E-cadherin regulates the function of the EphA2 receptor tyrosine kinase.

EphA2 is a member of the Eph family of receptor tyrosine kinases, which are increasingly understood to play critical roles in disease and development. We report here the regulation of EphA2 by E-cadherin. In nonneoplastic epithelia, EphA2 was tyrosine-phosphorylated and localized to sites of cell-cell contact. These properties required the proper expression and functioning of E-cadherin. In breast cancer cells that lack E-cadherin, the phosphotyrosine content of EphA2 was decreased, and EphA2 was redistributed into membrane ruffles. Expression of E-cadherin in metastatic cells restored a more normal pattern of EphA2 phosphorylation and localization. Activation of EphA2, either by E-cadherin expression or antibody-mediated aggregation, decreased cell-extracellular matrix adhesion and cell growth. Altogether, this demonstrates that EphA2 function is dependent on E-cadherin and suggests that loss of E-cadherin function may alter neoplastic cell growth and adhesion via effects on EphA2.

Dynamic interaction of PTPmu with multiple cadherins in vivo.

There is a growing body of evidence to implicate reversible tyrosine phosphorylation as an important mechanism in the control of the adhesive function of cadherins. We previously demonstrated that the receptor protein tyrosine phosphatase PTPmu associates with the cadherin-catenin complex in various tissues and cells and, therefore, may be a component of such a regulatory mechanism (Brady-Kalnay, S. M., D.L. Rimm, and N.K. Tonks. 1995. J. Cell Biol. 130:977- 986). In this study, we present further characterization of this interaction using a variety of systems. We observed that PTPmu interacted with N-cadherin, E-cadherin, and cadherin-4 (also called R-cadherin) in extracts of rat lung. We observed a direct interaction between PTPmu and E-cadherin after coexpression in Sf9 cells. In WC5 cells, which express a temperature-sensitive mutant form of v-Src, the complex between PTPmu and E-cadherin was dynamic, and conditions that resulted in tyrosine phosphorylation of E-cadherin were associated with dissociation of PTPmu from the complex. Furthermore, we have demonstrated that the COOH-terminal 38 residues of the cytoplasmic segment of E-cadherin was required for association with PTPmu in WC5 cells. Zondag et al. (Zondag, G., W. Moolenaar, and M. Gebbink. 1996. J. Cell Biol. 134: 1513-1517) have asserted that the association we observed between PTPmu and the cadherin-catenin complex in immunoprecipitates of the phosphatase arises from nonspecific cross-reactivity between BK2, our antibody to PTPmu, and cadherins. In this study we have confirmed our initial observation and demonstrated the presence of cadherin in immunoprecipitates of PTPmu obtained with three antibodies that recognize distinct epitopes in the phosphatase. In addition, we have demonstrated directly that the anti-PTPmu antibody BK2 that we used initially did not cross-react with cadherin. Our data reinforce the observation of an interaction between PTPmu and E-cadherin in vitro and in vivo, further emphasizing the potential importance of reversible tyrosine phosphorylation in regulating cadherin function.

Isolation and characterization of the promoter region of the chicken N-cadherin gene.

N-cadherin (CDH2) is a member of the cadherin family of Ca2(+)-dependent cell-cell adhesion molecules. To investigate mechanisms controlling CDH2 transcription, we isolated and analyzed a genomic DNA sequence containing 2.8 kb of 5' flanking region and the first two exons of chicken CDH2. Sequence analysis of the promoter region of CDH2 revealed no CCATT or TATA boxes, but showed a high overall GC content, high CpG dinucleotide content, and several consensus Sp1 and Ap2 binding sequences. When fused to the cat reporter gene in transient transfection experiments, the sequence from positions -3231 to -118 (relative to the translation start site) directed high-level expression in CDH2-expressing chicken primary retinal cells and mouse N2A cells, but was much less active in chicken embryonic fibroblast cells and mouse 3T3 cells which do not express CDH2. Similarly, this promoter fragment directed variable, but neuronal-specific, expression of reporter genes in adult transgenic mice, but failed to produce the correct pattern of expression in other tissues, implying that additional sequences further upstream and/or within introns of CDH2 may play important roles in the transcriptional control.

E-cadherin mediates adhesion and suppresses cell motility via distinct mechanisms.

