Tissue interactions and cell differentiation: neurone-sensory cell interaction during otic development

Neurones of the statoacoustic ganglion (SAG) develop a high degree of specificity within the mature inner ear, but exhibit plasticity during development (Whitehead & Morest, 1985; Pujol, 1986). It is unlikely that neuronal networks are wholly genetically determined since the genome could not possibly specify the myriad of connections that have to be made within an intact organism. To begin to understand the mechanism by which this specificity is achieved (e.g. tonotopic organization of the organ of Corti), an understanding at the cellular level of neuronesensory cell interactions is a prerequisite.

This article reviews recent findings that add to our knowledge of otic development and the interplay between SAG neurones and sensory cells that compose both the vestibular and auditory sensory receptors.

There is direct experimental evidence (Campenhout, 1935; Yntema, 1937) that the neurones that compose the amphibian SAG are wholly of placodal origin. There is no direct evidence showing the origins of these neurones in mammals and there are supporters for both placode (Halley, 1955; Batten, 1958) and neural crest (Bartelmez, 1922; Deol, 1967) contributions to the neuronal population of this ganglion. A comprehensive study of the origins of avian cephalic ganglia (D’ Amico Martel & Noden, 1983) using chickquail interspecies transplants has shown that all of the acoustic and most, if not all, vestibular neurones of the SAG are derived from the otic placode (Fig. 1). A later analysis of chimaeric chickquail labyrinths and statoacoustic ganglia that were not wholly derived from the placodal transplants define the site of origin of the SAG neurones (Noden, 1984). Both vestibular and auditory neurones originate from a common area of otic epithelium which goes on to form the medial wall of the utricule (Fig. 2). Therefore, most neurones of the SAG originate from a site quite different from their eventual peripheral targets (e.g. organ of Corti). This is significant when postulating a mechanism by which these SAG neurones send their dendritic processes to the appropriate sites within the developing inner ear. Tbese facts preclude the possibility that migrating neuroblasts establish tracts that later serve a substrate guidance role (Letourneau, 1979) for the nerve growth cones of these neurones as they innervate their appropriate inner ear sensory receptors.

Nerves can exert trophic effects on the sensory receptors that they innervate (e.g. taste buds; Guth, 1969). Ingrowth of SAG dendrites to their presumptive peripheral target sites precede overt cytodiffer-entiation of these target sensory cells in several species (chick, Knowlton, 1967; rat, Weibel, 1957; mouse, Sher, 1971; lamprey, Thornhill, 1972). These observations coupled with thymidine labelling of the developing mouse labyrinth (Ruben, 1967), which shows that the oldest hair cells of Corti’s organ are last to differentiate, suggested a causal relationship between ingrowth of SAG dendrites and cytodiffer-entiation of otic sensory epithelium. Observations of nerve fibre presence associated with otic sensory epithelium cytodifferentiation in cultured reaggregates of chick otocysts (Orr, 1968) strengthened the ‘fait accompli’ acceptance of the theory that ingrowth of SAG dendrites induces cytodifferentiation of otic sensory epithelium.

An in vitro study (Van De Water, 1976) utilizing mouse otocysts explanted either ‘with’ or ‘without’ SAG neurones directly addressed the question of whether or not these neurones and their neuritic processes exert a trophic effect on the differentiation of both vestibular and auditory sensory epithelia. Analysis of a total of 180 inner ear explants of three gestational age groups representing three distinct stages of otocyst-SAG interaction did not show any differences (light microscopy) in cytodifferentiation of either auditory or vestibular sensory epithelia (Fig. 3A, B). A similar series of ‘with’ or ‘without’ SAG otic explant (Dragone & Van De Water, 1988) excised from 13-day (gestational) mouse embryos have confirmed these findings (Van De Water, 1976) using immunity tochemical detection of neurone-specific enolase as a marker for both auditory (Fig. 4A, B) and vestibular hair cell differentiation (Dechesne, Sans & Keller, 1985; Dechesne & Pujol, 1986).

The hypothesis that inner ear sensory cells do not require an extrinsic trophic stimulus as could be supplied by SAG neurones to cyto differentiate is also supported by ultra structural findings from the developing macula utriculus of the mouse (Van De Water, Wersall, Anniko & Nordeman. 1978) and labelling studies of the embryonic rat vestibular epithelia (Sans & Chat, 1982). These findings are strengthened by the observations in the developing chick inner ear that differentiation of all labyrinthine hair cells is normal and their presynaptic specializations are maintained even though their nerve terminals are absent as a result of in ovo administered β-bungarotoxin (Hirokawa, 1977). Embryonic chick auditory receptors that are rendered aneural and grafted to the chorioallantoic membrane also develop normally in every respect (Fig. 5) when compared to chick cochlea grafts that were transplanted and developed with an intact SAG (Corwin & Cotanche. 1985).

