EPHA3 as an Eph receptor: Expression & Function
EPHA3 RTK was originally isolated from a pre-B acute lymphoblastic leukemia (ALL) cell line (LK63) over 2 decades ago. EPHA3 follows the structure of Eph RTK already described, with binding of its ligands (ephrin-A5 and -A2) to specific sites in the N-terminal ephrin-binding domain and adjacent sushi-like region, which correspond to receptor heterodimerization, heterotrimerization and higher-order clustering initiation points, respectively. All three binding sites are required for intact signaling and kinase activation. Clustering of ephrin ligands is necessary also, with monomeric ephrin-A5 antagonizing activation of EPHA3.[7,17] Ephrin binding phosphorylates three major tyrosines – Y596 and Y602 in the juxtamembrane region and Y779 in the kinase activation loop – with subsequent kinase domain activation. CrkII adapter protein is recruited to the activated kinase domain and is crucial for the ensuing signaling cascade, which results in activation of RhoA GTPases.[16,18–21] Guanine exchange factors, such as, ephexin, may also be involved in this signaling cascade, as with other EphA receptors; however, this has not been demonstrated in the case of EPHA3 specifically. The adapter protein Nck1 interacts with activated EPHA3 receptors by binding to Y602 tyrosine residue. Phenotypically, prior to Ephrin activation, EPHA3 expressing cells are spindle-shaped, adherent to matrix and exhibit extensive cellular processes, reminiscent of a mesenchymal phenotype.
In cell lines studied (melanoma and kidney epithelial cell lines expressing EPHA3), activation results in loss of cellular pro-trusions, cells rounding and membrane blebbing. The net result is detachment of EPHA3-expressing cell subsequent to increased contractility and loss of adhesion to the extracellular matrix.[7,20]
Other than the activation of RhoA, the molecular mechanisms effecting this phenotypic change in EPHA3-expressing cells are not very well elucidated. In general, Rho family GTPases are established as downstream signaling molecules of Eph receptors, often activated by guanine exchange factors, which shift the balance within the Rho family to preferential activation of RhoA and inhibition of Rac1 and Cdc42 (Figures 3 & 4).[22–25] This alters cell motility properties, with Rac1/cdc42 promoting cell motility and formation of lamellopodia and filopodia in migrating cells, and RhoA inhibiting cell motility.[26,27] Studies in EPHA3 cells indicate that Rho-associated protein kinase was involved in mediating the phenotypic change following EPHA3 activation, in keeping with known effects of RhoA activation. Further downstream signaling molecules that interact with RhoA and Rho-associated protein kinase, for example myosin light chain kinase, may be recruited to effect stress fiber formation and contraction of the cytoskeleton. RhoA signaling is also capable of modulating cell–cell and cell–matrix focal adhesion complexes, with dephosphorylation of focal adhesion kinase (FAK) and paxillin following Rho activation. Activation of EPHA3 in Jurkat T-cell leukemia line recapitulates these cell adhesion changes.
EPHA3 expressing cells: morphology and signaling in the absence of ephrin ligand binding. EPHA3-expressing cells are adherent to surrounding stroma in the absence of ephrin binding and may adopt a mesenchymal-type morphology. EPHA3 is involved in mediating epithelial–mesenchymal transition during embryological development of the heart and in the setting of malignancy; this morphological change has been noted with other Eph receptors too in the absence of ligand activation of the kinase domain. When the EPHA3 kinase is not activated, the Rac/Cdc42 subsets of the Rho family of GTPases dominate and mediate cellular adhesion, formation of cellular protrusions and promote an invasive phenotype. Furthermore, these receptors exhibit crosstalk with Akt and are involved in the promotion of proliferation as a result.
EPHA3 expressing cells: morphology and signaling following ephrin ligand binding. Binding of ephrin-A5 to EPHA3 results in cell rounding, retraction of cellular processes, such as, lamellopodia, and loss of adhesion. These effects are mediated by the activation of RhoA – following ligand binding, the kinase domain is phosphorylated and recruits CrkII. RhoA is activated and subsequently activates ROCK and other signaling molecules, which reorganize the cytoskeleton and mediate cell contractility and the morphologic changes noted. Cdc42 and Rac are inactive and additionally there is inhibition of pathways that would otherwise result in proliferation (Ras-Erk) and loss of interplay with Akt, which may also suppress proliferation. FAK: Focal adhesion kinase.
