The growth and differentiation of transitional epithelium in vitro

Bladder tissue formation from cultured bladder urothelium

Tissue recombination is a powerful method to evaluate the paracrine-signaling events that orchestrate the development of organs using the in vivo environment of a host rodent. Studies have reported the successful generation of primary cultures of rodent bladder urothelium, but none have reported their use to recapitulate bladder tissue with tissue recombination. We propose that primary cultured bladder urothelium, when recombined with inductive embryonic bladder mesenchyme, will form bladder tissue in a recombination model. Adult rat bladders were isolated and urothelium obtained. Sheets of bladder urothelium were re-suspended in collagen and maintained in tissue culture. After expansion (>20 passages), the urothelium was recombined with embryonic day-14 mouse bladder mesenchyme, then grafted beneath the renal capsule of immunocompromised mouse hosts. Grafts were harvested after 28 days. Control grafts were performed with bladder mesenchyme alone, cultured bladder urothelium alone, and collagen matrix alone. Final tissues were evaluated with staining and immunohistochemistry (H&E, Gomori's trichrome, broad-spectrum uroplakin, and smooth muscle actin alpha and gamma). Immunocytochemistry on cultured urothelium for broad-spectrum keratin, vimentin, and broad-spectrum uroplakin confirmed pure populations, void of mesenchymal contaminants. Staining of recombinant grafts demonstrated bladder tissue with mature urothelium and stromal differentiation. Control tissues were void of bladder tissue formation. We have successfully demonstrated that a chimeric bladder is formed from primary cultured bladder urothelium recombined with embryonic bladder mesenchyme. This is a powerful new tool for investigating the molecular mechanisms of bladder development and disease. Future applications may include the in vitro genetic manipulation of urothelium and examining those effects on growth and development in an in vivo environment.

[Current status of tissue engineering in urology. Review of the literature]

In the eighties a new field of the medicine appears wich applies the principles of cellular cultivation to synthetic biodegradable polymers scaffolds with the purpose of creating autologous biological substitutes that could improve, maintain or restore the function of organs or damaged tissues. The Tissue Engineering constitutes a new discipline in full phase of development especially in USA, with multiple potential applications in several medical specialities. Our speciality can't remain indifferent to interest and encouraging future originated by this new science. In this work we have made a wide bibliographical revision in the Medline to know the antecedents, current state and the possible future applications of Tissue Engineering in Urology.

[Experimental study about viability of autologous free graft in vitro cultivated urinary epithelium]

OBJECTIVE

The purpose of this study is to apply the in vitro keratinocyte culture techniques and the tissue engineering principles to urothelium, to obtain a three-dimensional autologous tissue suitable for grafting. We also showed the viability of free graft cultured urothelium in an experimental model.

MATERIAL AND METHODS

An animal experimental model was designed to apply the techniques of cellular culture and tissue engineering. Biopsy specimens of bladder mucosa were obtained, in vitro cultured and posteriorly implanted in each animal. We established three groups based on different follow-up periods (7, 14 and 30 days), and made a final histomorphological study to demonstrate the viability of the graft at the end of its respective follow-up period.

RESULTS

A three-dimensional in vitro tissue was obtained, composed of a bio-artificial submucosa (fibrin gel and fibroblast) where the uroepithelial cells were seeding; a biodegradable polyglycolic acid mesh was used to facilitate the tissue manipulation and implantation. In the morphological study all the implants appeared viable, but the grafts with longer implantations periods were better conformed, showing a tisular structure with multiple cellular layers.

CONCLUSIONS

In vitro keratinocyte culture techniques could be applied to other epithelial tissues as the urothelium. We obtained a three-dimensional in vitro tissue suitable for grafting in a relatively short time. The histological study demonstrated that free autologous urothelial graft is totally viable, opening future clinics applications.

Tissue-engineered urinary bladder wall using PLGA mesh-collagen hybrid scaffolds: a comparison study of collagen sponge and gel as a scaffold

PURPOSE

Tissue engineering of the urinary bladder using autologous cells and biodegradable scaffold is a promising method for augmentation. The authors developed 2 hybrid scaffolds by combining poly (DL-lactic-co-glycolic acid; PLGA) mesh for mechanical strength with collagen sponge or gel suitable for cell seeding. The aim of this study was to compare collagen as a scaffold between collagen sponge and gel and to construct a tissue-engineered urinary bladder wall utilizing these hybrid scaffolds.

