Immunohistochemical analysis of proteoglycan biosynthesis during early development of the chicken cornea.

Antibodies to core proteins of chicken corneal keratan sulfate proteoglycan and chondroitin sulfate proteoglycan were prepared and purified by use of an affinity column. Using these antibodies and monoclonal antibody 5-D-4 to keratan sulfate (commercial), the localization of proteoglycans in developing corneas (Days 5 to 17 of embryonic age and 2 days after hatching) was determined immunohistochemically. Keratan sulfate proteoglycan antigen was not detected in cornea on Day 5, but it was detected uniformly over the whole stroma on Day 6, ca. 12 h after invasion of the primary stroma by mesenchymal cells. The absence of the antigen in cornea of Day 5 was confirmed by Western blotting of the corneal extract. Immunohistochemistry with 5-D-4 antibody revealed that the keratan sulfate chain was undersulfated in corneas of Days 6 to 7, because the staining was much weaker than that in cornea of Day 8. In addition, keratan sulfate proteoglycan antigen was detected uniformly over the whole stroma on Days 7 to 17 and 2 days after hatching, but not in the epithelial layer on Day 13 and after: because the epithelial layer was clearly not observed on photomicrographs until Day 13, it is not known whether keratan sulfate proteoglycan was synthesized by the epithelium during Days 6 to 12. In contrast, chondroitin sulfate proteoglycan antigen was detected in cornea on Day 5 and also, like keratan sulfate proteoglycan, uniformly over the whole stroma on Day 6 through 2 days after hatching. Furthermore, the chondroitin sulfate proteoglycan was not detected in the epithelial layer on Day 13 and after. These results show that keratan sulfate proteoglycan is synthesized by the stromal cells which invade the primary stroma between Day 5.5 and 6, while chondroitin sulfate proteoglycan is synthesized by epithelial and/or endothelial cells before the invasion, and also by the stromal cells after the invasion.     

Keratan sulfate proteoglycan during embryonic development of the chicken cornea

Antibodies to corneal keratan sulfate proteoglycan (KSPG) were used to characterize the pattern of KSPG accumulation during differentiation of neural crest cells in the stroma of embryonic chick cornea. Immunohistochemistry with monoclonal antibody I22 to keratan sulfate found this KSPG antigen localized inside stromal cells at stage 29 (Day 6), ca. 12 hr after migration into the primary stroma. A 2- to 3-day lag then occurred before appearance of extracellular keratan sulfate, first seen on Day 9 (Stage 35) in the posterior stroma. Keratan sulfate antigen accumulated in a posterior to anterior direction during subsequent development. Uniform staining of the stroma for keratan sulfate did not occur until after Day 16. Among several tissues, only corneal stroma contained an extracellular matrix which stained for keratan sulfate, though intracellular staining of some cartilage cells was observed. Accumulation of KSPG antigens in developing cornea was measured in unfractionated guanidine extracts with a quantitative ELISA using three different antibodies against KSPG. Increases were first detected after Day 9 using monoclonal I22, and somewhat later with the other two antibodies. Assays with all three antibodies detected a sustained, exponential increase of KSPG throughout the 5 days prior to hatching. Keratan sulfate continued to accumulate after hatching, but an antibody with specificity to KSPG core protein, detected no relative increase in antigen after hatching. This suggests a modulation of KSPG primary structure late in development and after hatching. Overt differentiation of individual neural crest cells thus appears to begin ca. 12 hr after their arrival in the primary stroma; a lag of 2-3 days precedes active secretion of KSPG.     

Analysis of corneal development with monoclonal antibodies. II. Tissue autonomy in cornea-skin recombinants