Although Eph-mediated changes in the cytoskeletal structure are well recognized, survival pathways may be promoted or inhibited based on Eph RTK phosphorylation status. With phosphorylation of Eph receptors, Rac1 and Cdc42, GTPases are inactivated, with downregulation of survival pathways supported by interaction with Akt (Figure 3).[32,33]
Opposing actions elicited by EPHA3 signaling, as with other Ephs, are under the control of relative Eph and ephrin concentrations, composition of receptor signaling clusters and presence of kinase null receptors, crosstalk with growth factors and the activities of ADAM10 and protein tyrosine phosphatases (PTPs), with each potentially switching the outcome of Eph receptor activation (Figure 2).[31,34–36] ADAM10 has been specifically implicated in mediating cell repulsion on binding of EPHA3 to ephrin-A5. ADAM10 associates with EPHA3 at the cell membrane with its substrate recognition site located in the cysteine-rich region, directly binding the Eph ligand-binding domain. On binding with Eph, inhibitory interactions between the cysteine-rich and disintegrin regions of ADAM10 and its protease region are overcome; thus activating the protease with an unusual mechanism of cleavage of trans (opposite cell) ephrin mediating cell repulsion in this case. Cell repulsion appears to be the default response to ephrin-mediated Eph receptor forward signaling. Dephosphorylation of FAK and paxillin is also associated with this response, as discussed above in relation to EPHA3 activation (Figure 2a). By contrast, cell adhesion and migration may be the outcome of signaling complex assembly in the setting of low ephrin concentration or low Eph kinase activity, for example due to mutations of the Eph kinase domain or the activity of PTPs. PTP1b has been shown to negatively regulate EPHA3 forward signaling, promoting cell adhesion (Figure 2b). In the absence of Eph forward signaling, again, the Rho family GTPases Rac and Cdc42 are upregulated with consequent cellular protrusion formation, cell adherence and migration (Figure 3).
EPHA3 is highly expressed at various stages of embryonic development of the brain and spinal cord, lungs, kidney, heart and musculature. EPHA3 is virtually absent in normal adult tissues[40,41] and sparsely expressed in the CNS, with relatively high expression in retina,[40,41] and less marked expression reported in bladder, prostate, uterus and heart.
Although an early study reported possible expression of EPHA3 in hematopoietic cells of bone marrow and in the thymus, subsequent studies have indicated that EPHA3 is not highly expressed in these populations [41,42]
EPHA3 plays a prominent role in development of neurological pathways, for example retinotectal development, assisting in guidance of retinal ganglion axons to superior colliculus in higher organisms,[44,45] in keeping with high expression noted in retina in humans. Additional roles within the CNS in axon guidance have been established,[46,47] and EPHA3 is expressed in the medial motor column of the spinal cord. EPHA3 null mutants did not exhibit abnormality of axon guidance to muscle target in one study; however, another study reported disordered callosal axon projection with knockdown of EPHA3. Studies of EPHA3 null mice outline the importance of this RTK in the developing heart and suggest a role of EPHA3 in epithelial–mesenchymal transition (EMT). EPHA3 and ephrin-A1 are expressed in a complementary fashion in the developing heart, with EPHA3 expressed in mesenchymal endocardial cushions and ephrinA1 segregated to endocardial endothelium. With EPHA3 knockdown, the endocardial cushions are hypoplastic and insufficient for competent atrioventricular valve formation with consequent elevated cardiac filling pressures, biatrial enlargement and cardiac failure resulting in the death of 75% of the mice within 2 days of life. In knockout mice, there were fewer endocardial cushion cells undergoing EMT, and consequently ineffective migration of mesenchymal cells into cardiac jelly, thus impairing valve formation. Further studies examining the effects of increased expression of wild-type EPHA3 (which therefore has kinase activity) in the heart tube, atrioventricular valves and ventricular explants revealed EMT in all of these components of the developing heart, including ventricular explants that would not normally exhibit cells undergoing EMT. EPHA3 required kinase activity in order for EMT to proceed, with inhibition of EMT on overexpression of kinase null EPHA3. EPHA3 therefore appears to promote EMT in a kinase-dependent manner and this may be of relevance to its role in hematologic malignancies.
The distinct pattern of expression in postembryonic tissues may be of importance for EPHA3 as a novel focus for targeted therapy, as with scant receptor expression in normal cells, a satisfactory adverse effect profile may be anticipated.
Expert Rev Hematol. 2012;5(3):325-340. © 2012 Expert Reviews Ltd.