The first Eph receptor, which was found to be overexpressed in erythropoietin-producing hepatocellular carcinoma cell line (hence, the name Eph), was identified by homology cloning using the kinase domain of the viral oncogene v-fps as a probe for low-stringency hybridization (Hirai et al. 1987). Since then, other homologous members were cloned by low-stringency cross-hybridization or PCR with primers based on conserved sequences in the kinase domain. There are a total of fourteen homologous Eph receptors identified in mammals to date, comprising the largest family of RTK. They are generally categorized as either EphA or EphB receptor based on ligand specificity (see below).
During the early days when their ligands had not yet been identified, Eph receptors were often described as “orphan receptors” with unknown function. The first Eph receptor ligand, named at that time B61, was cloned as a cytokine-inducible gene (Holzman et al. 1990), although it was not recognized as a ligand of Eph receptor until several years later (Bartley et al. 1994). Using Fc-tagged soluble Eph receptors to screen expression libraries, other Eph receptor ligands have been cloned and collectively called ephrins (derived from Eph-interacting proteins; Eph Nomenclature Committee 1997). Although B61 was first described as a secreted protein, it turns out that all ephrins are membrane proteins. The eight different ephrins can be subdivided into two families based on the type of membrane anchorage: ephrinAs (ephrinA1 to A5) are linked to the plasma membrane via a glycosyl-phosphatidyl inositol (GPI) moiety, and ephrinBs (ephrinB1 to B3) anchor to the plasma membrane through the presence of a transmembrane domain. In general, ephrinAs bind to EphA receptors (EphA1 to A8 and A10), while EphB receptors (EphB1 to B4 and B6) are preferentially activated by ephrinBs. The notable exceptions are EphA4, which binds to both the A- and B-class ephrins with high affinity, and EphB2, which also binds to ephrinA5 in addition to the transmembrane ephrins.
Multiple Eph receptors and ephrins have also been identified in other vertebrates such as zebrafish and Xenopus. However, there is only a single Eph receptor being identified from the invertebrates C. elegans and Drosophila (Drescher 2002), and a partial sequence of Eph receptor has been cloned from the sea slug Aplysia californica. It therefore appears that Eph receptors and ephrins are ancient and evolutionarily conserved, but the divergence of the eph ancestor gene into multiple genes might occur during the evolution of vertebrates.
Characteristics of Ephrins and Eph Receptors
Full-length Eph receptors are ∼120 kD. The extracellular domain of Eph receptors can be subdivided into the amino-terminal ephrin-binding domain, the cysteine-rich region, and two fibronectin-type III repeats. This is followed by a single transmembrane domain rich in hydrophobic amino acids for membrane anchorage and the intracellular domain that can be functionally segregated into the juxtamembrane region, the kinase domain, the sterile alpha motif (SAM) protein-protein interaction domain, and the PDZ-binding motif. Almost all Eph receptors have tyrosine kinase activity, except EphB6 which lacks certain conserved amino acid residues in the kinase domain and therefore does not possess catalytic activity. However, it is possible that EphB6 can still function in a kinase- dependent manner, probably through heterodimerization with other EphB receptors or interaction with other tyrosine kinases such as Cbl.
