In this study, we tested the role of EphB signaling in the guidance of NM axons to dorsal versus ventral dendritic regions of NL and in the morphogenesis of these auditory nuclei. We found that reduction of EphB2 forward signaling, either by kiEphB2 transfection or by local inhibition with soluble fusion proteins, resulted in a significant increase in the number of mistargeted contralateral NM axons as compared to controls. In addition, there were significantly more targeting errors when forward signaling through both EphA and EphB receptor subclasses were inhibited compared to inhibition of EphB receptors alone. While EphB2 electroporation, performed at E2, did not alter axon targeting, we found that it resulted in abnormal NL morphology in which NL was significantly shorter along both rostrocaudal and mediolateral brain axes than in controls. Morphological defects were also seen with fusion protein treatment, performed from E6 to E9. EphB1-Fc injections resulted in markedly aberrant NM projections at E10 that exceeded the contralateral NL boundary and projected laterally. EphA4-Fc injections resulted in malformed NL lamina, and aberrant laterally projecting axons, though less prominent than in EphB1-Fc treated embryos. NL was significantly shorter along the rostrocaudal axis with no change in mediolateral length compared to IgG-Fc injected and untreated controls. Together, these studies demonstrate several distinct functions for Eph proteins in the formation and connections of NM and NL.
Forward EphB2 signaling guides NM axons
We previously showed that misexpression of EphA4 resulted in a significant number of targeting errors in which contralateral NM axons extended into the dorsal NL neuropil . The expression of EphA4 in NL and not NM, together with the observation that EphA4 and kiEphA4 had similar effects, suggested a non-cell-autonomous role in which EphA4 in the target stimulates NM attraction through reverse signaling in NM axons. While those studies were initially motivated by the asymmetric expression of EphA4 in NL neuropil regions , the finding that many NM axons retained their appropriate targeting after misexpression prompted us to consider the potential roles of other Eph proteins. While EphA4 was unique among these proteins in its asymmetry, both EphB2 and ephrin-B1 showed high expression levels in dorsal and ventral NL neuropil, and EphB2 and ephrin-B2 were seen in NM axons .
The electroporation studies carried out here examined the role of EphB2 in NM axons. We found that kiEphB2, but not EphB2, increased targeting errors by NM axon branches into contralateral dorsal NL. Because kiEphB2 decreases forward signaling while EphB2 increases it, our results suggest that forward signaling through EphB2 in NM axons is needed for the restriction of NM axons. The observation that EphB2 overexpression in NM axons did not have an opposite effect likely reflects the normally high precision in NM axon segregation. Together with our EphA4 study, these results suggest that bidirectional Eph signaling regulates the binaural segregation of NM axons in NL.
Based on our previous expression analysis , forward signaling through EphB2 in NM axons could be elicited through (1) ephrin-B1, which might provide chemorepulsive signals to NM axons entering the ventral NL neuropil, and/or (2) by ephrin-B2, which is expressed in NL cell bodies and could act as a barrier to axons to prevent them from reaching the dorsal neuropil layer. This latter possibility could provide an effective barrier for both ipsilateral and contralateral NM-NL projections. As ephrin-B2 is expressed in NM axons , an attractive reverse signaling cue from EphA4 in the dorsal NL neuropil would then be most effective in axons dorsal to NL (that is, from ipsilateral NM), which were not exposed to these chemorepulsive cues.
The expression of EphB receptors and ephrin-B ligands in both axon and target suggests several possible roles for EphB signaling. In addition to axon guidance, Eph-ephrin signaling, and in particular, EphB signaling, has a well-documented role in synaptogenesis and synaptic plasticity [48–61]. Because NM-NL synapses form at about E10, just after axons reach the neuropil area, the guidance of axons to these regions may be linked to formation of synapses in the correct regions. Moreover, forward and reverse signaling may influence each other due to interactions in cis, whereby ephrins can bind to Eph receptors within the same cell [62, 63] or segregate laterally into distinct signaling domains [64, 65]. These interactions can lead to downstream signaling and can alter the availability of either class for binding with proteins in trans[62, 63, 66, 67].
EphA and EphB signaling provide distinct axon guidance cues
In our fusion protein studies, we found that infusion of EphA4-Fc resulted in significantly more targeting errors in the NM-NL pathway than infusion of EphB1-Fc, suggesting that inhibition of EphA and EphB signaling is more effective than inhibition of EphB signaling alone. These results are consistent with the observation of extensive targeting errors with EphA4 electroporation  and suggest that in addition to EphB signaling and EphA4-ephrin-B2 interactions, EphA4 may facilitate axon guidance through interactions with ephrin-A ligands. We have previously demonstrated expression of ephrin-A2 in the auditory nerve and NL neuropil during the formation of NM-NL projections , indicating its feasibility as a candidate. Coordinated function between A and B classes has been shown to guide orderly projection patterns, notably in retinotectal projections [68–72]. Though we postulate here about loss of normal EphB2 repulsive cues, it is also likely EphA4 attractive cues are affected with EphA4-Fc injection and it would be interesting to explore this by analyzing ipsilateral NM-NL projections.
