Differential function of N-cadherin and cadherin-7 in the control of embryonic cell motility.

Similar amounts of N-cadherin and cadherin-7, the prototypes of type I and type II cadherin, induced cell-cell adhesion in murine sarcoma 180 transfectants, Ncad-1 and cad7-29, respectively. However, in the initial phase of aggregation, Ncad-1 cells aggregated more rapidly than cad7-29 cells. Isolated Ncad-1 and cad7-29 cells adhered and spread in a similar manner on fibronectin (FN), whereas aggregated cad7-29 cells were more motile and dispersed than aggregated Ncad-1 cells. cad7-29 cells established transient contacts with their neighbors which were stabilized if FN-cell interactions were perturbed. In contrast, Ncad-1 cells remained in close contact when they migrated on FN. Both beta-catenin and cadherin were more rapidly downregulated in cad7-29 than in Ncad-1 cells treated with cycloheximide, suggesting a higher turnover rate for cadherin-7-mediated cell-cell contacts than for those mediated by N-cadherin. The extent of FN-dependent focal adhesion kinase phosphorylation was much lower if the cells had initiated N-cadherin-mediated rather than cadherin-7-mediated cell adhesion before plating. On grafting into the embryo, Ncad-1 cells did not migrate and remained at or close to the graft site, even after 48 h, whereas grafted cad7-29 cells dispersed efficiently into embryonic structures. Thus, the adhesive phenotype of cadherin-7-expressing cells is regulated by the nature of the extracellular matrix environment which also controls the migratory behavior of the cells. In addition, adhesions mediated by different cadherins differentially regulate FN-dependent signaling. The transient contacts specifically observed in cadherin- 7-expressing cells may also be important in the control of cell motility.

Initial appearance and regional distribution of the neuron-glia cell adhesion molecule in the chick embryo.

This study represents a global survey of the times of the first appearance of the neuron-glia cell adhesion molecule (Ng-CAM) in various regions and on particular cells of the chick embryonic nervous system. Ng-CAM, originally characterized by means of an in vitro binding assay between glial cells and brain membrane vesicles, first appears in development at the surface of early postmitotic neurons. By 3 d in the chick embryo, the first neurons detected by antibodies to Ng-CAM are located in the ventral neural tube; these precursors of motor neurons emit well-stained fibers to the periphery. To identify locations of appearance of Ng-CAM in the peripheral nervous system (PNS), we used a monoclonal antibody called NC-1 that is specific for neural crest cells in early embryos to show the presence of numerous crest cells in the neuritic outgrowth from the neural tube; neither these crest cells nor those in ganglion rudiments bound anti-Ng-CAM antibodies. The earliest neurons in the PNS stained by anti-Ng-CAM appeared by 4 d of development in the cranial ganglia. At later stages and progressively, all the neurons and neurities of the PNS were found to contain Ng-CAM both in vitro and in vivo. Many central nervous system (CNS) neurons also showed Ng-CAM at these later stages, but in the CNS, the molecule was mostly associated with neuronal processes (mainly axons) rather than with cell bodies; this regional distribution at the neuronal cell surface is an example of polarity modulation. In contrast to the neural cell adhesion molecule and the liver cell adhesion molecule, both of which are found very early in derivatives of more than one germ layer, Ng-CAM is expressed only on neurons of the CNS and the PNS during the later epoch of development concerned with neural histogenesis. Ng-CAM is thus a specific differentiation product of neuroectoderm. Ng-CAM was found on developing neurons at approximately the same time that neurofilaments first appear, times at which glial cells are still undergoing differentiation from neuroepithelial precursors. The present findings and those of previous studies suggest that together the neural cell adhesion molecule and Ng-CAM mediate specific cellular interactions during the formation of neuronal networks by means of modulation events that govern their prevalence and polarity on neuronal cell surfaces.

Neural crest cell migration: requirements for exogenous fibronectin and high cell density.