The full-length ephrins contain ∼120 amino acids and anchor to the plasma membrane through the GPI-linkage (ephrinAs) or transmembrane domain (ephrinBs). The membrane attachment of ephrins is believed to be essential for mediating the cell contact-dependent functions of the Eph receptors (see below), although there are soluble isoforms of ephrins as a result of alternative splicing. Another unique feature of ephrins is their ability to transduce reverse signaling: upon interaction between ephrins and Eph receptors from opposing cells, not only the Eph receptors can trigger signal transduction via its catalytic kinase domain (“forward signaling”), but at the same time, both the ephrinA and ephrinB can also induce signal transduction in the ephrin-expressing cell, and this “reverse signaling” of ephrin is crucial in nervous system development and functioning, including axon guidance, synapse formation, and synaptic plasticity. The capability of transmembrane ephrinB to transduce reverse signaling is largely attributed to the presence of conserved tyrosine residues in the intracellular domain, which are phosphorylated after binding to Eph receptor by the Src family of tyrosine kinases. The phosphorylated tyrosine residues in ephrinB intracellular domain in turn recruit specific signaling molecules such as Grb4 via interaction with SH2 domain, leading to the propagation of signaling that regulates actin cytoskeleton dynamics (Cowan and Henkemeyer 2002). In addition to tyrosine phosphorylation, the carboxyl-terminal of ephrinB also contains PDZ-binding motif that is critically involved in synapse formation (Klein 2009; Lai and Ip 2009). Compared to the B-class ephrins, much less is known about how the GPI-anchored ephrinA transduces reverse signaling after binding to EphA receptor. The tyrosine kinases Fyn and Src can be activated by ephrinA5 reverse signaling to regulate cell adhesion, but the detailed mechanisms remain to be elucidated. EphrinA might also recruit the neurotrophin receptors p75 and TrkB as coreceptors for reverse signaling.
Function of Eph Receptors in the Nervous System
Ephrins and Eph receptors are crucial mediators of cell-cell communication in the nervous system, where they play indispensible roles both in neural development (e.g., axon guidance and formation of topographic mapping, synapse formation) and functioning of the matured brain (e.g., synaptic plasticity for learning and memory). Ephrin/Eph are also involved in a wide range of functions outside the nervous system, including cell sorting during embryonic segmentation, formation of blood vessels, maturation of the immune system, and bone formation (Pasquale 2008). Moreover, increasing evidence indicates that Eph receptors are critical determinants of cancer progression (Pasquale 2010). This chapter will focus on the function of ephrin/Eph in the nervous system.
Axon Guidance and Topographic Mapping
It has been known for long time that neurons from the retina connect to the optic tectum (or the superior colliculus in mammals) in a highly precise manner, such that retinal axons from the nasal side of the retina selectively project to the posterior side of the tectum, while the temporal retinal axons specifically innervate the anterior tectum. The identity of the molecules responsible to establish this precise retinotectal map remained obscure for many years until the discovery of the ephrin and Eph receptor. Three important lines of evidence indicate that ephrin represents the major repulsive guiding cue during retinotectal mapping. First, ephrins and Eph receptors are expressed in complementary gradients, with ephrinA2 and ephrinA5 expressed in a gradient from low (anterior) to high (posterior) across the tectum, while the expression of EphA3 is high in the temporal retinal and low in nasal retina. Second, in an in vitro stripe assay, ephrinA5 induces growth cone collapse of temporal retinal neurons, but not nasal retinal neurons. These observations therefore suggest that the countergradient of ephrinA and EphA in the tectum and retina is responsible for establishing the characteristic retinotectal map in a repulsive manner, such that the low expression of EphA3 in the nasal retina allows the projection of the axons into the posterior tectum where the expression of ephrinA2 and A5 is high, whereas the high expression of EphA3 in the temporal retina limits its axons to innervate the anterior tectum where the level of ephrinAs is low. Finally, knockout mice lacking ephrinA2 or ephrinA5 show disrupted retinotectal connectivity, thus indicating the essential role of ephrin/Eph in topographic mapping in the retina.
In addition to retinotectal mapping, ephrin/Eph also participates in axon guidance and topographic mapping in other brain areas, including the hippocampus-septum, and midline crossing by the commissural axons that connect the two halves of the hemisphere. It is noteworthy that in the latter situation, it is the ephrinB reverse signaling rather than Eph receptor forward signaling that mediates the axonal repulsion.