Expression of the calcium-dependent adhesion molecule E-cadherin suppresses the invasion of cells in vitro, but the mechanism of this effect is unknown. To investigate this mechanism, we analyzed the effects of expressing E-cadherin in mouse L-cells and rat astrocyte-like WC5 cells. Increased cellular adhesion mediated by E-cadherin reduced invasion in WC5 cells and in some L-cells, but not in others. In all cases, suppression of invasion was correlated with decreased cell movement as assessed in an in vitro wound-filling assay and a transwell motility assay. To define the relationship between adhesion mediated by E-cadherin and suppression of motility, we analyzed the effects of deleting different regions of the E-cadherin cytoplasmic domain. E-cadherin lacking the entire cytoplasmic domain did not mediate calcium-dependent adhesion and did not reduce cell motility when expressed in WC5 cells. E-cadherin lacking a portion of the catenin-binding domain did not associate with the cytoskeleton and did not promote adhesion, yet still suppressed the motility of WC5 cells. In addition, E-cadherin that retains an intact catenin-binding domain, but lacks a juxtamembrane portion of the cytoplasmic domain, mediated effective adhesion, but did not suppress motility. These results indicate E-cadherin mediates adhesion and suppresses cell motility via distinct of E-cadherin plays a key role in suppressing motility.

Schwann cells express NDF and SMDF/n-ARIA mRNAs, secrete neuregulin, and show constitutive activation of erbB3 receptors: evidence for a neuregulin autocrine loop.

Cultured Schwann cells secreted low levels (30 pg/ml/1.5 x 10(6) cells) of a 45-kDa neuregulin protein and showed constitutive activation of a neuregulin receptor, Erb-B3, suggesting the existence of an autocrine loop involving neuregulins in Schwann cells. RT-PCR analyses indicated that Schwann cells and fibroblasts in culture produced SMDF/n-ARIA and NDF but not GGF neuregulin messages. Schwann cell and fibroblast neuregulin messages encoded both beta and alpha domains; Schwann cell transcripts encoded only transmembrane neuregulin forms while fibroblast messages encoded transmembrane and secreted forms. SMDF/n-ARIA and NDF messages were also expressed in early postnatal rat sciatic nerve, suggesting a role for neuregulins in peripheral nerve development. An anti-neuregulin antibody inhibited the mitogenic response of Schwann cells to cultured neurons and to extracts of cultured neurons or embryonic brain, consistent with the accepted paracrine role of neuregulins on Schwann cells. Surprisingly, the same antibody inhibited Schwann cell proliferation stimulated by several unrelated mitogens including bFGF, HGF, and TGF-beta1. These data implicate both paracrine and autocrine pathways involving neuregulin form(s) in Schwann cell mitogenic responses.

Suppression of tumorigenicity, but not invasion, in glioblastoma/HeLa cell hybrids.

Somatic cell hybrids between SNB-19 human glioblastoma cells and human D98OR HeLa cells were produced and analyzed for their ability to form tumors in nude mice and to invade reconstituted extracellular matrix (Matrigel). Whereas both the SNB-19 and D98OR HeLa parental cells form tumors, four of six hybrid lines did not form tumors, even after periods up to six months, suggesting that each cell type can complement the tumorigenicity of the other. SNB-19 cells showed high rates of Matrigel invasion at all cell densities examined, whereas D98OR HeLa cells showed lower rates of invasion that were further reduced at high cell density. All six hybrid cell lines displayed a combination of these properties: at low cell density, the hybrids showed high rates of invasion, similar to the SNB-19 cells, but the invasion rate diminished at higher cell densities, similar to the D98OR HeLa cells. Taken together, these results provide new experimental evidence that several distinct genetic changes are involved in generating the tumorigenic and invasive phenotype of glioblastoma cells.

Induction of L1 mRNA in PC12 cells by NGF is modulated by cell-cell contact and does not require the high-affinity NGF receptor.

