The origin and evolution of the ectodermal placodes

Authors


Anthony Graham, MRC Centre for Developmental Neurobiology, King’s College London, London, UK. E-mail: anthony.graham@kcl.ac.uk or Sebastian Shimeld, Department of Zoology, University of Oxford, Oxford, UK. E-mail: Sebastian.shimeld@zoo.ox.ac.uk

Abstract

Many of the features that distinguish the vertebrates from other chordates are found in the head. Prominent amongst these differences are the paired sense organs and associated cranial ganglia. Significantly, these structures are derived developmentally from the ectodermal placodes. It has therefore been proposed that the emergence of the ectodermal placodes was concomitant with and central to the evolution of the vertebrates. More recent studies, however, indicate forerunners of the ectodermal placodes can be readily identified outside the vertebrates, particularly in urochordates. Thus the evolutionary history of the ectodermal placodes is deeper and more complex than was previously appreciated with the full repertoire of vertebrate ectodermal placodes, and their derivatives, being assembled over a protracted period rather than arising collectively with the vertebrates.

Introduction

Ectodermal placodes are focal thickenings of the cranial ectoderm that generate many different components of the sensory systems of the head. The emergence and utilisation of these embryonic structures have long been viewed as being important for the evolution of the vertebrates. Amongst the features that distinguish the vertebrates from other chordates is the accumulation of these placodal-derived sensory structures in the head. It was, therefore, hypothesised that the evolution of the ectodermal placodes was coincident with, and central to, the evolution of the vertebrates themselves. However, it has become apparent that the ectodermal placodes have a more complex and protracted evolutionary history. The aim of this review is to discuss how the different placodes are generated during development and to scrutinise the evidence for the existence of forerunners of some of the ectodermal placodes outside of the vertebrates.

The ectodermal placodes

A key feature of vertebrate cranial ectoderm is that during early development a series of thickenings, termed ectodermal placodes, arise within this tissue. These structures form at stereotypical positions reflecting the importance of localised cues in their development and subsequently undergo differential morphogenesis to generate key components of the cranial sensory apparatus. The ectodermal placodes comprise the adenohypophyseal, the olfactory, the lens, the ophthalmic and maxillomandibular trigeminal, the otic, lateral line and the epibranchial placodes (Graham & Begbie, 2000; Baker & Bronner-Fraser, 2001) (Fig. 1). The two most anterior placodes, the adenohypophyseal and olfactory, emerge from the anterior neural ridge, with the adenohypophyseal arising from the medial portion and the olfactory placodes from slightly more lateral regions. As development progresses, the relationship between these two placodes alters. With the closure of the neural folds and subsequent morphogenesis of the forebrain, the adenohypophyseal placode comes to lie ventrally, while the olfactory placodes occupy a more medial frontal position. The adenohypophyseal placode gives rise to the adenohypophysis, which will subsequently generate endocrine cells of the pituitary. The olfactory placode will form the olfactory epithelium, olfactory receptor cells, the primary sensory neurones that detect odorants, and the gonadotrophin-releasing hormone (GnRH) neuroendocrine cells. It was previously thought that, uniquely amongst the placodes, the olfactory was also a source of glial cells. However, a more recent study has shown that the olfactory glia are in fact neural crest-derived (Barraud et al. 2010).

Figure 1.

 A schematic showing the location of the ectodermal placodes in the vertebrate head. The olfactory placode (olf) coloured red is found at the tip of the forebrain, while the lens (L) placode shaded orange forms caudally as part of the eye. The adenohypophyseal (Ad) placodes, in yellow, can be seen to lie ventrally to the forebrain. The trigeminal placodes form alongside the anterior hindbrain at the levels of rhombomeres 1 and 2. The more anterior ophthalmic (To) placode is coloured dark blue and the maxillomandibular (Tmm) is shaded in a lighter blue. The large otic (Ot) placode forms opposite the central domain of the hindbrain and is in purple. The lateral line (LL) placodes form both anteriorly and posteriorly of the otic vesicle. The epibranchial placodes – geniculate (Eg), petrosal (Ep) and nodose (En) – form as part of the pharyngeal series. They are shaded green. Fore – forebrain, mid – midbrain, r1 – rhombomere 1, r2 – rhombomere 2, r3 – rhombomere 3, r4 – rhombomere 4, pa1 – pharyngeal arch1, pa2, pharyngeal arch 2, pa3 – pharyngeal arch 3, pa4 – pharyngeal arch 4, pa5 – pharyngeal arch s-somite.

