Ultrastructural and cytochemical observations on 5-fluorouracil-induced cleft-palate development in hamster

Sequential alterations in 5-fluorouracil-treated hamster fetal palate were studied by light and electron microscopy and by acid phosphatase cytochemistry. At an early stage in 5-fluorouracil-treated fetuses, when the palatal shelves were vertical, lysosomes first appeared in cells of the prospective fusion epithelium and then in the cells of subjacent mesenchyme. In contrast to controls, increasing numbers of both the epithelial and mesenchymal cells of the vertical palate showed lysosomal injury in 5-fluorouracil-treated fetuses as development progressed. Subsequently, the basal lamina in the vertical palate showed alterations, characterized initially by disturbances in lamina lucida, by fingerlike extensions of lamina densa, and ultimately by its complete breakdown. At a later stage, when shelves became horizontal, the lysosomes were absent in both the epithelial and mesenchymal cells, and the basal lamina continuity was restored. Unlike controls, however, 5-fluorouracil-treated horizontal shelves never contacted one another. Instead, the epithelia of the horizontal shelves underwent stratification. It appears that premature formation of lysosomes in palatal epithelial and mesenchymal cells following 5-fluorouracil treatment disrupts normal cytodifferentiation and affects the integrity of the basal lamina; both effects are associated with cleft-palate development.     

Structure of the epithelial - mesenchymal interface during early morphogenesis of the chick duodenum.

During the period of early morphogenetic folding of the intestinal epithelium, changes in the epithelial-mesenchymal interface were observed by light microscopy, scanning and transmission electron microscopy. The epithelium in cross-section, appears first as a circle, then an ellipse and finally by a triangle prior to the formation of the first three previllous ridges. The bases of all epithelial cells are flat at the circular stage. At the ellipse and triangle stages the bases of the epithelial cells occupying the sides possess lobopodia that do not penetrate the basal lamina. The immediate mesenchymal cells subjacent to those epithelial cells on the sides of the ellipse and triangle alter their orientation to being rounded-up or perpendicular to the plane of the basal lamina. Large numbers of fine mesenchymal pseudopodia in addition to many extracellular fibrils are revealed by transmission and scanning electron microscopy at the epithelial-mesenchymal interface. The fine mesenchymal pseudopodia come into close contact but do not penetrate the ruthenium red-staining basal lamina. The possible roles of close contact between epithelium and mesenchyme, the alteration in orientation of mesenchyme cells, and of the basal lamina in tissue interaction are discussed.     

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.     

Formation of cytoskeletal elements during mouse embryogenesis

Abstract. The ultrastructure of the day 8.5 mouse embryo has been studied by transmission electron microscopy, with special emphasis on the primary mesenchymal cells and their interaction with cells of the embryonic ectoderm and the proximal endoderm. The organization of the two polar epithelial cell layers (embryonic ectoderm and proximal endoderm), the isolated cells of the distal endoderm and the primary mesenchymal cells is described. Primary mesenchymal cells are different from embryonic ectoderm cells, from which they are derived, not only by the absence of desmosomes and intermediate-sized filaments of the cytokeratin type but also by their variable morphology not exhibiting stable polar architecture, and their numerous cytoplasmic processes which make contacts with the basal lamina of the ectoderm, the basal cell surface of the proximal endoderm, and other mesenchymal cells. Over most of the embryo the embryonic ectoderm is covered by a typical basal lamina, except for certain regions that are frequently characterized by cytoplasmic projections ('blebs') from the basal cell surface membrane. In contrast, the basal surface of the proximal endoderm is not covered by a continuous basal lamina and reveals mushroom-like protrusions of the cortical cytoplasm. Junctions between primary mesenchymal cells are numerous and include adhaerens-type formations of various sizes as well as gap junctions. Occasionally, a special type of junction between mesenchymal cells and embryonic ectoderm has been found, resulting in local interruptions of the basal lamina. The observations are discussed in relation to possible mechanisms of mesoderm formation and the drastic changes of cell character that accompany this process, including cytoskeletal changes such as the disappearance of cytokeratin filaments and the expression of vimentin.     

