An ultrastructural analysis of cell and matrix differentiation during early limb development in Xenopus laevis

Early development of the hind limb of Xenopus (stages 44–48) has been analyzed at the level of ultrastructure with emphasis on differentiation of extracellular matrix components and intercellular contacts. By stages 44–45, mesenchyme is separated from prospective bud epithelium by numerous adepidermal granules in a subepithelial compartment (the lamina lucida), a continuous basal lamina and several layers of collagen (the basement lamella). Tricomplex stabilization of amphoteric phospholipid demonstrates that each adepidermal granule consists of several membranelike layers (electron-lucent band 25–30 Å; electron-dense band 20–40 Å), which are usually parallel to the basal surface of adjacent epithelial cells. Collagen fibrils are interconnected by filaments (35 Å in diameter) which stain with ruthenium red. Epithelial cells possess junctional complexes at their superficial borders, numerous desmosomes at apposing cell membranes and hemidesmosomes at their basal surface. Mesenchymal cells predominantly exhibit close contacts (100–150 Å separation) with few focal tight junctions at various areas of their surface. By stages 47–48, adepidermal granules are absent beneath bud epithelium and layers of collagen in the basement lamella lose filamentous cross-linking elements. Filopodia of mesenchymal cells penetrate the disorganized matrix and abut the basal lamina. Hemidesmosomes disappear at the basal surface of the epidermis and mesenchymal cells immediately subjacent to epithelium exhibit focal tight junctions and gap junctions at their lateral borders. These structural changes may be instrumental in the epitheliomesenchymal interactions of early limb development. Degradation of oriented collagenous lamellae permits direct association of mesenchymal cell surfaces (filopodia) with surface-associated products of epithelial cells (organized into the basal lamina). Development of structural pathways for intercellular ion and metabolite transport in mesenchyme may coordinate events specific to limb morphogenesis.     

Response of epithelial and mesenchymal cells to culture on basement lamella observed by scanning microscopy

The basement lamella of Xenopus tadpole skin has been viewed in situ by scanning microscopy, then isolated by trypsin treatment and used as a substrate for cell culture. The basal lamina may also be viewed after EDTA treatment. Responses of epithelial and mesenchymal cells to the lamella have been compared. Mesenchymal cells from chick skin and heart ventricle flatten and attach between the plies of the lamella, then infiltrate it. Myoblasts appear to move less readily within the lamella. Embryonic Xenopus skin epithelium spreads over the surface. Isolated chick skin epithelial cells first begin to spread, then round up and eventually attach to each other in clusters which form a flat basal surface above the lamella. Thus epithelial and mesenchymal cells cultured on this isolated extracellular material mimic aspects of normal tissue organization.     

Hemidesmosome formation in vitro

Intact, viable sheets of adult rabbit corneal epithelium, 9 mm in diameter, were prepared by the Dispase II method (Gipson, I. K., and S. M. Grill, 1982, Invest. Ophthalmol. Vis. Sci. 23:269-273). The sheets, freed of the basal lamina, retained their desmosomes and stratified epithelial characteristics, but lacked hemidesmosomes (HD). Epithelial sheets were placed on fresh segments of corneal stroma with denuded basal laminae and incubated in serum-free media for 1, 3, 6, 18, or 24 h. Tissue was processed for electron microscopy, and the number of HD/micron membrane, the number of HDs with anchoring fibrils directly across the lamina densa from them, and the number of anchoring fibrils not associated with HDs were counted. After 6 h in culture, the number of newly formed HD was 82% of controls (normal rabbit corneas), and by 24 h the number had reached 95% of controls. At all time periods studied, 80-86% of HDs had anchoring fibrils directly across the lamina densa from them. Anchoring fibrils not associated with HDs decreased with culture time. These data indicate that the sites where anchoring fibrils insert into the lamina densa may be nucleation sites for new HD formation. Corneal epithelial sheets placed on two other ocular basal laminae, Descemet's membrane and lens capsule, had not formed HDs after 24 h in culture. These two laminae do not have anchoring fibrils associated with them. Rabbit epithelial sheets placed on the denuded epithelial basal lamina of rat and human corneas formed new HDs. Thus, at least in these mammalian species, HD formation may involve some of the same molecular components.     

