Regional and cellular distribution of ephrin-B1 in adult mouse brain

BR A IN RE S EA RCH 1 2 47 ( 20 0 9 ) 5 0 –61

a v a i l a b l e a t w w w. s c i e n c e d i r e c t . c o m

w w w. e l s e v i e r. c o m / l o c a t e / b r a i n r e s

Research Report

Regional and cellular distribution of ephrin-B1 in adult mouse brain Paolo Migani a,⁎, Carole Bartlett b , Sarah Dunlop b , Lyn Beazley b , Jennifer Rodger b a b

Istituto di Biochimica, Università Politecnica delle Marche, via Ranieri 66, 60131 Ancona, Italy School of Animal Biology, the University of Western Australia, Nedlands 6907, Australia

A R T I C LE I N FO

AB S T R A C T

Article history:

The membrane-bound proteins ephrins and their receptors, Eph receptor tyrosine kinases, are

Accepted 29 September 2008

known for their key role during development of the central nervous system (CNS). Ligand/

Available online 17 October 2008

receptor interactions as a result of cell–cell contacts activate intracellular signalling pathways which mediate specific cellular responses. Activation can occur bidirectionally in both the

Keywords:

receptor and the ligand-bearing cells. Eph receptor and ephrin families have been implicated

Ephrin

in synaptic plasticity in the mature brain: effects include long-term potentiation/depression of

Adult brain

excitatory transmission (LTP/LTD) and an action on the structure and number of synaptic

Immunoreactivity

contacts. However, due to the redundancy of binding between receptors and ligands, the role

Neurotransmission

of individual proteins has not yet been completely elucidated. Ephrin-B1 has been suggested

Plasticity

to play a role in synaptic plasticity in the hippocampus, but its expression and localization at pre- or post-synaptic sites has been poorly documented, most likely due to the apparent low activity of the corresponding gene in mature brain. Here we present immunohistochemical data demonstrating a broad but highly regulated cellular distribution of ephrin-B1 in the mature mouse brain. We show that ephrin-B1 is expressed post-synaptically on dendritic spines in the cortex, supporting a role in synaptic plasticity in this region. However, the prevalent extra-synaptic distribution in regions such as the hippocampus and cerebellum suggests an additional structural role, perhaps at the neuron/glia interface. © 2008 Elsevier B.V. All rights reserved.

1.

Introduction

Ephrins and their receptors, the Eph tyrosine kinases are membrane-bound proteins that are key elements in the regulation of and response to cell–cell interactions in the immature tissue environment. The primary function of ephrins during development consists of the patterning of cell and axonal populations. During the earliest stages of nervous system development, ephrin expression forms boundaries for tissue formation and cell migration (Klein,

1999; Coulthard et al., 2002). At later stages, as cell mobility is reduced, ephrins direct the formation of organized axonal projections, through the formation of expression gradients across interconnected brain regions (Palmer and Klein, 2003). In addition to their role as topographic guidance cues, there is strong evidence that Eph receptor–ephrin interactions contribute to the formation of synaptic contacts (Dalva et al., 2000; Rodenas-Ruano et al., 2006). Cellular responses are triggered by signalling events elicited via the receptor as a consequence of ligand binding.

⁎ Corresponding author. Fax: +39 071 2204398. E-mail address: [email protected] (P. Migani). 0006-8993/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.brainres.2008.09.100

BR A IN RE S E A RCH 1 2 47 ( 20 0 9 ) 5 0 –6 1

However, there is also evidence for receptor induced signalling via the ligand, known as reverse signalling (Lim et al., 2008). Ephrins are divided into two sub-families, namely the A-type and the B-type: ephrin-As are anchored to membranes by a phosphatidyl-inositol (PI) linkage, while ephrin-Bs are transmembrane proteins and are primarily responsible for reverse signalling events (Aoto and Chen, 2007). Within families, specificity of the ephrin/receptor interaction is relatively low, since receptors bind all ligands of the same group. In addition, there is evidence for some cross-group binding (EphA4 and EphB2; Pasquale, 2004; Himanen et al., 2004). In contrast to strong and widespread expression of ephrins and receptors during brain development, their levels are low in the adult, in accordance with the decrease of their guiding role, which is maintained only in the germinative subventricular zone (SVZ) and the corresponding path to the olfactory bulb, the rostral migratory stream (RMS) (Conover et al., 2000): however, some genes for B-type ligands and receptors maintain high expression in specific areas (Liebl et al., 2003), implying a role in adult brain function, the most likely being structural modification of mature synapses. In adult hippocampus, ephrin-A3 located on the membrane of glial cells could interact with the EphA4 receptor on dendritic spines to trigger a reduction in spine density (Murai et al., 2003). It has been also shown that, at hippocampal CA1–CA3 synapses, post-synaptic ephrin-B3 can have both a receptormediated influence on the pre-synapse protein composition and a receptor-independent (reverse) effect on the number of the excitatory contacts (Rodenas-Ruano et al., 2006). Another current idea on the contribution of ephrins/receptors to synaptic plasticity is related to the activity-driven molecular

Fig. 1 – Western blot analysis of ephrin-B1 in proteins from adult mouse brain. A: Proteins from the whole cerebral cortex were extracted, separated by electrophoresis and bound to a membrane. Membrane lanes from duplicate samples were cut into strips and separately incubated with the anti-ephrin-B1 antibody (first lane) or with the same antibody which had been immunoadsorbed (I ADS) with the ephrin-B1-Fc chimera protein, diluted one hundred times (second lane) or ten times (third lane) from the original solution (0.25 mg/ml). Marks for the molecular weight (Mr) were obtained from a standard protein mixture, run on a separate lane. Membrane lanes were probed with a peroxidase-linked secondary antibody and the immunopositive bands evidenced by peroxidase-driven luminescence. A similar system was used to detect β-actin, as a loading control (small panels). B: Proteins from mouse cortex were incubated with different amounts of a de-glycosylating enzyme, to yield different amounts of the de-glycosylated product. Labels above each lane refer to different amounts of de-glycosylation enzyme added, in International Units/ml. Samples were then processed for western blotting with the anti-ephrin-B1 antibody as described for panel A. C: The main zones of the adult mouse brain were separated and their proteins processed in separated samples. The band lanes of the different zones were immunostained in the same membrane with the anti-ephrin-B1 antibody, as described for panel A.