Formation of cytoskeletal elements during mouse embryogenesis. IV. Ultrastructure of primary mesenchymal cells and their cell-cell interactions

The ultrastructure of the day 8.5 mouse embryo has been studied by transmission electron microscopy, with special emphasis on the primary mesenchymal cells and their interaction with cells of the embryonic ectoderm and the proximal endoderm. The organization of the two polar epithelial cell layers (embryonic ectoderm and proximal endoderm), the isolated cells of the distal endoderm and the primary mesenchymal cells is described. Primary mesenchymal cells are different from embryonic ectoderm cells, from which they are derived, not only by the absence of desmosomes and intermediate-sized filaments of the cytokeratin type but also by their variable morphology not exhibiting stable polar architecture, and their numerous cytoplasmic processes which make contacts with the basal lamina of the ectoderm, the basal cell surface of the proximal endoderm, and other mesenchymal cells. Over most of the embryo the embryonic ectoderm is covered by a typical basal lamina, except for certain regions that are frequently characterized by cytoplasmic projections ("blebs') from the basal cell surface membrane. In contrast, the basal surface of the proximal endoderm is not covered by a continuous basal lamina and reveals mushroom-like protrusions of the cortical cytoplasm. Junctions between primary mesenchymal cells are numerous and include adhaerens-type formations of various sizes as well as gap junctions. Occasionally, a special type of junction between mesenchymal cells and embryonic ectoderm has been found, resulting in local interruptions of the basal lamina. The observations are discussed in relation to possible mechanisms of mesoderm formation and the drastic changes of cell character that accompany this process, including cytoskeletal changes such as the disappearance of cytokeratin filaments and the expression of vimentin.     

Basal lamina formation by porcine thyroid cells grown in collagen- and laminin-deficient medium

Summary Porcine thyroid cells isolated by dispase treatment were cultured in either (a) Matrigel, (b) agarose with the addition of different combinations of basic fibroblast growth factor and laminin, or (c) on agarose-coated dishes. The formation of follicles and the presence of a basal lamina was investigated by routine electron microscopy of Araldite-embedded material and by light and electron microscopical immunocytochemical detection of the basal lamina components, laminin and collagen type IV. After 10 days of culture in Matrigel or agarose, a basal lamina-like structure surrounded most follicles. Follicles of cells growing in agarose and overlaid with a medium containing thyrotropin and fibroblast growth factor showed a fluorescent band at the basal side of the follicles after immunocytochemical staining with anti-laminin and anti-collagen antibodies. Routine electron microscopy showed that a basal lamina-like structure lined the outside of the follicle. This structure could be subdivided into a lamina lucida and a lamina densa. Electron microscopical immunogold labelling revealed that immunologically detectable laminin was confined to the lamina densa. These findings suggest that even in the absence of basal lamina components in the culture medium, thyroid cells are able to form follicles with a regular basal lamina when they are cultured in a three-dimensional environment.     

Basal lamina formation on thyroid epithelia in separated follicles in suspension culture

When thyroid follicles are isolated by collagenase treatment of minced thyroid lobes, the basal lamina around each follicle is removed. The basal lamina does not reform when follicles are cultured in suspension in Coon's modified Ham's F-12 medium containing, in addition, 0.5% calf serum, insulin, transferrin, and thyrotropin. We have added acid soluble collagen and/or laminin to see if they would result in the formation of a basal lamina. An extended basal lamina did not form when follicles were embedded in a gel formed from acid-soluble rat tendon collagen or from calf skin collagen when added at a concentration of 100 micrograms collagen/ml. However, laminin at a concentration of 5.1 micrograms/ml gave rise to short segments of a basal lamina within 30 min. At longer time intervals, the segments lengthened and covered the base of many cells, and were continuous across the gap between cells and across the mouth of a coated pit. Not all basal surfaces were covered, and no exposed apical surfaces with microvilli had a basal lamina. There was no obvious difference in the appearance of the basal lamina if collagen was added in addition to laminin, but collagen, in contact with the plasma membrane when added alone, was lifted off the membrane in the presence of the basal lamina. The basal lamina appeared denser if formed in the presence of 5% serum instead of 0.5%.     

