Exogenous fibronectin is not required for organogenesis in vitro

Summary The biological effect of plasma fibronectin on the differentiation of embryonic mouse kidney and tooth was studied in organ cultures. Transferrin (50μg/ml) was a strong mitogen for kidney cells, whereas, the addition of soluble fibronectin (50 to 250 μg/ml) had no detectable effect on differentiation or proliferation. The same serum-free, transferrin-containing medium did not support tooth differentiation. however, fibronectin was not a necessary serum component because fibronectin-free serum supported tooth development. It was demonstrated with antibodies specific for human fibronectin that the exogenously added human fibronectin at 50 μg/ml did not become incorporated to the cultured organs. Only minimal incorporation to the kidney basement membrane area was observed when fibronectin concentration was 250μg/ml. The mesenchymal stroma and the basement membranes of the kidney and tooth rudiments cultured in fibronectin-free media stained intensely with conventional fibronectin antibodies, indicating endogenous production of fibronectin. Outgrowing epithelial cells from isolated kidney tubules produced fibronectin as well as laminin. The results suggest that the fibronectin found in the stroma and basement membranes is an endogenous product of the developing tissues and that plasma fibronectin is not required for in vitro organogenesis. The results also indicate that it is difficult to study the effect of fibronectin on morphogenetic processes because it may not penetrate the organ explants in vitro. This work was supported by grants from the Sigrid Jusélius Foundation, Finska L?kares?llskapet, Finnish Life Insurance Companies and Grant CA 28896 and Cancer Center Support Grant CA 30199 from the National Cancer Institute, Department of Health and Human Services, Washington, D.C.     

Cellular origin of fibronectin in interspecies hybrid kidneys

The cellular origin of fibronectin in the kidney was studied in three experimental models. Immunohistochemical techniques that use cross-reacting or species-specific antibodies against mouse or chicken fibronectin were employed. In the first model studied, initially avascular mouse kidneys cultured on avian chorioallantoic membranes differentiate into epithelial kidney tubules and become vascularized by chorioallantoic vessels. Subsequently, hybrid glomeruli composed of mouse podocytes and avian endothelial-mesangial cells form. In immunohistochemical studies, cross-reacting antibodies to fibronectin stained vascular walls, tubular basement membranes, interstitium, and glomeruli of mouse kidney grafts. The species-specific antibodies reacting only with mouse fibronectin stained interstitial areas and tubular basement membranes, but showed no reaction with hybrid glomeruli and avian vascular walls. In contrast, species-specific antibodies against chicken fibronectin stained both the interstitial areas and the vascular walls as well as the endothelial-mesangial areas of the hybrid glomeruli, but did not stain the mouse-derived epithelial structures of the kidneys. In the second model, embryonic kidneys cultured under avascular conditions in vitro develop glomerular tufts, which are devoid of endothelial cells. These explants showed fluorescence staining for fibronectin only in tubular basement membranes and in interstitium. The avascular, purely epithelial glomerular bodies remained unstained. Finally, in outgrowths of separated embryonic glomeruli, the cross-reacting fibronectin antibodies revealed two populations of cells: one devoid of fibronectin and another expressing fibronectin in strong fibrillar and granular patterns. These results favor the idea that the main endogenous cellular sources for fibronectin in the embryonic kidney are the interstitial and vascular cells. All experiments presented here suggest that fibronectin is not synthesized by glomerular epithelial cells in vivo.     

Laminin and fibronectin promote the haptotactic migration of B16 mouse melanoma cells in vitro

The migration of tumor cells through basement membranes and extracellular matrices is an integral component of tumor invasion and metastasis. Laminin and fibronectin are two basement membrane- and extracellular matrix-associated noncollagenous glycoproteins that have been shown to promote both cell adhesion and motility. Purified preparations of laminin and fibronectin stimulated the directed migration of B16 murine metastatic melanoma cells in vitro as assessed in modified Boyden chambers. The stimulation of migration occurred over a concentration range of 1-100 micrograms/ml of laminin or fibronectin, with a peak response occurring between 12.5 and 25 micrograms/ml. The maximal response of these cells was 80-120-fold higher than control migration. Affinity-purified antibody preparations specifically abrogated the migration of these cells in response to the respective proteins. Tumor cells in suspension were preincubated in physiologic levels of plasma fibronectin prior to assay to partially mimic what occurs when a metastasizing cell is in the blood stream. This preincubation with plasma fibronectin had no effect on the subsequent migration of cells in response to either laminin or fibronectin. Furthermore, experiments using filters precoated with fibronectin or laminin indicated that these cells could migrate by haptotaxis to these two proteins. We conclude that tumor cell migration in response to such noncollagenous adhesive glycoproteins could be an important aspect in the invasion and metastasis of certain malignant cell types.     

