Immunolocalization of basal lamina components during development of chick otic and optic primordia.код для вставкиСкачать
THE ANATOMICAL RECORD 235:443-452 (1993) lmmunolocalization of Basal Lamina Components During Development of Chick Otic and Optic Primordia S. ROBERT HILFER AND GWENDALYN J. RANDOLPH Department of Biology, Temple University, Philadelphia, Pennsylvania 19122 ABSTRACT Immunolocalization of laminin, fibronectin, and type IV collagen was examined during early morphogenetic shape changes of the avian inner ear and eye. The ear was studied from formation of the otic placode to invagination of the otic pit and the eye from the optic vesicle stage to formation of an optic cup. Distribution and intensity of immunoreactivity were compared in the two organ primordia and in adjacent epithelial layers. Laminin formed a continuous layer at the basal surface of the otic ectoderm and adjacent neural tube at all stages. The basal surfaces of the optic and lens epithelia also were continuously covered with laminin throughout development. The otic placode became attached to the neural ectoderm through a single layer of fibronectin and collagen IV between the layers of laminin. The ring-like attachment between the edges of the optic cup and lens primordium had the same structure. In addition, the central regions of the optic and lens primordia were attached by fibrils containing type IV collagen, whereas finer strands containing fibronectin and laminin also connected the otic epithelium and neural tube. The results are discussed in terms of models of invagination for the two primordia. 0 1993 Wiley-Liss, Inc. Key words: Otic placode, Optic vesicle, Organogenesis, Laminin, Fibronectin, Type IV collagen Invagination of epithelial sheets to form pit-like structures occurs during organogenesis of many embryonic primordia. The mechanisms that control invagination have been identified as comprising both intracellular and extracellular forces. Much attention has been paid to the role of the cytoskeleton in causing bending of epithelial sheets (reviewed in Hilfer and Searls, 1986; Trinkaus, 1984). An extensive literature also exists on the involvement of extracellular matrix in organ formation. For example, interference with collagen, proteoglycan, and glycoprotein synthesis inhibits branching of salivary, lung, and kidney primordia (Klein et al., 1989; Nakanishi et al., 1986; Spooner and Faubion, 1980; Spooner et al., 1985; Thompson and Spooner, 1983) and invagination of optic (Yang and Hilfer, 1982; Gerchman et al., 1991) and otic (Rausch and Hilfer, 1988; Gerchman, et al., 1991) primordia. The involvement of extracellular macromolecules in organotypic shape changes appears to be through their linkage with the cell surface. Integral membrane proteins, such as integrins (reviewed in Ginsberg et al., 1990) and the heparan sulfate proteoglycan syndecan (Hayashi et al., 19871, act as receptors principally for N-linked glycoproteins, including fibronectin and laminin but also for collagen, the major structural component of the basal lamina. Epithelial links t o the basal lamina and extended extracellular matrix may initiate shape changes or stabilize them once they have occurred. If such molecules have a role in the process of shape change, differences most likely exist in their 0 1993 WILEY-LISS. INC. presence, distribution, concentration, or molecular form a t critical developmental stages. In this study, the distributions of three major components of basal laminae are compared in primordia of the eye and inner ear. Since invagination of these organ primordia is controlled differently, it was expected that differences would exist in the temporal compositions of their basal laminae. In both primordia, invagination is inhibited by tunicamycin (Yang and Hilfer, 1982; Rausch and Hilfer, 1988), a drug that inhibits N-linked glycosylation (Mahoney and Duskin, 1979). However, the optic primordium can be stimulated to form precociously or inhibited from invaginating by agents that affect cytoskeletal function (Hilfer et al., 1981; Brady and Hilfer, 1982; Maloney and Wakely, 1982), while the otic primordium is unresponsive to these treatments (Hilfer et al., 1989). Fluorescent immunolocalization techniques were used to study temporal and regional differences in distribution of laminin, fibronectin, and type IV collagen. The similarities in developmental distribution outweigh the differences between the basal laminae of these two primordia. Some differences do exist in the connections between Received April 28, 1992; accepted August 13, 1992. Gwendalyn J. Randolph is now at the Department of Pathology, State University of New York at Stony Brook, Stony Brook, NY 11794. 