Cells of the neural crest participate in a major class of cell migratory events during embryonic development. From indirect evidence, it has been suggested that fibronectin (FN) might be involved in these events. We have directly tested the role of FN in neural crest cell adhesion and migration using several in vitro model systems. Avian trunk neural crest cells adhered readily to purified plasma FN substrates and to extracellular matrices containing cellular FN. Their adhesion was inhibited by antibodies to a cell-binding fragment of FN. In contrast, these cells did not adhere to glass, type I collagen, or to bovine serum albumin in the absence of FN. Neural crest cell adhesion to laminin (LN) was significantly less than to FN; however, culturing of crest cells under conditions producing an epithelioid phenotype resulted in cells that could bind equally as well to LN as to FN. The migration of neural crest cells appeared to depend on both the substrate and the extent of cell interactions. Cells migrated substantially more rapidly on FN than on LN or type I collagen substrates; if provided a choice between stripes of FN and glass or LN, cells migrated preferentially on the FN. Migration was inhibited by antibodies against the cell-binding region of FN, and the inhibition could be reversed by a subsequent addition of exogenous FN. However, the migration on FN was random and displayed little persistence of direction unless cells were at high densities that permitted frequent contacts. The in vitro rate of migration of cells on FN-containing matrices was 50 microns/h, similar to their migration rates along the narrow regions of FN-containing extracellular matrix in migratory pathways in vivo. These results indicate that FN is important for neural crest cell adhesion and migration and that the high cell densities of neural crest cells in the transient, narrow migratory pathways found in the embryo are necessary for effective directional migration.

E-cadherin expression determines the mode of replacement of normal urothelium by human bladder carcinoma cells.

The high recurrence rate of human bladder cancer can be attributed to intraepithelial expansion of tumor cells or shedding and subsequent implantation of tumor cells elsewhere in the bladder. E-Cadherin is a calcium-dependent cell-cell adhesion molecule, and loss of E-cadherin by tumor cells is associated with increased tumor aggressiveness. Here we demonstrate that E-cadherin is also an important determinant of the mechanisms which are involved in the recurrence rate of bladder cancer. In a recently developed in vitro cocultivation model, we studied the effect of E-cadherin expression on the intraepithelial expansion pattern of six different human bladder carcinoma cell lines into primary murine urothelium. Bladder carcinoma cells lacking E-cadherin infiltrate into the primary urothelium as individual cells (pagetoid pattern). In contrast, a sharp demarcation is observed between E-cadherin-positive bladder cancer cells and the primary urothelium (carcinoma in situ pattern). With the same model, we demonstrate that only E-cadherin-positive bladder carcinoma cell lines could attach to and colonize the intact primary urothelium. We hypothesize that it is the latter process that plays an important role in the high recurrence rate that is observed in some of the patients.

Decreased expression of keratinocyte growth factor receptor in a subset of human transitional cell bladder carcinomas.

Growth factors and growth factor receptors are involved in tumor progression. The fibroblast growth factor receptor 2 gene encodes distinct isoforms. The isoforms which bind KGF (keratinocyte growth factor or FGF-7) are called KGF-R or FGFR2b. KGF-R is expressed in different epithelia and is involved in the control of epithelial-mesenchymal interactions. Expression of KGF-R mRNA was examined in normal human bladder and transitional cell carcinoma of the bladder (TCC) by semi-quantitative RT-PCR using TFIID and GAPDH as internal standards. In normal bladder, the KGF-R mRNA was detected in the urothelium but not in the underlying stroma. In TCCs, the level of KGF-R mRNA was generally either normal or low. Eighteen out of 54 TCCs had a KGF-R mRNA level below 30% of that found in normal urothelium. This decrease in KGF-R mRNA was not accompanied by an increase in BEK (FGFR2c) mRNA, the other major splice variant of the fibroblast growth factor receptor 2 gene. Expression of the KGF-R was also monitored by immunohistochemistry using a functional KGF-immunoglobulin chimera. The receptor was uniformly expressed throughout the normal urothelium except for the umbrella cells. Immunoreactivity for KGF-R was found to be negative in tumors with low levels of KGF-R mRNA, while the peritumoral normal urothelium was positive. Among patients with muscle invasive tumors, those exhibiting a low level of KGF-R mRNA had a significantly higher proportion of cancer deaths. Our results suggest that decreased expression of KGF-R can be considered as a marker of tumor progression in muscle invasive TCCs.

Relationship between E-cadherin and fibroblast growth factor receptor 2b expression in bladder carcinomas.