Developmental autonomy of corneal epithelial and stromal components was assessed by their subsequent differentiation after recombination with feather-forming thigh dermis and epidermis, respectively. Work by others has shown that feather-forming dermis exhibits strong inductive ability when used in such epithelial-mesenchymal recombinations. After culture of the recombinants on the chorioallantoic membrane (CAM) of host embryos, differentiation as "cornea" was assessed immunohistochemically using the anti-corneal stromal matrix and anti-corneal epithelial antibodies described previously (Zak and Linsenmayer, Dev. Biol. 99, 373-381, 1983). Feather initiation and outgrowth and keratin synthesis served as markers for differentiation as skin. It has been found that corneal epithelia from 5-day embryos, when grown in association with feather-forming dermis from the thigh, will participate in feather formation. In such recombinants, when the corneal epithelium became incorporated into feathers it failed to express the corneal epithelial antigen, but in regions of the recombinant where feathers did not form, de novo expression of the antigen was sometimes detected. The limited liability of the epithelium is not present in corneal epithelia taken from embryos a day or two older. When such epithelia were used for making the recombinants, no feathers were formed and the corneal epithelial antigen was extensively produced. Thus epithelial determination occurs long before the epithelium would begin to overtly differentiate and express the epithelial antigen in vivo (about 12 days of development). In reciprocal recombinations of corneal stromas with feather-forming epidermis, the stromas proceeded to express the corneal stromal matrix specific antigen de novo after culture on the CAM. They did not, however, redirect differentiation of the epidermis which never expressed the corneal epithelial antigen and in some cases went on to keratinize. These results indicate that development of both the corneal epithelial and stromal components becomes autonomous at least several days before these tissues overtly differentiate. This suggests that the component tissues of the cornea may not interact in a manner typical of those of other organs which, in general, are thought to require continual interaction of their epithelial and mesenchymal components for normal development.     

Monoclonal antibody analysis of ocular basement membranes during development

As a first step in a study of the role(s) of basement membranes in ocular morphogenesis, we have produced a variety of monoclonal antibodies against native lens capsule from adult chicks, and have used these reagents to stain histological sections of ocular tissues from 4 1/2- to 18-day-old chicken embryos. Four different patterns of immunofluorescence were observed in sections of corneas of 18-day-old chicken embryos stained with these antibodies. The antibodies in group 1 stained the basement membranes of both the corneal epithelium and the endothelium (as well as Descemet's membrane). Those in groups 2 and 3 stained only the epithelial or endothelial basement membranes, respectively. The group 4 antibody stained the corneal stroma as well as Bowman's membrane and Descemet's membrane. The antibodies in group 1 could be further subdivided into groups 1a and 1b on the basis of temporal differences in the onset of staining in corneas from 4 1/2- to 7-day-old embryos. Thus, this series of monoclonal antibodies appears to recognize at least five different antigenic determinants. When these antibodies were used to stain sections of eyes at different stages of development, we found that the characteristic differential staining of some basement membranes was maintained throughout development, while the staining properties of others changed. This indicates that many of the ocular basement membranes may differ from one another in composition or conformation, and that at least some of them may undergo developmental changes. We also noticed a similarity in the pattern of fluorescence associated with the basement membranes of the limbal blood vessels and the corneal endothelium that is consistent with the hypothesis that the corneal endothelium is derived from the early periocular vascular endothelium. Our observations of developing corneas also revealed that the antigen recognized by the group 4 antibody may be produced by both the corneal epithelium and the stromal fibroblasts. The suitability of monoclonal antibodies for probing basement membrane heterogeneity is discussed.     

Cell-surface associated proteins of corneal fibroblasts: dissection with monoclonal antibodies

It is now generally accepted that the cell surface is involved in the interaction of the cells with the extracellular matrix. To identify and characterize cell-surface-associated components of corneal fibroblasts, several monoclonal antibodies were developed. Hybridomas were developed by fusing mouse myeloma cells SP2/OAg14 with spleen cells from mice immunized with membrane fractions of corneal fibroblasts grown in culture. Twenty-five hybridomas secreting monoclonal antibodies to cell-surface components were selected by an enzyme-linked immunosorbent assay using corneal fibroblasts grown in microtiter plates as the substrate. Immunohistochemical staining demonstrated that the antigenic determinants recognized by these antibodies were not present on corneal epithelial cells, but were present on skin fibroblasts. The antigenic determinants recognized by two of these antibodies, designated 10D2 and 716, were matrix components of the corneal stroma. Immunochemical characterization of the antigens was carried out by indirect precipitation of the radioactively labeled cellular proteins with the monoclonal antibodies and sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) analysis of the precipitates. Four antibodies were able to precipitate antigens from cell extract in detectable amounts. Antibodies designated 5E2, 9G2, and 10D2 recognized antigens consisting of polypeptides of approximate molecular weights 105K and 110K, while antibody 716 recognized an antigen of 100K molecular weight. However, based on the tissue distribution and cell-surface distribution, these antibodies reacted with different antigenic determinants. The antigen recognized by 716 was also secreted by cells in culture but consisted of 220K and 200K polypeptide chains. It was tentatively identified as cellular fibronectin, based on the reaction of this antigen with polyclonal antibodies to plasma fibronectin.     