444 S.R. HILFER AND G.J. RANDOLPH Figs. 1-4 445 OTIC AND OPTIC BASAL LAMINA basal laminae of these organs and the basal laminae of adjacent epithelial layers. MATERIALS AND METHODS Preparation of Specimens Fertilized white leghorn eggs (Truslow Farms, Chestertown, MD) were incubated a t 37°C in a humidified commercial egg incubator. Embryos ranging from stages 10 through 15 (Hamburger and Hamilton, 1951) were removed from the yolk, transferred to 0.1 M phosphate-buffered saline (PBS), pH 7.2, containing 1.0 mM calcium and magnesium as chloride salts, and cut just caudal to the otic placodes/vesicles. The embryos were fixed for 20 min at room temperature in freshly prepared 3.0% paraformaldehyde in PBS containing Ca2+ and Mg2+.Fixed embryos were embedded in acrylamide (Johnson and Blanks, 1984) to provide support for the delicate structure of these young embryos. Fixed embryos were placed on a pad of polymerized acrylam- Fig. 1. Fluorescence micrographs of chick embryos at the level of otic primordia, treated with antilaminin and fluorescein isothiocyanate (F1TC)-labeled secondary antibodies. Bar = 100 pm. a: At stage 10, laminin surrounds the neural tube (NT) and forms a continuous layer under the ectoderm, including the otic primordium (OP). Intensity of fluorescence is greater between neural tube and otic placode than in regions where they are not adjacent. b: At stage 11, the distribution is much the same as earlier. This thinner section allows laminin to be visualized as two lines between neural tube and otic primordium. c: At stage 14, the difference in intensity of laminin fluorescence between medial and lateral regions of the otic pit is not as distinct as earlier. The most dorsal region still reacts more intensely than the rest of the surface (arrowhead). d: By stage 15, the otic pit has separated from the surface of the neural tube, and differences in fluorescent intensity are no longer visible. At all stages, laminin immunoreactivity within the mesenchyme is punctate. Fig. 2. Phase-contrast micrographs of the same frozen sections embedded in acrylamide as in Figure 1. Bar = 100 pm. a: The primordium (OP) forms a slight depression along the neural tube (NT) a t stage 10. b: By stage 11, the otic primordium is folded (arrow) between the surface adjacent to the neural tube (NT) and the surface adjacent to somitomeric mesoderm. c: At stage 14,the ventral otic pit has separated partially from the neural tube (NT); mesenchymal cells (M) are in the gap. d By stage 15, a layer of mesenchymal cells lies between the otic pit (OP) and the neural tube (NT). Convergence of dorsal and ventral folds partially closes the pit. Fig. 3.Fluorescence micrographs of chick embryos at the level of otic primordia, treated with antifibronectin and FITC-labeled secondary antibodies. Bar = 100 pm. a: At stage 11, the neural tube (NT), otic primordium (OP), and pharynx (P) are surrounded by a layer of fibronectin, which appears to merge where the otic and neural primordia are apposed. The mesenchyme also is immunofluorescent (arrow). b At stage 13 the distribution of fluorescence has not changed. The gap (arrowhead) is a sectioning artefact. NT, neural tube. c: At stage 14, fibronectin immunoreactivity in the ventral separation of the apposed region shows as two lines, and adjacent mesenchyme is more intensely fluorescent than at earlier stages. NT, neural tube. d By stage 15, the intensity of fluorescence is diminished along the basal surface of the otic pit (OP). Fig. 4. Higher magnifications of the ventral apposition between the otic primordium and neural tube. Bars = 10 pm. a: Distribution of laminin in the region of the otic fold at stage 11 (enlargement of Fig. lb). Cells a t the fold are elongated (arrowhead) and maintain contact with the surface of the neural tube (NT).Some fine strands of laminin (arrows) bridge the gap between the two laminin layers. b Between the two separating layers at stage 14, the strands of laminin seen in the gap at stage 11are not visible. NT, neural tube. c: Distribution of fibronectin at the basal surface of the fold at stage 13. Strands of fibronectin stretch between the two cell layers (arrow) and surround mesenchyme cells (M). d The same region a t stage 14 contains mesenchyme cells surrounded by a fibronectin-containing matrix. ide and surrounded by acrylamide monomer, which was polymerized by addition of 10% ammonium persulfate at room temperature. The acrylamide blocks were trimmed, surrounded by OCT Compound (Miles Laboratory, Naperville, IL), and quick frozen in liquid nitrogen for sectioning. Sections of 10-15 p,m were prepared at -26°C and mounted on gelatin-coated slides. Antibodies Monoclonal mouse antibodies against avian laminin (Bayne et al., 1984) and fibronectin (Gardner and Fambrough, 1983) were obtained from the Developmental Studies Hybridoma Bank (maintained by the Department of Pharmacology and Molecular Sciences, Johns Hopkins University School of Medicine, Baltimore, MD, and the Department of Biology, University of Iowa, Iowa City, IA, under contract N01-HD-6-2915 from the NICHD). Immunolocalization of laminin and fibronectin also was studied using a polyclonal rabbit antimouse laminin antibody kindly provided by Dr. H.K. Kleinman and a polyclonal rabbit antimouse fibronectin antibody kindly furnished by Dr. J.M. Chen. Polyclonal rabbit antimouse collagen IV antibody was the gift of Dr. C. Little. Whole-molecule antirabbit IgG antibodies conjugated with fluorescein isothiocyanate (FITC; Atlantic Antibodies, Scarborough, ME) or antimouse IgG conjugated with tetramethylrhodamine isothiocyanate (RITC, Cappel, West Chester, PA) were used as secondary antibodies. lmrnunofluorescence The lowest dilutions were determined at which antilaminin (full strength, monoclonal; 1:3,200 polyclonal), anticollagen IV (1:200), and antifibronectin (full strength, monoclonal; 1:lOO polyclonal) antibodies could be used and still provide a bright fluorescent signal. Fluorescently labeled secondary antibodies were diluted 1:lOO relative to their concentrations a t purchase for use in immunostaining. Dilutions of all antibodies were made with PBS. Slides were incubated at 20°C for 15 min in PBS containing 0.2% bovine serum albumin (BSA) to block nonspecific binding. The slides were