We examined the effects of nerve growth factor (NGF) and cell-cell contact on expression of the neural cell adhesion molecule L1 in PC12 cells. After 7 d exposure to NGF, but not after exposure to EGF, FGF, TGF beta, or dibutyryl cAMP (dbcAMP), L1 mRNA levels increased fourfold. This increase was not blocked by K252a, an inhibitor of the high-affinity NGF receptor, although neurite extension was completely inhibited. L1 mRNA levels also increased in NGF-treated mutant PC12 cells (PC12nnr5) that lack the high-affinity NGF receptor. The effect of NGF on L1 mRNA was greatest in cells cultured at high density, but its effect on cells cultured at low density was augmented by antibody to L1 (to mimic L1 homophilic binding). Various extracellular matrix components had no differential effects on L1 mRNA levels in either the presence or absence of NGF. Together, these findings suggest that NGF regulates L1 expression by a mechanism that is independent of the high-affinity NGF receptor and that this regulation is modulated by cell-cell contact but not by cell-extracellular matrix interactions.

Myelin glycoprotein P0 is expressed at early stages of chicken and rat embryogenesis.

Although previous studies suggest that P0 is expressed only in myelinating Schwann cells, monoclonal antibody 1E8 reacts with P0, yet also stains early Schwann cell precursors and non-myelinating Schwann cells (Bhattacharyya et al.: Neuron 7:831-844, 1991). We therefore characterized the 1E8 epitope and analyzed P0 mRNA expression during development. Immunoblot analyses of P0 fusion proteins and of deglycosylated P0 indicated that the 1E8 epitope is polypeptide. Northern blot and polymerase chain reaction (PCR) analyses revealed that P0 is encoded by a single mRNA that is expressed in chicken embryos as early as E4 and in rat embryos as early as E14. These data indicate that the antigen recognized by 1E8 in early chicken embryos is P0 and that, during development of both chickens and rats, P0 mRNA is expressed long before myelination.

Axons arrest the migration of Schwann cell precursors.

The neural crest gives rise to a variety of cell types including Schwann cells of the peripheral nervous system. Schwann cell precursors begin to differentiate early and migrate along specific pathways in the embryo before associating with nerve trunks. To determine whether motor axons direct the migration of Schwann cell precursors along specific pathways, we tested the effect of ablating the ventral half of the neural tube, which contains motor neuron cell bodies. The ventral neural tube was removed unilaterally from lumbar regions of chicken embryos at stage 17, when neural crest cells are just beginning to migrate and before motor axons have extended out of the neural tube. At several stages after ventral tube ablation, sections of the lumbar region of these embryos were stained with anti-acetylated tubulin to label developing axons, HNK-1 to label migrating neural crest cells and 1E8 to label Schwann cell precursors. In many embryos the ablation of motor neurons was incomplete. The staining patterns in these embryos support the idea that some Schwann cells are derived from the neural tube. In embryos with complete motor neuron ablation, at stage 18, HNK-1-positive neural crest cells had migrated to normal locations in both control and ablated sides of the embryo, suggesting that motor axons or the ventral neural tube are not required for proper migration of neural crest cells. However, by stage 19, cells that were positive for HNK-1 or 1E8 were no longer seen in the region of the ventral root, nor ventral to the ventral root region. Because Schwann cell precursors require neural-derived factors for their survival in vitro, we tested whether neural crest cells that migrate to the region of the ventral root in ventral neural tube-ablated embryos then die. Nile Blue staining for dead and dying cells in ventral neural tube-ablated embryos provided no evidence for cell death at stage 18. These results suggest that motor axons arrest the migration of Schwann cell precursors during neural crest migration.

Cell adhesion molecules and the migration of LHRH neurons during development.

During embryogenesis, LHRH neurons arise in the olfactory epithelium, migrate along the olfactory nerve, and enter the forebrain. We have examined the distribution of several cell adhesion molecules (CAMs) in the developing chick olfactory system and brain to determine whether differential distributions of these adhesion molecules might be important in pathway choices made by migrating LHRH neurons. Single- and double-label immunocytochemical studies indicated that high levels of N-CAM and N-cadherin were expressed throughout the olfactory epithelium and not restricted to the medial half of the olfactory epithelium where most of the LHRH neurons originate. Further, high levels of N-CAM, Ng-CAM, and N-cadherin were uniformly expressed throughout the entire olfactory nerve while migrating LHRH neurons were confined to the medial half of the nerve. However, once LHRH neurons reach the brain, they migrate dorsally and caudally, tangential to the medial surface of the forebrain, along a region enriched in N-CAM and Ng-CAM. After this first stage of migration within the brain, LHRH neurons migrate laterally. At this stage, there is no correlation between the intensity of N-CAM and Ng-CAM immunostaining and the location of LHRH neurons. These results suggest that N-CAM, Ng-CAM, and N-cadherin do not play a guiding role in LHRH neuronal migration through the olfactory epithelium and olfactory nerve but that migrating LHRH neurons may follow a "CAM-trail" of N-CAM and Ng-CAM along the medial surface of the forebrain.