The lens placode, which generates the crystalline-containing cells of the lens, is also associated with the forebrain but forms more posteriorly alongside the region of the diencephalon that will give rise to the optic cup. More caudally, the trigeminal placodes, the ophthalmic (or profundal) and maxillomandibular, emerge at the level of the midbrain-hindbrain boundary. Although both of these structures will give rise to somatosensory neurones of the trigeminal ganglion that will innervate the jaws and facial region, these placodes differ from each other. The more anterior ophthalmic (or profundal) placode directly generates post-mitotic neurones within the ectoderm that then migrate to the forming ganglion, whereas the maxillomandibular placode produces neuroblasts that terminally differentiate at the site of ganglion formation (Begbie et al. 2002; McCabe et al. 2009). It is also important to note that although these two placodes contribute to a single trigeminal ganglion in amniotes, they generate distinct ganglia in many gnathostomes. For example, clearly distinct profundal and trigeminal placodes giving rise to correspondingly distinct ganglia are readily observable in shark embryos (O’Neill et al. 2007).

Further posteriorly, the otic placode forms alongside the central region of the hindbrain and it is from this structure that the entire inner ear forms. This includes the vestibular and auditory apparatus, their mechanosensory hair cells and the neurones of the vestibuloaccoustic ganglion. In anamniotes, lateral line placodes also form anterior and posterior of the otic placode, and although the number of lateral line placodes can vary it has been suggested that ancestrally there were six: three pre-otic and three post-otic (Northcutt, 1997). These placodes generate both the sensory receptive organs of the lateral line system and the sensory neurones that innervate these structures. Finally, the epibranchial placodes arise in more lateral and ventral regions in close association with pharyngeal arches, with each epibranchial placode arising close to the tips of the clefts between the pharyngeal arches. Indeed, the development of the epibranchial placodes occurs within the overall development of the pharyngeal metamere (Graham, 2008). These placodes give rise to neuroblasts that migrate, and contribute post-mitotic neurones, to the geniculate, petrosal and nodose ganglia, which in turn relay gustatory and viscerosensory information from the oro-pharyngeal cavity (Blentic et al. 2011). In all vertebrates, the most anterior epibranchial placode is the geniculate, the second most anterior the petrosal and the more posterior placodes, which vary in number from one in amniotes to four in shark embryos, is the nodose (Begbie et al. 1999; O’Neill et al. 2007).

Stepwise development of the ectodermal placodes

The generation of the ectodermal placodes and the formation of their specific derivatives involve a progression from an early naïve state through subsequent steps of commitment wherein placodal fates become progressively fixed. This was clearly demonstrated in the classic transplantation experiments carried out by Jacobson in the 1960s (Jacobson, 1963). He showed that if the cranial ectoderm flanking the open neural plate were rotated along the rostrocaudal axis, such that the presumptive otic region comes to lie at the position normally occupied by the olfactory placode, then this region would now form an olfactory apparatus. However, if the ectoderm were rotated later, at neurula stages, the ectoderm would not simply follow the cues of its new position but would continue to follow its normal fate. In keeping with these observations, more recent fate-mapping studies have shown that precursors of adjacent placodes are intermingled with each other and with epidermal precursors at early stages, and that as development progresses they resolve into discrete specific placodal primordia (Streit, 2002; Pieper et al. 2011).

A feature that unites all of the placodes is that they have a common origin in the cranial ectoderm; placodes contributing to sensory structures are not found in the trunk. It has therefore been suggested that a first step in the development of the placodes is the formation of a pre-placodal primordium, which comprises a region of ectoderm that flanks the anterior neural plate border that expresses a number of different transcription factors (Schlosser, 2006). However, of the genes that are expressed in this territory, which includes Dlx, GATA and Fox genes, it is the members of the Six and Eya gene families that are the most significant. The Six genes, of which there are generally six family members, encode transcription factors that act with their co-factors the Eya genes (which have four members that encode protein phosphatases) in a number of developmental contexts including placode development and, significantly, all placodes are united by their expression of members of these gene families.