While our study focused on targeting of NM axons to distinct dorsal versus ventral NL regions, selective inhibition of the EphB class of receptors revealed an additional dimension of axon guidance for contralaterally projecting NM axons along the mediolateral axis. In particular, EphB fusion proteins resulted in pronounced lateral overgrowth by NM axons. Unlike the electroporation studies where misexpression was generally limited to NM axons and led to aberrant dorsoventral targeting, injection of fusion proteins into the hindbrain produced a broader inhibition that likely included EphB receptors in the NL neuropil. The observation of lateral overgrowth of NM axons using this approach may thus indicate that EphB signals arising in or near NL normally provide chemorepulsive cues that limit lateral growth.
Morphogenesis of auditory nuclei
In addition to axon targeting errors, manipulations of Eph signaling also resulted in stereotyped morphological abnormalities. When EphB2 forward signaling was increased using plasmid electroporation at E2, prior to cellular migration and NL flattening into a monolayer, we observed a significant reduction in the size of NL. This effect may be a result of impaired migration of NL cells from the auditory anlage and/or from a reduction in total NL cells either by changes in cell fate specification or increased cell death. During normal development, NL undergoes extensive (84%) cell death as the monolayer forms . Further analysis of changes in nuclei density and cell movement over time would be required to evaluate this possibility rigorously. NL was often disorganized, but in many cases the laminar appearance of NL was normal. Given that the majority of transfected cells were in NM, these results suggest a non-cell-autonomous role for EphB2 that would implicate NM-NL interactions in generating the appropriate morphology.
Our observations are consistent with previous reports that EphB/ephrin-B signaling has been shown to guide normal cellular migration in mammalian neocortex, hippocampus and cerebellum , avian and Xenopus neural crest [74–76], and zebrafish notochord  and hindbrain , whereby a contact repulsion mechanism is implicated [79, 80]. EphB2 forward signaling in particular is responsible for lamination of hippocampal dentate gyrus cells, another brain region with distinct dorsal-ventral connectivity . Though Eph-ephrin signaling is bidirectional , the abnormal morphologies were not seen with kiEphB2 transfection, where forward but not reverse signaling was impaired. Similarly disorganized NM-NL nuclei were also observed when EphA4 but not kiEphA4 was overexpressed  further implicating Eph receptor forward signaling in normal development. Together, these data suggest that forward signaling through EphA and EphB, possibly through their common ligand ephrin-B2, is necessary for the normal separation and organization of NM and NL. Indeed, inhibition of EphB forward signaling alone during E6 to E9 typically did not result in malformed nuclei, whereas inhibition of both subclasses together during E6 to E9 was sufficient to produce malformed nuclei at E10. Because NL lamination occurs when NM axons approach NL, it remains unclear whether either process is dependent on the other. Such an interaction is consistent with the observation that EphA4-Fc treated embryos often resulted in an abnormal NL morphology and also had a significant increase in targeting errors compared to EphB1-Fc treated embryos. Since multiple axons were often found crossing together following EphA4-Fc treatment, the possibility exists that fasciculation cues may also have been affected. However, because embryos with severely malformed nuclei did not meet inclusion criteria for axon targeting analysis, it is difficult to provide more than a qualitative correlation. Our findings suggest that integrated actions of Eph receptor signaling are necessary for migration of auditory nuclei precursor cells during development and that in turn, appropriate migration may be necessary for axonal target specification.
Morphogenesis relies on Eph signaling at several developmental stages
When EphB forward signaling was inhibited using fusion proteins in ex ovo preparations from E6 to E9, though NL lamination appeared normal, we observed a tendency for NM axons to overshoot the lateral boundary of their contralateral NL target, suggesting that during normal development ephrin-B2 expressing axons may be limited to ventral NL neuropil by EphB2 forward signaling in NL cells. However, when forward signaling through both EphA and EphB was inhibited by EphA4-Fc injection, we observed more significant morphological defects. Similar to results for EphB2 electroporation, there was a reduced rostrocaudal extent of NL compared to controls. In contrast, the mediolateral extent was unchanged, suggesting that this axis is set earlier in development, while rostrocaudal extension may be more protracted.
The differences in morphological defects between treatment groups suggest that Eph proteins have distinct roles during different developmental phases. Effects seen with electroporation at E2 could result from early morphogenetic events, such as cell proliferation and formation of the auditory anlage from precursors in distinct regions [2, 46], as well as later events, such as separation of NM and NL from the anlage and flattening of the NL cell body layer . In contrast, effects of fusion protein infusion at E6 to E9 reflect only these later events. While electroporation targets mainly NM cells, fusion proteins diffuse broadly within the brainstem and may affect NM and NL as well as their surrounding regions. Thus the cell autonomy of these effects is difficult to determine. The observation that EphB2 receptor overexpression in NM led to morphologic defects in NL suggests a role for reverse signaling, but could also indicate changes in the levels of available ephrins in NM cells due to interactions with exogenous EphB2. Though the exact mechanisms involved here are unknown, perturbations to migratory pathways of neocortex, cerebellum and hippocampus are seen with loss of EphB2 and ephrin-B signaling and may be linked to changes in expression, recruitment and/or signaling of extracellular matrix proteins such as Reelin [73, 81]. Migrating cortical neurons appear to use EphB versus EphA signaling differentially in determining radial versus tangential movement, respectively . Likewise, our results suggest that Eph proteins have a significant role in the formation of the auditory brainstem circuit at several developmental time points, along discrete axes and for distinct developmental events.