Increasing N-CAM-mediated cell-cell adhesion does not reduce invasion of RSV-transformed WC5 rat cerebellar cells.

The WC5 rat cerebellar cell line, infected with a Rous sarcoma virus (RSV) that is temperature-sensitive for pp60v-src transformation, expresses high levels of the neural cell adhesion molecule, N-CAM, when grown at the non-permissive temperature for pp60v-src activity. At the permissive temperature, N-CAM expression is 4- to 10-fold reduced and the cells aggregate poorly. To evaluate the effects of variations in N-CAM expression, we compared the invasive ability of transformed WC5 cells that express low levels of N-CAM with transformed cells in which N-CAM-mediated adhesion was restored. WC5 cells were transfected with expression vectors containing cDNAs encoding the 120 or 180 kDa forms of chicken N-CAM linked to constitutive promoters. Several permanently transfected lines that expressed chicken N-CAM at the cell surface were isolated. These cell lines showed enhanced aggregation at the permissive temperature relative to untransfected WC5 cells or cells transfected with control constructs. By comparing the ability of control and transfected WC5 cells to invade reconstituted extracellular matrix, we tested the effect of variations in N-CAM-mediated adhesion on invasion. Clones that expressed high levels of N-CAM showed invasion rates that were similar to control cells, indicating that increasing N-CAM-mediated adhesion does not inhibit the invasiveness of RSV-transformed WC5 cells.

Neuron-Schwann cell signals are conserved across species: purification and characterization of embryonic chicken Schwann cells.

A monoclonal antibody, 1E8, which recognizes the peripheral myelin protein, P0, specific for chicken Schwann cells and their precursors (Bhattacharyya et al., Neuron 7:831-844, 1991), was used to immunoselect Schwann cells from embryonic day 14 (E14) chicken sciatic nerve. When cultured, these immunoselected cells displayed properties characteristic of perinatal rodent Schwann cells, including S100-immunoreactivity and O4 antigen-immunoreactivity. In addition, the purified chicken Schwann cells divided slowly when cultured alone, but when co-cultured with chicken or rat sensory neurons, they bound to axons and proliferated. Proliferation was also stimulated by the addition of bovine brain membrane extracts or chicken brain membranes. The 1E8 monoclonal antibody was also used to test the effect of axonal contact on P0 expression. Chicken Schwann cells purified using the 1E8 monoclonal antibody gradually lost P0 when cultured alone. These cells remained 1E8-negative even after prolonged co-culture with embryonic rat dorsal root ganglion neurons or chicken sensory ganglia. These results demonstrate that chicken Schwann cells behave like rodent Schwann cells in their expression of specific antigens, interactions with axons, and regulation of P0 expression. In addition, chicken Schwann cells respond to neuronal signals from the rat and cow, illustrating the cross-species conservation of these signals.

Conserved regulatory elements in the promoter region of the N-CAM gene.

Genomic clones containing 5'-flanking sequences, the first exon, and the entire first intron from the chicken N-CAM gene were characterized by restriction mapping and DNA sequencing. A > 600-bp segment that includes the first exon is very G + C-rich and contains a large proportion of CpG dinucleotides, suggesting that it represents a CpG island. SP-1 and AP-1 consensus elements are present, but no TATA- or CCAAT-like elements were found within 300 bp upstream of the first exon. Comparison of the chicken promoter region sequence with similar regions of the human, rat, and mouse N-CAM genes revealed that some potential regulatory elements including a "purine box" seen in mouse and rat N-CAM genes, one of two homeodomain binding regions seen in mammalian N-CAM genes, and several potential SP-1 sites are not conserved within this region. In contrast, high CpG content, a homeodomain binding sequence, an SP-1 element, an octomer element, and an AP-1 element are conserved in all four genes. The first intron of the chicken gene is 38 kb, substantially smaller than the corresponding intron from mammalian N-CAM genes. Together with previous studies, this work completes the cloning of the chicken N-CAM gene, which contains at least 26 exons distributed over 85 kb.