incubated with primary antibody in humid chambers at 20°C overnight, washed with PBS containing 0.2% BSA, and incubated for 1 h r at 20°C with secondary antibodies. Finally, slides were washed in PBS and mounted with Crystal/Mount (Biomeda, Foster City, CA). Immunostained slides were examined using a Leitz (Rockleigh, NJ) Ortholux I1 or Nikon (Melville, NY) Optiphot microscope equipped for epifluorescence and were photographed using Kodak TRIX Pan film. Digestions Several methods were used to test masking as a cause for absence or low intensity of immunoreactivity. Before antibody staining, sections were digested with 0.1% bovine testicular hyaluronidase (Sigma) in 0.1 M acetate buffer, pH 5, at 37°C for 1hr, to remove hyaluronate and chondroitin sulfates (Meier and Hay, 1973). To evaluate whether collagen IV immunoreactivity could be enhanced by degradation of heparan sulfates, Borenfreund's nitrous acid treatment (Cifonelli, 1968) was used, as described by Kosher and Searls (1973). 446 S.R. HILFER AND G.J. RANDOLPH Fig. 5. Fluorescence micrographs of frozen sections of chick embryos embedded in acrylamide, treated with anticollagen type IV antibody and FITC secondary antibody after mild proteolysis. Bar = 100 pm. Collagen type IV fluorescence forms a continuous layer at the basal surfaces of the otic primordium and neural tube a t stage 11 (a), stage 12 + (b),and stage 14 (cf. Fig. 6. Phase-contrast micrographs of the same sections and at the same magnifications as in Figure 5.Bar = 100 pm. a: Section through an otic primordium at stage 11. The placode has started to fold parallel to the neural tube. b: At stage 12+, the initial fold is well established, and additional folds are forming dorsal (DF) and ventral (VF) to it. c: By stage 14,continued ventral folding (VF) has shaped Mild proteolytic digestion just prior to antibody processing (Finley and Petrusz, 1985) was tested as a means of amplifying the intensity of antibody immunoreactivity. Slides pretreated with PBS for 15 min a t 4°C were incubated with 0.025% trypsin (Worthington, Freehold, NJ) for 1 min at 4°C and subsequently washed with 0.025% soybean trypsin inhibitor (Worthington). Control slides were treated with a 1:1 mixture of trypsin and soybean trypsin inhibitor. the primordium into a deep pit, and the dorsal fold (DF) is closing the opening. Fig. 7.Higher magnifications of the distribution of collagen IV in otic primordia. Bars = 10 pm. a: The upper region of the section shown in Figure 5a. At stage 11, the apposition point near the dorsal margin of the otic primordium is highly immunofluorescent. b: The ventral margin of the apposition between otic primordium and neural tube in the same section. Fibrils containing collagen IV span the gap between the two basal laminae. c: At stage 14,where mesenchyme separates the two epithelial layers ventrally, collagen is present in only a few, fine fibrils (arrow), which span the gap. Control Studies Nonspecific immunoreactivity of the secondary antibody was assessed by substituting PBS plus BSA, hybridoma cell media, nonimmune serum, or acetate buffer (for the testicular hyaluronidase digestions) for the primary antibody in the staining procedure. No immunoreactivity was detected. Specificity of the antimouse laminin antibody was surveyed by incubating the antilaminin antibody with laminin-impregnated OTIC AND OPTIC BASAL LAMINA nitrocellulose paper (purified laminin was a gift from Dr. J.M. Chen) or with 5% condensed milk. Solutions of antilaminin antibody preabsorbed with laminin did not react with the sections, whereas preabsorption with milk had no effect on immunoreactivity. To ascertain the relative differences in immunoreactivity between paraformaldehyde-fixed sections and sections fixed under milder conditions, several embryos were placed in OCT directly after removal from the yolk and immediately frozen in liquid nitrogen-cooled isopentane. After sectioning, the slide were fixed in acetone a t -20°C for 2-10 min and allowed to air dry in a cold chamber before reacting with antibodies. RESULTS General Considerations In this study, frozen sections of chick optic and otic primordia between stages 10 and 15 were treated with antibodies raised to fibronectin, laminin, and type IV collagen, Nine embryos were examined for each stage and antibody type. The photomicrographs shown in the figures have been selected as typical of the composite observations. When more than one source of antibody against a single basal lamina component was used (see Materials and Methods), the same results were obtained with each source. Frozen sections of embryos fixed in 3.0% paraformaldehyde for 20-25 min resulted in immunoreactivity of the same intensity as sections of embryos fixed with acetone when the antilaminin and antifibronectin antibodies were used. Sections of embryos fixed in 3.0% paraformaldehyde and reacted with the collagen type IV antibody resulted in immunofluorescence of very low intensity in basal laminae. Digestions of sections with bovine testicular hyaluronidase or nitrous acid did not enhance reactivity to anticollagen type IV. However, a mild trypsin digestion, performed on sections prior to antibody application, consistently uncovered bright collagen IV immunoreactivity without degradation of the tissue, when the embryos were fixed with paraformaldehyde for less than 30 min. Trypsin digestions partially decreased laminin immunof luorescence and resulted in complete loss of fibronectin localization. Frozen