51

mechanisms involved in long term potentiation and depression of excitatory transmission (LTP and LTD), particularly in the hippocampal area (rev. by Aoto and Chen, 2007). It has been shown that the genetic deletion of the EphB2 receptor affects D-aspartate postnatal LTP-related activity of the n-methyl(NMDA) receptor at CA1 hippocampal subfield and dentate gyrus synapses (Henderson et al., 2001); furthermore, postsynaptic B-type ephrins have been linked to LTP/LTD at the CA3/ CA1 connections (Grunwald et al., 2001), whereas post-synaptic EphB receptors and pre-synaptic ephrin-Bs have been implicated in NMDA-independent LTP at mossy fibres (Contractor et al., 2002). These findings demonstrate pre- and post-synaptic effects of forward and reverse signalling suggesting the existence of multiple mechanisms of ephrin/receptor interplay across specific synaptic structures. However, interpretation of these studies is complicated by the lack of specificity in ephrin/Eph receptor binding and a first step towards understanding this possibly redundant

52

BR A IN RE S EA RCH 1 2 47 ( 20 0 9 ) 5 0 –61

system is to establish a map of the cellular/subcellular distribution of individual proteins. Our previous study of ephrin-B2 immunoreactivity in the adult mouse brain suggested a novel role for the protein at somatic inhibitory synapses, confirming the relevance of this approach (Migani et al., 2007). Here, we describe the cellular distribution of ephrinB1, a protein which has been extensively used in its recombinant form as a tool to study the impact of EphB receptor/ephrin-B interactions on synaptic plasticity effects (Grunwald et al., 2001), but whose presence in the adult brain has been scarcely documented. We show that ephrin-B1 immunoreactivity is located in sites on neuronal somata/ main dendritic branches and in spine synaptic sites, in different proportions depending on the brain region.

2.

Results

2.1.

Western blot

We tested affinity-purified commercial antibodies raised against the ephrin-B1 protein; their specificity was determined by western blotting. The antibody we found to be the most specific (R&D Systems) labelled, even at high dilution of the original preparation (1:1000), a single broad band on western blot of protein extracted from adult mouse brain (Fig. 1 panels A–C). The band had an apparent maximum at 45 kDa of the molecular weight scale but its span was different in different preparations: we assumed that this variability was dependent on the presence of ephrin-B1 molecules with different poly-

saccharide attachment and branching, since the gene sequencing-derived peptide structure displays a potential glycosylation site (UniProtKB/Swiss-Prot P52795; residue 139). This view is supported by the fact that in vitro enzymatic deglycosylation lowered the apparent mass to a minimal value of 41 kDa (Fig. 1, panel B). These measured molecular weight values are compatible with those given in the original paper describing ephrin-B1 (Shao et al., 1994), since they are included in the range obtained from the amino acid sequence and measured by western assay of the cloned protein (38 and 45 kDa). As additional controls, we pre-adsorbed the antibody with recombinant ephrin-B1 chimera protein and signal was abolished in western (Fig. 1, panel A) and tissue experiments (not shown); we also found that the antibody was specific to ephrin-B1 since western signals were not reduced by the treatment with recombinant ephrin-B2 chimera (not shown). Expression of ephrin-B1 was at similar levels in olfactory bulb, cortex, striatum, hippocampus and cerebellum (Fig. 1, panel C).

2.2.

Immunohistochemistry

The results of the immunohistochemical analysis of the ephrin-B1 distribution in the adult mouse brain are summarized in Table 1.

2.3.

Olfactory structures

In the olfactory system, ephrin-B1 immunoreactivity was differentially expressed within the primary and secondary

Table 1 – Summary of ephrin-B1 immunoreactivity distribution and intensity in mouse brain

Areas with strong expression are highlighted in gray.

BR A IN RE S E A RCH 1 2 47 ( 20 0 9 ) 5 0 –6 1

olfactory structures (Figs. 2A–D). A very low signal was detected in the olfactory bulb, where immunoreactivity was slightly more intense superficially from the mitral cell layer to the glomerular layer (Fig. 2, panel A). The immunopositive tissue in the core of the olfactory bulb and peduncle corresponds to the rostral migratory stream (RMS; Figs. 2A and B, arrows); double label immunofluorescence revealed partial co-localization with the glial fibrillary acidic protein (GFAP) immunoreactivity, thus demonstrating ephrin-B1 expression in the population of standing glial elements of this zone (B cells; Conover et al., 2000) (not shown). The most intensely stained structure was the posterior part of the piriform cortex, where immunoperoxidase signal was uniformly distributed in the neuropil but spared several round profiles most likely corresponding to neuronal cell bodies (Fig. 2D). To support the neuronal identity of ephrin-B1 immunopositive cells, adjacent slices were immunostained for the

Fig. 2 – Distribution of the ephrin-B1 peroxidase-DAB immunoreactivity in the olfactory bulb and other brain olfactory areas. Ephrin-B1 was probed by a specific antibody in transverse slices (see Experimental procedures). Ephrin location was displayed by targeting the antigen–antibody complex with a peroxidase-linked secondary antibody, and by the DAB insoluble peroxidase product. Panel A: immunoreactivity in the layers of the olfactory bulb (Gl = glomerular layer; EPl = external plexiform layer; Mi = mitral cell layer; IPl = internal plexiform layer; GrO = granule layer). The arrow indicates the rostral migratory stream (RMS). Panel B: immunoreactivity in the olfactory peduncle (AOD, AOL, AOV = dorsal, lateral and ventral aspects of the olfactory nucleus). The arrow indicates the rostral migratory stream (RMS). Panel C: immunoreactivity in the superficial layers of the olfactory tubercle (Tu) and in the piriform cortex (Pir). Panel D: immunoreactivity in the ventrolateral aspect of the cortex (BLA = basolateral amygdaloid nucleus; Pir, I, II and III = cellular layers of the piriform cortex). Scale bars: A, B, D: 500 μm; C: 200 μm.

53

neuronal body marker NeuN, and revealed a similar distribution (not shown). A similar pattern of ephrin-B1 immunoreactivity, but with lower staining intensity, was observed for the anterior part of the piriform cortex itself and the olfactory tubercle (Fig. 2C), as well as for the cortical structures of the olfactory peduncle (Fig. 2B).

2.4.