METHODS

The PLGA mesh-collagen hybrid scaffolds were prepared by introducing collagen sponge or gel into the PLGA knitted mesh. Urothelial and smooth muscle cells were obtained from porcine urinary bladder wall and were cultured in their respective media. The cells were seeded on these hybrid scaffolds. These constructs were analyzed morphologically and immunohistochemically.

RESULTS

The urothelial layer was generated 3 dimensionally by culturing urothelial cells with PLGA mesh and collagen sponge. The smooth muscle layer was constructed by culturing smooth muscle cells with PLGA mesh and collagen gel. And a novel tissue-engineered urinary bladder wall was constructed laminating the urothelial and smooth muscle layers.

CONCLUSIONS

Ex vivo construction of urinary bladder wall using hybrid scaffolds prepared by combining PLGA mesh with collagen sponge or gel was successful. This tissue-engineered urinary bladder wall allows easy handling and may become a promising tool for bladder augmentation.

Modulation of corneal epithelial stratification by polymer surface topography

The topography and porosity of a polymer may affect the epithelialization of a corneal implant. We used an in vitro model to examine the effect of polymer surface topography on corneal epithelial tissue stratification and the deposition of proteins associated with epithelial adhesion. A range of topographies was provided by polycarbonate membranes with nominal pore diameters of 0.1, 0.4, 0.8, 1.0, 2.0, or 3.0 microm and a nonporous surface. Stratification of epithelial tissue outgrowth on these surfaces was evaluated using light and electron microscopy. Deposition of proteins associated with basement membrane and adhesion complex formation at the tissue-polymer interface was assessed using immunohistochemistry. Surfaces with pores in the 0.1-0.8-microm-diameter range supported superior stratification and protein deposition compared with those containing pores of > or = 1.0 microm. Cytoplasmic processes penetrated single pores 2.0 and 3.0 microm in diameter and fused pores 1.0 microm in diameter. Tissue on the nonporous surface had a lower level of stratification compared with surfaces with pores 0.1-0.8 microm in diameter. These results point to the significance of surface topography in biomaterial applications that require persistent epithelialization.

Primary uroepithelial cultures. A model system to analyze umbrella cell barrier function

Despite almost 25 years of effort, the development of a highly differentiated and functionally equivalent cell culture model of uroepithelial cells has eluded investigators. We have developed a primary cell culture model of rabbit uroepithelium that consists of an underlying cell layer that interacts with a collagen substratum, an intermediate cell layer, and an upper cell layer of large (25-100 micrometer) superficial cells. When examined at the ultrastructural level, the superficial cells formed junctional complexes and had an asymmetric unit membrane, a hallmark of terminal differentiation in bladder umbrella cells. These cultured "umbrella" cells expressed uroplakins and a 27-kDa uroepithelial specific antigen that assembled into detergent-resistant asymmetric unit membrane particles. The cultures had low diffusive permeabilities for water (2.8 x 10(-4) cm/s) and urea (3.0 x 10(-7) cm/s) and high transepithelial resistance (>8000 Omega cm2) was achieved when 1 mM CaCl2 was included in the culture medium. The cell cultures expressed an amiloride-sensitive sodium transport pathway and increases in apical membrane capacitance were observed when the cultures were osmotically stretched. The described primary rabbit cell culture model mimics many of the characteristics of uroepithelium found in vivo and should serve as a useful tool to explore normal uroepithelial function as well as dysfunction as a result of disease.

In vitro assessment of a collagen sponge for engineering urothelial grafts

OBJECTIVE

To investigate the deposition of urinary crystals and the growth characteristics of urothelial cells on a collagen sponge, as a preliminary step in engineering urothelial autologous grafts.