As development proceeds, two distinct multi-placodal territories emerge (Schlosser, 2006). The more rostral region lies adjacent to the anterior neural folds and will give rise to the adenohypophyseal, olfactory and lens placodes. This area can be defined through its expression of Pitx, Six 3/6 and FoxE genes, amongst others. It has been shown that the formation of the adenohypophyseal placode from within this territory is promoted by hedgehog signalling from the midline (Cornesse et al. 2005; Dutta et al. 2005; Zilinski et al. 2005). Thus hedgehog overexpression results in the expression of adenohypophyseal markers in the presumptive lens region, while loss of hedgehog signalling sees the expansion of the olfactory and lens placodes into the adenohypophyseal domain. Contrastingly, the decision between olfactory and lens placodes is mediated by FGF and Bmp signalling (Bailey et al. 2006; Sjodal et al. 2007). FGF signalling from the anterior neural plate promotes olfactory fates over lens, while Bmp signalling promotes lens over olfactory. The more posterior domain will generate the otic, lateral line and epibranchial placodes and can be defined via its expression of Pax2, Pax8 and Sox3 (Ladher et al. 2010). Numerous studies have shown the importance of Fgf signalling, from the hindbrain and/or endomesoderm, for the development of the otic placode. However, it has also been suggested that early Fgf signalling is not only required for the otic placode but is more generally involved in the development of the posterior placodal domain. Wnts are also important in otic development and it has been shown that Wnts act to promote otic development at the expense of epibranchial development (Ohyama et al. 2006; Freter et al. 2008). At even later stages, Bmps play a key role in promoting further epibranchial development including the neurogenesis within the placode (Begbie et al. 1999; Holzschuh et al. 2005; Kriebitz et al. 2009). Finally, the trigeminal placodes, the ophthalmic and maxillomandibular, would seem to emerge from ectoderm lying between these two territories and it has been shown that the development of both of these placodes requires FGF and wnt signalling, with PDGF signalling also being involved in ophthalmic development (Lassiter et al. 2007; Canning et al. 2008; McCabe & Bronner-Fraser, 2008).

Each individual placode will subsequently go on to generate quite distinct derivatives. Some, such as the olfactory and the otic placodes, will give rise to very complex structures (Graham & Begbie, 2000; Baker & Bronner-Fraser, 2001). The olfactory placode invaginates and will generate the sensory and respiratory epithelium of the olfactory apparatus. This further includes the formation of the olfactory receptor cells which project axons to the olfactory bulb, and the GnRH neuroendocrine cells that migrate from the placode into the forebrain. The otic placodes also develop as a thickening that invaginates but this structure goes further, forming a vesicle that pinches off from the surface ectoderm and subsequently goes on to form the very elaborate structures of the inner ear – the semicircular canals and auditory tube – and the hair cells and the sensory neurones that innervate them which relay vestibular and auditory information to the brainstem. The lateral line placodes of anamniotes also display a relatively complex morphogenesis and come to generate both receptor cells and the sensory neurones that innervate them. Other placodes, however, such as the trigeminal and epibranchial placodes are much simpler. These exist as thickenings of the ectoderm which do not invaginate, and from which neuroneal cells delaminate to generate the sensory neurones of their corresponding cranial ganglia. Finally, the lens placode will simply form the lens epithelium.