The lateral line is capable of regenerating its sensory receptors in the absence of lateral line sensory nerves (Speidel, 1947; Jorgensen & Flock. 1976), however, there is a limited but detectable trophic influence of these nerves on the number of sensory receptors that populate the regenerated tails which may take several months post-regeneration to express its effects (Speidel, 1948; Borden & Corwin, 1985). In the developing inner ear, analysis of in vitro developed sensory receptors did not indicate an effect on either sensory cell size or number (Van De Water, 1976). Definitive proof is provided by the SEM analysis of grafted neural and aneural embryonic chick cochlea; quantification of the overall sensory cell pattern which includes cell locations, cilia size and number per individual sensory cell (Fig. 5) as well as sensory hair bundle orientation shows no differences (Corwin & Cotanche, 1985). These observations show that the inner ear, unlike the lateral line, does not exhibit any demonstrable trophic response to the presence of SAG neuronal elements and that the programme for sensory cell differentiation is intrinsic.

Dependence of neurones on their target tissues during development is the basis of the theory that competition between dendrites or axons for connections determines survival during the developmental period of normal cell death (Landmesser & Pilar, 1978). Some natural cell death occurs in the developing chick SAG shortly after the establishment of synaptic contacts with both peripheral and central targets (Ard & Morest, 1984). In vitro experiments that manipulated gangliontarget tissue interactions in otic explants of both chick (Ard, Morest & Hauger, 1985) and mouse embryos (Zhou & Van De Water, 1987a) show that both peripheral (sensory epithelium) and central (brainstem) target tissues provide trophic support of SAG neurones (Fig. 6). Recent in vitro studies (Dragone & Van De Water, 1988) examining this question with the aid of antibodies raised against neuronespecific enolase show that peripheral target tissue (i.e. 100% of explants with NSE + neurones; average 70 neurones/explant) provides better trophic support of neuronal differentiation than central target tissue (i.e. 33 % of explants with NSE + neurones; average 2 neurones/explant). These findings are consistent with a trophic role for neuronetarget tissue interactions that support neuronal survival and maturation in both avian and mammalian statoacoustic ganglia. Both peripheral and central targets are implicated with the peripheral target providing more active trophic support of these neurones in vitro.

Regenerating lateral line receptors deprived of their natural sensory nerves attract nearby spinal nerves (Speidel, 1948). Van De Water (1976) proposed that presumptive sensory epithelium of the inner ear prior to overt cytodifferentiation undergoes chemodifferentiation and establishes attractant fields that act to direct ingrowing SAG neurites towards areas of differentiating sensory epithelia. The first experimental results to support this attractant field hypothesis was provided by cocultures of two otic explants that shared a single SAG (Van De Water & Ruben, 1983, 1984; Anniko & Van De Water, 1986). These otocyst pairs were oriented so that the ganglion of the ‘with’ SAG explant was adjacent to the site of SAG removal of the closely opposed ‘without’ SAG otocyst. Analysis of these cocultured otocysts show that SAG nerve fibres grew into the otic sensory epithelia within both of these inner ear explants (Fig. 7A,B). This phenomenon occurred in both 11- and 12·5-day otocyst cocultures, but was not observed in cocultures of inner ears derived from 14-day mouse embryos. Cultures of 12-day otocyst that had their SAGs excised and replaced by 10·5-day trigeminal ganglia provide evidence that these otic epithelium attractant fields are not specific for SAG neurones (Van De Water & Ruben, 1984). Trigeminal neurites were regularly seen in association with both vestibular and auditory sensory cells of these heterotypic ganglia implant cultures. Other studies where multiple targets are interacted with a ganglion explant have also demonstrated that target tissue need not be the natural target of a ganglion to stimulate directed outgrowth of neurites from that ganglion (Ebendal & Jacobson, 1977; Pollack, Muhlach & Liebig, 1981). This observation that otic sensory epithelium can attract neurites from an implanted heterotypic ganglion suggests that the attractant fields produced by this epithelium are of a general nature and actual specificity might rely more on spatial and temporal factors as a mechanism for the attraction of specific SAG neurites to their appropriate target sites.