MATERIALS AND METHODS

Collagen sponges were exposed to a continuous flow of urine at pH 5.3 and 6.3 for 1 week. The sponges were examined microscopically for crystal deposition and analysed for their calcium content. Two cell lines, RT112, derived from a well-differentiated transitional cell carcinoma, and UROtsa, an immortalized urothelial cell line, were seeded on the collagen sponges. Cells were cultured for 6, 12 and 21 days. The pattern of growth was analysed by histology and immunostaining with a pan-cytokeratin antibody. Growth was assayed to quantify cell proliferation on the sponges.

RESULTS

No crystals were evident on any of the collagen sponges. Calcium deposition was negligible at pH 5.3. Although calcium levels were measurable at pH 6.3, the levels were very low. Both cell lines attached and grew in a stratified manner on the collagen sponge, RT112 forming a layer 6-8 cells thick, and UROtsa a layer 4-6 cells thick; cell proliferation was maximal at 5-10 days. The sponge remained easy to handle after 3 weeks in culture.

CONCLUSION

These findings show that collagen sponges support the growth and stratification of urothelial cells, and indicate that the collagen sponge is a suitable substrate for developing urothelial autologous grafts.

Proteoglycan core protein syndecan in bladder biopsies

The etiology of interstitial cystitis (IC) may be related to a dysfunctional epithelium caused by an abnormal permeability barrier. The presence of deleterious urinary substances (quaternary amines) that alter an otherwise normal epithelium may also be contributory. IC disease could reflect an inability of the bladder to repair its protective surface-coat material (glycosaminoglycans and proteoglycans), which is constantly exposed to a toxic urine environment. Bladder biopsy tissue from IC patients and derived explant cells were investigated to determine if mRNA for a proteoglycan core protein could be extracted and evaluated by reverse transcriptase-polymerase chain reaction (RT-PCR). Syndecan was chosen for this investigation because the available sequence information permitted PCR primers to be synthesized. The results indicated that biopsy tissue and explant cells could be utilized for the isolation of syndecan core protein mRNA. This proteoglycan was also demonstrated in mouse bladders by immunostaining and immunoblotting (but not in human tissues) using a syndecan-specific monoclonal antibody (281-2). Quantitative differences in IC tissues versus normal bladder tissue with respect to gene expression for this proteoglycan core protein can now be determined.

An in vitro model of intra-epithelial expansion of transformed urothelial cells

Replacement of normal urothelium by pre-cancerous epithelium may explain the high recurrence rate of human bladder cancer. An in vitro model was designed in order to study the mechanisms of expansion of transformed urothelial cells at the expense of normal urothelium. For this purpose, mouse bladder explants were allowed to expand on a transparent porous membrane. Subsequently, cell sheets of the non-tumorigenic mouse urothelial cell line NUC-5, the tumorigenic mouse urothelial cell line NUC-5 Py and the human bladder-carcinoma cell line T24 were inoculated adjacent to the primary explant. Daily measurements of the outgrowth of the bladder explant were performed, and all cultures were terminated on day 24. At this time the post-confluent primary urothelial cell outgrowths still showed proliferative activity, as demonstrated by bromodeoxyuridine incorporation. In due course the non-tumorigenic NUC-5 cells were replaced by the bladder outgrowth. T24 and NUC-5 Py cells were able to inhibit the bladder outgrowth, or even infiltrate or replace the explant. This was confirmed by immunohistochemistry with (species-specific) anti-cytokeratin antibodies and by microscopic evaluation of cross-sections of the porous membrane. This co-cultivation model appears to be suitable for the in vitro study of the mechanisms of intra-epithelial expansion of transformed urothelial cell lines.

