Ephrin-B2 reverse signaling is required for topography but not pattern formation of lateral superior olivary inputs to the inferior colliculus

Abstract

Graded and modular expressions of Eph-ephrins are known to provide positional information for the formation of topographic maps and patterning in the developing nervous system. Previously we have shown that ephrin-B2 is expressed in a continuous gradient across the tonotopic axis of the central nucleus of the inferior colliculus (CNIC), whereas patterns are discontinuous and modular in the lateral cortex of the IC (LCIC). The present study explores the involvement of ephrin-B2 signaling in the development of projections to the CNIC and LCIC arising from the lateral superior olivary nuclei (LSO) prior to hearing onset. Anterograde and retrograde fluorescent tracing methods in neonatal fixed tissue preparations were used to compare topographic mapping and the establishment of LSO layers/modules in wild-type and ephrin-B2(lacZ/+) mice (severely compromised reverse signaling). At birth, pioneer LSO axons occupy the ipsilateral IC in both groups but are delayed contralaterally in ephrin-B2(lacZ/+) mutants. By the onset of hearing, both wild-type and mutant projections form discernible layers bilaterally in the CNIC and modular arrangements within the ipsilateral LCIC. In contrast, ephrin-B2(lacZ/+) mice lack a reliable topography in LSO-IC projections, suggesting that fully functional ephrin-B2 reverse signaling is required for normal projection mapping. Taken together, these ephrin-B2 findings paired with known coexpression of EphA4 suggest the importance of these signaling proteins in establishing functional auditory circuits prior to experience.

Figure 1

Diagram illustrating layered CNIC and modular LCIC arrangements. LSO sends bilateral layered projections to CNIC and ipsilateral modular inputs to LCIC.

Figure 2

Localized, frequency-matched LSO-IC topography in WT mice at birth (A,B) and hearing onset (C,D). At postnatal day 0 (P0), pioneer LSO axons have invaded the ipsilateral (A) and contralateral CNIC (B) and have trajectories that appear to recognize the resident fibrodendritic architecture. At hearing onset (C,D), localized LSO dye placements yield a consistent topography, with refined layers in frequency-matched regions of the target IC. LSO insets (A,C) illustrate quantified dye locales and relative spread in each case. IC insets demonstrate fields of view (rectangles) of anterograde CNIC labeling. Insets not to scale. Curved contours demarcate the ventromedial border of the CNIC (solid line) and the LCIC/CNIC boundary (dashed line). Scale bars = 200 µm. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Figure 3

Ephrin-B2^lacZ/+ mice lack a clear LSO-IC topography yet still form discernible afferent layers. At birth, LSO axons were consistently observed in the ipsilateral IC (A), but never contralaterally (B). By the onset of hearing (C,D), localized dye placements still result in widespread projection distributions. Although a tight topography was lacking, afferent layers were apparent in both the ipsilateral and the contralateral CNIC. Insets as in Figure 2. Scale bars = 200 µm. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Figure 4

LSO-IC projection differences in WT and ephrin-B2^lacZ/+ mice. A: Quantification of WT projection topography (circles, ipsilateral; squares, contralateral) as a function of TZ center (%CNIC) vs. dye placement center (%LSO). Similar positive linear regression slopes verify topographic mapping of both inputs, such that low-frequency dye placements yield low-frequency TZs and high-frequency dye placements yield high-frequency TZs. B: Unlike WTs, ephrin-B2^lacZ/+ mice have unrefined projection distributions, lack an obvious topography, and therefore show no correlation between relative TZ center and LSO dye placement location. C: Quantification of TZ size for both uncrossed and crossed LSO inputs in both WT and ephrin-B2^lacZ/+ mutants. Although no significant difference is observed between projections in the two groups (P > 0.05), there is a significant difference for both inputs between WT and ephrin-B2^lacZ/+ mice (*P < 0.001).