Development of olfactory nerve glia defined by a monoclonal antibody specific for Schwann cells.

Although there is considerable interest in the possible role of olfactory glia in the pathfinding abilities of olfactory nerve axons, the complete development of these glia in vivo has not been described. Using a specific Schwann cell marker, the 1E8 antibody, we have found that olfactory nerve glia can be identified throughout development. These glia appear to originate in the olfactory placode and migrate initially into the periphery of the olfactory nerve, and later into the center of the nerve. Olfactory nerve glia enter the presumptive olfactory bulb with the olfactory receptor neuron axons and distribute themselves along the edge of the olfactory nerve layer. The fact that olfactory nerve glia are specifically immunostained by the 1E8 monoclonal antibody, which recognizes the Schwann cell-specific protein P0, suggests that these cells more closely resemble Schwann cells than astrocytes or enteric glia. These results support and extend previous findings suggesting that olfactory nerve glia have distinctive developmental and anatomical features which may be important to the regenerative capacity of the olfactory system.

S100 is present in developing chicken neurons and Schwann cells and promotes motor neuron survival in vivo.

We used polyclonal antisera recognizing S100, a small acidic protein highly enriched in nervous tissue, to stain sections of embryonic chicken lumbosacral spinal cord and hindlimb. S100 immunoreactivity was detected in developing sensory neurons of the dorsal root ganglia (DRG) and motor neurons of the ventral spinal cord as early as embryonic day (E) 5, and staining persisted through hatching. In contrast, expression of S100 first became apparent in Schwann cells at E13, just before myelination, and was not detected in developing skin or muscle. Since S100 beta was present in motor and sensory neurons and is known to promote neuronal survival and neurite extension in vitro (Winningham-Major, Staecker, Barger, Coats, and Van Eldik, 1989), we tested the ability of S100 to promote neuron survival in an in ovo survival assay. Addition of S100 to chick embryos in ovo during the period of naturally occurring motor neuron cell death resulted in a significant increase in motor neuron survival, but had no effect on the in vivo survival of sensory neurons in the DRG. The findings that S100 is present in spinal motor neurons and that the addition of S100 enhances the survival of these cells in vivo are consistent with the possibility that S100 may act as a naturally occurring neuron survival factor during development.

P0 is an early marker of the Schwann cell lineage in chickens.

We have generated a monoclonal antibody, termed 1E8, that is specific for myelinating and nonmyelinating Schwann cells in mature chickens. 1E8 first stains cells at the edge of the neural crest; later, cells located between the neural tube and somites and in the sclerotome are immunopositive. Double labeling with HNK-1 indicates that these 1E8-positive cells represent a subset of neural crest cells in the ventral migratory pathways. 1E8-positive cells are later associated with the dorsal and ventral roots and with extending nerve trunks. In Western blots, 1E8 reacts with proteins comigrating with P0. Immunodepletion experiments establish that all P0 molecules carry the 1E8 determinant. The developmental distribution of P0, as determined by 1E8 immunoreactivity, differs from that reported for P0 in mammals and suggests that, in chicken, P0 is an early marker for the Schwann cell lineage.

Invasion by WC5 rat cerebellar cells is independent of RSV-induced changes in growth and adhesion.

The WC5 rat cerebellar cell line, which is infected with a Rous sarcoma virus that is temperature-sensitive for pp60src transformation, shows temperature-dependent expression of the neural-cell-adhesion molecule (N-CAM) and glial fibrillary acidic protein (GFAP). We found that WC5 cells maintained at the non-permissive temperature in both monolayer cultures and spheroids are subject to density-dependent inhibition of growth, whereas cells maintained at the permissive temperature continued to grow. The movement of isolated WC5 cells at both temperatures was similar, while the migration of WC5 cells out of 3-dimensional aggregates was faster at the non-permissive temperature. We tested whether the RSV-induced changes affect the invasion of the WC5 cells in 2 in vitro assays: the chorio-allantoic-membrane assay and the chick-heart-fragment assay. In both assays, WC5 cells grown at either temperature were invasive. These results indicate that growth rate is unrelated to invasion and that loss of N-CAM-mediated cell-cell adhesion is not necessary for invasion.