sections of unfixed embryos, which were subsequently fixed with acetone, were immunofluorescent for collagen type IV without resorting to trypsin digestions or other manipulations. Acetone fixation could not be used routinely in this study because of poor morphological preservation of the tissue. Ofic Primordium From stage 10 to stage 12, the otic epithelium becomes thicker and forms a fold parallel to the longitudinal axis of the embryo (Figs. la,b, 2a,b). By stage 14, the otic primordium forms a pit (Figs. l c , 2c), followed by narrowing of the opening a t later stages (Figs. Id, 2d). Laminin immunofluorescence (Fig. 1) is bright and spatially relatively uniform from stage 10 through stage 15. From stage 10 to stage 13, laminin localizes to distinct parallel lines along the length of the otic primordium and appears slightly brighter adjacent to the neural tube than adjacent to the underlying somitomere (fig. la,b). The basal surface of the neural tube is uniformly immunoreactive but also appears brighter 447 adjacent to the otic primordium than in the ventral region. At stage 14, the otic epithelium begins to separate from the neural tube (Fig. 2c), and the brighter immunofluorescence is lost in the region of separation (Fig. lc). By stage 15, the two epithelia are separated by mesenchymal cells (Fig. 2d). Laminin localization is uniformly along the entire basal surface of the otic pit and is brighter than along the neural tube (Fig. Id). The pattern of fibronectin localization is similar to that of laminin from stage 10 to stage 14 (Fig. 3a-c). At earlier stages, fibronectin forms a solid band between the basal surfaces of the otic and neural epithelia, but two separate layers are visible at stage 14 (Fig. 3 4 . The intensity of fibronectin reactivity along the otic pit and neural tube basal laminae is sharply reduced by late stage 15 (Fig. 3d). The same sections display bright fibronectin reactivity in the pharyngeal region, indicating that poor antibody coverage of the sections cannot be the cause of the observed reduction in intensity. Antifibronectin reactivity at stage 15 is not enhanced by treatment of sections with bovine testicular hyaluronidase, which digests hyaluronate and chondroitin sulfate (Meier and Hay, 1973). These proteoglycan components presumably could deter antibody accessibility to fibronectin due to their intimate interaction with fibronectin in vivo (Jilek and Hormann, 1979; Perkins et al., 1979). Comparison of laminin and fibronectin immunolocalization at higher magnification shows a difference in their distributions. From stage 10 to stage 13, when the medial otic placode is in close contact with the neural tube, laminin (Fig. 4a) appears as a distinct line at each epithelial basal surface, with fine cross connections between the two layers. By stage 14, the two layers of laminin show no cross connections (Fig. 2b). Fibronectin fills the gap between the laminin layers, and cross connections are not visible until the otic and neural layers separate at stage 14 (Fig. 412).At that time, cross connections that contain fibronectin become visible at the edge of the fused region, where the otic epithelium is deflected over somitomeric mesoderm. Thread-like connections no longer are seen where mesenchymal cells separate the two epithelial layers. Instead, fibronectin forms a fine meshwork surrounding the invading cells (Fig. 4d). From establishment of the placode to formation of the otic pit (Figs. 5, 6 ) , immunolocalization of collagen IV follows the pattern of laminin. Anticollagen antibody binds with uniform intensity along the entire basal length of the otic epithelium and neural tube from before stage 10 (Fig. 5a) to after stage 14 (Fig. 5c). Collagen IV immunofluorescence (Fig. 7a) merges where the two basal laminae are in close contact dorsally and forms cross connections ventrally (Fig. 7b). Upon mesenchymal invasion, little immunoreactivity is found in the space surrounding the mesenchymal cells (Fig. 7c). As described above, inconsistent localization occurs with anticollagen antibodies except after mild trypsinization. This procedure reduces laminin and eliminates fibronectin localization. Digestion of chondroitin sulfate or heparan sulfate does not improve collagen immunoreactivity, suggesting that fibronectin rather than proteoglycans occupies the major binding sites on collagen IV in this primordium. This is in contrast to cartilage formation, where proteoglycan 448 S.R. HILFER AND G.J. RANDOLPH Figs. 8-12. 449 OTIC AND OPTIC BASAL LAMINA masking of types I and I1 collagen has been demonstrated (von der Mark et al., 1976). Optic Prirnordiurn The optic primordium was investigated from the period prior to formation of the lens placode at stage 13 to the establishment of a deep optic cup at stage 15 (Figs. 9, 11, 14). Laminin is distributed over the entire basal surface of the optic primordium and adjacent ectoderm (Fig. 8). However, at all stages, the lateral one-half of Fig. 8. Immunofluorescence of frozen sections through the optic region of chick embryos, treated with antilaminin followed by FITClabeled secondary antibody. Bar = 100 pm. a: At stage 13,the basal laminae of presumptive lens ectoderm and lateral optic vesicle (OV) fluoresce more intensely than the medial optic surface and neural tube (NT). Little if any laminin is visible between the two layers. b At stage 14, a similar distribution is apparent even in this thick section. c: At stage 15,laminin is distributed along the basal