E-cadherin is a cell-cell adhesion molecule expressed predominantly by epithelial cells. Reduction or loss of E-cadherin immunoreactivity has been associated with tumour progression in many epithelial cancers, including bladder carcinomas. The fibroblast growth factor receptor 2b (FGFR2b) recognized specifically by FGF7 is expressed only by epithelial cells. Recently, decreased expression of FGFR2b protein and mRNA was found to be associated with tumour progression in bladder carcinomas. The purpose of this investigation was to look for a possible relationship between E-cadherin and FGFR2b expression in bladder carcinomas. As decreased E-cadherin immunoreactivity was found to correlate directly with decreased expression at the mRNA level, the possible relationship between E-cadherin and FGFR2b was investigated at the mRNA level using semi-quantitative RT - PCR in 92 transitional cell carcinomas (TCCs) and four lymph node metastases. All tumours with low E-cadherin expression had low expression of FGFR2b, whereas tumours with low FGFR2b mRNA could express any level of E-cadherin mRNA. The same observation was equally valid for bladder and colon cancer cell lines suggesting that, besides bladder tumours, this relationship could apply to other carcinomas types. These results suggest that a relationship exists between the transcription of the E-cadherin and FGFR2b genes preventing high expression of FGFR2b where expression of E-cadherin is low. We suggest that reduced expression of FGFR2b in conjunction with decreased expression of E-cadherin may contribute to the aggressive behaviour attributable to high grade TCCs.

Cell adhesion mechanisms in gangliogenesis studied in avian embryo and in a model system.

Neural crest cells undergo rapid changes in their cell-to-cell and cell-to-extracellular matrix adhesion during the ontogeny of the peripheral nervous system. The mechanisms of adhesion have been analyzed to assess the respective roles played by the cell adhesion molecules (CAMs) and the differentiated junctions. Crest cells which lose their terminal bar junctions after emigration from the neural tube contain only very few gap junctions during gangliogenesis. The calcium-dependent cell adhesion molecules, L-CAM, disappear from the neural crest and never reappear in crest cell derivatives. In contrast, the number of calcium-independent cell adhesion molecules, N-CAM, diminishes transiently during the migratory phase. In vitro, N-CAM is expressed de novo either just before or at the onset of aggregation into autonomic ganglion rudiments, whereas it is delayed in the dorsal root ganglion cells. In vitro, N-CAM mediates the calcium-independent aggregation mechanism; the rate of aggregation is, however, similar whether crest cells are derived from well-spread cultures or from two- and three-dimensional clusters. Crest cells also exhibit a calcium-dependent mechanism of adhesion controlled by molecules differing from N-CAM but which may codistribute on many different cell types during embryogenesis. These two classes of cell adhesion molecules are present on the surface of neural precursors prior to their differentiation into neurons and glial cells.

Involvement of epidermal growth factor receptor in chemically induced mouse bladder tumour progression.

This study was designed to investigate the role of the epidermal growth factor receptor (EGFR) and its ligands in chemically induced mouse bladder cancer. Bladder tumours were induced in C57Bl/6 and B6D2F1 mice by treatment with the carcinogen, N:-butyl-N:-(4-hydroxybutyl) nitrosamine (BBN). The levels of mRNA for EGFR and its ligands were analysed by reverse transcription-polymerase chain reaction (RT-PCR) in bladder tumours and in normal bladder urothelia. EGFR mRNA was detected in all tumours, transforming growth factor alpha (TGFalpha) mRNA levels were similar to those in normal bladder urothelia or were decreased and mRNA levels for amphiregulin, heparin-binding epidermal growth factor-like factor (HB-EGF) and betacellulin were significantly higher than those in normal urothelia. Seven cell lines were derived from chemically induced tumours. These cell lines were able to grow in serum-free conditions. All the cell lines tested expressed the genes encoding EGFR and at least one of its ligands. Proliferation of these cell lines was inhibited by AG1478, a specific EGFR tyrosine kinase inhibitor, strongly suggesting that EGFR was involved in cell growth. As expected, EGFR was found to be phosphorylated in serum-free medium, this phosphorylation being inhibited by AG1478. Conditioned medium of a bladder cancer cell line had EGFR-stimulating activity and an antibody directed against EGFR inhibited proliferation by 45%. This suggests that tumour cell growth is stimulated by an autocrine loop involving EGFR and secreted growth factors. AG1478 decreased the expression of genes for amphiregulin, HB-EGF and betacellulin, showing that EGFR activation induces up-regulation of the EGFR ligands. These results suggest that EGFR plays a critical role in bladder tumour progression.