Cortex

Ephrin-B1 immunoreactivity was moderate to high throughout the cerebral cortex (Fig. 3). Immunoreactivity was distributed across the entire thickness of the cortex and most prominently labelled structures in the neuropil. Close examination at high magnification (Figs. 3B, C, G and J), suggested localization to small-scale subcellular structures. There was no strict correlation between the distribution of immunoreactivity and functional cortical regions. However, while expression was relatively uniform in motor and associative zones (Figs. 3A–D and H), immunostaining within the primary somatosensory cortices was prominent in the most superficial and deep layers, revealing a band of lesser intensity corresponding to layer IV (Figs. 3H and K). Furthermore, a portion of the primary visual cortex appeared less stained than adjacent areas (Fig. 3K). Due to the widely distributed immunoreactivity in the neuropil, these differences are unlikely to be due to differences in cell density, which varies only weakly in the adjacent zones and layers, as verified by NeuN staining in duplicate slices (not shown). When analysed by immunofluorescence at mid-power magnification, immunoreactivity in the neuropil had a dotted appearance (Fig. 3L). At higher magnification, we observed regularly-spaced puncta partially or totally co-localized with synaptophysin-marked boutons (Figs. 3M, N, P and Q), presumably corresponding to dendritic spines. Ephrin-B1 positive puncta in the neuropil and within cell bodies were also close and occasionally co-localized with expression of the glial marker GLT-1 (Figs. 3O and R) which is preferentially distributed to portions of astrocytes close to synaptic structures (Minelli et al., 2001). Immunofluorescence was also indicative of a somatic distribution and revealed that ephrin-B1 was in fact expressed on the outermost portion of neuronal cell bodies and main dendritic branches (Figs. 3N and Q). The irregular, clustered appearance of this somatic labelling at higher magnification (Fig. 3Q) suggests a discontinuous distribution to small and contiguous membrane areas that rarely correspond to synaptophysin-marked puncta (synaptic terminals).

2.5.

Basal ganglia

Ephrin-B1 immunoreactivity in the basal ganglia (septum, striatum and amygdaloid nuclei) was similar in intensity and distribution to that in cortex. Among the septal structures, only the dorsal lateral nucleus and the septofimbrial nucleus were immunostained (Figs. 3D and E). In striatum, immunoreactivity was uniformly distributed among the dorsal and ventral part of the structure, as neuropilar patches mixed with unstained fibre bundles (Figs. 3D and F). Both peroxidase-DAB staining and immunofluorescence revealed small to mediumsized neuronal cell bodies as round, weakly stained or unstained figures in the neuropilar matter (Figs. 4A and B);

54

BR A IN RE S EA RCH 1 2 47 ( 20 0 9 ) 5 0 –61

several neuropilar structures displayed a similar distribution for the ephrin-B1 and the synaptophysin markers but colocalization was not frequent (Fig. 4B). Similar results were obtained for the amygdaloid nuclei (not shown).

2.6.

Hippocampus

Ephrin-B1 immunoreactivity was distributed with mediumlow intensity in all the layers of the hippocampus proper and

BR A IN RE S E A RCH 1 2 47 ( 20 0 9 ) 5 0 –6 1

dentate gyrus, as shown in Fig. 3H. At higher magnification, immunofluorescence also revealed a similar expression of the protein in the cell body layer and in the dendritic layers, with no significant difference between the different zones (Figs. 5A, D and G). The uniform ephrin signal was largely superimposed with that of synaptophysin in the dentritic layers (Figs. 5B, E and H) but, unlike than in cortex, the occurrence of profiles with close contacts between ephrin-marked spines and the synaptophysin-bearing boutons were limited in all the zones (Figs. 5C, F and I). We confirmed that expression of ephrin-B1 was on neurons by double labelling with neuronal excitatory amino acid transporter EAAC1 (EAAT3) (Figs. 5J–L).

2.7.

Thalamus and hypothalamus

Most of thalamic and hypothalamic areas displayed low or null levels of ephrin-B1 immunoreactivity. As an exception, staining in the medial epithalamic habenular nucleus was similar to that in the hippocampus (Fig. 3H). Furthermore, in the ventromedial hypothalamic nucleus (VMH) and arcuate nucleus, we observed sparse but regularly-distributed small stained structures (Figs. 6A and B). By immunofluorescence, VMH labelling appeared to be concentrated into one pole of

55

the cell body (Fig. 6C), presumably representing a specialized somatic membrane site for axo-somatic synapses. The occurrence of structures of this type was less visible in the arcuate nucleus where labelled structures mostly localized at the periphery of round cell body profiles and were frequently double-labelled with synaptophysin (Fig. 6D).

2.8.

Midbrain

Low or null levels of ephrin-B1 immunoreactivity were observed in the midbrain tectal areas (superior and inferior colliculi), as well as for most of the tegmental areas (not shown). Weak immunoreactivity highlighted the anterior tegmental nucleus and the pontine nuclei (not shown); the stain distribution was entirely neuropilar and no marked cell bodies were visible.

2.9.

Cerebellum

Ephrin-B1 immunoreactivity was distributed in all the structures of the cerebellar cortex. The labelling intensity appeared to increase from the medial, vermal portions of the cortex to the lateral, neocerebellar ones; the areas of the deep cerebellar nuclei were only weakly stained (Fig. 7A). Within the