Epithelial-Directed Mesenchyme Differentiation in vitro

To assess the requirement for specific or possibly non-specific epithelial instructions for mesenchymal cell differentiation, we designed studies to evaluate and compare homotypic with heterotypic tissue recombinations across vertebrate species. These studies further tested the hypothesis that determined dental papilla mesenchyme requires epithelial-derived instructions to differentiate into functional odontoblast cells using a serumless, chemically-defined medium. Theiler stage 25 C57BL/6 or Swiss Webster cap stage mandibular first molar tooth organs or trypsin-dissociated, homotypic epithelial-mesenchymal tissue recombinants resulted in the differentiation of odontoblasts within 3 days. Epithelial differentiation into functional ameloblasts was observed within 7 days. Trypsin-dissociated and isolated mesenchyme did not differentiate into odontoblasts under these experimental conditions. Heterotypic recombinants between quail Hamburger-Hamilton stages 22–26 mandibular epithelium and Theiler stage 25 dental papilla mesenchyme routinely resulted in odontoblast differentiation within 3 days in vitro. Odontoblast differentiation and the production of dentine extracellular matrix continued throughout the 10 days in organ culture. Ultrastructural observations of the interface between quail and mouse tissues indicated the reconstitution of the basal lamina as well as the maintenance of an intact basal lamina during 10 days in vitro. Quail epithelial cells did not differentiate into ameloblasts and no enamel extracellular matrix was observed. These results show that quail mandibular epithelium can provide the required developmental instructions for odontoblast differentiation in the absence of serum or other exogenous humoral factors in a chemically-defined medium. They also suggest the importance of reciprocal epithelial-mesenchymal interactions during epidermal organogenesis.     

Retention of epithelial basal lamina allows isolated mandibular mesenchyme to form bone

The initiation of bone formation in the avian mandible requires that neural crest-derived cells undergo an inductive interaction with mandibular epithelium. To examine the role of the epithelial basal lamina in that interaction, mandibles were separated into their epithelial and mesenchymal components following exposure to the chelating agent, EDTA. Transmission and scanning electron microscopy was used to show that the basal lamina was retained as a continuous layer over the mesenchyme. Osteogenesis was initiated when such EDTA-isolated mesenchyme was grafted to the chorioallantoic membranes of host embryos. In contrast, mesenchyme isolated using trypsin and pancreatin failed to form bone. It is concluded that the property of mandibular epithelium which permits osteogenesis resides within the basal lamina.     

An SEM analysis of the epithelial--mesenchymal interface in the mandible of the embryonic chick

In this paper the ultrastructural features of the epithelial-mesenchymal interface in mandibular processes of embryonic chicks have been examined using scanning electron microscopy. Mandibular epithelium is required for the mesenchyme to differentiate as osteoblasts and to deposit the membrane bones of the mandible. The surface morphology of the epithelium changes from the lateral to the medial face of the mandible from rounded cells, each with a central cilium to flattened cells with numerous microvilli. Treatment with trypsin and pancreatin was used to digest the basal lamina so as to separate epithelium from mesenchyme. This exposed a thick, fibrillar basement membrane (reticular lamina), which was thicker underlying the caudal epithelium than under the cephalad epithelium. Addition of collagenase to the trypsin/pancreatin solution degraded some of the basement lamella, especially that underlying epithelium on the caudal portion of each mandibular process. Selective degradation of basement lamella is postulated as one means of regulating inductive epithelial-mesenchymal interactions. EDTA was used to isolate basal laminae on mandibular mesenchyme. SEM was used to confirm the integrity of the basal lamina, its structure, and its association with overlying epithelial cells and underlying basement lamella.     

牦牛输卵管上皮细胞分离培养和纯化鉴定

为建立牦牛输卵管上皮细胞原代培养及纯化方法,通过选取牦牛输卵管,运用机械刮取法和0.25%胰蛋白酶消化两种方法分离上皮细胞进行体外培养。对不同分离方法的培养效果比较,培养细胞进行形态学观察与传代培养、MTT比色检测细胞活力并制定生长曲线,原代及传代上皮细胞的免疫组织化学鉴定,冷冻解冻后经台盼蓝排斥试验检测活细胞数。结果表明该试验分离出的原代细胞,纯化后传代培养,经鉴定为牦牛输卵管上皮细胞,培养的细胞生长状况良好,建立了一套牦牛输卵管上皮细胞分离培养及纯化鉴定的方法。     

Effect of 4-methylumbelliferyl-β-d-xyloside on collagen synthesis in chick limb bud mesenchymal cell cultures