Isolation, characterization and immunological determination of basement membrane-associated heparan sulfate proteoglycan

Basement membrane-associated heparan sulfate proteoglycan (HSPG) was extracted from isolated porcine glomerular basement membranes and purified by ion-exchange chromatography. The proteogycan was characterized by specific enzymatic digestions, by amino-acid analysis, by SDS-polyacrylamide gel electrophoresis and by density gradient centrifugation. Polyclonal antibodies were raised against the purified HSPG in rabbits. Antibodies were characterized by enzyme immunoassays, immunoprecipitation and immunohistological methods. They were shown to recognize specifically the core protein of HSPG from porcine, human and rat glomerular basement membrane but did not recognize HSPG from guinea pig or rabbit kidney. The affinity-purified antibodies did not cross-react with other basement membrane proteins like laminin, fibronectin or collagen type IV nor with chondroitin sulfate-rich or keratan sulfate-rich proteoglycans from human or bovine tissue. Using these antibodies an enzyme immunoassay was developed for determination of HSPG in the range of 1-100 ng/ml. Studies with cultured porcine endothelial cells showed that subendothelial basement membrane-associated HSPG may be determined with the enzyme immunoassay.     

Comparison of the ability of basement membranes produced by corneal endothelial and mouse-derived Endodermal PF-HR-9 cells to support the proliferation and differentiation of bovine kidney tubule epithelial cells in vitro

The proliferation and morphological differentiation of bovine kidney collecting-tubule epithelial cells has been examined as a function of substrata and plasma factors. Collecting kidney tubule explant maintained in vitro gave rise to two distinct cell populations; one was composed mostly of fibroblastic cells whereas the other was epithelioid (EP cells). The proliferation of fibroblastic cells when exposed to serum-supplemented medium was best expressed when cells were maintained on a basement membrane produced by bovine corneal endothelial cells. This basement membrane has a composition, which in previous studies has been shown to favor the proliferation of mesenchymal cells. In contrast, the proliferation of EP cells was best expressed when cells were maintained on a basement membrane produced by the mouse-derived endodermal cell line PF-HR-9 (HR-9-BM). This basement membrane has a biochemical composition very similar to the basement membrane underlying the kidney tubules. Although the fibroblast confluent monolayer maintained on bovine corneal endothelial cell extracellular matrix did not undergo morphogenesis, the confluent monolayer of EP cells maintained on HR-9-BM shows hemicyst formation, suggesting that they were capable of vectorial fluid transport. They also built a complex three-dimensional kidney tubulelike network. Some tubules became grossly visible and floated into the tissue culture medium, remaining tethered to the cell monolayer at either end of the tubule. On an ultrastructural level, the tubules consisted of cells held together with junctional complexes arranged so as to form a lumen. The smallest lumen were bordered by 2-3 cells, and the largest ones by 8-15 cells. The lumens of the larger tubules did contain granular fibrillar and amorphous debris. Low-density EP cell cultures maintained on HR-9-BM could be induced to proliferate at a rate approaching that of cultures exposed to serum when they were exposed to medium supplemented with high-density lipoprotein (HDL, 750 micrograms protein/ml) and transferrin (50 micrograms/ml). When exposed to HDL concentrations equal or lower than 250 micrograms protein/ml, low-density cultures proliferated at a slow rate and readily formed tubulelike structures. This observation indicates that EP cells do not need to reach confluence to undergo morphogenesis, and that HDL, which in the presence of transferrin supports the cell proliferation, can favor their differentiation into tubulelike structures once its concentration becomes limiting for mitogenesis.     

