External information is relayed to the brain through precisely organized neural circuits that allow for sensory perception of the immediate environment (Cang and Feldheim, 2013; Fritzsch et al., 2019). The development of this neural wiring involves the growth and navigation of axons to their proper synaptic targets, often over large distances. Navigating axons integrate attractive and repulsive guidance information from the environment for directed extension (Kolodkin and Tessier-Lavigne, 2011). This guidance information comes in the form of both secreted and extracellular matrix-associated molecular cues of varying spatial distribution. Interpretation of these signals by growing axons into an appropriate chemotropic response allows for navigation and the formation of functional synapses with their targets (Maler et al., 1974; Chilton, 2006; Chai et al., 2015; Russell and Bashaw, 2017). This synaptic connection between sensory neurons and their centrally located targets generates a sensory map capable of faithfully transmitting information from the afferent divisions of sensory organs to their respective nuclei (Fritzsch et al., 2019).

The afferent neurons of the inner ear make precise wiring decisions in the hindbrain reflective of the topographic distribution of hair cells in the periphery (Maklad and Fritzsch, 2003; Appler and Goodrich, 2011). These specific decisions are essential for assembling the neuronal circuits dedicated for processing auditory or vestibular input. Auditory and vestibular signals are relayed from the periphery to the hindbrain by spiral ganglion neurons (SGNs) and vestibular afferents, respectively. SGNs send peripheral projections into the cochlea, where they innervate tonotopically organized hair cells. Within the cochlea, frequencies are encoded across a gradient with higher frequencies detected by hair cells in the base and lower frequencies detected by hair cells in the apex (Warr and Guinan, 1979; Bourk et al., 1981; Luo et al., 2009; Meyer et al., 2009). SGNs send their central processes to the cochlear nucleus (CN) located within the hindbrain, maintaining their tonotopic organization. Upon entering the hindbrain, SGNs will bifurcate, sending descending processes toward the posteroventral cochlear nucleus (PVCN) and dorsal cochlear nucleus (DCN), as well as ascending processes toward the anteroventral cochlear nucleus (AVCN), where they synapse on second order neurons within these nuclei (Wickesberg and Oertel, 1988; Ryugo and Parks, 2003). Vestibular afferents innervate the various peripheral sensory organs of the vestibular system and send central projections to specific areas of the vestibular nucleus (VN) and cerebellum (Dickman and Fang, 1996; Maklad and Fritzsch, 2003). While vestibular afferents and SGNs project together toward the hindbrain as the vestibulocochlear nerve, within the nerve, these two populations of neurons remain segregated, maintaining topographic fidelity to their respective peripheral sensory organs (Fritzsch, 1981).

The mechanisms that guide inner ear afferents to their peripheral hair cell targets has been a long-standing question (Fekete and Campero, 2007; Coate and Kelley, 2013; Coate et al., 2015). However, relatively little is known about the molecular guidance associated with central pathfinding. Remarkably, across all vertebrate species, auditory and vestibular afferents target the same respective dorsoventral columns of the hindbrain, indicating that central pathfinding may be occurring through a set of conserved guidance mechanisms (Duncan and Fritzsch, 2012). In support of this, transplantation studies in Xenopus and chick have suggested that inner ear afferents respond to conserved guidance cues during central circuit assembly (Elliott et al., 2013, 2015; Elliott and Fritzsch, 2018; Gordy et al., 2018). Furthermore, mouse SGNs project to the CN even in the absence of a majority of their second order synaptic targets (Maricich et al., 2009; Elliott et al., 2017), implying that the target cells themselves are not necessary for proper guidance. Taken together, these studies indicate that inner ear afferents may be responding to conserved spatiotemporal gradients of diffusible guidance cues for proper central pathfinding. Wnt/PCP signaling in the CNS is conserved across vertebrates and may be a promising avenue (Goodrich, 2008). Two Wnt/PCP signaling components, Prickle1 (Yang et al., 2017) and Frizzled3 (Fzd3) (Wang et al., 2002; Duncan et al., 2019), have been shown to be involved in central circuit assembly. In Prickle1 mutants, apical SGNs aberrantly cover a more expansive area of the CN and overlap with basal fibers (Yang et al., 2017). A similar phenotype is observed in Fzd3 null mutants, in which apical SGNs project in a disorganized manner within the CN (Duncan et al., 2019). Additionally, in Fzd3 null mutants, some vestibular afferents will project out of the VN and into the CN in mice or into the lateral line nucleus in frogs, indicating a reliance on Fzd3 within vestibular afferents as well (Duncan et al., 2019). Transplantation of ears between control and Fzd3 knockdown Xenopus embryos demonstrated that a loss of Fzd3 within the hindbrain did not result in aberrant vestibular afferent projections, whereas afferent projections were aberrant when Fzd3 expression was downregulated within the inner ear (Duncan et al., 2019). This suggests that it is Fzd3 expression within the ear and not in the hindbrain that is necessary for central guidance of auditory and vestibular afferents. Fzd3 is expressed throughout the inner ear, including inner ear afferents (Lu et al., 2011; Ghimire et al., 2018) and hair cells. Within the inner ear afferents, Fzd3 is expressed in a gradient, with a higher expression in the apex compared to the base (Duncan et al., 2019). This distinctive expression pattern suggests that the differential level of Fzd3 expression may contribute to the proper segregation of SGNs to their respective regions of the CN; however, it remains unclear if Fzd3 is acting in a cell autonomous manner within inner ear afferents or if Fzd3 expression within their peripheral hair cell targets is the root of the central defects observed in mutants (Duncan et al., 2019). Either scenario seems likely as Fzd3 has been shown to have a cell autonomous role in some instances (Wang et al., 2002; Chai et al., 2014), while in others, it is the expression of Fzd3 within other cell types that is critical for proper axon pathfinding (Qu et al., 2014).