surfaces of both sides of the optic cup and around the lens vesicle. Little if any immunoreactivity is detectable between the basal laminae of retinal disc and lens, but these two basal laminae fluoresce more intensely than those of the future retinal pigment epithelium, and outer lens surface, and the future cornea. Fig. 9. Phase-contrast micrographs of the same sections as in Figure 8. Bar = 100 pm. a: At stage 13, the pseudostratified cells of the lateral (RD) and medial (P) walls of the optic vesicle are the same height. The overlying ectoderm (E) has not formed a lens placode. b: In this thick section of a stage 14 primordium, the invaginated retinal disc (RD) can be recognized as thicker than the medial future retinal pigment epithelium (PI. The adjacent ectoderm has formed a thickened lens pit (L). c: At stage 15, the retinal disc is folded inward against the medial wall, and the lens has formed a vesicle (LV) and separated from the retinal and ectodermal surfaces. The surface ectoderm is the future cornea (CO). Fig. 10. Immunofluorescence of optic primordia after treatment with antifibronectin and FITC-labeled secondary antibody. Bar = 100 pm. a: At stage 13 + , fibronectin appears as separate, continuous layers along the basal surfaces of the retinal disc and lens placode. A small amount of immunoreactivity in the form of particles or fine filaments is between the layers. The basal lamina of the future retinal pigment epithelium is less intensely fluorescent. The mesenchyme also has slight fluorescence. b At stage 14,the same pattern is seen with additional fluorescence between mesenchyme cells. c: At stage 15, the lens surface is less immunoreactive than the point of retinal disciectodermal apposition. The central region of the retinal disc, where it separates from the lens, shows weak fluorescence (arrowhead). The future retinal pigment epithelium shows uneven reactivity. Fig. 11. Phase-contrast micrographs of the sections in Figure 10. Bar = 100 pm. a: At stage 13 + , a lens placode (LP) has formed, and the retinal disc is recognizable by its elongated cells. b: This section is from the edge of the optic primordium of a stage 14 embryo. The shallow optic cup and adjacent lens are visible. c: This section, from a slightly younger stage 15 embryo than that shown in Figure 8c, was covered with some displaced acrylamide. The lens and retinal disc are closely apposed except at their centers. Fig. 12. Higher magnifications of selected regions of optic primordia. Bar = 10 pm. a: The ventral region of a stage 13 + optic primordium. Laminin is distributed along the retinal disc (RD) and ectoderma1 surfaces but not between. b The dorsal fusion point of the same optic primordium. Laminin is present as two unconnected layers. RD, retinal disc. c: Enlargement of the ventral region of the retinal disc (RD) and lens placode shown in Figure 10a. The ventral fusion is shown (arrowhead). At the center of the retinal disc and lens, fibronectin is present as particles between the separated basal laminae as well as within these layers. d Enlargement of the dorsal junction of retinal disc, lens vesicle, and cornea shown in Figure 1Oc. A sheet of fibronectin-containing material connects the lens vesicle and margin of the retinal disc (arrow). A single, highly fluorescent band of fibronectin lies between the margin of the cup and the apposed ectoderm. the primordium (the future neural retina or retinal disc) stains more intensely than the medial surface (the future retinal pigment epithelium). The basal laminae of the retinal disc and lens placode become fused a t the margins just prior to the initiation of invagination (Johnston et al., 1979). Laminin occurs a t two distinct lines through this peripheral band, with no indication of joining (Fig. 12a,b). This separation differs from the otic primordium, in which fine strands traversed the laminin layers of otic and neural basal laminae (Fig. 1). Immunolocalization of fibronectin also shows greater intensity in basal laminae of the retinal disc and lens placode than of the future retinal pigment epithelium (Fig. 10). As early as stage 13+ , fibronectin forms a single fluorescent line between the retinal disc and the dorsal and ventral margins of the newly formed lens placode (Figs. 10a, 12c).By stage 15, a subtle reduction occurs in fibronectin intensity, with only the margin of the optic cup remaining as brightly reactive as the lateral basal laminae a t earlier stages (Figs. lOc, 12d). This position represents the point of fusion of the ectodermal and neuroectodermal basal laminae. A thin and weakly fluorescent connection exists between the region of fused basal laminae, the surface ectoderm, and the lens vesicle (Figs. lOc, 12d). Fibronectin immunofluorescence is weakest where the lens is separated from the center of the retinal disc (Fig. 1Oc). Crossconnecting fibers of laminin or fibronectin are not seen at the center of the retinal disc in the thinnest sections in which ectodermal and retinal disc basal laminae can be resolved as separate layers (Figs. 8,10,12a-c);however, some particles of fibronectin localize between the two layers. This distribution contrasts with the localization of these constituents in the otic primordium. The distribution of collagen type IV is similar to that of laminin and fibronectin only in that the basal lamina of the future retinal pigment epithelium is less reactive than that of the retinal disc and lens primordium (Fig. 13a,b). These differences in