Keeping sensory cells and evolving neurons to connect them to the brain: molecular conservation and novelties in vertebrate ear development

Abstract

The evolution of the mechanosensory cellular module and the molecular details that regulate its development has included morphological modifications of these cells as well as the formation of larger assemblies of mechanosensory cell aggregates among metazoans. This has resulted in a wide diversity of mechanosensory organs. The wide morphological diversity of organs, including the associated morphological modifications of the mechanosensory cells, suggests parallel evolution of these modules and their associated organs. This morphological diversity is in stark contrast to the molecular conservation of developmental modules across phyla. These molecular data suggest that the evolution of mechanosensory transduction might have preceded that of distinct cellular differentiation. However, once a molecular network governing development of specialized cells involved in mechanosensory transduction evolved, that molecular network was preserved across phyla. Present data suggest that at least the common ancestor of triploblastic organisms, perhaps even the common diploblastic ancestor of bilaterian metazoans, had molecular and cellular specializations for mechanosensation. It is argued that the evolution of multicellular organs dedicated to specific aspects of mechanosensation, such as gravity and sound perception, are evolutionary transformations that build on this conserved molecular network for cellular specialization, but reflect distinct morphological solutions. We propose that the sensory neurons, connecting the craniate ear with the brain, are a derived feature of craniates, and possibly chordates, that came about through diversification of the lineage forming mechanosensory cells during development. This evolutionarily late event suggests a heterochronic shift, so that sensory neurons develop in mammals prior to mechanosensory hair cells. However, sensory neuron development is connected to hair cell development, likely in a clonal relationship. The theme of cellular conservation is reiterated in two examples of chordate otic diversification: the evolution of the horizontal canal system and the evolution of the basilar papilla/cochlea. It is suggested that here again, cellular multiplication and formation of a special epithelium predates the functional transformation to an 'organ' system for horizontal angular acceleration and sound pressure reception, respectively. Overall, evolution of the vertebrate ear needs to be understood as an interplay between and utilization of two gene networks or modules. One is at the level of the molecularly and developmentally conserved mechanosensory cellular module. The other is an increased complexity in the morphology of both adult mechanosensory cells and organs by the addition of end-stage and novel features and associated gene networks to detect specific aspects of mechanosensory stimuli.

Fig. 1

The morphological evolution of the craniate ear is shown. It is assumed that the outgroup had mechanosensory cells, but no ear. The hagfish ear shows a single torus with only three sensory cell patches, two rings of hair cells forming the anterior and posterior sensory canal crista and the common macula. The sensory cristae of hagfish have no cupula, a unique and likely primitive feature of chordates. Evolution results in multiplication of endorgans through developmental segregation, culminating in a total of 9 endorgans in certain limbless amphibians. In parallel, the ear becomes a labyrinth of as many as three distinct semicircular canals and three distinct recesses harboring the otoconia/otolith bearing saccular, lagenar and utricular macula. These recesses form two distinct patterns: one pattern is found among chondrychthians and lungfishes; the second pattern is found in actinopterygian and sarcopterygian fish. Sarcopterygian fishes have evolved a separate organ, the basilar papilla, that exists in most tetrapods and that becomes the mammalian cochlea. AP, amphibian papilla; AVC, anterior vertical crista; BP, basilar papilla; HC, horizontal crista; L, lagena; PN, papilla neglecta; PVC, posterior vertical canal; S, saccule; U, utricle. [Modified from Fritzsch and Beisel, 2001; Fritzsch, 2003].

Fig. 2

The independent formation of morphologically distinct sensory organs in the three clades of triploblastic animals is contrasted with the evolution of a molecular network involved in mechanosensory cell formation, which might have already evolved in diploblastic animals. Note also that organs specialized for gravity perception evolved, apparently independently, in diploblastic organisms, supporting the idea of the ancestry of mechanosensory transduction. Note that numerous genes known to affect auditory sensation are present in both ecdysozoans and chordates and therefore likely evolved in the common triploblastic ancestor. [Modified from Fritzsch and Beisel, 2003; Kozmik et al., 2003; O’Brien and Degnan, 2003]. See table 1 for details regarding the genes.

Fig. 3

The evolution of bHLH genes involved in mechanosensory cell formation and the implications of this molecular evolution for the evolution of different mechanosensory morphologies in insects and craniates are shown. Note that insects and craniates have paralogous members of each of the four families. However, although craniates have both neurogenin and atonal family members playing a role in the ear (ngn1 governs sensory neuron formation, Atoh1/ Math1 that of hair cells), insects have only one bHLH gene member involved which governs the formation of the mechanosensory cell (with axon). The distribution of insect and craniate genes of the bHLH family suggests that bHLH multiplication predated the evolution of two distinct cell types in the craniate ear, each one associated with a conserved (atonal) and a novel (ngn1) gene associated with mechanosensory development. Insect genes are in italics. [Modified from Fritzsch et al., 2000; Bertrand et al., 2002].

Fig. 4

The distribution of some of the species and the taxa they belong to are displayed to provide the context of our current understanding of mechanosensation in metazoan evolution. Note that 1,2,3, and 4 indicate steps at which certain molecular, cellular and histological specializations must have existed, and these specializations thus represent shared derived characters of the organisms constituting the crown groups. (1) A mechanosensory channel must have already existed in unicellular organisms and is thus considered ancestral to metazoans. Data on poriferans showing the existence of Pax and Pou domain factors support this notion. (2) Diploblasts share mechanosensory channels that can be experimentally exchanged between C. elegans and M. musculus. These channels belong to the TRP, ENaC/BNac or TMC channels and are in cells specialized for mechanosensation. Cells containing these channels may require Pou domain factors for their differentiation and Pax for their assembly into larger organs. (3) bHLH genes of the atonal family are essential for mechanosensory cell/hair cell differentiation. Other genes such as senseless, pannier, decapentaplegic, Eya are involved in organ formation and are also conserved across phyla, as they are found in both insects and chordates. The presence of a single bHLH gene in C. elegans might be a secondary reduction. (4) Auditory organs evolved independently in chordates and insects several times, always modifying the mechanosensory system used in gravity and chordotonal sensation in those two lineages, respectively. [Modified from Fritzsch and Beisel, 2003; Kozmik et al., 2003; O’Brien and Degnan, 2003].