Study of an elasmobranch inner ear which continues to grow and add new sensory cells throughout its life shows that the auditory neurones of this labyrinth expand their terminal arbours continually and preferentially orient toward the site of postem bryonic hair cell production in the macula neglecta (Corwin, 1985). This finding also supports the hypothesis of attractant fields as a mechanism for guiding the growth cones of these continually growing neurones to the areas of newly produced hair cells within this auditory receptor.

To define better the temporal aspects of the SAG ability to respond to attractant fields, a staged series of SAG were implanted into a common age (i.e. 12-day) otic explant. This stage otocyst is in an active stage of SAG neuritic ingrowth to the as yet undiffer-entiated otic epithelium (Sher, 1971) and at a stage where neurites from both homo- and heterotypic ganglia implants are attracted to areas of differentiating sensory cells (Van De Water & Ruben, 1984). A homochronic control (i.e. 12-day) and heterochronic series of SAG (i.e. excised from 13-, 14-, and 15-day embryos) were implanted into 12-day otic explants and then allowed to develop for 8 days in vitro. Analysis of the resultant specimens show that all of these SAG implants sent fibres to areas of both vestibular (Fig. 8) auditory sensory receptors. These preliminary results suggest that SAG neurones retain their ability to extend neurites to differentiating otic sensory cells for an extended period during development of the inner ear.

Fibronectin (Rogers et al. 1983), laminin (Rogers et al. 1983; Hammarback, Palm, Furcht & Letourneau, 1985) and collagen (Roufa, Johnston & Bunge, 1983) have all been shown to promote neurite outgrowth. A cell adhesion molecule, N-CAM, has been shown to increase in the innervated sensory areas of the chick inner ear while L-CAM increased in the nonsensory areas (Richardson, Crossin, Chuong & Edelmann, 1987), This study also noted changes in the expression of both neuroneglial cell adhesion molecules, Ng-CAM and cytotactin during SAG nerve fibre invasion of the otic sensory epithelia. These authors suggest that the spatiotemporal expression of extracellular matrix molecules are significant factors in establishing patterns of innervation and specificity of neuronal connections in the chick inner ear. A preliminary study of SAG explants cultured on the surface of HEMA hydrogels which contained matrix molecules (Zhou & Van De Water, 1987b) shows that the mouse SAG responds to the two extracellular matrix molecules tested (i.e. collagen type I and fibronectin) by extending neurites.

Fig. 9 summarizes the origins of the SAG and the interplay between neurones and sensory cells of the inner ear that occur during the development of this complex sensory receptor (Van De Water, 1986). The neurones that compose the SAG are primarily of placodal origin with a minimal contribution of neural crest cells. All of the SAG supporting cells are derived from the cephalic neural crest. The hypothesis that cytodifferentiation of inner ear sensory hair cells (i.e. both vestibular and auditory) is intrinsic to these cells and does not depend upon the extrinsic stimulus that could be provided by the neuronal elements of the SAG to trigger its expression is supported by both in vivo, in vitro and in ovo studies. In contrast, the neurones that compose the SAG are dependent upon the trophic support provided by both peripheral and/or central target tissues for survival and maturation. Recent in vitro and in vivo studies support the hypothesis that differentiating inner ear sensory epithelia produce attractant fields that guide the ingrowth of SAG neurites. The neurones of the SAG appear to be able to respond to these attractant fields for an extended period during labyrinthine development. The exact mechanism by which these attractant fields are produced (e.g. biochemical extracellular matrix: Varón, Manthorpe & Williams, 1984; electrical: Patel & Poo, 1982; substrate guidance: Letourneau, 1979; Carney & Silver, 1983) by the differentiating sensory cells of the inner ear is currently not known. The attractant fields appear to have broad effects and are not specific for SAG neurites, therefore, temporalspatial information and the mechanism by which they are generated will be vital to understanding how such a process can establish specificity of neuronal connections within the inner ear. The spatiotemporal patterns of extracellular matrix molecules (ECM) are suggested to be a significant factor in the establishment of the patterns of innervation within the inner ear. These ECM results are provocative but await more definitive proof of their actual role in the establishment of patterns of innervation within the developing inner ear.

Many of the studies cited in this paper that were performed by the author were supported by grants from NIH (NS 08365), March of Dimes-Birth Defects Foundation (1-372) and The Deafness Research Foundation. Thanks for the typing of this paper are given to Mina Pozzuto. Vera Galinovic-Schwartz is thanked for all of her excellent nerve fibre stains and other histological preparations. The author also thanks Drs D’ Amico Martel and Noden for the use of Fig. 1, Dr Noden for Fig. 2 and Drs Corwin and Cotanche for Fig. 5 of this paper.