Synapse Formation and Synaptic Plasticity
Both ephrinA and ephrinB regulate synapse formation and synaptic function, but their effects and the underlying mechanisms appear to be very different (Klein 2009; Lai and Ip 2009). The importance of EphB in synapse formation is illustrated by the observation that triple knockout mice lacking EphB1/B2/B3 neurons in the hippocampus fail to form dendritic spines, the tiny specialized dendritic protrusions where majority of excitatory neurotransmission takes place. EphrinB-EphB interaction promotes synapse formation by means of several mechanisms: (1) their high-affinity binding provides adhesion to stabilize the initial axon/dendrite contact during synapse formation; (2) both ephrinB and EphB receptor contain PDZ-binding motif at their C-termini that allows simultaneous clustering of PDZ domain-containing scaffold proteins at nascent pre- and postsynaptic specialization; (3) EphB forward signaling increases the motility of immature spines (filopodia), which in turn facilitates the axon-dendrite contact during synapse formation; (4) EphB enhances the function of NMDA receptor, the subtype of ionotropic glutamate receptor that mediates activity-dependent synapse formation during development; and (5) transsynaptic activation of ephrinB reverse signaling also promotes maturation of dendritic spines as well as formation of presynaptic termini.
In contrast to EphB forward signaling which promotes the formation and maturation of dendritic spines, activation of EphA forward signaling results in spine retraction and loss of functional synapses in cultured hippocampal neurons (Fu et al. 2007). It has been proposed that ephrinA3 localized on astrocytic processes that are in close proximity of dendritic spines activates EphA4 in the postsynaptic neuron to organize dendritic spines and prevent their unlimited growth (Murai et al. 2003). This was supported by the observations that hippocampal neurons in EphA4 knockout organotypic slices display disorganized dendritic spines. Activation of EphA4 forward signaling also leads to loss of functional synapses by triggering the polyubiquitination of AMPA receptors, the major subtype of glutamate receptors for excitatory neurotransmission. This results in degradation of AMPA receptors by the proteasome and subsequent reduction of glutamatergic neurotransmission (Fu et al. 2011).
Adhesion Versus Repulsion: Modulation of Ephrin/Eph Interaction and Function
Most of the functions performed by ephrin/Eph within or outside the nervous system involve repulsive cell-cell interaction (e.g., growth cone collapse, spine retraction, segmentation). However, ephrin/Eph interaction can also be adhesive in specific cellular context (e.g., synaptogenesis, endothelial cell attachment, and capillary assembly). One interesting topic on the study of ephrin/Eph signaling therefore involves understanding how the same family of proteins can generate both adhesive and repulsive signals in a context-dependent manner and how the initial high-affinity trans-interaction between ephrin/Eph on opposing cells switch to become repulsive. Two major mechanisms that modulate the ephrin/Eph interaction have been proposed to potentially explain the switch from adhesion to repulsion. One is internalization of the intact ephrin/Eph complex into their expressing cells by trans-endocytosis. Another mechanism involves extracellular cleavage of ephrin by the metalloprotease ADAM10/Kuzbanian. Both endocytosis and proteolytic cleavage are expected to terminate the interaction between ephrin and Eph receptor and hence allow cell retraction subsequent to the initial cell-cell contact.
Besides endocytosis and proteolytic cleavage, ephrin/Eph interaction can also be modulated by several other mechanisms that can fine-tune the adhesive and repulsive effect. For example, cis-binding between EphA3 and ephrinA5 expressed on the same cell can inhibit transactivation of the receptor by ephrinA5 expressed in neighboring cells, and this modulatory effect of ephrinA5 cis-binding is involved in retinotectal mapping. In addition, it has been shown that alternatively spliced isoforms of EphA7 that lack kinase domain can determine the outcome of adhesion and repulsion during neural tube formation in early development. Similarly, soluble forms of ephrin due to alternative splicing also exist, although it is not clear if they can compete with membrane-bound ephrin and inhibit the transactivation of Eph receptor. Taken together, ephrin/Eph cis-binding and the generation of inhibitory isoforms of ephrin/Eph by alternative splicing can potentially increase the diversity of interactions between ephrin and Eph receptor and hence fine-tune the adhesive and repulsive effect of ephrin. This property might be particularly advantageous for their function as positional cues (e.g., during topographic mapping) to guide cell position and connections in a highly precise manner.