Long term growth and maintenance of stratified rat urothelium in vitro

Culture conditions that allow long term growth and maintenance of rat urothelium have been determined using short (3 to 8 days) and long (14 to 60 days) term measurements of cell density and tritiated thymidine incorporation as indices. The basal nutrient medium utilized was a mixture of 199 plus Ham's F 12 (1:1) supplemented with insulin (1 microgram/ml) and hydrocortisone (1 microgram/ml). Long term culture of urothelium seems to require porous collagen. Porous albumin, or plastic dishes thinly coated with albumin, collagen, fibronectin or mixtures thereof, did not support long term maintenance. Serum was required at a concentration of 5%, independent of other additives. Decreasing Ca++ levels below that normally found the basal medium (approximately 1 X 10(-3] to as low as 1 X 10(-4), resulted in increased short term proliferation, but decreased long term maintenance by causing a loss of stratification of the urothelium. Even a slight increase in Ca++ concentration from 1.0 to 1.5 X 10(-3) resulted in an inhibition of proliferation and an increase in the number of large flat cells which subsequently sloughed off in sheets. The deletion of either insulin, hydrocortisone or both, inhibited growth. The addition of epidermal growth factor (EGF) or its homologue, transforming growth factor (TGF-alpha), increased cell proliferation markedly and caused a variable increase in stratification. However, epithelium induced to rapid growth and proliferation with EGF, eventually exhausted its growth potential and died. TGF-beta 1, alone or in combination with either EGF or alpha-TGF, had no additional effect upon urothelial growth. Repeated transfers of urothelium by enzymatic dissociation led to decreased growth and maintenance potential. The data indicates that long term maintenance of stratified urothelium in culture requires a porous collagen substrate and fetal bovine serum together with hormonal requirements and concentrations of Ca++ that neither greatly stimulate nor inhibit growth.

A biphasic chamber system for maintaining polarity of differentiation of cultured respiratory tract epithelial cells

A simple, disposable, biphasic cultivation chamber has been developed for respiratory tract epithelial cells. This chamber, the Whitcutt chamber, contains a movable, transparent, permeable gelatin membrane that can be employed either submerged in the culture medium, thereby feeding the cells by the traditional immersion method, or raised to the surface of the culture medium, to bring the apical surfaces of the cells into contact with air and provide nutrients only from below (basal feeding). The effects of biphasic cultivation on the growth and differentiation of respiratory tract epithelial cells from different sources have been studied in Whitcutt chambers. Primary hamster tracheal epithelial (HTE) cells grown to confluence with basal feeding developed a ciliated columnar morphology, with differentiated features (cilia and mucous granules) located in the apical region of the epithelial layer. These cells secreted mucinlike molecules from the apical surface (i.e. the surface in contact with air). Although the apical localization of differentiation features was greater, mucous cell differentiation achieved by basal feeding was quantitatively not greater than that achieved by continuous immersion feeding. Similarly, basal feeding did not alter the degree of epithelial cell differentiation in cultures derived from rat, rabbit, and monkey tracheas or from human bronchial and nasal tissues. In contrast, the differentiation of guinea pig tracheal epithelial cells in culture was significantly influenced by the feeding method employed. When fed basally, guinea pig tracheal epithelial cell cultures expressed various mucociliary functions with resemblance to mucociliary layers in vivo, whereas constantly immersed cultures seemed stratified and squamous. These results suggest that, at least for guinea pigs, the combination of feeding methods provided by the Whitcutt chamber can be used to achieve differentiated cultures of tracheal epithelial cells with a polarity of differentiation that is similar to that observed in intact airways in vivo.

Epithelial-stromal interactions in the adult bladder: urothelial growth, differentiation, and maturation on culture facsimiles of bladder stroma

Analysis of the responsiveness of isolated adult urothelium to a series of different stromal cell-extracellular matrix combinations demonstrated the capacity of stromal cells to induce and maintain normal patterns of urothelial growth, differentiation, and maturation in vitro. By incorporating embryonic mesenchymal derived (Swiss 3T3) cells into type I collagen matrices, simplified three-dimensional tissue-like facsimiles of bladder stroma were derived. When recombined with sheets of isolated urothelium these facsimiles could approximately reproduce the capacity of natural stromal tissue to support the expression of normal urothelial tissue specific characteristics. In contrast cocultures between urothelia and monolayers of 3T3 cells, applied to the surface of planar collagen substrata could only permit urothelial cell attachment but not growth or differentiation whereas lethally irradiated 3T3 (feeder) cells, under similar experimental conditions, could support the maintenance of an immature or incompletely differentiated urothelium. Conditioned medium elaborated by cultured 3T3 cells could not stimulate further differentiation in urothelia cultured alone on planar collagen substrata. These studies indicate that a significant portion of the regulatory capacity of the stroma in stromal-urothelial interactions can be accounted for by the activities of a closely applied population of stromal cells, provided the cells are viable and presented to the urothelium in a three-dimensional context in combination with collagen. The capacity of embryonic mesenchymal cells to express properties appropriate to the development of a multilayered terminally differentiated urothelium suggests that normal interactions between adult urothelium and stroma are of limited specificity with the urothelium requiring an essential input of permissive signals only.