Fig. 3 – Ephrin-B1 immunoreactivity in the main cerebral cortex areas. A–C: peroxidase-DAB immunoreactivity in the frontal cortex. Panel A shows the immunoreactivity distribution in superficial and deep layers of the primary and secondary anterior motor cortex (M1, M2). Panels B and C show enlargements of the superficial portion of the primary motor cortex (M1) and that of the agranular insular cortex (Ai), in the ventrolateral aspect of the frontal cortex (not shown in A). Scale bars: A: 500 μm; B, C: 100 μm. D–G: peroxidase-DAB immunoreactivity in the anterior cortex and in sub-cortical areas of the anterior forebrain. Panel D shows the immunoreactivity in the anterior cortex, striatum and septum. The immunostain signal is distributed with continuity and uniformity in the motor cortex, primary and secondary (M1, M2), and in the areas 1 and 2 of the cingulate cortex (CG1, CG2). Across the layers (I–IV) of the primary somatosensory cortex (S1), the signal is weakest in layer IV. (LSD = lateral septal nucleus, dorsal; CPu = caudate putamen, striatum; cc = corpus callosum). The superficial layers of the primary motor cortex are shown, at medium power magnification, in panel G. Panel E shows at medium power magnification the immunoreactivity in the lateral septal nucleus, dorsal (LSD) and intermediate (LSI) aspects; panel F shows at medium power magnification of the immunoreactivity, from the caudate putamen (CPu), across the corpus callosum (cc) to the lower layer (VI) of the somatosensory cortex. Scale bars: D: 500 μm; E, F: 200 μm; G: 100 μm. H–K: Ephrin-B1 immunoreactivity in the mid-forebrain cortex and posterior (visual) cortex. Panel H shows peroxidase-DAB immunoreactivity in the motor areas (M1, M2), in the posterior parietal association area (PPtA) and in the primary somatosensory area (S1) of the mid-forebrain cortex; also shown are the retrosplenial agranular and granular cortex (RSA, RSG). Also displayed for immunoreactivity are the hippocampus proper subfield (CA1, CA2, CA3), the fascia dentata (dentate gyrus, DG), and the epithalamus with the medial habenular nucleus (MHb). Panel I shows immunoreactivity in the layers (I–IV) of the primary somatosensory cortex, at medium power magnification; the superficial layers are shown at higher magnification in panel J. Panel K shows a low power magnification image of the primary and secondary visual areas of the cortex (V1 and V2), and in the posterior hippocampus, as well as the lack of immunoreactivity in the dorsal midbrain (S = subiculum; SC = superior colliculus). Scale bars: H, K: 1 mm; I: 500 μm; G: 200 μm; J: 100 μm. L–R: Cellular distribution of ephrin-B1 in adult mouse cortex. Panel L shows fluorescein-tagged ephrin-B1 immunoreactivity in the superficial dorsal cortex (layers I–III). Note the punctate appearance of the immunofluorescence signal. Panel M: High power magnification of the superficial part of the dorsal cortex (layer I), with fluorescence signal for ephrin-B1 (green) and synaptophysin (red). Note that puncta of both signals share average dimensions and distribution, and are often close to each other or even partially superimposed (merge colour, yellow), representing dendritic spines and synaptic boutons, respectively. These structures are also shown in the close-up picture in panel P, where some examples of their relative positions are given. Panel N: High power magnification of intermediate parts of the dorsal cortex (layers II/III), with fluorescence signal for ephrin-B1 (green) and synaptophysin (red). Indications as for M. Note also the dotted appearance of the ephrin-B1 signal on the surface of the empty profiles for cell bodies and main dendritic structures. Examples of this feature are given in the close-up figure in panel Q. Panel O: High power magnification of intermediate parts of the dorsal cortex (layers II/III), with fluorescence signals for ephrin-B1 (green) and the glial glutamate transporter GLT-1 (red). Note the dotted appearance of the GLT-1 signal, marking peripheral portions of the astrocyte cells in the vicinity of synaptic contacts. In the close-up of this picture (panel R), note the proximity and the occasional superimposition of these figures with those of the dendritic spines, marked by the ephrin signal. Scale bars: L: 50 μm; O: 15 μm; M, N: 10 μm.

56

BR A IN RE S EA RCH 1 2 47 ( 20 0 9 ) 5 0 –61

Fig. 4 – Ephrin-B1 immunoreactivity in striatum. Panel A: Peroxidase-DAB immunoreactivity. Panel B: Superimposition of the fluorescein-tagged ephrin-B1 immunoreactivity (green) with the Texas Red-tagged synaptophysin (red). Scale bar: A: 200 μm; B: 100 μm.

cerebellar cortical structure the molecular and granular layers displayed immunoreactivity in the neuropil, and no marked cell bodies were detected (Fig. 7B); the Purkinje cell layer was visible at low magnification as a row of empty or very weakly stained cell bodies (Fig. 7C). Due to the diffuse neuropilar distribution of the ephrin signal in the molecular and granular layers (Fig. 7D), double labelling immunofluorescence showed a high degree of ephrin-B1 and synaptophysin overlap. However, there were few examples of co-localization at identifiable spine/bouton figures in the molecular layer (eg. Figs. 7E and F) and at synaptic glomeruli in the granular layer (Fig. 7G). Similarly to the disposition in cerebral cortex, the ephrin signal displayed a close contiguity with the glial GLT-1 immunomarker throughout the cerebellar cortex (not shown).

3.

Discussion

3.1. Ephrin-B1 cellular distribution: peripheral and somatic cellular sites The main finding of the present study regards the wide distribution of the ephrin-B1 immunoreactivity on neurons in the adult mouse brain. This result was not unexpected, since different B-type ephrins and Eph receptors have recently been detected in parts of the adult mammalian brain (Henderson et al., 2001; Grunwald et al., 2001; Liebl et al., 2003; Xiao et al., 2006; Migani et al., 2007; see also Yamaguchi and Pasquale, 2004). However, our results contrast with the weak ephrin-B1 signal detected by in situ hybridization in previous studies (Liebl et al., 2003; Allen Brain Atlas, ABA), which suggests that expression was restricted to the pyramidal/granule cell layers of the hippocampus/fascia dentata and in the cerebellar Purkinje cell layer. The difference in expression levels between mRNA and protein could be explained by mRNA being transported to fine dendritic branches, rather than concentrated in neuronal cell bodies. In agreement with the possibility, protein synthesis at peripheral rather than central cellular sites may be a mechanism for rapid long-term changes in synaptic plasticity (Martin, 2004). Although western blotting suggested that ephrin-B1 expression was uniform throughout brain regions, immuno-

histochemistry revealed a highly regulated distribution of ephrin-B1 within the adult mouse brain. The highest expression is observed in the cerebral cortex: the main signal appears to be located on dendritic spines while a weak signal is located on the neuronal cell bodies, a distribution which fits the idea of prominent protein production at the cell periphery. In cortex, ephrin-B1 is expressed, with minor differences, in all regions and layers: however, detectable decreases in signal intensity are associated with specific zones, such as a portion of the primary visual area (V1), and layer IV of the primary somatosensory area, most likely representing differences in the density of ephrin-bearing synapses. Compared to that in cortex, immunoreactivity in the hippocampus was almost exclusively distributed to somatic and main dendritic structures with expression on spines being weak. This distribution matches that of mRNA in the cell body profiles (Allen Brain Atlas, ABA): it is quite possible, therefore, that unlike in cortex, much of the ephrin-B1 protein synthesis, trafficking and membrane incorporation in the hippocampus occur at the somatic level. In the cerebellar cortex, ephrin-B1 is also expressed in the whole neuronal structure. Labelling of Purkinje cell bodies and dendrites matches that of the specific mRNA (ABA; Liebl et al., 2003) and supports a central production of the ephrin-B1 protein. Labelling of intracellular sites, however, is intriguing since it is clearly detected by immunofluorescence and only weakly by DAB/immunoperoxidase labelling (cft. Fig. 7, panels C and D). Due to similarity in the tissue treatment, we argue that this discrepancy is not an artefact but rather could be related to a difference in sensitivity of the two methods. The uniformity of signal throughout the molecular layer suggests that protein production and/or transport extends along the branches of the Purkinje cell dendritic tree. Whether the distribution extends to the dendritic spines is not clear from our work, since synaptophysin immunoreactivity appears to be superimposed on large ephrin-B1positive dendritic profiles, rather than co-localized into small synaptic profiles. A uniform distribution among the Purkinje cell structures, which apparently extends to intracellular sites, has been described in the adult rat for the EphA4 receptor (Martone et al., 1997): this result is likely extendable to the mouse and is suggestive of interactions

