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

The auditory system’s functionally organized topographic maps preserve spatial representations and signal attributes received by the periphery. Tonotopic maps of best frequency are the principal organizational feature exhibited by all auditory structures (Merzenich and Reid, 1974; Roth et al., 1978; Semple and Aitkin, 1979; Schreiner and Langner, 1988; Kandler et al., 2009). It has been proposed that, within patterned tonotopic arrangements, secondary ‘‘nucleotopic’’ maps exist in mosaic form in which discrete neuronal compartments or modules receive varying input arrays (Oliver and Heurta, 1992). Accurate alignment of inputs defines functional zones necessary for processing additional stimulus features (Schreiner and Langner, 1997; Fathke and Gabriele, 2009). Given the significance of topographic maps and the spatial precision necessary in defining functional auditory circuits, surprisingly little is known about the mechanisms that guide such connections early in development. The present study examines protein interactions thought to play a role in the establishment of orderly connections in the auditory midbrain or inferior colliculus (IC). The IC receives numerous converging inputs terminating within a single topographical framework (Fathke and Gabriele, 2009), making it an excellent model for studying targeting questions. Previous studies from our laboratory revealed discrete afferent patterns in the central nucleus and lateral cortex of the IC (CNIC and LCIC, layered and modular, respectively) in a variety of species (Gabriele et al., 2000a,b, 2007, 2011; Henkel et al., 2005; Fathke and Gabriele, 2009). Similarly to inputs arising from the cochlear nuclei and nuclei of the lateral lemniscus (Oliver, 1984, 1987; Kandler and Friauf, 1993; Oliver et al., 1997; Gabriele et al., 2000a,b; Fathke and Gabriele, 2009), the lateral superior olive (LSO) sends bilateral layered projections to the CNIC (Shneiderman and Henkel, 1987; Gabriele et al., 2007; Fathke and Gabriele; 2009). Though initially diffuse, uncrossed and crossed LSO terminal fields segregate into clear, interdigitating axonal layers by hearing onset (Gabriele et al., 2007). In comparison with the CNIC, the LCIC receives less subcollicular input. However, recently we reported a patchy projection in rat and mouse to deep portions of the LCIC arising from the ipsilateral LSO (Gabriele et al., 2011). The input is robust and terminates in a series of discontinuous modules that span the rostrocaudal dimension of the LCIC. This distribution mimics that of the previously described LCIC modular organization (Chernock et al., 2004). Similar to their layered counterparts in the CNIC, LSO axonal modules emerge within the LCIC during the first postnatal week and are fully defined by hearing onset (Gabriele et al., 2011). The spatial resolution necessary to establish an early topographic registry and the described LSO-IC patterning likely requires close cell-to-cell signaling via membrane-tethered guidance molecules. The Eph family of receptor tyrosine kinases, and their corresponding ligands, the ephrins, exhibit bidirectional attractant (adhesive) or repulsive (de-adhesive) binding behaviors that are known to serve as positional labels for guiding topographic map formation and precise patterning of spatially complex connections (Flanagan and Vanderhaeghen, 1998; Wilkinson 2001; Cowan and Henkemeyer, 2002; Kullander and Klein, 2002). Recently, we reported graded (CNIC) and modular (LCIC) ephrin-B2 expression that correlates temporally and spatially with developing LSO projection patterns (Gabriele et al., 2011). Furthermore, the LSO is EphA4 positive and is known to have a high binding affinity for ephrin-B2. To understand better the role of ephrin-B2 in establishing order in the auditory midbrain, we examined the projection from the LSO to the IC in control and ephrin-B2 mutant mice. Though maintaining the ability for full activation of forward signaling (ephrin-to-Eph), our ephrin-B2 lacZ mutant’s ability to reverse signal (Eph-to-ephrin) is highly compromised (Dravis et al., 2004). We show that CNIC axonal layers and LCIC modules still form in ephrin-B2lacZ/+ mutants prior to experience, despite the diminished capability for reverse signaling. However, accurate LSO-IC topographic mapping was consistently disrupted in our mutants. Overall this suggests ephrin-B2 is essential for positional information and guidance of developing IC inputs but is perhaps not necessary for the subsequent segregation into characteristic axonal patterns.