intensity remain until late in stage 14, when collagen type IV immunofluorescence of the central retinal disc is reduced to that of the future retinal pigment epithelium, while the lens vesicle continues to react more intensely (Fig. 14c). Collagen IV localizes as a single band in the fused basal laminae at the optic/lens primordia margins from stage 13+ to stage 15, while collagen type IV immunofluorescence forms two parallel bands between the rest of the retinal disc and lens primordium (Fig. 15a). Fibrillar connections that contain type IV collagen become visible between the retinal disc and lens beginning with formation of the lens placode and increase in number during invagination. By stage 15, an extensive fibrillar meshwork of collagen type IV is formed between the optic cup and the lens pit in the central region (Figs. 14c, 15b). DISCUSSION This study shows that laminin, fibronectin, and type IV collagen exist as continuous sheets a t the basal surfaces of the avian otic primordium when it first forms. All three extracellular macromolecules also are present along basal surfaces of the optic primordium and lens epithelium even before the initiation of invagination. The data extend an earlier study on chick 450 S.R. HILFER AND G.J. RANDOLPH Fig. 13. Immunofluorescence of optic primordia after mild proteolysis and treatment with anticollagen IV and FITC-labeled secondary antibody. Bar = 100 pm. At all stages, the ectodermalioptic interface fluoresces more intensely than other parts of the preparation. a: At stage 12, type IV collagen localizes as separate layers a t the basal surfaces of both the ectoderm and optic vesicle. The ventral region of apposed layers reacts more intensely than the rest of the primordium. b: In contrast, a single intense band of collagen fluorescence is seen between the retinal disc and lens placode at stage 13 + . c: At stage 14 + , the margins of the optic cup and the gap between lens and retinal disc are intensely immunoreactive. otic development (Richardson et al., 1987) and show that the distribution of these basal laminar components in the chick optic primordium is similar to that of the mouse (Svoboda and O’Shea, 1987; Haloui et al., 1988).This distribution of the three components is consistent with the observation that all these embryonic structures have complete basal laminae. The distinctive patterns of invagination of the otic placode and optic vesicle suggest that points of folding may be established by differential expression of extracellular macromolecules. These differences could be qualitative or quantitative. Qualitatively, a hinge point could be determined either by availability of a unique adhesive component or by removal of a component that prevents adhesion. Alternatively, folding could occur either as a result of increased adhesiveness at a hinge point or as decreased adhesiveness along a fold. Qualitative differences were not found in the distribution of laminin, fibronectin, and type IV collagen but may exist in other extracellular components that have not been examined so far. A more subtle qualitative difference could occur as in the control of laminin A chain expression in the developing kidney (Klein et al., 1988; Ekblom et al., 1990). It is possible that similar control mechanisms operate in other organ primordia, including the eye and inner ear. Several types of quantitative differences do exist in the distributions of these three extracellular components in otic and optic primordia. The lateral surface of Fig. 14. Phase-contrast micrographs of the sections shown in Figure 13. Bar = 100 pm. a: Optic vesicle at stage 12, cut through the optic stalk (0s). b: Horizontal section of a stage 13 optic vesicle. c: Nearmidline section of a stage 14 + optic cup and lens pit. + Fig. 15. High magnifications of collagen IV fluorescence. Bars = 10 pm. a: At stage 12 immunoreactivity is more intense a t the ectoderma1 surface than at the optic vesicleioptic stalk surface, but the basal laminae form distinct layers. Some collagen type IV-containing material lies between the two basal laminae. b: At stage 14 +, the space between the central optic cup (RD)and lens is filled with a meshwork of collagen type IV-containing fibrils and sheets. the optic primordium is more immunoreactive for laminin, fibronectin, and collagen type IV than the medial surface, and the otic primordium-neural tube interface is somewhat more immunoreactive than otic epithelium adjacent to somitomeric mesoderm. These differences as well as decreased fibronectin immunoreactivity after completion of invagination are difficult to relate to the process of invagination. The observed points of fusion and cross connections between the basal laminae of each primordium and adjacent cell layers, rather than differences in concentration, are more likely to play a role in invagination. The otic epithelium is closely apposed to the adjacent neural tube, based on immunolocalization of both fibronectin and collagen in this region prior to invasion of mesenchyme. Similarly, basal laminae are fused at the margins of the retinal disc and lens placode, based on immunolocalization of collagen type 1V and fibronectin in this study and earlier silver staining studies (Johnston et al., 1979). These points could act to stabilize cell layers and prevent them from bending while active or passive movements in other parts of the primordia cause the invagination. There are several interesting differences as well as similarities in the