Blood plasma fibronectin does not influence fibroblast-induced collagen gel contraction

Human embryo fibroblasts were cultured within three-dimensional collagen gel in the Eagle medium supplemented with 10% calf serum. Addition of fibronectin (Fn) to the nutrient medium neither stimulated nor inhibited gel contraction by fibroblasts. Gel contraction also proceeded in the medium containing ultroser G or Fn exhausted calf serum. Addition of 10 or 50 micrograms/ml of Fn to the medium with ultroser G exerted no effect on the process of gel contraction. Addition of antibodies prepared against Fn (70 micrograms/ml) and tetrapeptide RGDS (100 and 500 micrograms/ml) to the medium supplemented with Fn exhausted calf serum did not change the rate of collagen gel contraction. The results obtained suggest that the rate of collagen gel contraction by fibroblasts is not dependent on the presence of exogenous fibronectin.     

Light microscopic immunolocalization of type IV collagen, laminin, heparan sulfate proteoglycan, and fibronectin in the basement membranes of a variety of rat organs

Immunohistochemical methods were used to determine whether type IV collagen, laminin, fibronectin, and heparan sulfate proteoglycan were present in diverse basement membranes. Antisera or antibodies against each substance were prepared, tested by enzyme-linked immunosorbent assay, and exposed to frozen sections of duodenum, trachea, kidney, spinal cord, cerebrum, and incisor tooth from rats aged 20 days to 34 months. Bound antibodies were then localized by indirect or direct peroxidase methods for examination in the light microscope. Immunostaining for type IV collagen, laminin, fibronectin, and heparan sulfate proteoglycan was observed in all of the basement membranes encountered. Fibronectin was also found in connective tissue. In general, the intensity of immunostaining was strong for type IV collagen and laminin, moderate for heparan sulfate proteoglycan, and weak for fibronectin. The pattern was similar in the age groups under study. Very recently the sulfated glycoprotein, entactin, was also detected in the basement membranes of the listed tissues in 20-day-old rats. It is accordingly proposed that, at least in the organs examined, type IV collagen, laminin, fibronectin, heparan sulfate proteoglycan, and entactin are present together in basement membranes.     

Interleukin 6 promotes epithelial migration by a fibronectin-dependent mechanism.

We investigated the effect of interleukin 6 (IL-6) on the migration of rabbit corneal epithelium in vitro and on the attachment of dissociated corneal epithelial cells to a fibronectin matrix. When corneal blocks were cultured with IL-6 for 24 hours, the length of the path of epithelial migration over exposed corneal stroma increased significantly (p less than 0.005 at the concentration of 10 ng/ml) in proportion to the concentrations of IL-6 (0.1-10.0 ng/ml). The addition of antiserum against fibronectin or of GRGDSP abolished the stimulatory effect of IL-6 on epithelial migration. When corneal epithelial cells were cultured with various concentrations of IL-6, suspended, and plated on wells coated with fibronectin (10 micrograms/ml), the number of cells attached to the wells increased in a dose-dependent manner. The presence of antibody against fibronectin or of GRGDSP during the attachment assay decreased the number of cells attached to the fibronectin matrix, regardless of the fact that the cells had been cultured with IL-6 or not. IL-6 stimulated the attachment of corneal epithelial cells to collagen type IV and to laminin matrices. However, the presence of GRGDSP did not affect the cell attachment to collagen type IV and to laminin. These findings strongly indicate that IL-6 stimulates epithelial migration in the cornea by a fibronectin-dependent mechanism, presumably the increased expression of fibronectin receptors.     

Formation of basement membranes in the embryonic kidney: an immunohistological study

Specific antibodies to laminin, type IV collagen, basement-membrane proteoglycan, and fibronectin have been used in immunofluorescence microscopy to study the development of basement membranes of the embryonic kidney. Kidney tubules are known to form from the nephrogenic mesenchyme as a result of an inductive tissue interaction. This involves a change in the composition of the extracellular matrix. The undifferentiated mesenchyme expresses in the composition of the extracellular matrix. The undifferentiated mesenchyme expresses fibronectin but no detectable laminin, type IV collagen, or basement-membrane proteoglycan. During the inductive interaction, basement-membrane specific components (laminin, type IV collagen, basement membrane proteoglycan) become detectable in the induced area, whereas fibronectin is lost. While the differentiation to epithelial cells of the kidney requires an inductive interaction, the development of the vasculature seems to involve an ingrowth of cells which throughout development deposits basement-membrane specific components, as well as fibronectin. These cells form the endothelium and possibly also the mesangium of the glomerulus, and contribute to the formation of the glomerular basement membrane. An analysis of differentiation of the kidney mesenchyme in vitro in the absence of circulation supports these conclusions. Because a continuity with vasculature is required for glomerular endothelial cell differentiation, it is possible that these cells are derived from outside vasculature.     