BR A IN RE S E A RCH 1 2 47 ( 20 0 9 ) 5 0 –6 1

57

Fig. 5 – Ephrin-B1 distribution in hippocampus, probed by fluorescence immunoreactivity. Panels A, D and G document the distribution of the fluorescein-tagged ephrin-B1 immunoreactivity in CA1, CA3, and in the fascia dentata (FD), respectively. Note that the fluorescence signal is distributed with similar intensity in the cell body layers (pyramidal, in the hippocampus proper and granule cell in FD) and in the dendritic layers. The fluorescence signal corresponding to ephrin-B1 (green) is poorly superimposed to that corresponding to synaptophysin (red), in all the zones (panels B, E and H); this is due to the low density of ephrin-marked puncta in the dendritic field, as shown at high magnification (panels C, F and I). The distribution of ephrin-B1 to cell body and the main dendritic tree is documented (for the CA3 subfield) by the substantial superimposition of the ephrin fluorescence signal (green, panel J) with that of the neuronal marker EAAC1 (EAAT3; red, panel K), as shown by the merge picture (panel L). Scale bars: A, D, G: 50 μm; J: 20 μm; G, F, I: 10 μm.

between these proteins. In addition, a functional redundancy for ephrin-B1 and EphA4 receptor is possible, since a substitute effect of B-type ephrins on the function of this receptor has been documented in the developing cerebel-

lum (Karam et al., 2002). The subcellular distribution of ephrin-B1 in cerebellum could be further investigated by immunohistochemistry on dissociated/cultured preparations and by observations at the electron microscopy level.

58

BR A IN RE S EA RCH 1 2 47 ( 20 0 9 ) 5 0 –61

Fig. 6 – Ephrin-B1 immunoreactivity in the ventromedial hypothalamus. Panels A and B show, at low and medium-high power magnification, the appearance of peroxidase-DAB immunoreactivity in the ventromedial hypothalamic nucleus (VMH). Note the dotted figures at a pole of the cell body profiles (Arc = arcuate nucleus). Panel C: Distribution of the fluorescein-tagged immunoreactivity in the VMH. Note the presence of the fluorescence signal in ring-like profiles, with asymmetrical disposition on some neuronal cell body profiles. Panel D: Merge image of ephrin-B1 immunoreactivity (green) and synaptophysin (red) in the arcuate nucleus. Note the signal superimposition at dotted figures on the cell body profile, most likely corresponding to axo-somatic synapses. Scale bars: A: 500 μm; B: 200 μm; C: 50 μm; D: 25 μm.

3.2.

Putative roles of ephrin-B1 in synaptic plasticity

The overall distribution of ephrin-B1 in the mature mouse brain supports the possibility of a role for this protein in synaptic plasticity, but only for selected areas. Thus expression was concentrated at a high number of synaptic structures in the cortex but a limited number in hippocampus and cerebellum, where actions at extra-synaptic sites seem prominent. To our knowledge, no studies have examined the role of Eph receptors and ephrins in cortical plasticity and it thus remains unclear whether the protein's role is to activate signalling pathways involved in LTP/LTD, or to more directly induce structural/morphological modification of synapses through cytoskeletal changes. However, our results show a similar post-synaptic distribution of ephrin-B1 in the hippocampus and cortex, suggesting that the model of action studied for the hippocampal LTD/LTP, involving post-synaptic ephrin-Bs and pre-synaptic EphB receptors (Grunwald et al., 2004), could also be occurring in the cortex. Furthermore, the reduced immunoreactivity in layer IV of the primary

somatosensory areas is consistent with the decline of LTP/ LTD at thalamocortical synapses in the adult (Crair and Malenka, 1995; Feldman et al., 1998). The variability detected within adjacent visual areas suggests that it would be possible to use this structure to test the correlation between the ephrin-B1 expression and specific types of synaptic plasticity. On the other hand, Eph receptor-mediated or ‘reverse’ effects on synapse structure and number, which have been demonstrated for other ephrin components (Murai et al., 2003; Rodenas-Ruano et al., 2006) cannot be ruled out. The possible involvement of ephrin-B1 in mechanisms of this kind is indirectly supported by the labelling pattern we show in the arcuate and ventromedial hypothalamic nuclei. The occurrence of rapid, hormone-stimulated changes in the density of synapses has been documented in these nuclei (Naftolin et al., 1996; Flanagan-Cato, 2000), suggesting that rapid modifications at membrane sites may underlie the formation/disruption of synapses. This action could involve ephrin-B1 signalling and correspond to the polarized, ephrinB1 positive profiles in the ventral hypothalamus. The idea could be tested in further research in this area, by probing the effects of manipulating the ephrin/receptor system on hormone-directed changes of synapse morphology and activity.

3.3.

Possible effects of ephrin-B1 at neuron–glia interface

We also detected ephrin-B1 positive extra-synaptic sites, at different densities, in several areas of the mature brain. Their subcellular distribution needs to be further investigated, particularly in view of the heterogeneous cell labelling in the cerebellum. A possible function suggested by the contiguity with a glial marker is an interaction with Eph receptors on glial cells, to regulate stability or adaptive dynamics of ion/ metabolite exchange at the neuron/glia interface. Apparently, there are no preliminary data on this matter, since the few studies of the involvement of ephrin/Eph receptor components in glia are focussed on the tripartite glial–neuronal interaction at the synapse (Murai et al., 2003; Nestor et al., 2007). However, some data displayed the influence of the astrocytic EphA4 receptor activation on the glial glutamate release at extra-synaptic sites, a phenomenon which is most likely connected with internal calcium oscillation in the same cells (Fellin et al., 2004; Nestor et al., 2007). Even if the molecular base and physiological significance of these phenomena are unclear, a connection with ephrin-B1 on neuronal extra-synaptic sites seems to represent a workable hypothesis for future studies.