In the present work, we utilized two Fzd3 conditional knockout models to selectively eliminate Fzd3 from individual cell populations within the inner ear. We used Neurod1-cre (Li et al., 2012) to recombine floxed Fzd3 (Hua et al., 2013) within inner ear afferents and Atoh1-cre (Matei et al., 2005) to recombine Fzd3 within inner ear hair cells (Wang et al., 2005; Fujiyama et al., 2009; Yang et al., 2010). Our results show that Fzd3-mediated regulation of inner ear afferent central pathfinding relies on the expression of Fzd3 within inner ear afferents.

The results presented here expand upon our previous study implicating a possible conserved role of Fzd3 in inner ear afferent targeting of the correct dorsoventral column in the hindbrain (Duncan et al., 2019). Furthermore, using unique ear transplantation experiments in Xenopus, this previous study determined that Fzd3 was dispensable within the hindbrain but was required by cells originating from the ear for proper afferent guidance centrally. Consequently, our present study was designed to determine which cell type(s) within the inner ear required Fzd3 expression for appropriate targeting of inner ear afferent central projections. By utilizing conditional knockout mouse lines of Fzd3 in hair cells (Atoh1-cre, Figures 1, 4) and in inner ear afferents themselves (Neurod1-cre, Figures 3, 4), we demonstrate that Fzd3 acts in a cell autonomous manner within the inner ear afferents to influence proper central pathfinding of both SGNs and vestibular afferents (Figure 5). Conditionally knocking out Fzd3 from the hair cells (Figures 1, 4) has no apparent effect on the central pathfinding of either inner ear afferent populations. Conditional deletion of Fzd3 from the inner ear afferents leads to central pathfinding defects in both SGNs and vestibular afferents (Figures 3, 4), comparable to the central phenotype observed in Fzd3 null mutants (Duncan et al., 2019). Consistent with the null mice, Neurod1-cre:Fzd3^f/f mice display aberrant SGN pathfinding, especially by apical SGNs, as well as complete mistargeting of some vestibular afferents to the cochlear nucleus. We cannot rule out a minor non-autonomous effect from another cell type not examined in the current study.

Despite multiple studies looking into cell-specific requirements for Fzd3 in axon guidance, variation exists between different cell types, with Fzd3 having a cell autonomous role in some instances of neuronal circuit assembly but a non-cell autonomous role in others. While our data supports that Fzd3 has a cell autonomous role in inner ear afferent central pathfinding, it is important to note that Fzd3 does not always have a cell autonomous role in axon guidance within the central nervous system. A study examining the role of Wnt/PCP pathways in development of facial branchiomotor neurons suggested that Fzd3 may have both cell autonomous and non-cell autonomous roles, but that they do not have to be considered separately (Vivancos et al., 2009). Their rationale, supported by the differing roles of Fzd3 determined in other studies (Wang et al., 2002; Qu et al., 2010, 2014; Chai et al., 2014), proposes that Fzd3 may function in the response of developing axons to chemoattraction cues (cell autonomous), but that Fzd3 may act in surrounding cell types to influence axon guidance. The inner ear afferent neurons have been shown to have defects in their peripheral projections due to loss of Fzd3 in a non-cell autonomous manner (Ghimire et al., 2018; Ghimire and Deans, 2019). Thus, the loss of Fzd3 has a non-autonomous effect on the peripheral projections and our study shows an autonomous effect on the central projection of inner ear afferents. The cell autonomous nature of Fzd3 in controlling the central pathfinding of inner ear afferent neurons is similar to that of 5-HT and mdDA neurons of the hindbrain (Fenstermaker et al., 2010), and commissural axons in the neural tube (Lyuksyutova et al., 2003; Onishi et al., 2013; Onishi and Zou, 2017). In both of these locations it is thought that Wnts are the ligand binding FZD3.