An improved method for the preparation and culture of urothelial cells

Urothelial cells have been prepared by a new method involving collagenase treatment of the lumen of a ureter. These cells have been identified as epithelial and successfully subcultured. In addition, we have observed that growth rate is significantly increased by the inclusion of an extract of bovine hypothalamus in the growth medium. This system for cell preparation and culture should greatly facilitate studies involving urothelial cells.

Serial cultivation of normal rat bladder epithelial cells in vitro

Recent advances in culture techniques for human urothelial cells have led to the development of an improved method for growing primary rat bladder epithelial cells. We report here the conditions developed for large-scale in vitro growth and serial cultivation of normal diploid rat bladder epithelial cells. Primary cultures were initiated by attachment of bladder mucosal explants to type I collagen gels. A rapid outgrowth of epithelial cells from the explants occurred when cultured in a hormone-supplemented medium with epidermal growth factor. These primary outgrowths were passaged by nonenzymatic dispersion with 0.1 per cent ethylenediaminetetracetic acid and replating onto new gels. The capacity for routine serial passaging and maintenance of rat bladder epithelial cells required the presence of epidermal growth factor, a requirement not observed with human urothelial cells. The characteristics of the cultured rat bladder epithelial cells were similar to human urothelial cells in: ultrastructural and phase-contrast morphologic properties, showing junctional complexes, desmosomes, stratification and an apical glycocalyx; the absence of stromal cell contamination; and the ability to be serially passaged. Spontaneous cell-line formation was observed with the rat bladder epithelial cells, but has not been found with the human urothelial cells. With the method that we have developed, the number of rat bladder epithelial cells generated from a single bladder of a 4 to 6 week old rat was increased 100-fold from about 7 X 10(5) cells to 7 X 10(7) viable cells within 3 weeks of culture. The capability of culturing normal, primary rat bladder epithelial cells on this scale has not been reported previously and will facilitate comparative studies of the biological and molecular characteristics of the mammalian urothelium. Furthermore, this culture system will be useful for carcinogenesis studies, including metabolic activation of carcinogens and cellular transformation in vitro.

Effect of a collagen substrate on the growth and development of normal and tumorigenic rat urothelial cells

The growth and morphology of 4 tumorigenic rat urothelial cell lines grown on collagen-coated nylon discs was characterized and compared to normal cells. In contrast to cells cultured on a plastic substrate with or without a thin "nonporous" collagen coating, tumor cells grown on porous collagen-coated nylon discs: 1) grew to greater protein densities; 2) formed tissue structures characteristic for the type of tumor they developed upon back-transplantation; and 3) could be grown and cultured indefinitely without subculturing. Thus, similarly to normal urothelial stratification and differentiation in vitro, tumorigenic cells apparently require a "porous" collagen substrate to allow differentiation analogous to that observed in vivo.

Altered growth patterns in vitro of human papillary transitional carcinoma cells

In vitro growth patterns and morphologic characteristics of five low-grade human papillary transitional cell carcinomas (TCCs) were compared and contrasted with those of normal human urothelial cells in culture. Biopsies of TCC were performed by transurethral resection. Specimens of normal human ureters were obtained surgically. Singly dispersed TCC cells grew in 0.3% agarose semisolid medium with a cloning efficiency ranging from 0.02% to 0.71%. Singly dispersed normal ureteral urothelial cells under the same conditions did not form colonies in 0.3% agarose. Neither singly dispersed TCC nor normal urothelial cells formed colonies when plated on collagen-gel substrates. In primary explant culture, normal human urothelial cells grew rapidly, to form tightly adherent flat sheets of apparently nonmotile cells. Autoradiographic labeling with 3H-thymidine of growing cultures of normal urothelial cells showed cell division primarily in the zones of growth near the explant. Outgrowth of TCC from primary explants was loosely adherent. One TCC explant culture gave rise to a continuous suspension culture. Numerous multilayered cellular formations of fronds, nodules, and "walls" were seen around the periphery of TCC explant colonies. Autoradiography showed that these multilayered areas of TCC growth contained actively dividing cells. The altered ability of papillary TCC to form superficial multilayered formations in vitro distinguishes them from normal human urothelium and reflects the morphologic characteristic of this tumor type in vivo.