Previous studies showed that cultures of chick limb bud mesenchymal cells plated at high density, to maximize chondrogenic expression, had a much reduced extracellular matrix around chondrocytes when exposed to 4-methyl-, umbelliferyl-β-d-xyloside. The majority of newly synthesized chondroitin sulfate chains were found in the culture medium presumably bound to the xyloside as opposed to their normal deposition on the core protein of proteoglycan. The question remained open as to whether the development of an abnormal matrix affected the synthesis of extracellular deposition of other cartilage-specific macromolecules. We have analyzed, both morphologically and biochemically, the synthesis and deposition of Type I and Type II collagen by β-d-xyloside-treated cultures of limb mesenchymal cells. While the rate of collagen synthesis per plate and its extracellular accumulation after 8 days in culture were reduced to some extent, the ratios of Type II to Type I collagen and the morphological distribution of these macromolecules were not affected by exposure to β-d-xyloside. We conclude that the expression of the cartilage-specific Type II collagen during chondrogenic differentiation is, although reduced, qualitatively not dependent on the amount of extracellular chondroitin sulfate chains attached to matrix-associated proteoglycan core protein. However, prolonged exposure of limb bud cells to xylosides leads to the formation of a chondroitin sulfate- and collagen-deficient matrix which, in turn, reduces the capacity of limb bud cells to synthesize Types I and II collagen.     

Maintenance and Distribution of Epithelial Stem/Progenitor Cells after Corneal Reconstruction Using Oral Mucosal Epithelial Cell Sheets

We assessed the maintenance and distribution of epithelial stem/progenitor cells after corneal reconstruction using tissue-engineered oral mucosal cell sheets in a rat model. Oral mucosal biopsy specimens were excised from green fluorescent protein (GFP) rats and enzymatically treated with Dispase II. These cells were cultured on inserts with mitomycin C-treated NIH/3T3 cells, and the resulting cell sheets were harvested. These tissue-engineered cell sheets from GFP rats were transplanted onto the eyes of a nude rat limbal stem cell deficiency model. Eight weeks after surgery, ocular surfaces were completely covered by the epithelium with GFP-positive cells. Transplanted corneas expressed p63 in the basal layers and K14 in all epithelial layers. Epithelial cells harvested from the central and peripheral areas of reconstructed corneas were isolated for a colony-forming assay, which showed that the colony-forming efficiency of the peripheral epithelial cells was significantly higher than that of the central epithelial cells 8 weeks after corneal reconstruction. Thus, in this rat model, the peripheral cornea could maintain more stem/progenitor cells than the central cornea after corneal reconstruction using oral mucosal epithelial cell sheets.     

Expression of J1/tenascin in the crypt-villus unit of adult mouse small intestine: implications for its role in epithelial cell shedding.

The localization of the extracellular matrix recognition molecule J1/tenascin was investigated in the crypt-villus unit of the adult mouse ileum by immunoelectron microscopic techniques. In the villus region, J1/tenascin was detected strongly in the extracellular matrix (ECM) between fibroblasts of the lamina propria. It was generally absent in the ECM at the interface between subepithelial fibroblasts and intestinal epithelium, except for some restricted areas along the epithelial basal lamina of villi, but not of crypts. These restricted areas corresponded approximately to the basal part of one epithelial cell. In J1/tenascin-positive areas, epithelial cells contacted the basal lamina with numerous microvillus-like processes, whereas in J1/tenascin-negative areas the basal surface membranes of epithelial cells contacted their basal lamina in a smooth and continuous apposition. In order to characterize the functional role of J1/tenascin in the interaction between epithelial cells and ECM, the intestinal epithelial cell line HT-29 was tested for its ability to adhere to different ECM components. Cells adhered to substratum-immobilized fibronectin, laminin and collagen types I to IV, but not to J1/tenascin. When laminin or collagen types I to IV were mixed with J1/tenascin, cell adhesion was as effective as without J1/tenascin. However, adhesion was completely abolished when cells were offered a mixture of fibronectin and J1/tenascin as substratum. The ability of J1/tenascin to reduce the adhesion of intestinal epithelial cells to their fibronectin-containing basal lamina suggests that J1/tenascin may be involved in the process of physiological cell shedding from the villus.     