4.

Experimental procedures

The research was conducted on adult (2–9 month-old) female mice, from the C57Bl/6J and Balb/c strains. The results were qualitatively similar for the two strains. The animals were maintained and sacrificed following the guidelines of the European Community Commission for the Care and Use of Laboratory Animals. Anaesthesia was with 150 mg/kg Pentobarbitone sodium (Delvet PTY, NSW, Australia) i.p. The

BR A IN RE S E A RCH 1 2 47 ( 20 0 9 ) 5 0 –6 1

research was approved by the Ethics Commission for Animal Experimentation of the University of Ancona, and the Animal Ethics Committee of the University of Western Australia.

4.1.

59

Western blot

For analysis with an anti-ephrin-B1 antibody, tissue samples were separated by a standard method of sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE) and transferred on polyvinylidene difluoride (PVDF) membranes. Briefly, animals were anaesthetized (as described above) and perfused with ice-cold 0.1 M phosphate buffered (pH 7.4) saline (PBS); brains were removed, dissected on ice into transverse slices and separated into required regions. Samples were homogenized in a Tris(hydroxymethyl) aminomethane (TRIS)buffered protein extraction medium, containing a commercial mixture of protease inhibitors (Roche Diagnostic Australia, Castle Hill, NSW, Australia) and centrifuged at 10,000 ×g (30 min, 4 °C); total protein content was measured by the Bradford method. Protein aliquots (30 μg) were diluted in Laemmli buffer and heated at 90 °C (5 min), loaded in a 10% SDS-PAGE gel (Bio-Rad Laboratories, Hercules, CA USA) and separated by a field of approx. 17 V/cm. Proteins were then transferred from the gel to a PVDF, methanol-primed membrane, in a Tris–glycine–methanol buffer, by a 125 mA current (2 h). After a short treatment of staining/destaining with Sudan Red solution, to visualize the protein lanes, the membrane was soaked for 1 h at room temperature in a blocking solution, made-up by dissolving dry, defatted milk (5%, w/v) in Trisbuffered saline with 0.1% (v/v) Tween 20 (TBS-T). We used the anti-ephrin-B1 polyclonal antibody (R&D Systems, Minneapolis, MN, USA, AF473), raised in goat against a peptide whose sequence corresponds to the extracellular domain of mouse ephrin-B1, and affinity-purified. The original antibody solution (0.1 mg protein/ml) was subjected to a 1:500 dilution in the blocking medium: 200 μl aliquots were applied to membrane strips corresponding to individual lanes and incubated for 2 h at room temperature. For the immunoadsorbance test, ephrinB1-Fc chimera protein (R&D Systems) was added to the Fig. 7 – Ephrin-B1 immunoreactivity in cerebellum. Panel A: Peroxidase-DAB immunoreactivity in cerebellar cortex folia and base nuclei, at low power magnification (PFL = paraflocculus; Crus1 = crus 1, ansiform lobule; Sim = simple lobule; 3–5 = cerebellar lobules; Lat = lateral cerebellar nucleus; Int = interposed cerebellar nucleus; Med = median cerebellar nucleus). Panel B: peroxidase-DAB immunoreactivity of the cerebellar cortex (lobules 4 and 5) (Mol = molecular layer; Pu = Purkinje cell layer; gr = granule cell layer). Panel C: medium-high power magnification appearance of immunoreactivity in cerebellar cortex (crus 1). Scale bars: A: 1 mm; B: 500 μm; C: 100 μm. Panel D: Distribution of the fluorescein-tagged ephrin-B1 immunoreactivity in layers of the cerebellar cortex. Panel E: Merge image of the fluorescence signals for ephrin-B1 (green) and synaptophysin (red) in the molecular layer, at mid-power magnification. Panel F: Merge image of the fluorescence signals for ephrin-B1 (green) and synaptophysin (red) in the Purkinje cell layer, at mid-power magnification. Note distribution of the ephrin signal to the cell body and main dendritic tree profiles. Panel G: Merge image of the fluorescence signals for ephrin-B1 (green) and synaptophysin (red) in the granule cell layer. Scale bars: D: 50 μm; E–G: 20 μm.

60

BR A IN RE S EA RCH 1 2 47 ( 20 0 9 ) 5 0 –61

antibody solution and agitated for 2 h at room temperature prior the membrane strip application. For detection, membrane strips were washed three times 10 min in blocking buffer and incubated with a anti-goat, horseradish peroxidase (HRP)conjugated antibody (Jackson ImmunoResearch Laboratories, West Grove, PA, USA) at 1:10,000 dilution in blocking medium (90 min). After washing (TBS-T, 5 min, three times), signal was revealed by a HRP-driven chemiluminescence reaction (ECL, Amersham Biosciences, Piscataway, NJ, USA) and exposure to autoradiographic film. For the glycosylation test, protein samples were treated with a N-de-glycosylation enzyme before the above-described procedure. The enzyme peptide N-glycosidase F (PNGase F) (enzyme and buffered solution GlycoProfile II kit, Sigma Chemical Co., St. Louis, MO, USA) was added in reduced amounts and incubated for 15 min, or in excess for 1 h at 37 °C, to produce partial or more complete de-glycosylation. Deglycosylated samples were run on SDS-PAGE gels and analysed as described above.

4.2.

Immunohistochemistry

Anaesthetized animals were perfused through the left ventricle and the aorta with 25 ml of phosphate (0.1 M) buffered (pH 7.4) saline (PBS) with heparin (0.05%, w/v) followed by 50 ml of 4% (w/v) paraformaldehyde (Merck, Darmstadt, Germany), in PBS. Brains were immediately removed and immersed overnight in 4% (w/v) paraformaldehyde in PBS, at 4 °C. Brains were then cryoprotected by soaking in 15% (w/v) sucrose (in PBS), embedded in cutting medium (OCT) and rapidly frozen on dry-ice. Cutting was performed at − 20 °C in a Leitz cryostat (Leica Microsystems, Wetzlar, Germany). Transverse (coronal) 20 μm-thick brain slices were collected in 0.5% (w/v) NaN3 (as preservative) in PBS, and stored at 4 °C. Peroxidase-DAB immunochemistry was performed on sets of free-floating serial transversal slices, representing the entire brain, in 24-well plates. After a washing step (5 min, with PBS at room temperature, three times), slices were treated with a solution of 10% (v/v) methanol and 3% (v/v) H2O2 in PBS (20 min, room temperature), to quench endogenous peroxidase activity; they were then washed with PBS (10 min, three times). To block unspecific antibody binding sites, the slices were incubated for 1.5–2 h at room temperature with a PBS-diluted (10%, v/v) normal serum of the species in which the secondary antibody was raised (rabbit); 5% (w/v) bovine serum albumin (BSA) was also added to the solution. To prevent bacterial or fungal growth the solution was sterilized by filtration prior the use. Incubation with the anti-ephrin-B1 antibody was conducted by gentle rocking the slices at room temperature for 4 h and then leaving them at 4 °C overnight (8–12 h); the solution of commercial anti-ephrin-B1 goat antibody (R&D Systems) was diluted (1:200) with blocking solution. For the immunoadsorbing test, an excess amount of ephrin-B1-human Fc chimera protein (1:100–1:10 dilution of the commercial solution, R&D Systems) was added to the diluted antibody solution and agitated at room temperature for 2 h before applying to slices. After washing to remove the primary antibody (blocking solution, 20 min, room temperature, twice), slices were incubated with the secondary, anti-goat, biotin-conjugated antibody (Sigma Chemical Co., B-7024, produced in rabbit,