Ectodermal placodes and the evolution of the vertebrates

Almost 30 years ago, Northcutt & Gans (1983) put forward a very influential hypothesis for the emergence of the vertebrates from protochordate ancestors. They suggested that the evolution of the vertebrates was associated with the assumption of an active predatory lifestyle, and that the major differences between the vertebrates and other chordates lay in a number of modifications to the head. They proposed that these modifications had their developmental origins in the muscularisation of the hypomere and the emergence of two novel embryonic cell types, the neural crest and the ectodermal placodes. The muscularisation of the hypomere underpinned alterations to the pharyngeal region that would have promoted a shift from this structure acting to aid filter feeding towards the pharyngeal apparatus being more involved in respiration and thus supporting a more active lifestyle. Neural crest cells were viewed as being of importance, as they are the source of much of the cartilage and bone and dentine of the vertebrate head, and it was thought that these cells evolved with the vertebrates. Finally, the emergence of the ectodermal placodes was key to the amplification of the sensory systems of the vertebrates. Gans & Northcutt (1983) noted that in comparison with the sense organs of protochordates, which show little apparent specialisation, vertebrates display paired special sense organs – olfactory, optic, and otic – as well as cranial ganglia that relay gustatory information from the taste buds and electro- and mechanosensory information from the lateral line organs. These structures, which develop in the head region, are central to the predatory lifestyle of the vertebrates and it was therefore suggested that the ectodermal placodes were also a vertebrate innovation.

Since that time, however, there has been a huge expansion in the general area of evolutionary developmental biology and in particular in our understanding of the modifications to the developmental programme that likely underpinned the evolution of the vertebrates. An emergent theme from studies of this sort has been the identification of the stepwise evolution of features that were previously thought to have appeared suddenly. For example, it is now clear that the ascidians possess cells that display key neural crest cell characteristics; these cells migrate from the neural primordium in an anteroposterior sequence, populate the periphery and generate a differentiated cell type, pigment cells, which is associated with vertebrate neural crest cells (Jeffery et al. 2004). These cells, however, probably lack the ability to generate the full repertoire of vertebrate neural crest derivatives, such as skeletogenic cell types. Thus, the key features that we associate with vertebrate neural crest cells are likely to have been assembled over a protracted period during evolution (Donoghue et al. 2008). Similarly, there has also been an expansion in our understanding of how the ectodermal placodes develop, particularly with respect to the key genes that are important for their formation. This in turn has aided our understanding of how the placodes have evolved. In particular, it is now clear that the ectodermal placodes also did not arise collectively with the evolution of the vertebrates, but that the full repertoire of vertebrate placodes and their derivatives was put together over a longer evolutionary history than previously supposed.

The evolutionary origins of the ectodermal placodes

If one is to gain any insights into the evolutionary origins of the placodes, studies of our nearest extant relatives, the cephalochordates and the urochordates, are of great importance (Fig. 2). For many decades, it was generally accepted that the cephalochordates were the sister group to the vertebrates, whereas the urochordates occupied a more basal position. However, it has now become apparent from both molecular phylogenetic and developmental studies that it is the cephalochordates that are basal, and that the urochordates are the sister group to the vertebrates (Delsuc et al. 2008). This has had a profound impact upon attempts to dissect the evolutionary history of vertebrate characteristics. In the older phylogeny, if placodal forerunners were identified in vertebrates and urochordates but not in amphioxus then it was possible that these structures had evolved with the chordates and may have been secondarily lost in cephalochordates. However, given the same results, the current phylogeny would imply that these placodal forerunners were not present in early chordates but evolved in the last common ancestor of the urochordates and vertebrates, a group called the olfactores. This would also reinforce the view that urochordates and vertebrates are sister groups. Interestingly, this is very much what is seen with forerunners of particular vertebrate placodes being more readily identified in urochordates than in amphioxus.

Figure 2.

 Phylogenetic tree of the major taxonomic groups in the Deuterostomia. The echinoderms and hemichordates form a group known as the Ambulacraria. Within the chordates, the cephalochordates are the earliest diverging lineages with the urochordates most closely related to the vertebrates.

Placodes in urochordates

Urochordate embryos possess structures which have been proposed to be homologous to placodes (Fig. 3). The first such suggestion, over 100 years ago, was that the oral siphon primordium and some of the structures that develop from it were homologous to the vertebrate adenohypophysis (Julin, 1881; Willey, 1893). This was essentially based on the position of the primordium just anterior to the brain and because, although it definitely formed the oral siphon during metamorphosis, it was also believed to form the adult neural complex, a binary structure with distinct neural and secretory components as found for the vertebrate anterior and posterior pituitary. Jefferies added to this by proposing that the ascidian atria are homologous to the otic placode (Jefferies, 1986). The atria start to form at the tadpole stage as one or two invaginations (depending upon the species) and are positioned alongside the posterior part of the brain. They develop into the atrium itself, which at metamorphosis forms the chamber into which filtered water is passed before expulsion via the atrial siphon. Jefferies’ original proposal was based on interpretation of fossil data, as well as on their position relative to the brain and their formation from the ectoderm via invagination, both characters shared with the vertebrate otic placode.