Antisera to basal lamina and glial endfeet disturb the normal extension of axons on retina and pigment epithelium basal laminae

In order to determine the role of the extracellular matrix in regulating the directed growth of embryonic neurites, antisera to retina (a-RBL I and II), to pigment epithelium (a-PBL) and to glomerular (a-GBL) basal lamina were probed for an effect on the ordered extension of neurites. In the assays, retina explants from chick and quail were cultured on basal lamina from embryonic chick retina and pigment epithelium either in the presence of anti-basal lamina antisera or in the presence of the corresponding preimmune sera. In the presence of all anti-basal lamina antisera, normal extension of axons was greatly inhibited both on retina and on pigment epithelium basal lamina. The antisera affected the growth pattern and the morphology of the individual axons in two ways: in the presence of a-RBL I the short axons were less directed, developed more and longer side branches, and the lamellipodia of the growth cones were reduced in size compared to axons from control cultures. In the presence of a-RBL II and a-GBL, axons grew slowly out from the explants as very thick bundles, strikingly different from axons in control cultures. The antiserum to pigment epithelium basal lamina induced both strong fasciculation and disorganization of the linear fiber extension, being intermediate between the two types of effects observed after antiserum addition. The results suggest that adhesive matrix molecules in basal laminae have important functions in elongation, fasciculation and in the morphology of growing axons.     

Collagen synthesis and deposition during mammary epithelial cell spreading on collagen gels

Mouse mammary epithelial cultivated on collagen gels demonstrate active spreading as the cells form monolayers. In this novel system, initiation of cell spreading is preceded by de novo synthesis of type IV collagen. The newly synthesized collagen is partitioned such that after 48 hr, approximately 24% is found in the culture medium, 35% is intracellular, and 41% is deposited in the extracellular matrix of the developing epithelium. Cultures deprived of serum failed to spread and to synthesize collagen. Proline analogues were shown to inhibit cell spreading and to suppress collagen synthesis in a dose-dependent manner. Cytochalasin D inhibition of F-actin elongation was shown to prevent cell spreading but not to suppress total collagen synthesis. During cytochalasin D treatment, inhibition of cell spreading was shown to result from failure to deposit or to maintain deposited collagen in the epithelium extracellular matrix. The data indicate that synthesis and extracellular deposition of a major basal lamina component (viz. type IV collagen) must precede and then accompany epithelial cell spreading in collagen gel culture. It is suggested that the microfilament apparatus, through some hypothetical integral membrane protein, can anchor extracellular type IV collagen, which then provides a necessary condition for cell spreading.     

Ectodermal-mesenchymal interspace during the formation of the chick leg bud

Summary The ectodermal-mesenchymal interspace of the chick leg bud was studied at stages leading to the formation of the apical ectodermal ridge (A.E.R.) (stages 14 to 19 HH), using scanning and transmission electron microscopy. The main findings were: 1. a continuous basal lamina under the ectoderm; 2. extracellular fibrils interconnecting the basal lamina and mesenchymal cell processes; 3. an increase in the number of the fibrils during these stages, with the highest number under the A.E.R.; 4. branching mesenchymal cell processes that spread over the basal lamina, making contact with it in all stages. The morphology of the interspace and the changes in it suggest that extracellular material may be significant in the ectodermal-mesenchymal interactions in the limb bud.     

Morphogenetic determination of the dermoepidermal interface during tail regeneration in Rana catesbeiana.

During regeneration of the amputated tadpole tail, reconstruction of the epithelial basal lamina and basement lamella occurs only after the other major morphogenetic processes are well established. At 4 days after tail transection of the bullfrog tadpole, electron microscopy of the internal surface of the basal cell layer of the blastemal epithelium reveals it to be relatively free of extracellular matrix. By 11 days a basal lamina of distinct regularity has formed, and the first rodlets and fibers signaling the replacement of the collagenous basement lamella are identified. At 15 days the basal cells of the epithelium start to exhibit specialization of their internal cell surfaces: Hemidesmosomes and associated tonofilaments appear, and the adepidermal globular layer is formed. Orthogonal packing of collagen plies begins by 19 days after transection, the number of layers exceeding 22 in the latter stages of regeneration.     

The isolation and characterization in vitro of normal epithelial cells,endothelial cells and fibroblasts from rat urinary bladder