diluted 1:100 with blocking solution), for 1 h at room temperature. After a further washing step (PBS, 10 min, room temperature, three times), the antigen/antibody complex was detected by a biotin–streptavidin–peroxidase system (Dako Denmark, Glostrup, Denmark), and visualized by a 5 min incubation in diaminobenzidine (DAB)–metal complex (Pierce Biotechnology, Rockford, IL, USA). Slices were rinsed in PBS and mounted on glass slides, dehydrated in ascending alcohols, xylene treated and mounted with Depex (Sigma-Aldrich Chemie, Buchs, Switzerland). Images were recorded by a digital camera and a computer image processing software (Olympus Corporation, Tokyo, Japan): illumination and recording setting were maintained through the observations. The fluorescence immunohistochemistry was conducted with a similar procedure. Free-floating slices were washed, treated with H2O2-methanol and then subjected to a 15 min treatment of 0.3% (v/v) Triton X100, in PBS. Following the blocking step, the incubation with the primary anti-ephrin-B1 antibody (1:100 dilution of the original solution, R&D Systems) was extended to 32–36 h, at 4 °C. After antibody removal and washing (blocking solution, 10 min, three times), slices were incubated with a secondary, biotin-conjugated antibody (Vector Laboratories, Burlingame, CA. USA; produced in rabbit, diluted 1:500 with blocking solution), for 3.5 h at room temperature. After antibody removal (blocking solution, 10 min, once; PBS, 10 min, three times), slices were then incubated with fluorescein-conjugated avidin (Vector Laboratories, diluted 1:500 with PBS), for 1 h at room temperature. After washing (PBS, 10 min, four times), the wet slices were placed on slides and coverslipped with anti-fluorescence fading fluid (Vectashield, Vector Laboratories). For ephrin/synaptophysin double labeling, a mouse monoclonal anti-synaptophysin antibody was used (ABCAM, Cambridge, UK, SY38, 1:10) and the blocking step was extended to avoid nonspecific binding to the tissue (mouse-on-mouse blocking medium, Vector Laboratories; 1:100 dilution with the normal blocking solution; 1 h, room temperature). After a washing step (normal blocking solution, 10 min, twice), slices were incubated with both primary antibodies (anti-ephrin-B1, 1:100; anti-synaptophysin, 1:10), in the normal blocking solution, for 32–36 h at 4 °C. Both secondary antibodies (from Vector Laboratories) (anti-goat, biotin-conjugated, 1:500; anti-mouse, Texas Red-conjugated, 1:500, blocking solution) were also used in the same incubation step (3.5 h). The slices were then subjected to the fluorescein– avidin step, washed and coverslipped as described before. Double labeling for ephrin and the glial marker (glutamate transporter) GLT-1 was conducted with a polyclonal anti-GLT1 antibody produced in guinea pig (Chemicon International, Temecula, CA, USA, AB1783). Primary antibodies were used together (anti-ephrin, 1:100; anti-GLT-1, 1:400) in the normal blocking solution, for 32–36 h, at 4 °C. The anti-GLT-1 antibody was detected with an anti-guinea pig, Texas Red-conjugated secondary antibody produced in goat (Vector Laboratories, 1:500, blocking medium): to avoid interference, this was used after the anti-goat secondary antibody to detect the antiephrin antibody. Each secondary step lasted for 3.5 h. Slices were finally subjected to the above-described avidin-fluorescein and coverslipping steps. Double labeling for ephrin and the neuronal marker EAAC1 (glutamate transporter, also known as EAAT3) was conducted

BR A IN RE S E A RCH 1 2 47 ( 20 0 9 ) 5 0 –6 1

with a polyclonal anti-EAAC1 antibody produced in goat (Chemicon International, AB1520). Primary antibodies were used together (anti-ephrin, 1:100; anti-EAAC1, 1:50), in the normal blocking solution, for 32–36 h, at 4 °C. Both secondary antibodies (anti-goat, biotin-conjugated, 1:500, Vector Laboratories; anti-rabbit, Texas Red-conjugated, 1:20, Santa Cruz Biotechnology, Santa Cruz, CA, USA, blocking solution) were also used in the same incubation step (3.5 h). The slices were then subjected to the fluorescein-avidin step, washed and coverslipped as described before. Fluorescent images were obtained by a Bio-Rad MRC 1024 confocal scanning microscope (Bio-Rad Microscience, Hemel Hempstead, UK); images were reconstructed, magnified and merged from distinct signals collected for the fluorophores (fluorescein and Texas Red) in separated optic channels.

Acknowledgments The authors would like to thank Michael Archer of the School of Animal Biology, University of Western Australia, for technical assistance, and Simone Bellagamba of the Istituto di Biochimica, Università Politecnica delle Marche for image and text processing.