Figure 5

Precise topography in WT mice evidenced by narrow bands of retrogradely labeled LSO neurons at birth (A,B) and hearing onset (C,D). Focal dye placements in the CNIC (insets A,C) reliably produce restricted bands of retrogradely labeled cells in matching frequency aspects of the ipsilateral (A,C) and contralateral LSO (B,D). Similar to anterograde findings, bilateral LSO-IC connections are established at birth (A,B), already displaying considerable projection topography. As experience ensues (C,D), an even finer topographic connectivity is apparent, with highly restricted bands of retrogradely labeled LSO cells. IC insets (A,C) illustrate quantified dyes locales and relative spread in each case. CNIC dye placements often resulted in some ipsilateral, retrogradely labeled MSO and/or periolivary neurons. MSO, medial superior olive. Scale bars = 200 µm. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Figure 6

Ephrin-B2^lacZ/+ mice exhibit widely distributed retrograde LSO labeling despite focal IC dye placements. In contrast to the case in WT mice, localized CNIC dye placements result in retrogradely labeled neurons throughout the tonotopic axis of the LSO (A,C,D). Similar to anterograde findings in ephrin-B2^lacZ/+ mutants at birth, diffuse LSO connections were observed ipsilaterally (A) but not contralaterally (B). An absence of topography persists for the uncrossed (C) and crossed (D) LSO projections leading up to hearing onset, with localized dye placements resulting in extensive labeling throughout the LSO tonotopic axis. Evidence of ipsilateral MSO and superior periolivary nuclear (SPON) labeling was not uncommon. Insets as in Figure 5. Scale bars = 200 µm. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Figure 7

Brightness profiles and autocorrelation functions confirm periodic CNIC layering in both WT and ephrin-B2^lacZ/+ mutants leading up to hearing onset. Single-channel images of anterograde labeling (A,C,E,G) for WT (A–D; cases with large dye placements) and mutants (E–H) were analyzed for periodicity (i.e., afferent layering). Brightness profiles (B,D,F,H) and corresponding autocorrelation function maxima (0.7709, 0.7894, 0.6817, 0.6830, respectively; all cases >0.6) show periodic signal components that were not significantly different (Student’s t-test results for all comparisons, P > 0.05) and closely match the observed biological layering data. Though lacking a strict topography, LSO projections still form discernible layers in the ipsilateral and contralateral CNIC of ephrin-B2^lacZ/+ mice. Boxed areas depict orientation of sampling areas for brightness profiles. Scale bars =200 µm.

Figure 8

Brightness profiles and autocorrelation functions of uncrossed LSO inputs in WT and ephrin-B2^lacZ/+ mutants at P12. Curved contour sampling (dashed) of LSO axonal labeling in LCIC (A: WT; C: ephrin-B2^lacZ/+) and corresponding brightness profiles (B,D; autocorrelation maxima = 0.774, 0.957, respectively). Modular axonal labeling (arrowheads, A,C) corresponds to peaks on matching brightness profiles (arrows, B,D). Strong autocorrelation maxima confirm modular compartmentalized LSO inputs to the LCIC in both WT and ephrin-B2^lacZ/+ mutants, with no statistically significant difference between groups (Student’s t-test, P > 0.05). Scale bars = 200 µm.

Figure 9

Summary of known Eph-ephrin IC expression and potential signaling model for LSO-IC projections. A: Schematic of Eph-ephrin protein expression in neonatal mouse leading up to the onset of hearing. EphA4 and ephrin-B2 are coexpressed in LCIC modules and graded across the CNIC. Additional experiments are ongoing to determine whether their LCIC modular expression is superimposed or complementary. In contrast, preliminary observations suggest that ephrin-B3 is absent in the CNIC and most concentrated in nonmodular LCIC zones, similar to what has been reported for EphA7 (Torii et al., 2012). Though not noted here, expression of each of these proteins is significantly downregulated as experience ensues (see Miko et al., 2007; Gabriele et al., 2011). B: Potential Eph-ephrin interactions in ordering developing LSO axons with target IC neurons. EphA4-expressing LSO axons likely encounter target IC neurons that express ephrin-B2 (CNIC and LCIC). Our ephrin-B2^lacZ/+ mutant is capable of forward signaling (into the Eph-expressing axons; bottom) yet, because of a truncated cytoplasmic domain, is incapable of reverse signaling (into the ephrin-expressing target IC cell; top). Topographic mapping is disrupted in these mutants, but layered and modular IC pattern formation remains unaffected.