SFL production by carcinoma cells induces the aggressive properties of nonproducing cells in vivo via a community effect.

A metastatic rat bladder carcinoma cell line, M-NBT-II, produces and secretes a dissociating factor called SFL, whereas the tumorigenic parental cell line from which it originated (E-NBT-II) does not. In this work, we report that SFL production is correlated with an invasive phenotype in three-dimensional collagen gels or organotypic cocultures. This invasiveness may be related to the production of gelatinolytic activity. We have also investigated the behavior of SFL-producing cells within an NBT-II solid tumor. Here we report that the presence of 14% of SFL-factor-producing cells are sufficient to increase the tumorigenicity, and subsequently the metastatic behavior, of the entire cell population, indicating that there is no clonal dominance of the SFL-producing cells for tumorigenicity and metastatic spreading, but rather a community effect. SFL factor may contribute to cell-cell cooperativity by paracrine or other indirect mechanisms.

Implication of scatter/growth factors in tumor progression.

Several steps during cancer progression have been distinguished on the basis of anatomo-pathological observations and experimental data. The first step, which consists of the detachment of the cancer cells from the primary tumor prior to their migration, has received much attention. Several lines of evidence have indicated that inducer molecules of tumor cell dispersion are scatter factors which are similar or identical to some growth factors. Our studies have focused on the dispersing effect of growth factors, such as acidic FGF (aFGF) on a rat bladder carcinoma cell line. These studies demonstrated that specific extracellular matrix components might contribute to the scattering effect of soluble growth factors. Additionally, our results indicated that the dispersing action of aFGF is counterbalanced by its mitogenic effect, since these two functions of aFGF cannot be observed simultaneously for the same cell. Depending on its location in the cell collective, a given cell chooses to enter mitosis or to scatter in response to aFGF. The choice between the two responses is apparently driven by molecules belonging to the transducing pathways of aFGF signaling. Finally, our data indicated that aFGF-induced tumor cell scattering leads to increased in vitro invasiveness and in vivo metastasis. Interestingly, the presence of few aFGF-producing tumor cells in a population of non-producing cells dramatically enhances the growth rate and the metastatic properties of the whole tumor, suggesting that a low proportion of highly metastatic cells in a heterogeneous cell population might modify the behavior of the tumor mass.

Involvement of cell motility in tumor progression.

Since one crucial step in tumor progression consists of the acquisition of invasive and metastatic properties, it is important to analyze the mechanisms used by cancer cells to disperse. Among the possible mechanisms of cell dispersion, cell motility appears as a central phenomenon that still needs to be understood at the molecular level. Our experimental approach to the contribution of cell motility in carcinoma cell dissemination is based on the study of the NBT-II rat bladder carcinoma cell line. The epithelial cell line gives rise to isolated, actively migrating, fibroblast-like cells in response to specific stimuli (collagens and acidic fibroblast growth factor [aFGF]). Analysis of the scattering response indicates that the different stimuli can synergize, leading to increased motility and invasiveness. NBT-II cells have two types of response to aFGF: they can either proliferate or scatter. In addition, the two responses are mutually exclusive, suggesting that the cell status can dictate whether or not tumor cells will disperse after exposure to a scatter factor. Finally, recent studies on the involvement of epithelial-specific cadherins in the process of aFGF-induced cell scattering indicate that a sustained expression of E-cadherin is not sufficient to protect cells from dispersing. In conclusion, our experimental model offers the opportunity to dissect the molecular events leading to tumor cell dissemination.

Early epochal maps of two different cell adhesion molecules.