Radiation-induced changes of fibroblasts in the squamous cell carcinoma of the head and neck

The regrowth of mesenchymal tissue (stroma) surrounding the malignant epithelium is an important step in tissue remodelling during and after irradiation. The radiation-induced fibroblastic changes were studied on tissue samples taken before, during and after the radical irradiation of the squamous cell carcinoma of the head and neck. Elongated fibroblasts with large amount of rough endoplasmic reticulum were seen around the tumor epithelium before radiation. The fibrosis increased during irradiation and at the same time the shape of the fibroblasts changed so that they became more triangular and nuclear structures became more prominent together with hyperchromasia. The amount of cell organelles declined although there was a large amount collagen present. Epithelial cells invaded through the basal lamina. In most samples the basal lamina could not be seen at all and the tumor cells were dispersed between stromal elements. On the other hand there were close contacts between epithelial and mesenchymal cells throughout the study in places where the basal lamina was broken, which might indicate epithelio-mesenchymal interaction. Also the connective tissue formed by fibroblasts and collagen might be part of the radiation induced healing and destruction of the tumor cells.     

Ultrastructure of neural crest formation in the midbrain/rostral hindbrain and preotic hindbrain regions of the mouse embryo

In the mouse embryo, neural crest mesenchyme associated with the first and second pharyngeal arches escapes from the epithelium that forms the tips of the midbrain/rostral hindbrain and preotic hindbrain neural folds. To investigate the ultrastructure of crest formation, embryos with four to eight pairs of somites were processed for transmission electron microscopy. In the earliest event related to crest formation, crest precursors in the midbrain/rostral hindbrain elongated and moved all or most of their contents to the basal region of the epithelium. Elongation was probably mediated by apical bands of microfilaments and longitudinally oriented microtubules. Elongated cells then relinquished apical associations while nonelongated cells maintained those associations and withdrew from the basal lamina. This resulted in an epithelium stratified into apical and basal (crest precursor) layers. The coalescence of enlarging extra-cellular spaces opened a delaminate gap between the two layers. Additional crest precursors entered this gap from the apical layer. From the time crest precursors began moving basally, some formed microfilament- and/or microtubule-containing processes, which penetrated the basal lamina. Some of these cells moved their contents into the larger, microtubule-containing processes, perhaps thereby escaping from the epithelium. Soon after elongating cells appeared, the basal lamina beneath the epithelium began to degrade in a pattern unrelated to process formation. This ultimately resulted in disruption of the lamina, dispersal of the basal layer of the epithelium, and release of the crest precursors in the delaminate gap. Once crest formation was complete, the apical layer reformed a basal lamina on a patch-by-patch, cell-by-cell basis. In the preotic hindbrain, elongating crest precursors apparently forced their basal faces through the basal lamina and then relinquished apical association to escape. As a result, the lamina was disrupted before the epithelium could stratify, and enlarged extracellular spaces appeared among mesenchymal cells rather than creating a delaminate gap. The failure of elongation to disrupt the basal lamina in the midbrain/rostral hindbrain and its success in the preotic hindbrain might be due to less-vigorous, less-concerted elongation in the midbrain/rostral hindbrain or to earlier, more rapid degradation of the lamina in the preotic hindbrain.     

Ultrastructural analysis of basal neuroepithelial cells in dysraphic mice

Ultrastructural aspects of the cellular pathology in the basal neuroepithelium of the hindbrain and spinal cord were analyzed in dysraphic loop-tail mice at nine days of gestation. Whereas the basal cytoplasm of the neuroepithelium in normal littermates showed a consistent electron density, the neuroepithelium in abnormal embryos was characterized by "light" and "dark" cells scattered randomly along the basal aspect of the hindbrain and spinal cord. In the abnormals, gaps occurred in the neural basal lamina, and the neuroepithelial cells often were in direct contact with cytoplasmic processes from mesenchymal cells and from notochordal cells; in normal littermates, contact was observed only between the intact and continuous neural basal lamina and mesenchymal cells and notochordal cells. Thus, it is possible that the pathological features observed ultrastructurally in the basal neuroepithelium in dysraphic embryos may represent faulty tissue interaction with adjacent notochordal and mesenchymal cells.     