Cell adhesion molecules in stromal corneal dystrophies

The aim of the present study was to investigate the expression pattern of different cell adhesion molecules in corneal stromal dystrophies. Fifteen corneal buttons from patients diagnosed with three different types of stromal corneal dystrophies and healthy corneas were investigated. Paraffin embedded sections were stained immunohistochemically with monoclonal antibodies against human intercellular adhesion molecule-1 (ICAM-1), endothelial selectin (E-selectin) and endothelial cadherin (E-cadherin) using the avidin-biotin-peroxidase-complex technique. The sections were compared to normal eye bank controls. In corneas from granular dystrophy patients ICAM-1 was expressed focally in epithelial cells and in keratocytes, and expressed diffusely in endothelial cells. In corneas from macular dystrophy patients diffuse epithelial staining was observed and the stromal and endothelial expression was found to be similar to that of granular dystrophy. In lattice dystrophy, only the epithelial cells and endothelium were intensively positive for ICAM-1. E-selectin was not present on any layer of the corneal specimens. E-cadherin was observed only in the epithelium of all three types of corneal dystrophies. Normal corneas did not express any of the investigated adhesion molecules. We found different expression patterns of adhesion molecules in corneas from stromal dystrophies. Our results suggest that adhesion molecules may be involved in the pathogenesis of corneal stromal dystrophies.     

Distinct roles for neuropilin1 and neuropilin2 during mouse corneal innervation

Trigeminal sensory innervation of the cornea is critical for protection and synthesis of neuropeptides required for normal vision. Little is known about axon guidance during mammalian corneal innervation. In contrast to the chick where a pericorneal nerve ring forms via Npn/Sema signaling, mouse corneal axons project directly into the presumptive cornea without initial formation of an analogous nerve ring. Here we show that during development of the mouse cornea, Npn1 is strongly expressed by the trigeminal ganglion whereas Npn2 is expressed at low levels. At the same time Sema3A and Sema3F are expressed in distinct patterns in the ocular tissues. Npn1(sema-/-) mutant corneas become precociously and aberrantly innervated by nerve bundles that project further into the corneal stroma. In contrast, stromal innervation was not affected in Npn2(-/-) mutants. The corneal epithelium was prematurely innervated in both Npn1(sema-/-) and Npn2(-/-) mutants. These defects were exacerbated in Npn1(sema-/-);Npn2(-/-) double mutants, which in addition showed ectopic innervation of the region between the optic cup and lens vesicle. Collectively, our data show that Sema3A/Npn1 and Sema3F/Npn2 signaling play distinct roles and both are required for proper innervation of the mouse cornea.     

Secretion of collagen types I and II by epithelial and endothelial cells in the developing chick cornea demonstrated by in situ hybridization and immunohistochemistry

Cells involved in the synthesis of collagen types I and II in the cornea of developing chick embryos have been studied by using in situ hybridization and immunohistochemistry. Corneas processed for in situ hybridization with the type I and II collagen probes demonstrated specific mRNAs in the epithelium of embryos at stage 18 with an increase at stages between 26 and 31, and then gradual decrease to the background level in the next several days. In the endothelium, a small amount of specific mRNA was recognized through these stages. In the stroma, only sections hybridized with the type I probe demonstrated mRNA in fibroblasts. Immunostaining demonstrated specific collagen types in the stroma at sites which were closely associated with cells containing specific mRNAs. Both collagens type I and II were present beneath the epithelium as narrow bands at stage 18; as the thicker primary stroma at stages 20 and 26; and as subepithelial, subendothelial and stromal staining at stage 31. Thereafter, type I collagen was increased in the stroma but it was also noted in the subepithelial and, to a lesser degree, subendothelial regions, whereas type II collagen was gradually confined to the subendothelial matrix. Electron microscopic examination of sections from 5-day-old (stage-27) embryo corneas using antibodies against the carboxyl propeptides of type I and II procollagens revealed the presence of these procollagens within the cisternae of the endoplasmic reticulum and Golgi vesicles in both epithelial and endothelial cells. In the epithelial cells both the periderm and basal cells contained these procollagens within the cytoplasmic organelles. These results indicate that not only the epithelial cells, but also the endothelial cells secrete collagen types I and II during the formation of the primary corneal stroma and for several days after invasion of fibroblasts.     