Figure 3.

 Relationships between vertebrate placodes and putative homologues in a urochordate such as Ciona intestinalis. The top panel is replicated from Fig. 1. In the middle is a schematic diagram of a urochordate tadpole larva with anterior to the right. The notochord is in red and the two siphon primordia are colour coded in green and orange, with putative relationships to subsets of vertebrate placodes shown. At the bottom is an adult urochordate, with blue arrows and colour coding showing what happens to the siphon primordia after metamorphosis.

With the advent of molecular approaches to the study of development, several authors have explored these homologies in solitary ascidians such as Ciona intestinalis at the level of gene expression and the molecular control of development. To summarise these studies, both oral and atrial siphon primordia do express genes also expressed in vertebrate placodes. In Ciona, both primordia express members of the generic placode markers of the Six and Eya gene families (Mazet et al. 2005). Significantly, however, the oral siphon primordium also expresses the anterior placode markers Pitx and Dlx (Boorman & Shimeld, 2002; Irvine et al. 2007), whereas the atrial siphon primordia express the posterior placode markers FoxI and Pax2/5/8 (Wada et al. 1998; Mazet et al. 2005). Other studies have further shown that the development of the atrial siphon primordia requires FGF signalling. In embryos treated with the FGF receptor antagonist, SU5402, or blockers of the MEK signalling pathway, atrial siphon primordia development is blocked (Kourakis & Smith, 2007). However, it should also be noted that some of the genes associated with the atrial primordia are also expressed in the developing pharyngeal slits of vertebrates, amphioxus and hemichordates (Schlosser, 2005; Kozmik et al. 2007; Gillis et al. 2012). Yet the atrial primordia differ in that they also come to express neurogenic markers at later stages (Mazet et al. 2005). Thus the overlapping expression of these genes in the buccal cavity and the atrial primordial with other genes associated with the anterior and posterior vertebrate placodal domains and their relative positions in the embryo have been used to support the hypothesis of homology of ascidian siphon primordia with vertebrate placodes.

The precise fate of cells within the two siphon primordia is less well studied. Bone & Ryan (1978) first noted that the atria contain clusters of sensory cells comprising cilia lodged in a gel matrix, presumed to be mechanosensory and hence possibly related to the mechanosensory hair cells of vertebrate otic and lateral line placodes (although the latter are secondary sensory cells, whereas the former are primary sensory cells). High quality ultrastructural work has also shown that the neural gland of the colonial ascidian Botryllus schlosseri produces neurones by delamination (a placode-like character) (Burighel et al. 1998), and that the oral and atrial siphons of many ascidians may include clusters of sensory cells inferred to be mechanosensory (Burighel et al. 1998; Mackie & Singla, 2003). At least some of these are secondary sensory cells, and the coronal organ in the oral siphon in particular includes secondary sensory cells with similarities to vertebrate hair cells, leading to the suggestion that they represent a shared character with similar cell types deriving from placode-like tissues in both vertebrates and tunicates, despite the apparent difference in positioning of the placodes themselves (Manni et al. 2006). Although the sensory cell embryological origin in B. schlosseri remains unclear, some lineage tracing has been conducted in C. intestinalis and has shown that the cerebral ganglion (the neural part of the neural complex), and the ciliated funnel that connects the complex to the oral siphon lumen, are in fact derived from the larval brain (Horie et al. 2011). However, these studies do not demonstrate a brain origin for the siphon sensory cells and hence it remains possible they are derived from the oral and atria primordia ectoderm, as per vertebrate placode sensory cells.