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.     

Fibroblasts promote the formation of a continuous basal lamina during myogenesis in vitro

Analyses were made of the requirements for the formation of a continuous basal lamina during myogenesis of quail muscle in vitro. A culture system was developed in which mass cultures of differentiating muscle cells were embedded in a native gel of rat tail collagen. Fibroblastic cells, which were also present in the cultures, migrated into the gel and within a few days surrounded the newly formed myotubes. In this environment, a continuous basal lamina was formed at the surface of the myotubes as demonstrated by immunofluorescent staining with monoclonal antibodies against type IV collagen, laminin, and heparan sulfate, as well as by electron microscopic immunolocalization. To distinguish between the role of the fibroblasts and the collagen gel in promoting basal lamina formation, clones of quail muscle cells lacking fibroblasts were subsequently embedded in a native rat tail collagen gel. Under these conditions, only very limited fluorescent staining for basement membrane components was observed associated with the myotubes. However, the introduction of chick muscle or skin fibroblasts into the clonal cultures just before gel formation resulted in the formation of an extensive basal lamina on the surface of the myotubes. Conditioned medium from fibroblast cultures by itself was not effective in promoting basal lamina formation. These results clearly show that during myogenesis in vitro fibroblasts must be in close proximity to the myotubes for a continuous basal lamina to form. These results probably relate closely to the interactions that must occur during myogenesis in vivo between the muscle cells and the surrounding connective tissue including the developing tendons.     

Ultrastructural and immunofluorescence studies of basal-lamina alterations during mouse-lung morphogenesis

Epithelial differentiation during lung development appears to be influenced by mesenchyme-derived instructions coupled with hormonal regulations. The basal lamina which is associated with progenitor and differentiating epithelia during mouse embryogenesis (Theiler-stages 16-28) was examined by transmission electron microscopy and indirect-immunofluorescence microscopy. During the embryonic phase of lung development, progenitor epithelia for the pulmonary acinus projected microvilli or cytoplasmic "feet" through the basal lamina, which resulted in discontinuities and a close approximation of the adjacent mesenchymal-cell processes. These changes were also associated with the transitory polarization of mesenchymal cells perpendicular to the plane of the basal lamina, which resulted in a sheet of cuboidal mesenchymal cells adjacent to the developing acinar-tubule epithelium. During the embryonic phase of lung development, these specific interstitial or mesenchymal cells stained for heparan-sulfate proteoglycans; no other cell types were immunostained. By Theiler-stage 25, the acinar-tubule epithelia had differentiated into type-II pneumonocytes which contained lamellar bodies and significant amounts of glycogen. Fibronectin, laminin, and heparan-sulfate proteoglycan were localized in the basement membranes during the embryonic, canalicular, and terminal sac phases of lung morphogenesis. A diffuse localization of fibronectin of the interstitial cell surfaces was observed. These observations indicate that major changes in the structure and composition of basal lamina occur during the embryonic and fetal phases of pulmonary-acinus epithelial-cell differentiation and the production of pulmonary surfactant. The major changes in the basal lamina may be partly mediated by mesenchyme-derived instructions for type-II epithelial-cell differentiation.     

Epithelial-mesenchyme interactions during odontogenesis : IV. Morphological evidence for direct heterotypic cell-cell contacts

During embryonic and neonatal mouse incisor tooth morphogenesis, direct epithelial-mesenchymal cell contacts were observed by electron microscopy. These direct contacts were evident along the epithelial-mesenchymal interface in the differentiation zone in which inner enamel epithelium was as yet a dividing cell population which had not as yet synthesized and secreted the enamel organic matrix. This region of cell differentiation was also characterized by the appearance of cell processes which extended from the epithelia through the basal lamina. Following the appearance of epithelial cell processes penetrating through the basal lamina, ectomesenchymal cell processes extended across the extracellular matrix and penetrated through the basal lamina and resulted in the formation of contact zones. Following degradation of the basal lamina, the mesenchymal cell processes penetrated into clefts within the preameloblast cells and formed cell contacts. By a combination of tannic acid and uranium acetate staining we observed that the tannic acid stain penetrated through intercellular spaces formed between the apposing mesenchymal and epithelial plasma membrane surfaces. We speculate that direct heterotypic cell contacts, which occur prior to the cessation of preameloblast cell division and precede the secretion of enamel proteins, may be instructive in the induction of enamel protein biosynthesis.     

