Coordinated Eph-ephrin signaling guides migration and axon targeting in the avian auditory system

Abstract

Background: In the avian sound localization circuit, nucleus magnocellularis (NM) projects bilaterally to nucleus laminaris (NL), with ipsilateral and contralateral NM axon branches directed to dorsal and ventral NL dendrites, respectively. We previously showed that the Eph receptor EphB2 is expressed in NL neuropil and NM axons during development. Here we tested whether EphB2 contributes to NM-NL circuit formation.

Results: We found that misexpression of EphB2 in embryonic NM precursors significantly increased the number of axon targeting errors from NM to contralateral NL in a cell-autonomous manner when forward signaling was impaired. We also tested the effects of inhibiting forward signaling of different Eph receptor subclasses by injecting soluble unclustered Fc-fusion proteins at stages when NM axons are approaching their NL target. Again we found an increase in axon targeting errors compared to controls when forward signaling was impaired, an effect that was significantly increased when both Eph receptor subclasses were inhibited together. In addition to axon targeting errors, we also observed morphological abnormalities of the auditory nuclei when EphB2 forward signaling was increased by E2 transfection, and when Eph-ephrin forward signaling was inhibited by E6-E8 injection of Eph receptor fusion proteins.

Conclusions: These data suggest that EphB signaling has distinct functions in axon guidance and morphogenesis. The results provide evidence that multiple Eph receptors work synergistically in the formation of precise auditory circuitry.

Figure 1

Schematized NM-NL circuit. (A) Illustration of a dissected chick embryo brainstem at E10 with auditory nuclei (including NM and NL) outlined bilaterally and the coronal plane of section for (B) indicated with a dashed line. (B) Coronal section of an E10 chick auditory brainstem with bisbenzimide nuclear staining. Bilateral projections from a single representative right-sided NM cell are schematized. Ipsilateral NM axon contacts dorsal NL neuropil with divergent branches of similar length, while contralateral NM axon contacts ventral NL neuropil with a delay line. Note the absence of cells in the neuropil surrounding NL cells where NM axons synapse. Scale bar = 0.5 mm in (A) and 100 μm in (B). VIII = eighth cranial (vestibulocochlear) nerve.

Figure 2

Transfection is limited to auditory nuclei as seen by EGFP fluorescence. (A) Dissected chick brainstem at E10 following in ovo electroporation at E2. Bilateral auditory nuclei are EGFP positive, as are the axons connecting NM to contralateral NL (white arrowheads). (B) Coronal section of E10 chick brainstem (left side) at approximately the level shown by a black dashed line in (A) shows transfection is extensive in but not limited to NM cells and their axons. In the section shown here, three NL cells and their bipolar tufts of dendrites within the cellular monolayer are visible by EGFP fluorescence. (A, B) White arrowheads delineate the margins of contralateral NM axons, and are oriented in (B) to indicate anterograde direction from NM origin to contralateral target. Scale bar = 1 mm in (A) and 100 μm in (B).

Figure 3

Targeting errors are increased when EphB2 is misexpressed during development. (A-C, A’-C’) RDA trace of left-sided NL in E10 embryos following E2 transfection of EGFP control (A, A’), full-length EphB2 (B, B’), or kiEphB2 (C, C’) plasmid. (A, B, C) are original 20X images with the outlined box expanded and color-inverted in (A’, B’, C’) to highlight fine processes of single axons crossing the NL cell body layer inappropriately (white arrows). (D) Quantification of the mean number of axon targeting errors per group with SEM bars. (E) Percentage of embryos in each group demonstrating atypical features of NL morphology, including a reduced length and/or a disorganized cellular layer. Scale bar = 100 μm in (A-C) and 35 μm in (A’-C’).

Figure 4

Eph receptor fusion protein binds endogenous ephrin ligands. (A) and (B) are adjacent coronal sections of an E10 embryo, incubated with either IgG-Fc (A, A’) or EphB1-Fc (B, B’) fusion proteins. Bisbenzamide nuclear staining of the same section (A’ and B’ for A and B, respectively) is used to identify NM and NL. White arrows indicate the ventral border of NL neuropil, and black arrows indicate the margins of the midline, both are locations where ephrin-B ligands are highly expressed at this age. EphB1- Fc binds NL and midline regions (B), whereas IgG-Fc does not (A). Scale bar = 100 μm.

Figure 5

Treatment with Eph receptor fusion protein results in increased targeting errors and atypical NL morphology. (A-D) RDA tracings of contralateral-projecting NM axons at E10 in untreated embryos (A) or embryos treated during E6 to E9 with IgG-Fc control (B), EphB1-Fc (C), or EphA4-Fc (D). Axonal processes inappropriately cross the NL cell body layer, terminating in the dorsal NL neuropil in embryos treated with EphB1-Fc or EphA4-Fc (white arrows, C, D). (E) Quantification of mean axon targeting errors and (F) percentage of embryos with an atypical NL (including either atypical NL morphology or atypical NL border or both). Scale bar = 25 μm.

Figure 6

Development of NL is affected by impaired Eph receptor signaling. Representative 40X images of lateral NL following RDA midline tracing of E10 control embryos (A), embryos treated with EphB1-Fc (B), EphA4-Fc (C), or transfected with full-length EphB2 (D). Top panels are a merge of bisbenzamide (BIS, blue) nuclear counterstaining and RDA (red) fluorescence from labeled contralateral NM axons; both channels are shown individually in respective middle and lower panels. NL is normally an elongated, flattened layer with a lateral border (leftmost on these images) well defined by contralateral NM axons (white arrowhead in A). Treatment with EphB1-Fc (B) results in disorganized NM axons projecting lateral to the NL border (white arrowhead in B). Treatment with EphA4-Fc results in an abnormal NL morphology, where the mediolateral length is reduced and the lateral aspect is disorganized and not flattened (C). NM is normally found dorsomedial to NL and not within the same imaging field at 40X magnification, but the positioning of NM relative to NL is altered following transfection with full-length EphB2 at E2 (D) or treatment with EphA4-Fc from E6 to E9 (C). Measurements of NL (E, F) demonstrate the reduced extent of NL along both rostrocaudal (E) and mediolateral (F) axes when EphB2 is overexpressed from E2, and along the rostrocaudal axis (E) when all Eph receptors are inhibited with EphA4-Fc injection from E6 to E9. Scale bar = 100 μm, left is lateral and dorsal is up.

Figure 7

Illustration of Eph-ephrin driven events important for normal development of the NM-NL circuit. Left-sided NM and NL are schematized in a coronal view similar to sections shown in Figures1, 2, 3, 4, 5, 6 with dorsal up and medial to the right. Migratory and axon targeting events during embryonic days 5 through 10 (E5 to E10) are described. The black line represents ipsilateral NM projections (that is, those arising from the left side) and the red line represents contralateral NM projections (that is, those arising from the right side, not shown).