Differential expression of the extra domain-containing form of cellular fibronectin in human placentas at different stages of maturation

The distribution of the extra domain-containing form of cellular fibronectin was studied in human placentas at different stages of maturation by using the monoclonal antibody 52DH1 in indirect immunofluorescence. In early chorionic tissue (7 to 10 weeks post menstruationem) cellular fibronectin was codistributed with laminin and type IV collagen in the trophoblastic basement membranes. At weeks 11 to 12 the trophoblastic basement membranes were negative but positivity was typically revealed in distinct aggregates in the stromal tissue. In second-trimester and term placentas the immunoreactivity was confined to the vessel endothelia of villous stroma. Extravillous trophoblast cells seen in placentas at different stages did not show positivity. Double staining with the 52DH1 monoclonal antibody and polyclonal fibronectin antibodies showed that both in the early and term placentas there was much fibrillar positivity only revealed with the polyclonal antibodies. The present results show that cellular fibronectin is a prominent component of early trophoblastic basement membranes and may thus play a special role in the maturation of chorionic villi.     

Differential expression of the extra domain-containing form of cellular fibronectin in human placentas at different stages of maturation

Summary The distribution of the extra domain-containing form of cellular fibronectin was studied in human placentas at different stages of maturation by using the monoclonal antibody 52DHl in indirect immunofluorescence. In early chorionic tissue (7 to 10 weeks post menstruationem) cellular fibronectin was codistributed with laminin and type IV collagen in the trophoblastic basement membranes. At weeks 11 to 12 the trophoblastic basement membranes were negative but positivity was typically revealed in distinct aggregates in the stromal tissue. In second-trimester and term placentas the immunoreactivity was confined to the vessel endothelia of villous stroma. Extravillous trophoblast cells seen in placentas at different stages did not show positivity. Double staining with the 52DHl monoclonal antibody and polyclonal fibronectin antibodies showed that both in the early and term placentas there was much fibrillar positivity only revealed with the polyclonal antibodies. The present results show that cellular fibronectin is a prominent component of early trophoblastic basement membranes and may thus play a special role in the maturation of chorionic villi.     

Expression of fibronectin and laminin in fetal male gonads in vivo and in vitro with and without testicular morphogenesis

Sertoli cell differentiation occurs in vitro, even when testicular morphogenesis is inhibited by addition of serum to the culture medium (Magre, S. and A. Jost: Proc. Natl. Acad. Sci. USA 81, 7831-7834 (1984]. Using indirect immunohistochemical technique, we have studied the expression of fibronectin and laminin in gonads lacking testicular morphogenesis, as compared to in vivo controls and gonads cultured in synthetic medium. In undifferentiated gonads in vivo, fibronectin and laminin are distributed uniformly in the blastema. If testicular differentiation occurs in vivo, laminin is detected only in the basement membranes; when it occurs in vitro, laminin is found both in the basement membranes and among the stromal tissue. In gonads without seminiferous cords (cultured in serum-supplemented medium), fibronectin and laminin are both present, they are uniformly distributed among the gonadal cells.     

Regulation of cell attachment and cell number by fibronectin and laminin

We have examined the effect of laminin and fibronectin on the attachment and growth on type IV collagen of a line of mouse epithelial cells and a strain of adult human fibroblasts. Laminin stimulated attachment of the epidermal cells and fibronectin stimulated fibroblast attachment. At high concentrations (100 micrograms/ml), the attachment proteins altered the growth of cells in culture. The epidermal cells grew better in media containing fibronectin-free serum supplemented with laminin. Fibroblasts, on the other hand, grew best in media containing serum supplemented with fibronectin. These data suggest that laminin promotes epithelial cell growth whereas fibronectin promotes fibroblast growth. This observation was confirmed when these cells were cocultured in the presence of the attachment proteins or of their respective antibodies. The mouse epidermal cells grew best when laminin was added to cocultures of fibroblasts and epithelial cells. Fibroblasts grew best in the presence of antibody to laminin and poorly in the presence of antibody to fibronectin. Thus, fibronectin and laminin may participate in the regulation of cell populations in vivo and may be involved in epithelial-mesenchymal interactions.     

