinsic sig- double-walled optic cup, to the most severe disruption, in which portions of the epithelial layers of the eye rudiment appeared to nal molecules, it has become possible to address spe- have become dissociated, losing continuity with the optic stalk. cific questions concerning the mechanisms that regulate cell and tissue development. In this article, we describe experiments that examined the role of the antibody 1-G, and none of these showed any develop- cell surface ECM receptor integrin9 in the early stages mental effects. Likewise, eight often embryos survived of eye development. injection with concentrated SP20 hybridoma superna- The cascade of events that follow the inductive tant with no apparent effects to any of the survivors. interactions between the heterotypic preoptic cell Injection of PBS as a control also caused no effects in types include both morphologic changes and alter- all three survivors. The injection of the RGD tripep- ations in gene expression. The morphologic events of tide showed similar effects to those of the JG22 and eye organogenesis have been well documented.1 CSAT antibodies in all eight embryos that survived Changes in gene expression, which mark the early on- treatment with this probe. All embryos that survived set of eye development, are less well documented. In treatment with RGES peptide (seven of ten) showed this study, we examined standard morphologic no developmental effects. changes and the expression and localization of two proteins that are normally expressed very early in eye DISCUSSION development. The lens-specific protein delta crystallin is an early marker of successful induction of lens devel- After neurulation is complete in the chick embryo, the opment.21 In this report, we also show that CA-II is a early diencephalon begins to evaginate into contralat- useful marker of early eye development, which is first eral eye rudiments. As the NE continues to expand expressed at 20 somite pairs in the chick embryo. Un- laterally, it contacts the SE. An ECM is deposited be- like delta crystallin, CA-II is expressed in both the SE tween the NE and the SE, and by approximately ten and NE derivatives of the optic primordium. In later somite pairs of development, the NE and SE are tightly ocular development, CA-II continues to be expressed 2 adherent to one another through the ECM. Inductive in the SE and NE tissues of the eye, which include the interactions, which cue biochemical and morphologic lens, neural and pigmented retina, and the ciliary epi- development of the eye, are believed to occur during thelium.26 During early chick development (ie, until 12 this stage of development. After this, cell differen- approximately day 6), CA-II expression is limited to tiation is evident because the cells of the NE and SE erythrocytes, scattered cells in the mesenchyme that assume a thickened cuboidal status and the NE invagi- border epithelial organ rudiments, and the SE and NE nates into a double lamina, which conforms to the tissue layers of the eye primordium. During this devel- shape of the developing lens vesicle. The two layers of opmental period, ocular epithelia appear to contain the resulting optic cup continue to differentiate as the the highest concentration of CA-II in the embryo.2632 neural retina and pigmented epithelium and the lens At this time, the role of the enzyme during early eye rudiment detaches from the skin to become the closed development is not certain. Recent studies have shown lens vesicle. During the fourth day of development, that, after the closure of the choroid fissure, inhibitors

Downloaded from iovs.arvojournals.org on 09/30/2021 1782 Investigative Ophthalmology & Visual Science, April 1993, Vol. 34, No. 5