While most studies in this area have focused on ascidians, another lineage of urochordates, the larvaceans such as Oikopleura, have also been investigated. This lineage probably diverged early in urochordate evolution, and does not undergo the radical metamorphosis of ascidians, instead retaining the tadpole body plan throughout their life. Both morphological and molecular work has identified structures in Oikopleura that share gene expression and morphological characteristics with vertebrate placodes (Bassham & Postlethwait, 2005). It has been shown that there is a region of ectoderm at the anterior margin of the developing CNS that expresses an eya gene as well as Six genes. This region also expresses the anterior vertebrate placodal marker, Pitx. A more posterior Eya and Six expressing territory was also identified. This was the area from which the Langerhans organs will develop, which are believed to be sensory structures (Bollner et al. 1986).

Combined, these studies show that at least some placodal structures existed in the common ancestor of vertebrates and urochordates, with homology supported by similarities in position relative to major landmarks such as the neural tube, by cellular aspects of development such as invagination, by gene expression and by the type of cells that develop from them (Mazet et al. 2005). Urochordates and vertebrates have diverged considerably, making determination of the ancestral role of these placodes problematic, though we can infer this included the formation of sensory cell types. Furthermore, the gene expression profiles of the oral and atrial siphons match the profiles of the anterior (adenohypophyseal and olfactory) and posterior (otic and lateral line) placodes of vertebrates, respectively, leading to the proposal that the common ancestor of urochordates had at least two distinct territories homologous to the anterior and posterior multiplacodal groups of vertebrates (Mazet et al. 2005).

This, however, leaves the evolutionary origin of several placodes unaccounted for; namely, the lens, profundal, maxillomandibular and epibranchial placodes. The lens placode appears to be confined to vertebrates. In C. intestinalis, however, three cells known as lens cells have been described. These relatively transparent cells lie in a stack over the ocellus, the simple light-receptive organ in the tadpole brain, but are unlikely to be true lens homologues as they do not focus light and develop from the brain itself, not the non-neural ectoderm (Taniguchi & Nishida, 2004). Some insight into the evolution of the lens, however, has been gained by the study of the crystallin genes, which encode proteins present in very high concentrations in the lens and are in part responsible for the lens’ biophysical properties (Shimeld et al. 2005). The identification of a C. intestinalis homologue of the vertebrate βγ-crystallins and the discovery that its regulatory elements could drive expression in the lens of a vertebrate suggest that while the lens itself may be a genuine vertebrate innovation, much of the genetic circuitry underlying its evolution was already present in the common ancestor of vertebrates and urochordates. One possible scenario for the evolution of the vertebrate lens is that the mechanisms present in the vertebrate ancestor for regulating cell thickening and invagination (as seen in the vertebrate otic placode and ascidian atria) and for regulating βγ-crystalin regulation via lens-associated transcription factors were brought together, uniting these properties in a new lens placode (Riyahi & Shimeld, 2007). Epibranchial placode homologues have also not been identified outside the vertebrates. Indeed, taste buds, one of the structures innervated by epibranchial neurones, are believed to be vertebrate-specific. It should be noted, however, that the epibranchial placodes develop adjacent to the branchial slits, structures which in ascidians form after metamorphosis and are relatively poorly studied. Finally, to our knowledge no invertebrate homologues of the profundal or maxillomandibular placodes have been proposed.

Clues to placode origins from cephalochordates and other animals

The third taxon sharing Phylum Chordata with the urochordates and vertebrates are the cephalochordates, including amphioxus. Definitive placode homologues have yet to be identified in amphioxus, although several proposals have been made. Perhaps the most convincing of these is that the structure known as Hatschek’s pit is homologous to the adenohypophysis. In adult amphioxus this structure is found as a pit in the roof of the pharynx which makes contact with the base of the brain, and it is presumed to be secretory in nature. It develops from a larval structure called the pre-oral pit, which includes an ectodermal component and expresses the Pitx gene (Yasui et al. 2000; Boorman & Shimeld, 2002). A second candidate placode is the rostral ectoderm either side of the anterior notochord, where an organ known as the corpuscles of de Quatrefages is located (Baatrup, 1982). These are small clusters of neurones suggested to be mechanosensory or chemosensory (although evidence for either is lacking) and hence on the basis of position and innervations perhaps homologous to the olfactory placode. However, although there have been some reports of relevant genes expressed in this area (for example Msx; Sharman et al. 1999), in general, gene expression studies are inconclusive, as putative placode markers such as Eya, Six1/2 and Six3/6 do not seem to be expressed in this region, at least at the developmental stages examined to date (Kozmik et al. 2007).