Epithelial cells, microvascular endothelial cells, and fibroblasts have been isolated in culture from normal urinary bladders of Fischer rats. Normal epithelial cells were cultured most efficiently when transitional epithelial sheets were plated on to collagen-coated roller flasks. The epithelial sheets were obtained by two micro-dissection techniques. In the first method, the epithelium was peeled as a large coherent sheet from the submucosal connective tissue following subepithelial injection of a collagenase solution, and after incubation of the bladders in the same enzyme solution. Epithelial sheets with intact basal cell layers were essential for culture success. On collagenous matrices, epithelial differentiation was similar to that in vivo. The in vitro transitional epithelium was composed of three cell layers, namely superficial, intermediate, and basal cells. Basal cells were attached to newly synthesized basal lamina by means of hemidesmosomes. Superficial cells were sealed at their apical lateral membranes by a junctional complex, i.e. a terminal bar. Asymmetric luminal membrane plaques were not apparent. In the second method, the epithelium was separated from the underlying connective tissue after collagenase-trypsin digestion of everted urinary bladders. Although the digest consisted mainly of epithelial cells, these rarely survived the first passage when plated on conventional plastic growth surfaces. After the third culture week, epithelial cells usually died and slowly growing colonies of fibroblasts or large flattened epitheloid cells became apparent. Epitheloid cells were identified by their typical ultrastructure as endothelial cells, showing Weibel-Palade bodies and pinocytotic caveolae. These cells were reactive with antiserum against factor VIII. The free surface of monolayer cultures was non-thrombogenic when incubated in the presence of platelets. Fibroblasts were isolated from heavily contaminated epithelial cell cultures after differential trypsinization. These three cell types represent the normal control cells of an in vitro tumor model for the study of invasiveness. All three cell types are involved in the formation and functional maintenance of the epithelial-stromal junction. The study of cell-cell and cell-matrix interactions may provide important clues for the understanding of tumor invasiveness, a process that starts at the epithelial-stromal junction and proceeds with its destruction.     

The turnover of basal lamina glycosaminoglycan correlates with epithelial morphogenesis

The mouse embryonic submandibular epithelium begins as a single bud from the floor of the mouth which, under the influence of its surrounding mesenchyme, grows and forms lobules that subsequently branch repetitively. The lobular morphology of the 13-day epithelium is maintained by its basal lamina which is a continuous layer on the interlobular clefts but is interrupted on the distal aspects of the lobules. The structural integrity of this lamina is dependent upon its glycosaminoglycan (GAG) which, by histochemistry, is more abundant on the interlobular clefts than on the distal lobules. We have investigated the basis for these regional differences in the lamina by examining the synthesis and degradation of total GAG as well as the accumulation and loss of laminar GAG during the morphogenesis of the 13-day gland. Autoradiography and histochemistry show that laminar GAG is rapidly turning over. Although it is relatively stable in the interlobular clefts, GAG is rapidly degraded on the distal lobules. This difference can account for the regional variation in basal laminar GAG accumulation. The results of incorporation kinetics and precursor pool specific activities of total epithelial GAG show that the rate of GAG synthesis is greater than its rate of degradation in the base of the lobules, which includes the interlobular clefts. In contrast, during morphogenesis, the rate of GAG degradation becomes greater than its rate of replacement in the distal lobules. The epithelial stalk appears to be in the steady state regarding GAG metabolism. We propose (a) that the rapid laminar GAG degradation on the distal lobules produces the interruptions in the lamina, allowing epithelial growth and expansion, and (b) that the metabolic stability of laminar GAG on the interlobular clefts maintains the integrity of this lamina which serves as a cellular constraint. The results are consistent with a model for epithelial morphogenesis in which the mesenchyme remodels the lamina, which in turn, dictates epithelial morphology. Regulation of basal lamina turnover may be a general mechanism for controlling the behavior of epithelial cell populations.     

Ultrastructural changes in the intestinal connective tissue of Xenopus laevis during metamorphosis

Ultrastructural changes in the intestinal connective tissue of Xenopus laevis during metamorphosis have been studied. Throughout the larval period to stage 60, the connective tissue consists of a few immature fibroblasts surrounded by a sparse extracellular matrix: few collagen fibrils are visible except close to the thin basal lamina. At the beginning of the transition from larval to adult epithelial form around stage 60, extensive changes are observed in connective tissue. The cells become more numerous and different types appear as the collagen fibrils increase in number and density. Through gaps in the thickened and extensively folded basal lamina, frequent contacts between epithelial and connective tissue cells are established. Thereafter, with the progression of fold formation, the connective tissue cells become oriented according to their position relative to the fold structure. The basal lamina beneath the adult epithelium becomes thin after stage 62, while that beneath the larval epithelium remains thick. Upon the completion of metamorphosis, the connective tissue consists mainly of typical fibroblasts with definite orientation and numerous collagen fibrils. These observations indicate that developmental changes in the connective tissue, especially in the region close to the epithelium, are closely related spatiotemporarily to the transition from the larval to the adult epithelial form. This suggests that tissue interactions between the connective tissue and the epithelium play important roles in controlling the epithelial degeneration, proliferation, and differentiation during metamorphic climax.     