of CA-II activity can decrease the rate of spheric ex- the RGD peptide with embryos treated with vehicle pansion of the embryonic eye.23 This presumably re- solution, SP20 supernatant, or the noninhibitory inte- flects an influence of CA-II on intraocular pressure, a grin antibody 1-G indicated the validity of our results. force that has been shown to drive early eye growth.33 Only when probes that have been shown to block the Furthermore, defects in eye development have been interaction of integrin with ECM molecules were used reported in mammalian species dosed with CA-II in- did we note the range of morphologic changes in eye hibitors during specific periods of embryonic develop- development described herein. Thus, it appears that ment.34 Hence, it is important to understand the regu- the integrin complex is directly involved in the com- lation of CA-II expression. plex tissue remodeling that occurs during early eye Delta crystallin, however, is believed to serve pri- development. Chicken integrin is a heterodimeric marily a structural function in the formation of the complex and can exist in multiple forms with multiple 10 unique cytoarchitecture of the lens.21 Recently, delta ligand-binding specificities. Our results show that an crystallin has been identified as being either identical integrin possessing the /3, subunit is involved in eye or closely related to the enzyme argininosuccinate morphogenesis. However, our results do not specify lyase.35 Whether or not the potential enzymatic activ- which a subunit(s) was involved. ity of delta crystallin plays a role in the lens primor- Our findings suggest that an RGD-containing dium is unknown. For our purposes, we used delta ECM molecule is also important in cell-ECM interac- crystallin as a marker of successful lens induction. tions during ocular morphogenesis. However, there The main objective of this study was to test are several possible such molecules.10 Indeed, al- whether or not inhibitory probes of cell-ECM interac- though suggestive, our results do not show that block- tion could influence either morphologic or biochemi- ade of the integrin-ECM interaction with antibody in- cal development of the eye. We used monoclonal anti- hibits precisely the same receptor-ligand interaction bodies to chicken integrin. Two antibodies that block as is inhibited by the RGD tripeptide. Other possible integrin binding to components of the ECM (JG22 and interactions have been documented and should be ex- CSAT) and a third antibody (1-G) that is specific for amined in the future. integrin but does not block its binding to ECM were Among the general changes in eye morphogenesis tested. In addition, we tested the RGD tripeptide, that we were able to produce by blocking the integrin- which represents an amino acid sequence common to ECM interaction was an apparent reduction in the size several ECM components that bind to integrin.1518 In of the eye rudiment. The volume of identifiable SE- all cases, we observed similar severe inhibition of the and NE-derived tissue in the treated eyes was reduced morphogenesis of ocular structures as follows: (1) fail- by as much as 80% compared with that in control eyes, ure of the optic vesicle to invaginate and form the either on the contralateral side in test embryos or in optic cup; (2) delayed or inhibited formation of the sham-treated controls. This suggests that the exposure lens vesicle and its separation from the SE; (3) break- to the blocking antibodies or RGD peptide either in- down of the integrity of the optic NE; (4) reduced size terfered with cell proliferation in these tissues or was of the optic prirnordium; and (5) reduction in the cytotoxic. Some pycnotic cells mixed in with appar- amount of immunodetectable fibronectin between the ently normal cells were usually evident in the most se- NE- and SE-derived structures. verely affected embryos. This was most noticeable in In this type of in situ analysis, it is difficult to dem- areas where the integrity of epithelial layers was onstrate the specificity of the inhibition. This is partic- disrupted. Thus, it is possible that some of the effects ularly true because it is impossible to control the quan- that we have described may have resulted from cell tity of the probe accurately that is delivered to the death. Nevertheless, the 1-G antibody, which also tissue layers. We attempted to assess antibody delivery binds to integrin, did not cause similar disruption of by immunostaining at intervals after injection. How- development and cell death. This supports our belief ever, as described earlier,17 injected antibodies dilute that the results described here were the result of below the threshold of immunostaining within 24 hr, blockade of ECM-cell interactions and not of some and our experiments were designed to examine mor- generalized immunolytic problem. The effectiveness phology 48 hr after the treatment of embryos. As did of the RGD peptide in these studies also supports this the earlier study,17 we attempted to introduce hybrid- argument. oma cells directly into the chick embryo and trace Although eye morphogenesis was severely disrup- these cells by prior vital labeling. Analyses at the termi- ted by ECM-cell interaction blockade, the expression nation of such experiments were inconclusive, and our of the early markers of optic biochemical differentia- success with injections of antibodies alone caused us to tion, CA-II and delta crystallin, was not prevented. In discontinue hybridoma cell injections. A comparison a few cases, CA-II expression seemed to occur even in of embryos treated with the inhibitory antibodies and cells of the NE that had completely lost cell-cell associ-