A more interesting observation pertinent to vertebrate placode evolution is that amphioxus sensory neurones may develop from quite a broad area of surface ectoderm. Individual cells, identifiable by the expression of such genes as Trk and COE (Mazet et al. 2004; Benito-Gutierrez et al. 2005) and by SEM, delaminate from the ectoderm, sprout sensory projections and establish neural connections with the central nervous system. These cells may be primary and/or secondary sensory cells, both of which have been described in the epidermis of later developmental stages (Lacalli & Hou, 1999) and some may also express Six and Eya genes (Kozmik et al. 2007). Although a definitive connection between gene expression and cell type has yet to be made, these data suggest that the generation of individual sensory neurones from the ectoderm in this manner is probably an ancient character of chordates, and likely reflects the ancestral mode of neuroneal generation in animals. Indeed, we can conclude that the vertebrate neurogenic placodes are not innovations in the sense that they produce sensory neurones outside of the CNS, but perhaps are innovations in the sense that this ability has become concentrated in focused, discrete cranial territories producing defined organs or ganglia.

Studies of other invertebrates have also provided insights. The functional relationship between Six, Eya and Pax genes, first identified in Drosophila, is found in most if not all animals and is frequently involved in sensory system development (see for example Posnien et al. 2010). The connection between this small gene network and sensory system development is hence clearly very ancient. Some of the individual cell types that develop from placodes may also be ancient, for example chemosensory neurones are often marked by COE gene expression, irrespective of whether they form from placodes (vertebrates) (Caenorhabditis elegans) (Dubois & Vincent, 2001; Pang et al. 2004). Neurosecretory cells that share a gene expression signature with similar vertebrate cells are found in the brains of invertebrates (Tessmar-Raible et al. 2007) and an anterior placode-like structure expressing the adenohypophyseal/olfactory/lens marker gene Pitx has been described in the beetle Tribolium castaneum (Posnien et al. 2010). This structure also expresses Six and Eya genes and is a good candidate for a protostome placode homologue.

Summary

Vertebrate placodes share several common features but are clearly distinct in both their development and evolution. Current evidence suggests that different placodes evolved at different times (Fig. 4), with the adenohypophyseal, olfactory and otic placodes having evolved prior to the separation of the vertebrate and urochordate lineages, and the epibranchial, profundal, maxillomandibular and lens placodes having evolved in the vertebrate lineage. The adenohypophyseal placode in particular may be much more ancient. Many of the cell types, cell/tissue behaviours and genetic mechanisms involved can also be traced to more distantly related animals. These data suggest a step-wise increase in the complexity of vertebrate cranial sensory systems, with new placodes acquired over time, and novelty arising via the integration of more ancient molecular and cellular building blocks. We envisage a scenario where ancient cell types, many of which were present in the ancestral bilaterian or even before that, have been incorporated into two specialised ectodermal fields – the forerunners of placodes – that have evolved in the ancestor of the vertebrates and urochordates. Additional placodes have then evolved in the vertebrate lineage, perhaps by diversification of the ancestral pair of placodes, for example the lens placode from the primitive anterior placode domain and the epibranchial placodes from the primitive posterior placode domain.

Figure 4.

 Phylogenetic tree of the Deuterostomia, as shown in Fig. 1. Onto this are marked inferred ancestral placode characters. At the base of the tree, the development of isolated cells from the ectoderm is likely to be very ancient. The anterior pituitary territory is also old; we show two possible placements as it is clearly a chordate wide character (black arrow) but may also predate deuterostome radiation (grey arrow). We suggest paired anterior and posterior thickenings evolved before the separation of the vertebrate and urochordate lineages, with homologies as shown in Fig. 3, with further elaboration of these domains and origin of lens and trigeminal placodes within the vertebrate lineage.

Ancillary