Changes in the distribution of type IV collagen,laminin, proteoglycan,and fibronectin during mouse tooth development

The distribution of certain basement membrane (BM) components including type IV collagen, laminin, BM proteoglycan, and fibronectin was studied in developing mouse molar teeth, using antibodies or antisera specific for these substances in indirect immunofluorescence. At the onset of cuspal morphogenesis, type IV collagen, laminin, and BM proteoglycan were found to be present throughout the basement membranes of the tooth. Fibronectin was abundant under the inner enamel epithelium at the region of differentiating odontoblasts and also in the mesenchymal tissues. After the first layer of predentin had been secreted by the odontoblasts at the epithelial-mesenchymal interface, laminin remained in close association with the epithelial cells whereas type IV collagen, BM proteoglycan, and fibronectin were distributed uniformly throughout this area. Later when dentin had been produced and the epithelial cells had differentiated into ameloblasts, basement membrane components disappeared from the cuspal area. These matrix components were not detected in dentin while BM proteoglycan and fibronectin were present in predentin. The observed changes in the collagenous and noncollagenous glycoproteins and the proteoglycan appear to be closely associated with cell differentiation and matrix secretion in the developing tooth.     

Changes in serum and tissue fibronectin during ontogeny

It has been proved that both in bovine and rat serum fibronectin concentration increases with age. While in foetal bovine the serum fibronectin concentration was 77 +/- 5 micrograms/ml, in adult cows the concentration of this protein reached 150.0 +/- 5.5 micrograms/ml. In newborn rats, the corresponding fibronectin concentration in the serum was 227 +/- 25 micrograms/ml as compared to 455 +/- 47 micrograms/ml in adult individuals. On the contrary, a drop in fibronectin concentration was found in the bovine glomerular basement membrane: fibronectin represents 3.8 +/- 0.6% of total protein in the newborn, while in animals older than two months this protein was not detectable. Accordingly, in newborn rats fibronectin represents 5.1 +/- 0.6% of total protein in the glomerular basement membrane and drops dramatically to 0.4 +/- 0.1% after one year of life with no detectable changes thereafter. In the alveolar basement membrane the changes are similar as in the glomerular basement membrane. While in newborn rats this protein represents 13.2 +/- 0.5% of total basement membrane protein, it drops to 3.2 +/- 0.3% after two years of life.     

Use of anti-idiotypic antibodies as probes for the interaction of TSH subunits with its receptor

TSH is a heterodimeric glycoprotein hormone, whose dissociated subunits are without biological activity. This has precluded the assessment of the relative contribution of each subunit to hormone action. We have raised anti-idiotypes to monoclonal antibodies specific, respectively, for the alpha and beta hTSH subunits. The anti-beta anti-idiotype inhibited 125I-hTSH binding to the beta subunit-specific monoclonal quantitatively, whereas 125I-hTSH binding to the alpha subunit-specific monoclonal was not inhibited by anti-alpha anti-idiotypes, suggesting that only the former is an "internal image" anti-idiotype. Neither of the two anti-idiotypes nor equimolar mixtures thereof inhibited 125I-bTSH binding to thyroid membranes, even though radiolabelled anti-idiotypes showed saturable binding to thyroid plasma membrane which was inhibited 41-65% by bTSH. Each anti-idiotype alone caused 9% inhibition (compared to 50% by NRIgG) of thyroid plasma membrane adenylate cyclase. Equimolar mixtures (125 micrograms/ml IgG of each anti-idiotype) induced enzyme activity equivalent to 85% of that of 250 mU/ml of TSH. The TSH-like action of the two anti-idiotypes was also reflected in their capacity to increase (450% by 250 micrograms/ml IgG compared to normal rabbit IgG) the uptake of 131I into isolated thyrocytes and to promote the organization of such cells into follicular structures. At 250 micrograms/ml, anti-beta anti-idiotype promoted the organization of small follicles and only at a concentration of 500 micrograms/ml did it enhance 131I uptake.     