Invasion of Rous sarcoma virus-transformed retinal cells: role of cell motility.

Transformation of retinal neuro-epithelial cells by Rous sarcoma virus (RSV) leads to many alterations in cell phenotype, including changes in cell movement, cell-cell adhesion and protease secretion. To define and quantitate the alterations in cell movement, we analyzed video recordings of cultured cells using the computer-assisted Dynamic Morphology System (DMS). Control neuro-epithelial cells showed very low levels of translocation and membrane activity. After transformation, neuro-epithelial cells exhibited increased membrane activity, although directed cell translocation remained low. Developing retinas also contain a small proportion of Müller glial cells, which were purified by repeated passaging of control cultures. In contrast to neuro-epithelial cells, both control and RSV-transformed glial cells showed high levels of translocation and membrane activity. To analyze how different kinds of cell movement affect invasive behavior, we compared the ability of control and RSV-transformed cells to invade the chorio-allantoic membrane of developing chicken embryos. Control neuro-epithelial cells were not invasive. RSV-transformed neuro-epithelial cells, which showed low levels of translocation as revealed by DMS, were invasive. Similarly, RSV-transformed glial cells were invasive while control glial cells, which translocated, were not invasive. These results suggest that high levels of cell translocation are not necessary for invasion. In addition, the results suggest that elevated membrane activity in neuro-epithelial cells may be important for their invasion.

Plasminogen activator gene expression is induced by the src oncogene product and tumor promoters.

Secretion of urokinase-type plasminogen activator (uPA) by chicken embryo fibroblasts (CEF) is increased approximately 50-fold following transformation by Rous sarcoma virus (RSV). Using a cloned and fully sequenced chicken uPA cDNA probe, we have established that this increase in plasminogen activator production can be largely accounted for by an increase in cellular uPA mRNA. CEF contained on average less than 1 molecule of uPA mRNA/cell, whereas RSV-CEF contained 25-60 molecules/cell. The increase in cellular uPA mRNA levels was dependent on the activity of the RSV-encoded transforming protein, protein-tyrosine kinase pp60v-src. Cells infected with an RSV mutant encoding a temperature-sensitive form of the src protein (ts-NY68) contained low uPA mRNA levels when cultured at the nonpermissive temperature and high uPA mRNA levels when maintained at the permissive temperature. Temperature shift studies with tsNY68-CEF demonstrated that changes in pp60v-src activity rapidly altered uPA mRNA levels; the uPA mRNA content of total RNA extracts increased and decreased with half-time kinetics of 3-5 h. Serine/threonine-specific protein kinases also appear to modulate uPA mRNA levels in CEF cultures. Exposure of CEF and RSV-CEF for 24 h to the protein kinase C activating agent phorbol myristate acetate (PMA) increased cellular uPA mRNA levels to 20 and 260 molecules/cell, respectively. These data are consistent with the previously observed synergism between RSV and PMA in increasing plasminogen activator secretion. Nuclear run-on transcription analyses established that both RSV and PMA increase cellular uPA mRNA levels by way of increased uPA gene expression.

Expression of neural cell adhesion molecules in normal and pathologic tissues.

Because cell-cell and cell-matrix interactions directly affect the growth, differentiation, and morphogenesis of neural tissue, abnormal changes in these processes could have severe pathologic consequences. Over the last few years, it has become possible to investigate these interactions at the molecular level due to advances in the identification and characterization of matrix components and receptors and cell adhesion molecules (CAMs). Emerging evidence suggests that two broad classes of CAMs are represented by the neural CAM, N-CAM, and the epithelial CAM, L-CAM. N-CAM and several other neural adhesion molecules contain immunoglobulin-like domains and do not require calcium for binding. In contrast, L-CAM, N-cadherin, and P-cadherin depend on calcium for activity and share structural features that differ from those of the N-CAM family. All of these CAMs are expressed in early embryos and in a variety of tissues throughout development, although each has a characteristic pattern. Initial studies suggest that injury, oncogenic transformation, and some genetic neurologic disorders are accompanied by changes in CAM expression that alter the adhesive or migratory behavior of cells.