Involvement of C-C chemokine ligand 2-CCR2 interaction in monocyte-lineage cell recruitment of normal human corneal stroma

Bone marrow-derived cells (BMCs) reside in the anterior stroma of the central and paracentral cornea, as well as all stromal layers of the peripheral cornea, in normal human eyes. We investigated the factors regulating the constitutive distribution of BMCs in normal human corneal stroma. Cultured human corneal keratocytes expressed several chemokines (growth-related oncogene/CXCL1-3, IL-8/CXCL8, and MCP-1/CCL2) in the Ab array study. CCR2 and CCR7 mRNAs were detected in BMCs by multiplex RT-PCR. Keratocytes/corneal epithelial cells and BMCs selected from normal human donor corneas by using magnetic beads expressed MCP-1/CCL2 and CCR2 protein, respectively. BMCs isolated from human corneal stroma showed a chemotactic response to MCP-1/CCL2 in the Boyden chamber assay. The chemotactic effect of keratocyte supernatant was inhibited by blockade of MCP-1/CCL2. This is the first work on constitutive expression of CCR2 by BMCs from the corneal stroma and MCP-1/CCL2 by keratocytes/epithelial cells. Our findings suggest that the interaction between MCP-1/CCL2 and CCR2 determines the distribution of constitutive BMCs in normal human corneal stroma.     

Thrombospondin-1 accelerates wound healing of corneal epithelia

To investigate a role of thrombospondin-1 (TSP-1), a multifunctional extracellular matrix protein, in corneal epithelial wound healing, we analyzed the expression of TSP-1 in the normal and wounded mouse corneal epithelia and the effect of exogenous TSP-1 on the wound healing. In immunohistochemical analyses of unwounded corneas, TSP-1 was only detectable in endothelial cells. In contrast, TSP-1 appeared on the wounded corneal surface and on the corneal stroma, at 30 min and 8-16 h, respectively, after making an abrasion on the corneal epithelium. This expression of TSP-1 disappeared after 36-48 h, when re-epithelialization was completed. The TSP-1 mRNA level in the wounded corneas increased as much as three fold compared with that in the unwounded corneas. In organ culture, exogenous TSP-1 stimulated the re-epithelialization of corneal epithelial wounds whereas anti-TSP-1 antibody significantly inhibited the re-epithelialization. These findings suggest the possibility that epithelial defects in the corneas stimulate the expression of TSP-1 in the wound area, resulting in the accelerated re-epithelialization of the cornea.     

Experimental prostate epithelial morphogenesis in response to stroma and three-dimensional matrigel culture.

To reproduce the structural and functional differentiation of human prostatic acini in vivo, prostatic epithelial and stromal cells derived from human primary cultures were cocultured in Matrigel. In the absence of stroma and serum, epithelial spheroids composed of solid masses of stratified and cuboidal cells formed. Outer cells of the spheroid expressed cytokeratins 1, 5, 10, and 14, whereas the inner cells expressed cytokeratin 18. The addition of 2% serum induced formation of a lumen surrounded by a layer of one or two cuboidal and columnar epithelial cells. The further addition of stromal cultures, dihydrotestosterone, and estrogen induced polarization of the epithelium and increased spheroid-forming efficiency. Epithelium expressed either cytokeratin 18 alone or additionally cytokeratins 1, 5, 14, and 10. All spheroid epithelium expressed prostate-specific antigen and prostate-specific membrane antigen. Androgen receptor was only detected in the presence of stroma, serum, and hormones. Thus, development of a functional and morphologically correct prostate gland in vitro is dependent on extracellular matrix, steroid hormones, and factors from stromal cells and serum.     

Immunofluorescent localization of collagen types I, II, and III in the embryonic chick eye.

The appearance and distribution of type I, II, and III collagens in the developing chick eye were studied by specific antibodies and indirect immunofluorescence. At stage 19, only type I collagen was detected in the primary corneal stroma, in the vitreous body, and along the lens surface. At later stages, type I collagen was located in the primary and secondary corneal stroma and in the fibrous sclera, but not around the lens. Type II collagen was first observed at stage 20 in the primary corneal stroma, neural retina, and vitreous body. It was particularly prominent at the interface of the neural retina and vitreous body and, from stage 30 on, in the cartilaginous sclera. The primary corneal stroma consisted of a mixture of type I and II collagens between stages 20 and 27. Invasion of the primary corneal stroma by mesenchyme and subsequent deposition of fibroblast-derived collagen corresponded with a pronounced increase of type I collagen, throughout the entire stroma, and of type II collagen, in the subepithelial region. Type II collagen was also found in Bowman's and Descemet's membranes. A transient appearance of type III collagen was observed in the corneal epithelial cells, but not in the stroma (stages 20–30). The fully developed cornea contained both type I and II collagens, but no type III collagen. Type III collagen was prominent in the fibrous sclera, iris, nictitating membrane, and eyelids.     