N-CAM, the neural cell-adhesion molecule, has previously been found to be expressed during several epochs of development and function, first as an early marker in embryo-genesis, later during organogenesis, and finally in adult life. L-CAM, the liver cell-adhesion molecule, has now been localized in embryonic and adult tissues of the chicken by fluorescent antibody techniques. In the early embryonic epoch, L-CAM and N-CAM appeared in epiblastic and hypoblastic tissues. L-CAM was distributed thereafter across all three germ layers. By the onset of neurulation, however, L-CAM disappeared in the region of the neural plate and N-CAM increased in amount in that region. L-CAM appeared strongly on all budding endodermal structures (liver, pancreas, lung, thyroid, parathyroid, thymus, and bursa of Fabricius) whereas N-CAM appeared most strongly in the neural plate, neural tube, and in cardiac mesoderm but was not found in endodermal derivatives. In placodes, both L-CAM and N-CAM were present until the formation of definitive neural structures, at which time L-CAM disappeared. In kidney precursors, the two CAMs followed a complex reciprocal pattern of appearance and disappearance. For the most part, however, the distributions of the two molecules did not overlap during organogenesis. Like N-CAM, L-CAM persisted in a distinctive pattern of expression in adult tissues. During embryonic development, the two different CAMs were distributed on tissues derived from more than two-thirds of the early embryonic surface. Interpretation of maps summarizing CAM distributions over a defined developmental epoch suggested a key role for both L-CAM and N-CAM in embryonic induction. Consistent with this interpretation and with the fact that the continuity of germ layers is lost when organ rudiments are formed, neither of the CAMs was limited in distribution to a single germ layer. The regions of the early epochal maps that lacked both L-CAM and N-CAM comprised some portions of the splanchnopleure and somatopleure. Certain adult tissues that derive from this lateral plate mesoderm such as smooth muscle also lacked L-CAM and N-CAM. Such observations suggest that at least one more CAM may exist in these and similarly derived tissues.

Ontogenetic expression of cell adhesion molecules: L-CAM is found in epithelia derived from the three primary germ layers.

Immunofluorescence techniques using specific antibodies against the liver cell adhesion molecule, L-CAM, were used to explore the appearance of L-CAM during early embryogenesis and organogenesis, as well as in adult tissue. Immunoblots of L-CAM from embryonic and adult organs indicated that molecules detected in each tissue were L-CAM, and that the antibodies were not simply detecting cross-reacting molecules. L-CAM was found in low levels on pregastrulation embryos. During gastrulation, the molecule remained present on ectoderm but was not detected on mesodermal and definitive endodermal cells. During neurulation, L-CAM disappeared from the neural ectoderm, in which staining for the neural cell adhesion molecule, N-CAM, had previously been shown to increase markedly. During organogenesis, L-CAM appeared in all endodermal structures, in ectoderm other than neural derivatives, in placodes, in extraembryonic ectoderm and endoderm, and in some mesodermal structures such as Wolffian ducts, oviduct, and kidney epithelium. Other mesodermal derivatives were not stained and the molecule was not detected in hemangioblastic areas of the lateral plate mesoderm nor in splanchnopleural derivatives such as spleen, adrenal glands, and gonads. During embryonic induction, for example, neurulation and in early kidney development, changes in L-CAM distribution were correlated with both locations and times of induction events. Analysis of distribution in the adult revealed that L-CAM was present in the stratum germinativum of the skin, in endodermally derived epithelia, in the female reproductive tract, and in the kidneys. In several fully differentiated glandular organs, L-CAM staining was restricted to basal or apical parts of the cell surface. When correlated with previous results obtained for N-CAM, these findings support the idea that local cell surface modulation of a small number of cell adhesion molecules may regulate other primary processes of development to yield specific patterns, both in early development and in organogenesis. Reflections of these patterns remain in adult life.

NBT-II carcinoma behaviour is not dependent on cell-cell communication through gap junctions.

To study the mechanism(s) underlying the proliferation of heterogeneous cell populations within a solid tumour, the NBT-II rat bladder carcinoma system was used. It has been first investigated whether the different cell populations are coupled through gap junctions (GJIC). Cells overexpressing the Cx43 were generated to test for any tumour suppressive activity in vivo. To determine whether GJIC is essential for tumour proliferation and the establishment of a cooperative community effect, NBT-II cells that are incompetent for cell coupling were generated. The data report that (i) carcinoma cells expressing or not FGF-1 are coupled through GJIC in vitro and in coculture and express the gap junction protein Cx43, (ii) overexpression of Cx43 in these cells does not affect their in vitro coupling capacities and in vivo tumourigenic growth properties, (iii) inhibition of GJIC through antisense strategy has no in vivo obvious consequence on the tumour growth properties of the carcinoma, and (iv) the community effect between two carcinoma cell populations does not critically involve cell coupling through gap junctions.