Downloaded from iovs.arvojournals.org on 09/30/2021 Integrins Affect Eye Morphogenesis 1783

ations and epithelial organization. Thus, it is likely that 10. Albelda SM, Buck CA. Integrins and other cell adhe- the factors that regulate the timely expression of CA- sion molecules. FASEBJ. 1990;4:2868-2880. II and delta crystallin in SE- and NE-derived optic tis- 11. Horwitz A, Duggan K, Greggs R, Decker C, Buck C. sues can be separated from some of the factors that The cell substrate attachment (CSAT) antigen has control eye morphogenesis. In this study, we exam- properties of a receptor for laminin and fibronectin. J Cell Biol. 1985; 101:2134-2144. ined only one specific class of ECM-cell interaction. 12. Tamkun JW, DeSimone DW, Fonda D, Patel RS, Buck There are numerous other possible interactions of C, Horwitz AF, Hynes RO. Structure of integrin, a ECM molecules with cells that might also be important glycoprotein involved in the transmembrane linkage in this developmental system. Therefore, we cannot between fibronectin and actin. Cell. 1986;46:271- predict whether or not other such ECM-cell interac- 282. tions might influence differential gene expression in 13. Greve JM, Gottlieb DI. Monoclonal antibodies which eye development and, even more specifically, the ex- alter the morphology of cultured chick myogenic cells. pression of CA-II and delta crystallin. In addition, dif- J Cell Biochem. 1982; 18:221-229. fusible factors are known to play a role in eye induc- 14. Wylie DE, Damsky CH, Buck CA. Studies on the function.3-4 Additional studies that attempt to address tion of cell surface glycoproteins. J Cell Biol. these other potential sources of inductive cues may 1979;80:385-402. increase our understanding of eye development specif- 15. Ruoslahti E, Pierschbacher MD. New perspectives in ically and the mechanisms of embryonic induction in cell adhesion: RGD and integrins. Science. 1987; 238:492-497. general. 16. Bronner-Fraser M. Alterations in neural crest migration by a monoclonal antibody that affects cell adhe- Key Words sion. J Cell Biol. 1985;101:610-6l7. integrin, carbonic anhydrase induction, morphogenesis, ex- 17. Bronner-Fraser M. An antibody to a receptor for fi- tracellular matrix bronectin and laminin perturbs cranial neural crest development in vivo. Dev Biol. 1986; 117:528-536. Acknowledgments 18. Boucaut J-C, Darribere T, Poole TJ, Aoyama H, Ya- mada KM, Thiery JP. Biologically active synthetic pep- The authors thank Lynn Milstead for artistic assistance, tides as probes of embryonic development: A competi- Louise McDonald for clerical aid, and Margaret Perkins for tive peptide inhibitor of fibronectin function inhibits technical assistance. gastrulation in amphibian embryos and neural crest cell migration in avian embryos. J Cell Biol. References 1984;99:1822-1830. 19. Krotoski DM, Domingo C, Bronner-Fraser M. Distri- 1. Coulombre AJ. The eye. In: DeHaan RL, ed. Organo- bution of a putative cell surface receptor for fibronec- genesis. New York: Academic; 1965:219-251. tin and laminin in the avian embryo. J Cell Biol. 2. McKeehan MS. Cytological aspects of embryonic lens 1986:103:1061-1071. induction in the chick. J Exp Zool. 1957; 117:31-64. 20. Naidet C, Semeriva M, Yamada KM, Thiery JP. Pep- 3. McKeehan MS. Induction of portions of the chick lens tides containing the cell-attachment recognition sig- without contact with the optic cup. Anat Rec. nal arg-gly-asp prevent gastrulation in Drosophila em- 1958; 132:297-305. bryos. Nature. 1987;325:348-350. 4. Karkinen-Jaaskelainen M. Transfilter lens induction 21. Piatigorsky J. Lens differentiation in vertebrates. A in avian embryo. Differentiation. 1978; 12:31-37. review of cellular and molecular features. Differentia- 5. Hendrix RW, Zwaan J. Changes in the glycoprotein tion. 1981; 19:134-153. concentration of the extracellular matrix between lens 22. Dunn BE, Boone MA. Growth of the chick embryo in and optic vesicle associated with early lens differentia- vitro. Poult Sci. 1975;55:1057-1063. tion. Differentiation. 1974; 2:357-362. 23. Linser PJ, Plunkett JA. A role for carbonic anhydrase 6. Hendrix RW, Zwaan J. The matrix of the optic vesicle- in early eye morphogenesis. Invest Ophthalmol Vis Sci. presumptive lens interface during induction of the 1989:30:783-785. lens in the chicken embryo. J Embryol Exp Morphol. 1975;33:1023-1049. 24. Linser PJ, Perkins MS, Fitch FW, Moscona AA. Com- 7. Hilfer RS, Yang JW. Accumulation of CPC-precipita- parative characterization of monoclonal antibodies to ble material at apical cell surfaces during formation of carbonic anhydrase. Biochem Biophys Res Commun. the optic cup. Anat Rec. 1980; 197:423-433. 1984; 125:690-697. 8. Kurkinen M, Alitalo K, Vaheri A, Stenman S, Saxen L. 25. Linser PJ. Multiple marker analysis in the avian optic Fibronectin in the development of embryonic chick tectum reveals three classes of neuroglia and carbonic eye. Dev Biol. 1979;69:589-600. anhydrase-containing neurons. / Neurosci. 1985; 9. Buck CA, Shea E, Duggan K, Horwitz AF. Integrin 5:2388-2396. (the CSAT antigen): Functionality requires oligomeric 26. Linser PJ, Moscona AA. Carbonic anhydrase C in the integrity./ Cell Biol. 1986; 103:2421-2428. neural retina: Transition from generalized to glia spe-