Circulating cornea specific antibodies and keratoplasty

The presence of corneal antibodies in blood sera was investigated in 130 patients by immunoblotting analysis. Sera were obtained from patients before the first keratoplasty, repeated corneal transplantation or keratoplasty of the other eye. Baseline levels of antibodies in 15 healthy blood donors served to establish a positivity threshold. Patients were divided according to diagnosis and number of keratoplasties. Corneal antigens were prepared from donor corneas not suitable for surgery. Hyperimmune sera obtained by immunization of rabbits were used as positive controls. Significant increase in corneal antibodies specific to epithelial antigens was found in patients (62.5% positive) with keratitis of microbial origin. The presence of circulating antibodies to endothelial or stromal proteins was without significant changes. Depending on the number of keratoplasties no statistical difference in corneal antibodies production was observed. The group of patients with increased anti-corneal antibodies already before transplantation had to take immunosuppressive and anti-inflammatory therapy after keratoplasty.     

Embryonic origin of avian corneal sensory nerves.

Sensory nerves play a vital role in maintaining corneal transparency. They originate in the trigeminal ganglion, which is derived from two embryonic cell populations (cranial neural crest and ectodermal placode). Nonetheless, it is unclear whether corneal nerves arise from neural crest, from placode, or from both. Quail-chick chimeras and species-specific antibodies allowed tracing quail-derived neural crest or placode cells during trigeminal ganglion and corneal development, and after ablation of either neural crest or placode. Neural crest chimeras showed quail nuclei in the proximal part of the trigeminal ganglion, and quail nerves in the pericorneal nerve ring and in the cornea. In sharp contrast, placode chimeras showed quail nuclei in the distal part of the trigeminal ganglion, but no quail nerves in the cornea or in the pericorneal nerve ring. Quail placode-derived nerves were present, however, in the eyelids. Neural crest ablation between stages 8 and 9 resulted in diminished trigeminal ganglia and absence of corneal innervation. Ablation of placode after stage 11 resulted in loss of the ophthalmic branch of the trigeminal ganglion and reduced corneal innervation. Noninnervated corneas still became transparent. These results indicate for the first time that although both neural crest and placode contribute to the trigeminal ganglion, corneal innervation is entirely neural crest-derived. Nonetheless, proper corneal innervation requires presence of both cell types in the embryonic trigeminal ganglion. Also, complete lack of innervation has no discernible effect on development of corneal transparency or cell densities.     

Development of a hemicornea from human primary cell cultures for pharmacotoxicology testing

We report the reconstruction and characterization of a hemicornea (epithelialized stroma), using primary human cells, for use in research and as an alternative to the use of animals in pharmacotoxicology testing. To create a stromal equivalent, keratocytes from human corneas were cultured in collagen–glycosaminoglycan–chitosan foams. Limbal stem cell-derived epithelial cells were seeded on top of these, giving rise to hemi-corneas. The epithelium appeared morphologically similar to its physiological counterpart, as shown by the basal cell expression of p63 isoforms including, in some cases, the stem cell marker p63ΔNα, and the expression of keratin 3 and 14-3-3σ in the upper cell layers. In addition, the cuboidal basal epithelial cells were anchored to a basement membrane containing collagen IV, laminin 5, and hemidesmosomes. In the stromal part, the keratocytes colonized the porous scaffold, formed a network of interconnecting cells, and synthesized an ultrastructurally organized extracellular matrix (ECM) containing collagen types I, V, and VI. Electron microscopy showed the newly synthesized collagen fibrils to have characteristic periodic striations, with diameters and interfibril spacings similar to those found in natural corneas. Compared to existing models for corneal pharmacotoxicology testing, this new model more closely approaches physiological conditions by including the inducing effects of mesenchyme and cell–matrix interactions on epithelial cell morphogenesis.     

Atomic force microscopy analysis of normal and photoablated porcine corneas

We showed the capabilities and accuracy of atomic force microscopy (AFM) techniques for imaging and analyzing the corneal epithelium and the photoablated corneal stroma. Eight normal porcine corneas, half of which were ablated using a scanning-spot excimer laser, were examined. All the corneas were imaged in balanced salt solution after fixation in glutaraldehyde. In the normal untreated corneas we observed the epithelial surface showing the typical polygonal cells and presenting numerous microprojections. The superficial epithelial cells were classified in three types as a result of the anterior-surface roughness measurements. AFM images of the photoablated corneal specimens showed undulations and granule-like features on the ablated stromal surface, specific to 193-nm ArF laser irradiation. Nevertheless, the quantitative analysis confirmed the precision of excimer laser surgery in removing sub-micrometric amounts of tissue. AFM showed to be a high-resolved imaging tool for the scanning of both native as well as photoablated corneal specimens. Also, this technique permits precise topographic analysis of the corneal plane, in the nanometric scale, of which smoothness is an important physical characteristic and necessary to achieve an optimal optical quality of the eye.     