Affiliation of large numbers of fibroblasts with larval muscle fibers

Muscle fibers from fourth and fifth instar caterpillars were examined with scanning and thin section electron microscopy. Scanning micrographs showed that early fifth instar specimens had a population of cells lying beneath the basal lamina over the surface of the muscle fiber and in conjunction with tracheoles and nerves. At least two cell types were present. One type could be categorized as tracheoblasts of their close association with the tracheoles and the presence of taenidia within the tracheoblast cytoplasm in sectioned material. A second cell type, characterized by long filamentous processes, contained extensive rough endoplasmic reticulum and cisternae swollen with an electron-dense substance similar in appearance to the basal lamina. This ultrastructural appearance is characteristic of vertebrate fibroblasts and certain types of insect hemocytes. Early and late fourth instar specimens had few cells on their muscle fiber surfaces. Measurements of the basal lamina thickness were taken from thin sections of nondigested muscle fibers of early fourth, late fourth, and early fifth instar animals. The results showed that the basal lamina underwent a large increase in thickness between the fourth and fifth instars. The proliferation of cells which appeared to be in an actively synthesizing state paralleled the increase in basal lamina thickness. This suggests the hypothesis that these cells are active in connective tissue formation, and contribute to the formation of the basal lamina that lies over both them and the muscle fiber.     

Light and electron microscopic observations on hydrocortisone-induced cleft palate in hamsters.

Palatal histogenesis in hydrocortisone-treated hamster fetuses was studied by both light and electron microscopy. At an early stage in the hydrocortisone-affected fetuses, when the palatal shelves hung vertically on either side of the tongue, necortic changes could be seen in some of the basal epithelial cells which lay adjacent to the fragmented basal lamina. The normal looking cells lay on an intact basal lamina and were attached to the contiguous necrotic cells by desmosomes. With horizontal reorientation of the palatal shelves and their approach to the midline, cellular necrosis and fragmentation of the basal lamina increased. When compared with normal cells, the hydrocortisone-affected ones were seen to be lighter, to contain fewer ribosomes and no lysosomes. At a later stage, when midline palatal fusion was lacking, the epithelium underwent stratification and keratinization while the necrotic debris was removed by mesenchymal macrophages. It appears that the normal process of protein synthesis is inhibited following hydrocortisone administration and that this, in turn, during palatogenesis, disrupts normal cellular differentiation and the integrity of the basal lamina, which are associated with the production of a cleft palate.     

Temporary disruption of the retinal basal lamina and its effect on retinal histogenesis.

An experimental paradigm was devised to remove the retinal basal lamina for defined periods of development: the basal lamina was dissolved by injecting collagenase into the vitreous of embryonic chick eyes, and its regeneration was induced by a chase with mouse laminin-1 and alpha2-macroglobulin. The laminin-1 was essential in reconstituting a new basal lamina and could not be replaced by laminin-2 or collagen IV, whereas the macroglobulin served as a collagenase inhibitor that did not directly contribute to basal lamina regeneration. The regeneration occurred within 6 h after the laminin-1 chase by forming a morphologically complete basal lamina that included all known basal lamina proteins from chick embryos, such as laminin-1, nidogen-1, collagens IV and XVIII, perlecan, and agrin. The temporary absence of the basal lamina had dramatic effects on retinal histogenesis, such as an irreversible retraction of the endfeet of the neuroepithelial cells from the vitreal surface of the retina, the formation of a disorganized ganglion cell layer with an increase in ganglion cells by 30%, and the appearance of multiple retinal ectopias. Finally, basal lamina regeneration was associated with aberrant axons failing to correctly enter the optic nerve. The present data demonstrate that a transient disruption of the basal lamina leads to dramatic and probably irreversible aberrations in the histogenesis in the developing central nervous system.     