Expression of extracellular matrix components is regulated by substratum

Reconstituted basement membranes and extracellular matrices have been demonstrated to affect, positively and dramatically, the production of milk proteins in cultured mammary epithelial cells. Here we show that both the expression and the deposition of extracellular matrix components themselves are regulated by substratum. The steady-state levels of the laminin, type IV collagen, and fibronectin mRNAs in mammary epithelial cells cultured on plastic dishes and on type I collagen gels have been examined, as has the ability of these cells to synthesize, secrete, and deposit laminin and other, extracellular matrix proteins. We demonstrate de novo synthesis of a basement membrane by cells cultured on type I collagen gels which have been floated into the medium. Expression of the mRNA and proteins of basement membranes, however, are quite low in these cultures. In contrast, the levels of laminin, type IV collagen, and fibronectin mRNAs are highest in cells cultured on plastic surfaces, where no basement membrane is deposited. It is suggested that the interaction between epithelial cells and both basement membrane and stromally derived matrices exerts a negative influence on the expression of mRNA for extracellular matrix components. In addition, we show that the capacity for lactational differentiation correlates with conditions that favor the deposition of a continuous basement membrane, and argue that the interaction between specialized epithelial cells and stroma enables them to create their own microenvironment for accurate signal transduction and phenotypic function.     

Fibronectin blocks p38 and jnk activation by cyclic strain in Caco-2 cells

Diverse repetitive forces deform the intestinal epithelium and basement membrane. Such repetitive deformation induces intestinal epithelial proliferation, differentiation, and intracellular signaling. Although at least some deformation-induced signals probably involve integrins, the matrix-dependence of these signals is poorly understood. We compared rapid strain activation of p38 and jnk in human Caco-2 intestinal epithelial cells cultured on collagen I, collagen IV, laminin, and tissue fibronectin. These signals were inhibited in cells on a fibronectin substrate, but activated by strain on collagens and laminin. Furthermore, adding 300 microg/ml plasma fibronectin (approximately the concentration found in plasma) to the culture medium inhibited strain activation of p38 and jnk in cells cultured on collagen. Since tissue and plasma fibronectin levels vary in acute or chronic inflammatory or infectious conditions, these results suggest that tissue or plasma fibronectin may modulate the intestinal epithelial response to repetitive deformation.     

In vitro neurite extension by granule neurons is dependent upon astroglial-derived fibroblast growth factor

When grown in the absence of astroglial cells, purified mouse cerebellar granule neurons survive less than 36 hr and do not extend neurites. Here we report that low concentrations of basic fibroblast growth factor (bFGF, 1-25 ng/ml) maintained the viability and promoted the differentiation of purified granule neurons. The effect of bFGF on granule cell neurite outgrowth was dose dependent. Neurite outgrowth was stimulated markedly in the presence of 1-25 ng/ml bFGF, but effects were not seen below 1 ng/ml or above 50 ng/ml. When affinity-purified antibodies against bFGF (1-5 micrograms/ml) were added either to purified granule cells or to co-cultures of neurons and astroglial cells, process extension by granule neurons was severely impaired. The inhibition of neurite outgrowth in the presence of anti-bFGF antibodies was reversed by the addition of 25 ng/ml of exogenous bFGF. In addition to neuronotrophic effects, bFGF influenced the rate of growth of the astroglial cells. This result depended on whether the astroglia were grown in isolation from neurons, where low doses of bFGF (10-25 ng) stimulated glial growth, or in coculture with neurons, where much higher doses of bFGF (100-250 ng/ml) were needed for glial mitogenesis. Immunoprecipitation of lysates from 35S-labeled cerebellar astroglial cells with anti-bFGF antibodies revealed a single band after SDS-PAGE at 18,000 Da, the molecular weight of bFGF. These results indicate that glial cells synthesize bFGF and are possibly an endogenous source of bFGF in cerebellar cultures. Thus, astroglial cells synthesize soluble factors needed for neuronal differentiation.     

Localization of basement membrane glycoproteins in rat kidney during foetal development

The distribution of basement membrane glycoproteins (type IV collagen, laminin, fibronectin, and proteoglycans) was studied in foetal rat kidney by immunohistochemical techniques using polyclonal antibodies. From the first stages of nephron differentiation, all these glycoproteins were detectable by immunofluorescence in the tubular and glomerular basement membranes and in the mesangial matrix. As differentiation proceeded, labelling of glycoproteins progressively intensified, except for that of fibronectin, which gradually decreased in the glomerular basement membrane (GBM) and was barely observable at full differentiation. With immunoperoxidase staining in electron microscopy, all glycoproteins were seen to be widely dispersed in the spaces between the epithelial and endothelial glomerular cells so long as the GBM remained a loose structure. However, after it became a compact, 3-layered formation, type IV collagen and laminin were distributed throughout the GBM, whereas proteoglycans and anionic sites appeared as 2 rows of granules confined to the laminae rarae.