Some studies suggest that the role of Fzd3, at least in some instances, may be to establish pioneer neurons or to organize guidepost cells (Qin et al., 2020). Therefore, it may be the loss of these critical guides in absence of Fzd3 that leads to disruption of normal projection patterns for developing axons. It is plausible that without Fzd3, an interaction that would normally occur between basal and apical fibers is missing, or a new interaction has been introduced, that is leading to the misprojections observed when Fzd3 is conditionally deleted from inner ear afferent neurons. Conversely, there may be missing or newly present cues between SGN fibers and vestibular afferents, contributing to the presence of vestibular afferents in the CN. It is unclear if the defects seen by eliminating Fzd3 using Neurod1-Cre is due to the loss from all neurons, or if the effects seen are due to elimination of Fzd3 in subpopulations of neurons. This subpopulation effect of Fzd3 will need to be clarified in future studies.

The graded expression of Fzd3 across SGNs (Duncan et al., 2019) may explain why we observed greater central pathfinding defects in apical SGNs in the Fzd3 mutants. Since Fzd3 is expressed at higher levels in apical SGNs as compared to basal SGNs (Duncan et al., 2019), loss of Fzd3 may have a more severe impact on the ability of apical afferents to properly navigate within the CN. This suggests that gradients of Wnt/PCP signaling components may be restricting the projection of a given SGN to a specific region of the CN, possibly with additional repulsive/attractive ques in the extracellular environment, as has previously been shown for the visual system. In the visual system, a gradient of another Wnt/PCP signaling component, Ryk, expressed across retinal ganglion cells, mediates repulsion from the graded expression of Wnt3 ligand in the tectum. Fzd expression in the retinal ganglion cells mediates attraction that, together with Ryk repulsion, helps establish the dorsoventral retinal mapping to the tectum (Schmitt et al., 2006). Whether SGNs are directed to a particular dorsoventral region of the CN based on the FZD3 receptor concentration and the concentration of Wnt ligands released from the dorsal hindbrain remains to be explored. Furthermore, the extent that additional Wnt/PCP components, such as RYK or other FZD receptors, are involved in inner ear pathfinding is not yet known. Fzd3 has been shown to work together with Celsr3, another PCP protein, in development of both the peripheral and central nervous system (Wang et al., 2002; Zhou et al., 2008; Vivancos et al., 2009; Qu et al., 2010, 2014; Chai et al., 2014). Wnt-mediated axon guidance has been well-established in other major neuronal tracts (Lyuksyutova et al., 2003; Liu et al., 2005; Keeble and Cooper, 2006; Fenstermaker et al., 2010) beyond the ear and eye and the expression of Wnts in the dorsal hindbrain is conserved across vertebrate species (Hernandez-Miranda et al., 2017; Elliott and Fritzsch, 2018). However, the issue of which Wnt ligand(s) play a role in central pathfinding of SGNs is complicated by the existence of 19 different Wnts in mammals that can bind to 10 different FZD receptors (Wang et al., 2012), as well as ROR (Minami et al., 2010) and RYK (Yoshikawa et al., 2003) receptors. Some promising candidates are Wnt3a and Wnt5a, based on their expression in hindbrain regions that the inner ear afferents are targeting (Louvi et al., 2007; Di Bonito and Studer, 2017) and their ability to bind to FZD3 (Dijksterhuis et al., 2014). Wnt1 is also expressed in the same hindbrain region; however, it has not been shown to bind to FZD3 (Dijksterhuis et al., 2014). Identification of Wnt ligands/receptors involved in inner ear afferent central pathfinding will be critical in beginning to unravel the role Wnt/PCP signaling plays in this process.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The animal study was reviewed and approved by the IACUC Western Michigan University.

ZS and EK carried out lipophilic dye and tdTomato experiments, data collection, and wrote the initial manuscript. KE carried out the RNAscope experiments. SS-K and PB aided in tissue collection and performed experiments. All authors conceived and designed the experiments, and edited the final manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.