Signal Transduction of Eph Receptor
Regulation of Actin Cytoskeleton
On the other hand, activation of EphA forward signaling leads to inhibition of Rac1 and activation of RhoA, both of which are central to the effect of ephrinA on growth cone collapse and dendritic spine loss. Activation of EphA4 reduces Rac1 activity via the GTPase-activating protein (GAP) α2-chimaerin, which is phosphorylated by EphA4 and becomes activated, leading to the increased GAP activity toward Rac1 and hence the reduction of Rac1 activity during EphA4-dependent growth cone collapse (Shi et al. 2007). In addition, stimulation by ephrinA1 triggers tyrosine phosphorylation and activation of the proline-directed serine/threonine kinase Cdk5, which then phosphorylates the RhoA-specific GEF ephexin1, leading to increased RhoA activity. The significance of Cdk5 and ephexin1 in mediating EphA4 signaling is verified by the observation that ephrinA1 stimulation fails to induce spine loss in neurons lacking Cdk5 or ephexin1 (Fu et al. 2007). EphA4 forward signaling also regulates actin cytoskeleton through phospholipase Cγ1, which becomes activated after tyrosine phosphorylation and induces membrane dissociation of the actin depolymerization factor cofilin during EphA4-mediated loss of dendritic spines.
Cross Talk with Adhesion Molecules
Besides actin cytoskeleton, another major target of ephrin/Eph signaling is adhesion molecules. Cross talk between Eph signaling and adhesion molecules has been implicated in the control of cell migration, cell sorting, and spine morphology (Arvanitis and Davy 2008). Depending on cellular context and modes of stimulation, both enhancement and inhibition of integrin-mediated cell adhesion by ephrin/Eph signaling have been reported. Signaling molecules that link ephrin or Eph receptor activation to the integrin pathway have also been identified. For example, reverse signaling of ephrinA2 or ephrinA5 increases cell adhesion to laminin that requires Src tyrosine kinase. EphA8 also enhances integrin-mediated cell adhesion via phosphatidylinositol-3 kinase. Moreover, EphA4 interacts directly with integrin in platelets to enhance cell attachment. Intriguingly, another study showed that integrin signaling is inhibited upon EphA4 activation during dendritic spine remodeling in neurons and involves reduced phosphorylation of downstream targets of integrin signaling such as Crk-associated substrate (Cas) and the tyrosine kinases FAK and Pyk2.
Ephrin/Eph also regulates cell sorting through the adhesion molecules cadherin and claudins. EphB activation regulates membrane localization of E-cadherin, which is required for the compartmentalization of colorectal cancer cells by ephrinB1 expressed on the intestinal cells. On the other hand, claudins present at the tight junction interact directly with EphA2 and ephrinB1, and EphA2-mediated tyrosine phosphorylation of claudin-4 regulates its localization and affects paracellular permeability.
Cross Talk with NMDA Receptor and Regulation of AMPA Receptor Trafficking
Glutamate is the major neurotransmitter of excitatory neurons in the brain. There are multiple subtypes of glutamate receptors in neurons, including the ionotropic NMDA and AMPA receptors, as well as the G-protein-coupled metabotropic glutamate receptor. One major mechanism underlying how Ephrin/Eph regulates synapse formation and synaptic function involves regulating the function of NMDA receptor and localization of AMPA receptor. EphB receptor physically interacts with NMDA receptor via the extracellular domain, and this interaction is crucial to mediate the enhancing effect of ephrinB on synapse formation. Activation of EphB also facilitates the function of NMDA receptor, as indicated by the increased Ca2+ influx and resulting induction of gene expression. In addition, ephrinB reverse signaling triggers Src-mediated tyrosine phosphorylation of NMDA receptors during LTP of hippocampal neurons.