distribution of these three macromolecules in the otic and optic primordia. In the chick, both primordia form so that neural crest cells do not penetrate between the ectoderm and neural ectoderm (Anderson and Meier, 1981). In both, fibronectin is a 451 OTIC AND OPTIC BASAL LAMINA significant component of the gap between the cell layers even in the absence of mesenchymal cells. Both primordia make contacts with adjacent epithelia by fusion of basal laminae. In both, fibronectin and type IV collagen localize to the fused region, but laminin forms distinct layers a t each epithelial surface. However, the shapes of the fused zones differ in the two primordia; the entire medial region of the otic placode is bound to the neural tube, while the retinal disc and lens placode are attached only a t their margins. Cross connections external to the fused regions also differ. The otic placode has only a few cross connections with the neural tube at the margin of the fused region, whereas the retinal disc and lens are connected by a n extensive series of fibrils within the confines of the marginal attachment zone. Otic cross connections consist of all three components, whereas cross connections between lens and retinal disc consist primarily of type IV collagen. These differences in organization of the basal laminae in the two primordia may reflect differences in the methods by which folding occurs. Preliminary studies support the conclusion that glycoproteins play a role in invagination of both primordia. Treatment of otic (Rausch and Hilfer, 1988; Hilfer et al., unpublished) and optic (Yang and Hilfer, 1982) primordia with tunicamycin inhibits invagination. Injection of antilaminin antibody has a similar effect (Visconti et al., unpublished). However, the mechanisms that control folding appear to be entirely different. Otic invagination is not affected by agonists or antagonists of Ca2+ transport (Hilfer et al., 1989), but inhibition of proteoglycan synthesis or injection of enzymes that degrade glycosaminoglycans prevents or delays folding (Gerchman et al., 1991, unpublished). Optic cup formation, in contrast is dependent on adenosine triphosphate (ATP), Ca2' , and calmodulin (Brady and Hilfer, 1982). In both, however, anchorages to adjacent cell layers appear to play important roles in folding. The data from this study are consistent with the following models for formation of the otic pit and optic cup. Formation of the otic vesicle begins when the medial region of the otic placode becomes held to the adjacent neural tube by fusion of the outer surfaces of their apposed basal laminae. Increase in volume of the somitomeric mesoderm ventral to the lateral region of the otic placode causes the primordium to fold along a hinge point adjacent to the ventral margin of the neural tube. Ventral migrations of neural crest cranial and caudal to the placode cause additional folding (Meier, 1978), which results in the box-like shape of the otic pit (Hilfer et al., 1989). Completion of folding to produce a n otic vesicle results from detachment of the otic epithelium from the neural tube by invasion of mesenchymal cells. In contrast, formation of the optic cup begins with fusion of the basal lamina of the optic vesicle a t the margin of the retinal disc with the basal lamina at the margin of the newly formed lens placode. This ring acts to prevent expansion in diameter of the optic primordium during invagination. Folding is initiated by a combination of constriction of cell apices at the region of fused basal laminae and expansion of cell apices within the central region of the retinal disc (Hilfer and Hilfer, 1983). Formation of a lens vesicle and vitreous cavity depends on release of the central region of the retinal disc from the lens epithelium and possibly a shift in position of the fused basal laminae to the new margin of the expanding optic cup. These two models raise a number of questions about the ways in which morphogenetically active epithelial layers produce and interact with their surface coats and extracellular matrices. Answers to these questions are necessary for a n understanding of the way in which organogenesis is controlled. ACKNOWLEDGMENTS This study was supported by NSF grant DCB8812287 and a grant from the Deafness Research Foundation. The technical assistance of Joyce W. Brown is greatly appreciated. We especially thank Drs. Edward Gruberg, Joel Sheffield, and Jinq-mai Chen for generously allowing the use of their equipment and for their excellent technical advice. LITERATURE CITED Anderson, C.B., and S. Meier 1981 The influence of the metameric pattern in the mesoderm on migration of cranial neural crest cells in the chick embryo. Dev. Biol., 85t385-402. Bayne, E.K., M.J. Anderson, and D.M. Fambrough 1984 Extracellular matrix organization in developing muscle: Correlation with acetylcholine receptor aggregates. J . Cell Biol., 99t1486-1501. Brady, R.C. and S.R. Hilfer 1982 Optic cup formation-A calcium regulated process. Proc. Natl. Acad. Sci. USA, 79:5587-5591. Cifonelli, J.A. 1968 Reaction of heparatin sulfate with nitrous acid. Carbohydrate Res., 8:233-242. Ekblom, M., G. Klein, G. Mugrauer, L. Fecker, R. Deutzmann, R. Timpl, and P. Ekblom 1990 Transient and locally restricted expression of laminin A chain mRNA by developing epithelial cells during kidney organogenesis. Cell, 60:337-346. Finley, J.C.W., and P. Petrusz 1985 The use of proteolytic enzymes for improved localization of tissue antigens with immunocytochemistry. In