Embryonic corneal epithelial interaction with exogenous laminin and basal lamina is F-actin dependent

Between the third and sixth day of embryonic development, the avian corneal epithelium produces both a basal lamina and the primary corneal stroma composed of 20 orthogonally arranged layers of collagen fibrils. If the epithelium is removed by enzyme treatment from the basal lamina and stroma, the basal cell surface extends cell processes (blebs) which contain disorganized actin filaments and the epithelium decreases production of collagen. When placed on extracellular matrix or on Millipore filters in media containing soluble matrix molecules, the epithelium retracts the blebs, forms an organized basal actin cortical mat, and doubles its production of collagen. In the current investigation, we provide evidence for the hypothesis that organization of the RER by the actin cytoskeleton mediates this stimulation of collagen production. Laminin-treated epithelia and epithelia isolated with the basal lamina intact were treated with an actin-disrupting drug, cytochalasin D. Actin aggregates occur throughout the epithelium, the RER becomes disorganized, and the increase in collagen production expected to result from addition of laminin does not take place. Morphometrical analysis of the distribution of RER in the basal compartment of control and cytochalasin-treated epithelia shows that the decrease in collagen production is accompanied by displacement of the RER from the basal area of the cells, suggesting that attachment of RER to the intact actin cytoskeleton is essential to maintenance of normal RER organization and function. We also found that laminin-mediated bleb retraction requires intact actin microfilaments, whereas bleb extension does not, and that nocodazole does not inhibit bleb extension or retraction.     

Neuronal pathfinding in developing wings of the moth Manduca sexta

The neural pattern of the moth wing is a simple two-dimensional network nestled between the two epithelial monolayers that form the upper and lower surfaces of the wing. All neural elements within the wing blade are sensory and their axons grow proximally toward the mesothoracic ganglion. The sensory nerves of the wing are intimately associated with the basal lamina of the upper epithelial layer; and the molding of neural pattern is coupled with cues in the basal lamina. The global landscape of the basal lamina can be altered by exchange of epithelial grafts. Axons generally cross control grafts as well as grafts that have been displaced distally. However, axons generally avoid grafts that have been transposed proximally. This asymmetric response of growing axons implies that directional cues in the substratum are also asymmetric along the length of the wing. The asymmetric, graded distribution of extracellular matrix molecules associated with the basal lamina of the wing's upper epithelium could provide the short-range cues that guide sensory axons in a particular direction.     

Effects of the extracellular matrix on fetal choroid plexus epithelial cells: changes in morphology and multicellular organization do not affect gene expression.

We have developed a primary culture system for fetal mouse choroid plexus epithelial cells which maintains their differentiated phenotype. When grown on a reconstituted basement membrane substrate (Matrigel) epithelial cells formed aggregates which became embedded in the matrix and developed into characteristic and highly reproducible multicellular vesicular structures. These vesicles consisted of a squamous layer of epithelial cells with extensive attachment to the matrix substrate, surrounding a fluid-filled lumen. Electron microscopy showed that cells comprising these vesicles had a high degree of membrane specialization and polarized morphology which in many respects mimicked the in vivo morphology. Biochemical analyses demonstrated that under these culture conditions the tissue-specific pattern of gene expression of fetal choroid plexus epithelium was maintained. After 6 days in culture these cells contained approximately the same amount of transthyretin mRNA as the 12.5-day choroid plexus in vivo, and the level of total RNA per cell, which is proportional to the protein synthetic capability of the cells, was also maintained. The pattern of protein secretion was also very similar to that generated by fetal mouse choroid plexus cells in vivo. In contrast choroid plexus epithelial cells attached poorly to collagen I gels. Heterogeneous aggregates were formed in which cell-cell interactions were more extensive than cell-substrate interactions, and in no cases was a central lumen observed. Cells on the surface of large aggregates showed some evidence of membrane polarization, while the majority of cells in the cultures exhibited little evidence of polarized morphology. Despite the striking difference in morphology and multicellular organization these cells still expressed high levels of transthyretin mRNA and maintained the same pattern of protein synthesis as cells cultured on Matrigel. These results indicate that the basement membrane is important for the organization of choroid plexus epithelial cells into a functional epithelium in vitro and thus presumably the maintenance of the integrity of the blood-brain barrier in vivo. In contrast to several other epithelial systems which have been studied, the type of extracellular matrix does not appear to directly influence tissue-specific gene expression by choroid plexus epithelial cells. Thus the level of gene expression is not dependent on the cytoarchitecture and multicellular organization of this cell type.