Ephrin/Eph signaling directly modulates neurotransmission by regulating the localization and trafficking of AMPA receptor. Exogenous expression of EphB2 induces clustering of AMPA receptors in primary cortical neurons, which requires the PDZ-binding motif in the carboxyl-terminus of EphB2. Furthermore, ephrinB2 reverse signaling also regulates AMPA receptors trafficking by two different mechanisms: (1) increasing AMPA receptor interaction with the synaptic scaffold protein GRIP and thus preventing their internalization and (2) counteracting the internalization of AMPA receptors by reducing Ser880 phosphorylation of GluR2 subunit of AMPA receptor.
Regulation of Protein Expression Via Transcription, Protein Synthesis, and Degradation
Emerging evidence suggests that ephrin/Eph also controls the abundance of specific proteins by regulating gene transcription, protein synthesis, and protein degradation. For example, both EphA forward and ephrinB reverse signaling can activate the transcription factor STAT3, which might have an implication in cancer progression (Lai et al. 2004). EphrinA also regulates the expression of acetylcholinesterase in muscle fibers, which hydrolyses the neurotransmitter acetylcholine at the neuromuscular junction. On the other hand, ephrinA-induced growth cone collapse involves inhibition of local translation of mRNA such as ß-actin in the growth cone via phosphorylation of the GTPase-activating protein Tsc2, which leads to depression of the protein kinases ERK and mTOR (Nie et al. 2010). Finally, recent studies revealed that Eph receptor forward signaling activates specific ubiquitin ligases such as APC2/Cdh1 and Ube3A. This results in the ubiquitination and proteasome-dependent degradation of specific proteins such as AMPA receptor and the Rho-GEF ephexin5, which is essential for homeostatic plasticity and excitatory synapse formation, respectively (Fu et al. 2011; Margolis et al. 2010).
Since their identification in the early 1990s, there has been substantial interest in studying this unusual class of receptor tyrosine kinases (RTK) (Lackmann and Boyd 2008). As a result, significant progress has been made on elucidating their molecular properties and how they participate in different biological functions. The studies on ephrin/Eph have undoubtedly generated many surprises and greatly increased the horizon of our understanding of cell-cell communication in a wide range of biological systems. It is also becoming clear that our knowledge on ephrin/Eph interaction and signaling might prove to be useful in the design of therapeutic agents against various diseases. For example, abnormal expression of ephrin/Eph has been reported in different types of tumors. Given their capability of regulating cell adhesion, cell migration, and vascular formation, it is generally believed that ephrin/Eph plays important role in tumor formation and invasion. Their role in tumor formation, however, appears to be highly complicated: they can either promote or suppress tumorigenesis, depending on the types of tumor (Pasquale 2008). Nonetheless, inhibition of ephrin/Eph function might be useful for antiangiogenic therapies for cancer, and small molecules that inhibit Eph receptor kinase activity or ephrin-Eph interaction have been designed and shown to reduce tumor growth in animal models (Pasquale 2010).
In addition to tumor growth and malignancy, recent study also raises the exciting possibility of targeting EphB receptor in Alzheimer’s disease therapy. EphB2 interacts with NMDA receptor and regulates its function in synapse formation and plasticity. It was recently reported that ß-amyloid binds to the fibronectin-type III repeat of EphB2 and promotes its degradation by the proteasome, and EphB2 expression in the hippocampus is reduced in hAPP transgenic mice, an animal model of Alzheimer’s disease. Remarkably, reversing EphB2 depletion by exogenous expression of EphB2 using lentivirus recues the synaptic and cognitive deficits of hAPP mice (Cisse et al. 2011). This study therefore suggests that small peptide that blocks the interaction between ß-amyloid and EphB2 might represent a potential pharmaceutical agent for the disease.