Techniques in Immunocytochemistry. G.R. Bullock and P. Petrusz, eds. Academic, London, Vol. 1, pp. 239-250. Gardner, J.M. and D.M. Fambrough 1983 Fibronectin expression during myogenesis. J . Cell Biol., 96t474-485. Gerchman, E.M., S.R. Hilfer, and D.M. Simmons 1991 Differential involvement of glycosaminoglycansin organogenesis of the avian inner ear and eye. J . Cell Biol., 114:293a (abstract). Ginsberg, M.H., J.C. Loftis, S. D'Souza, and E.F. Plow 1990 Ligand binding to integrins: Common and ligand specific recognition mechanisms. Cell Differ. Dev., 32t203-214. Haloui, Z., J.C. Jeanny, L. Jonet, Y. Courtois, and M. Laurent 1988 Immunochemical analysis of extracellular matrix during embryonic lens development in the cat fraser mouse. Exp. Eye Res., 46t463-474. Hamburger, V. and H. Hamilton 1951 A series of normal stages in the development of the chick embryo. J . Morphol., 88t49-92. Hayashi, K., M. Hayashi, M. Jalkanen, R.L. Firestone, R. Trelstad, and M. Bernfield 1987 Immunocytochemistry of cell surface heparan sulfate proteoglycan. A light and electron microscopic study. J . Histochem. Cytochem., 35t1079-1088. Hilfer, S.R., R.C. Brady, and J.j.W. Yang 1981 Intracellular and extracellular changes during early ocular development in the chick embryo. In Ocular Size and Shape: Regulation during Development. S.R. Hilfer and J.B. Sheflield, eds. Springer-Verlag, New York, pp. 47-78. Hilfer, S.R., R.A. Esteves, and J.F. Sanzo 1989 Invagination of the otic placode: Normal development and experimental manipulations. J . Exp. Zool., 251:253-264. Hilfer, S.R. and E.S. Hilfer 1983 Computer simulation of organogenesis: an approach to the analysis of shape changes in epithelial organs. Dev. Biol., 97t444-453. Hilfer, S.R. and R.L. Searls 1986 Cytoskeletal dynamics in animal morphogenesis. In Developmental Biology, A Comprehensive Synthesis. L. Browder, ed. Plenum, New York, pp. 3-26. Jilek, F. and H. Hormann 1979 Fibronectin (cold insoluble globulin) VI. Influence of heparin and hyaluronate acid on the binding of native collagen. Hoppe-Seylers Arch. Physiol. Chem., 360t597603. Johnson, L.V. and J.C. Blanks 1984 Application of acrylamide as an 452 S.R. HILFER AND G.J. RANDOLPH embedding medium in studies of lectin and antibody binding in the vertebrate retina. Curr. Eye Res., 3:969-974. Johnston, M.C., D.M. Noden, R.D. Hazelton, J.L. Coulombre, and A.J. Coulombre 1979 Origins of avian ocular and periocular tissues. Exp. Eye. Res., 29:27-43. Klein, D.J., D.M. Brown, A. Moran, T.R. Oegema, Jr., and J.L. Platt 1989 Chondroitin sulfate proteoglycan synthesis and reutilization of P-D-xyloside-initiated chondroitinidermatan sulfate glycosaminoglycans in fetal kidney branching morphogenesis. Dev. Biol., 133515-528. Klein, G., M. Langegger, R. Timpl, and P. Ekblom 1988 Role of laminin A chain in the development of epithelial cell polarity. Cell, 55:331-341. Kosher, R.A. and R.L. Searls 1973 Sulfated mucopolysaccharidesduring the development of Rana pzpzens. Dev. Biol., 3 2 . 5 0 4 8 . Mahoney, W.C. and D. Duskin 1979 Biological activities of the two major components of tunicamycin. J . Biol. Chem., 254:65726576. Maloney, C. and J . Wakely 1982 Analysis of tissue interactions in chick eye morphogenesis using cytochalasin B. Exp. Eye Res., 35:77-87. Meier, S. 1978 Development of the embryonic chick otic placode. I. Light microscopic analysis. Anat. Rec., 191 :447-458. Meier, S. and E.D. Hay 1973 Synthesis of sulfated glycosaminoglycans by embryonic corneal epithelium. Dev. Biol., 35:318-331. Nakanishi, Y., F. Sugiura, J . Kishi, and T. Hayakawa 1986 Local effects of implanted Elvax chips containing collagenase inhibitor and bacterial collagenase on branching morphogenesis of mouse embryonic submandibular glands in vitro. Zool. Sci., 3:479-486. Perkins, M.E., T.H. Ji, and R.O. Hynes 1979 Cross-linking of fibro- nectin to sulphated proteoglycans at the cell surface. Cell, 16; 941-952. Rausch, D.A. and S.R. Hilfer 1988 Putative role for extracellular matrix in otic vesicle formation. J . Cell Biol., 107:795a. Richardson, G.P., C.L. Crossin, C-M. Chuong, and G.M. Edelman 1987 Expression of cell adhesion molecules during embryonic induction 111. Development of the otic placode. Dev. Biol., 119:217-230. Spooner, B.S., K. Bassett, and B. Stokes 1985 Sulfated glycosaminoglycan deposition and processing a t the basal epithelial surface in branching and in P-D-xyloside-inhibited embryonic salivary glands. Dev. Biol., 109t177-183. Spooner, B.S. and J . Faubion 1980 Collagen involvement in branching morphogenesis of embryonic lung and salivary gland. Dev. Biol., 77:84-102. Svoboda, K.K.H. and K.S. Shea 1987 An analysis of cell shape and the neuroepithelial basal lamina during optic vesicle formation in the mouse embryo. Development, 100:185-200. Thompson, H.A. and B.S. Spooner 1983 Proteoglycan and glycosaminoglycan synthesis in embryonic mouse salivary glands: Effects of P-D-xyloside, a n inhibitor of branching morphogenesis. J. Cell Biol., 96t1443-1450. Trinkaus, J.P. 1984 Cells Into Organs, the Forces That Shape the Embryo, 2nd ed. Prentice-Hall, Englewood Cliffs, NJ. von der Mark, K., H. von der Mark, and S. Gay 1976 Study of differential collagen synthesis during development of the chick embryo by immunofluorescence. 11. Localization of type I and type I1 collagen during long bone development. Dev. Biol., 53:153-170. Yang, J-W. and S.R. Hilfer 1982 Interference with ocular development by inhibitors of glycoconjugate synthesis. Dev. Biol., 92: 41-53.