Growth and characterization of normal human urothelium in vitro

A method for initiating rapidly growing cultures of normal human transitional cells from ureter and embryonic bladder specimens has been developed and quantified. A new microdissection technique was used to nonenzymatically separate the urothelium. The use of enriched medium containing 10 micrograms/ml insulin, 5 micrograms/ml transferrin, and 1 microgram/ml hydrocortisone resulted in improved growth. The use of thin collagen gel substrates (0.6 ml/60 mm petri dish) resulted in 97% attachment of explants compared to 77% attachment on plastic. Explants grown on thicker collagen (2 ml/60 mm petri dish) showed, in addition to better attachment, enhanced growth of cells as determined both by measurements of colony size and cell density. Cultures of transitional cells that were initiated using explants could be passed three to five times using 0.1% EDTA for dispersion. Autoradiography of [3H]thymidine-labeled cells showed an initial phase of rapid cell division in primary explant cultures and restimulation of cell division in passaged cultures. Transmission electron microscopy showed that the cells growing out from the explants were continuous with the stratified urothelium maintained in the original explant. Stratification of transitional cells occurred in cultures of both ureter and embryonic bladder cells. Surface cells were joined near their apices by junctional complexes. Desmosomes and Golgi vesicles were present in all cells. Passage in culture did not alter the morphological characteristics of cells.

Polarization of thyroid cells in culture: evidence for the basolateral localization of the iodide "pump" and of the thyroid-stimulating hormone receptor-adenyl cyclase complex

When cultured in collagen gel-coated dishes, thyroid cells organized into polarized monolayers. The basal poles of the cells were in contact with the collagen gel, whereas the apical surfaces were facing the culture medium. Under these culture conditions, thyroid cells do not concentrate iodide nor respond to acute stimulation by thyroid-stimulating hormone (TSH). To allow the free access of medium components to the basal poles, the gel was detached from the plastic dish and allowed to float in the culture medium. After release of the gel, the iodide concentration and acute response to TSH stimulation were restored. Increased cAMP levels, iodide efflux, and formation of apical pseudopods were observed. When the thyroid cells are cultured on collagen-coated Millipore filters glued to glass rings, the cell layer separates the medium in contact with the apical domain of the plasma membrane (inside the ring) from that bathing the basolateral domain (outside the ring). Iodide present in the basal medium was concentrated in the cells, whereas no transport was observed when iodide was added to the luminal side. Similarly, an acute effect of TSH was observed only when the hormone was added to the basal medium. These results show that the iodide concentration mechanism and the TSH receptor-adenylate cyclase complex are present only on the basolateral domain of thyroid cell plasma membranes.

Morphological differentiation of an embryonic epithelium in culture

This communication reports the results of a morphological study of three-day old cultures of epiblast tissue from the early chick embryo. The most striking feature of these cultures was the appearance of "domes" or elevated blister-like structures, composed of a single layer of cells which were morphologically distinct from the remaining cells in the culture. The domes arose in high-density areas of the culture. Their roofs were lined by basal laminae that did not develop in other areas of the culture. In several morphological respects, the cells of the dome roof closely resembled the epiblast in vivo. This was in contrast to the cells spread on the substratum in sparse regions of the culture, which did not. Each dome was surrounded by a dense ring of multilayered ruffling cells which appeared to give rise to both the dome roof and to fibroblast-like cells that spread on the substratum beneath the dome. Fibroblast-like cells also developed in discrete patches in other regions of the culture. In other tissues, dome formation has been attributed to fluid transport by the epithelium; in the present case it is also possible to invoke the capacity of the epiblast to fold, as contributing to the mechanism of dome formation.

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 cells 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.