Amplification of invasiveness in organotypic cultures after NBT-II rat bladder carcinoma stimulation with in vitro scattering factors.

Acidic fibroblast growth factor (aFGF) or transforming growth factor-alpha (TGF-alpha), in addition to being mitogenic, induce individual scattering of NBT-II rat bladder carcinoma cell clusters on tissue culture dishes, suggesting that they may contribute to tumor cell dissemination. To assay their scattering potential and their effect on cell invasiveness in a more complex and physiologically relevant model, we analyzed the behavior of NBT-II spheroids confronted with urinary bladder in organotypic cultures. NBT-II spheroids progressively replaced the urothelium at the site of contact with the bladder explant. In the absence of aFGF or TGF-alpha, inserted cells grew in a pattern suggestive of local hyperplasia, with occasional invasive cell protrusions. Exogenous scattering growth factors elicited a more rapid appearance of these protrusions, which were also more numerous. NBT-II cells transfected with cDNA constructs bearing the gene of aFGF, TGF-alpha or the oncogene hst/KFGF were also used. After exogenous or autocrine stimulation of NBT-II cells with the growth factors, a deeper penetration of the bladder wall in the form of nodular outgrowths and clusters of infiltrating cells was always observed. Altogether these observations suggest that the stimulation of NBT-II clusters by scattering/growth factors can promote cell shedding and amplify invasiveness in the complex extracellular environment of bladder tissues.

Direct role of the carboxy-terminal cell-binding domain of fibronectin in neural crest cell motility.

We have analyzed the interaction of neural crest cells with fragments of fibronectin corresponding to the different spliced variants of the COOH-terminal cell-binding domain (COOH-ter CBD). We have shown that this domain can support cell adhesion and migration and that both the IIICS and HepII regions are involved in these events. The rate of locomotion is high, although undirectional, compared to that of whole fibronectin. Interactions with the COOH-ter CBD are controlled by alpha4beta1 and maybe other beta1 integrins and cell-surface proteoglycans. These receptors act cooperatively to mediate attachment, spreading, and migration on fibronectin.

[Adhesion and mobility of embryonic and tumoral cells]. / Adhérence et mobilité des cellules embryonnaires et tumorales.

Regulation of a number of adhesion molecules during neural crest cell migration was studied. The neural crest, a transient embryonic neural epithelium structure, undergoes mesenchymal transformation (epithelial-mesenchymal transition). The cells then migrate, giving rise to a variety of elements including the peripheral nervous system and melanocytes. During migration, neural crest cells do not express functional cell Adhesion Molecules but interact specifically with cell-binding domains in fibronectin molecules. A rat bladder carcinoma cell line was used as an in vitro model to study conversion of epithelial cells to a migratory fibroblast-like state. Conversion can be induced by culture on collagen or exposure to acidic Fibroblast Growth Factor (aFGF). Furthermore, constitutive fibroblast-like transformation can be induced by transfection with cDNA encoding aFGF. Growth factor-producing clones exhibit increased invasive and metastatic properties as compared with non-FGF-producing control cells. This model may provide increased understanding of the role of the different adhesion molecules in processes involving cell remodeling, such as tumor spread and development of metastases.

Impaired mechanical stability, migration and contractile capacity in vimentin-deficient fibroblasts.

Loss of a vimentin network due to gene disruption created viable mice that did not differ overtly from wild-type littermates. Here, primary fibroblasts derived from vimentin-deficient (-/-) and wild-type (+/+) mouse embryos were cultured, and biological functions were studied in in vitro systems resembling stress situations. Stiffness of -/- fibroblasts was reduced by 40% in comparison to wild-type cells. Vimentin-deficient cells also displayed reduced mechanical stability, motility and directional migration towards different chemo-attractive stimuli. Reorganization of collagen fibrils and contraction of collagen lattices were severely impaired. The spatial organization of focal contact proteins, as well as actin microfilament organization was disturbed. Thus, absence of a vimentin filament network does not impair basic cellular functions needed for growth in culture, but cells are mechanically less stable, and we propose that therefore they are impaired in all functions depending upon mechanical stability.