Formation of the endothelium of the avian cornea: A study of cell movement in vivo

In the analysis of endothelial morphogenesis reported here, scanning and transmission electron microscopes and the Nomarski light microscope were used to study both untreated and manipulated eyes of chick embryos. We found that migration of the cells into the corneal area is preceded at stage 22 by a movement of macrophages between the lens and posterior surface of the corneal stroma. At stage 23, endothelial cells move out mainly from the nasal and temporal edges of the eye where they were associated with vascular (primary) mesenchyme. Initially, they migrate through a fibrous matrix which occupies the space between lens and optic lip. When the endothelial cells reach the stroma and capsule of the lens, they can use both these surfaces as substrata, even though they seem to be more adherent to the stroma. By stage 25, the endothelium is complete and covered with fibrous matrix, which now fills and may help form the anterior chamber. The cells, initially mesenchymal, now differentiate to become epithelial (a characteristic of primary mesenchyme). The migrating endothelial cells have extended lamellipodia and filopodia along their leading edges; they show no evidence of ruffling. Moreover, contact inhibition alone does not cause them to monolayer; the presence of the lens is essential to prevent multilayering of the newly formed endothelium. In the discussion, the role of extracellular matrix and tissue boundaries in directing cell migration in vivo is emphasized.     

Stromal assemblies containing collagen types IV and VI and fibronectin in the developing embryonic avian cornea

The morphogenesis of type IV collagen-containing structures in the stromal matrix of the developing avian cornea was investigated using immunofluorescence and immunoelectron microscopic histochemistry. Two forms of type IV collagen-containing structures were seen; these differed in their probable origin, structure, molecular composition, and developmental fate. The major form of stromal type IV collagen-containing material, termed "strings," was observed only after swelling of the primary stroma and the onset of mesenchymal invasion. These strings are presumed to be products of the stromal cells. In immunofluorescence histochemistry they appeared as linear segments of type IV collagen-specific immunoreactivity. In immunoelectron microscopy, they appeared initially as electron-dense sausages of variable length and orientation. They frequently were associated with cell surfaces and, in fortuitous sections, appeared to connect adjacent cells. The strings also contained type VI collagen and fibronectin, but very little, if any, of the basement membrane components laminin and heparin sulfate proteoglycan (HSPG). As the stroma continued to expand in thickness, more of these structures were observed in a radial orientation, becoming quite long and less tortuous. Later in development, as stromal condensation proceeded, they disappeared. We suggest that the strings function to stabilize the stromal matrix, and perhaps to limit the rate and/or extent of stromal expansion, during a phase of rapid swelling and matrix deposition. The other form of type IV collagen-containing stromal material appeared as irregularly shaped plaques of basement membrane-like material identical to those previously described in mature corneas. These are likely derived from the corneal endothelial cells. They contained other basement membrane-associated components (laminin, HSPG) and fibronectin, but not type VI collagen. This material persists in mature corneas as sparse irregular stromal plaques and as matrix in the interface between Descemet's membrane and the corneal stroma.     

Fibronectin in the development of embryonic chick eye.

The time course of appearance and distribution of fibronectin in the developing eye have been studied in chick embryos by indirect immunofluorescence. At the 12-somite stage, fibronectin was detected as a layer under the ectodermal cells overlying the forebrain vesicle; it was also present in the head mesenchyme. During formation of the lens placode and its invagination, a zone containing fibronectin persisted around the lens as a component of the capsule. The fibronectin-containing layer was separated from the corneal epithelial cells during the formation of the acellular stroma. The migrating corneal endothelial cells were seen posterior to the fibronectin layer. The secondary stroma was strongly positive for fibronectin. Fibronectin disappeared from the cornea starting from its posterior part along with the corneal condensation. In the newborn chicken cornea, fibronectin was present only in Descemet's membrane. In addition, the embryonic vitreous body had a network of fibronectin-containing material. The distribution of fibronectin in the developing cornea, as well as other data available on this glycoprotein, is consistent with the proposed role of fibronectin in positioning and migration of cells and in organization of the extracellular matrix.