Downloaded from iovs.arvojournals.org on 09/30/2021 1784 Investigative Ophthalmology 8c Visual Science, April 1993, Vol. 34, No. 5

cific cell localization during embryonic development. regulation of glutamine synthetase and carbonic anhy- Proc Natl Acad Sci U S A. 1981; 78:7190-7194. drase II in neural retina. Proc Natl Acad Sci USA. 27. Rogers JH, Hunt SP. Carbonic anhydrase-II messen- 1986;83:9060-9064. ger RNA in neurons and glia of chick brain: Mapping 32. Linser PJ, Cohen JL. The roles and regulation of car- by in situ hybridization. Neuroscience. 1987; 23:343- bonic anhydrase in the vertebrate neural retina. In: 361. Dodgson SJ, Tashian RE, Gros G, Carter ND, eds. The 28. Hamburger V, Hamilton HL. A series of normal Carbonic Anhydrases: Cellular Physiology and Molecular stages in the development of the chick embryo. J Mor- Genetics. New York: Plenum; 1991:309-317. phoi 1951; 88:49-92. 33. Coulumbre AG. The role of intraocular pressure in the development of the chick eye. / Exp Zool. 29. Linser PJ, Moscona AA. Variable CA-II compartmen- 1956;133:211-225. talization in vertebrate retina. Ann N Y Acad Sci. 34. Scott WJ, Lane PD, Randall JL, Schreiner CM. Mal- 1984; 429:430-446. formations in nonlimb structures induced by acetazol- 30. Linser PJ, Sorrentino M, Moscona AA. Cellular com- amide and other inhibitors of carbonic anhydrase. partmentalization of carbonic anhydrase-C and gluta- Ann NY Acad Sci. 1984; 429:447-456. mine synthetase in developing and mature mouse 35. Wistow G, Piatigorsky J. Recruitment of enzymes as neural retina. Dev Brain Res. 1984; 13:65-71. lens structural proteins. Science. 1987; 236:1554- 31. Vardimon L, Fox LE, Moscona AA. Developmental 1556.

Downloaded from iovs.arvojournals.org on 09/30/2021