Mesenchyme formation from the trigeminal placodes of the mouse embryo

The trigeminal placode is a thickened region of ectodermal epithelium located along the side of the embryonic head. Mesenchyme escapes from the placode to form neurons of the trigeminal (V) ganglion. To further our knowledge of the morphogenesis of this escape, plastic thick sections were cut from mouse embryos and stained for light microscopy by using a technique which revealed escaping mesenchyme. The escape of trigeminal mesenchyme began at approximately 12 somites of age and was substantially complete by 30 somites. These results provided spatial/temporal orientation for a subsequent electron microscopic study. The first ultrastructural manifestation of escape was the penetration of an otherwise continuous basal lamina by small cell processes. The presence of longitudinally oriented microtubules within these processes suggests that mesenchymal cells escape through the basal lamina by using microtubules to direct/move their contents (e.g., the cell nucleus) into an enlarging process. Nuclei were distorted as they passed into these processes. This distortion suggests that basal lamina, together with a possible contribution from basal microfilaments, forms a rigid obstruction which is disrupted in the region from which a process is formed. In some cases a collar of basal lamina was observed around the necks of processes, but their distal membranes were invariably lamina-free. This lamina-free membrane is possibly that which is newly formed to accommodate the growing process. In later stages of escape, instances were observed in which the lamina was completely absent beneath an escaping cell and partially degraded beneath adjacent cells as well. These instances suggest that enzymatic digestion may play a role in degrading the lamina during mesenchymal escape. Apical desmosomes were often retained beyond the initial stages of escape. Mechanisms involved in their disruption are thus not among those which initiate escape.     

Relationship between the saccus endolymphaticus and the fourth ventricle of the brain of the domestic fowl (Gallus gallus f. domestica)

The fine structure of the wall of the SE was determined exactly and its relationship to the cisternae (the evaginations of the roof of the fourth ventricle extending to the SE) was defined. The way in which the cisterna is formed was defined and the development of its fine structure was described by comparing serial sections from 19-day embryos and adult fowls. Like the SE, the cisternae are lodged in the angle between the cerebellum and the medulla oblongata, in the subarachnoid space. The terminal segment of the cisterna lies in the immediate vicinity of the mesenchymal epithelium bordering the basal labyrinth of the SE cells. Collagen trabeculae keep the SE and the cisternae suspended in the subarachnoid space. The cisternae and trabeculae are wrapped in mesenchymal epithelium. The cisterna is avascular and does not communicate with the SE. The cisterna is lined internally with simple squamous epithelium (modified neural epithelium of the roof of the fourth ventricle). The bodies of the cells bulge into the lumen of the cisterna in the region of localization of their nucleus. The epithelium is seated on a pronounced basal lamina. The surface turned towards the subarachnoid space is lined continuously with mesenchymal epithelium without a basal lamina. The cells of the cisternal epithelium are connected by tight junctions of the type of zonulae occludentes and desmosomes. The basal lamina is continuous and distinct. The mesenchymal epithelium of the subarachnoid space has no basal lamina, as on the subarachnoid surface of the SE, the cisternae, the trabeculae, the pia mater and the arachnoidea.(ABSTRACT TRUNCATED AT 250 WORDS)     

Computational modeling of epithelial–mesenchymal transformations

An epithelial–mesenchymal transformation (EMT) involves alterations in cell–cell and cell–matrix adhesion, the detachment of epithelial cells from their neighbors, the degradation of the basal lamina and acquisition of mesenchymal phenotype. Here we present Monte Carlo simulations for a specific EMT in early heart development: the formation of cardiac cushions. Cell rearrangements are described in accordance with Steinberg's differential adhesion hypothesis, which states that cells possess a type-dependent adhesion apparatus and are sufficiently motile to give rise to the tissue conformation with the largest number of strong bonds. We also implement epithelial and mesenchymal cell proliferation, cell type change and extracellular matrix production by mesenchymal cells. Our results show that an EMT is promoted more efficiently by an increase in cell–substrate adhesion than by a decrease in cell–cell adhesion. In addition to cushion tissue formation, the model also accounts for the phenomena of matrix invasion and mesenchymal condensation. We conclude that in order to maintain epithelial integrity during EMT the number of epithelial cells must increase at a controlled rate. Our model predictions are in qualitative agreement with available experimental data.