REFERENCES

Aoto, J., Chen, L., 2007. Bidirectional ephrin/Eph signaling in synaptic functions. Brain Res. 1184, 72–80. Conover, J.C., Doetsch, F., Garcia-Verdugo, J.M., Gale, N.W., Yancopoulos, G.D., Alvarez-Buylla, A., 2000. Disruption of Eph/ephrin signaling affects migration and proliferation in the adult subventricular zone. Nat. Neurosci. 11, 1091–1097. Contractor, A., Rogers, C., Maron, C., Henkemeyer, M., Swanson, G.T., Heinemann, S.F., 2002. Trans-synaptic Eph receptor-ephrin signalling in hippocampal mossy fiber LTP. Science 296, 1864–1869. Coulthard, M.G., Duffy, S., Down, M., Evans, B., Power, M., Smith, F., Stylianou, C., Kleikamps, S., Oates, A., Lackmann, M., Burns, G.F., Boyd, A.W., 2002. The role of the Eph-ephrin signalling system in the regulation of developmental patterning. Int. J. Dev. Biol. 46, 375–384. Crair, M.C., Malenka, R.C., 1995. A critical period for long-term potentiation at thalamocortical synapses. Nature 375, 325–328. Dalva, M.B., Takasu, M.A., Lin, M.Z., Shamah, S.M., Hu, L., Gale, N.W., Greenberg, M.E., 2000. EphB receptors interact with NMDA receptors and regulate excitatory synapse formation. Cell 103, 945–956. Feldman, D.E., Nicoll, R.A., Malenka, R.C., Isaac, J.T.R., 1998. Long-term depression at thalamocortical synapses in developing rat somatosensory cortex. Neuron 21, 347–357. Fellin, T., Pascual, O., Gobbo, S., Pozzan, T., Haydon, P.G., Carmignoto, G., 2004. Neuronal synchrony mediated by astrocytic glutamate through activation of extrasynaptic NMDA receptors. Neuron 43, 729–743. Flanagan-Cato, L.M., 2000. Estrogen-induced remodeling of hypothalamic neural circuitry. Front Neuroendocrinol. 21, 309–329. Grunwald, I.C., Korte, M., Adelmann, G., Plueck, A., Kullander, K., Adams, R.H., Frotscher, M., Bonhoeffer, T., Klein, R., 2004.

61

Hippocampal plasticity requires postsynaptic ephrinBs. Nat. Neurosci. 7, 33–40. Grunwald, I.C., Korte, M., Wolfer, D., Wilkinson, G.A., Unsicker, K., Lipp, H., Bonhoeffer, T., Klein, R., 2001. Kinase-independent requirement of EphB2 receptors in hippocampal synaptic plasticity. Neuron 32, 1027–1040. Henderson, J.T., Georgiou, J., Jia, Z., Robertson, J., Elowe, S., Roder, J.C., Pawson, T., 2001. The receptor tyrosine kinase EphB2 regulates NMDA-dependent synaptic function. Neuron 32, 1041–1056. Himanen, J.P., Chumley, M.J., Lackmann, M., Li, C., Barton, W.A., Jeffrey, P.D., Vearing, C., Geleick, D., Feldheim, D.A., Boyd, A.W., Henkemeyer, M., Nikolov, D.B., 2004. Repelling class discrimination: ephrin-A5 binds to and activates EphB2 receptor signaling. Nat. Neurosci. 7, 501–509. Karam, S.D., Dottori, M., Ogawa, K., Henderson, J.T., Boyd, A.W., Pasquale, E.B., Bothwell, M., 2002. EphA4 is not required for Purkinje cell compartmentation. Brain Res. Dev. Brain Res. 135, 29–38. Klein, R., 1999. Bidirectional signals establish boundaries. Curr. Biol. 9, R691–R694. Liebl, D.J., Morris, C.J., Henkemeyer, M., Parada, L.F., 2003. mRNA expression of ephrins and Eph receptor tyrosine kinases in the neonatal and adult mouse central nervous system. J. Neurosci. Res. 71, 7–22. Lim, B.K., Matsuda, N., Poo, M.M., 2008. Ephrin-B reverse signaling promotes structural and functional synaptic maturation in vivo. Nat. Neurosci. 11, 160–169. Martin, K.C., 2004. Local protein synthesis during axon guidance and synaptic plasticity. Curr. Opin. Neurobiol. 14, 305–310. Martone, M.E., Holash, J.A., Bayardo, A., Pasquale, E.B., Ellisman, M.H., 1997. Immunolocalization of the receptor tyrosine kinase EphA4 in the adult rat central nervous system. Brain Res. 771, 238–250. Migani, P., Bartlett, C., Dunlop, S., Beazley, L., Rodger, J., 2007. Ephrin-B2 immunoreactivity distribution in adult mouse brain. Brain Res. 1182, 60–72. Minelli, A., Barbaresi, P., Reimer, R.J., Edwards, R.H., Conti, F., 2001. The glial glutamate transporter GLT-1 is localized both in the vicinity of and at distance from axon terminals in the rat cerebral cortex. Neuroscience 108, 51–59. Murai, K.K., Nguyen, L.N., Irie, F., Yamaguchi, Y., Pasquale, E.B., 2003. Control of hippocampal dendritic spine morphology through ephrin-A3/EphA4 signaling. Nat. Neurosci. 6, 153–160. Naftolin, F., Mor, G., Horvath, T.L., Luquin, S., Fajer, A.B., Kohen, F., Garcia-Segura, L.M., 1996. Synaptic remodeling in the arcuate nucleus during the estrous cycle is induced by estrogen and precedes the preovulatory gonadotropin surge. Endocrinology 137, 5576–5580. Nestor, M.W., Mok, L., Tulapurkar, M.E., Thompson, S.M., 2007. Plasticity of neuron–glia interactions mediated by astrocytic EphARs. J. Neurosci. 27, 12817–12828. Palmer, A., Klein, R., 2003. Multiple roles of ephrins in morphogenesis, neuronal networking, and brain function. Genes Dev. 17, 1429–1450. Pasquale, E.B., 2004. Eph–ephrin promiscuity is now crystal clear. Nat. Neurosci. 7, 417–418. Rodenas-Ruano, A., Perez-Pinzon, M.A., Green, E.J., Henkemeyer, M., Liebl, D.J., 2006. Distinct roles for ephrinB3 in the formation and function of hippocampal synapses. Dev. Biol. 292, 34–45. Shao, H., Lou, L., Pandey, A., Pasquale, E.B., Dixit, V.M., 1994. cDNA cloning and characterization of a ligand for the Cek5 receptor protein–tyrosine kinase. J. Biol. Chem. 269, 26606–26609. Xiao, D., Miller, G.M., Jassen, A., Westmoreland, S.V., Pauley, D., Madras, B.K., 2006. Ephrin/Eph receptor expression in brain of adult nonhuman primates: implications for neuroadaptation. Brain Res. 1067, 67–77. Yamaguchi, Y., Pasquale, E.B., 2004. Eph receptors in the adult brain. Curr. Opin. Neurobiol. 14, 288–296.