- Top of page
- MATERIALS AND METHODS
- LITERATURE CITED
- Supporting Information
The trigeminal, the fifth cranial nerve of vertebrates, represents the rostralmost component of the nerves assigned to pharyngeal arches. It consists of the ophthalmic and maxillomandibular nerves, and in jawed vertebrates, the latter is further divided into two major branches dorsoventrally. Of these, the dorsal one is called the maxillary nerve because it predominantly innervates the upper jaw, as seen in the human anatomy. However, developmentally, the upper jaw is derived not only from the dorsal part of the mandibular arch, but also from the premandibular primordium: the medial nasal prominence rostral to the mandibular arch domain. The latter component forms the premaxillary region of the upper jaw in mammals. Thus, there is an apparent discrepancy between the morphological trigeminal innervation pattern and the developmental derivation of the gnathostome upper jaw. To reconcile this, we compared the embryonic developmental patterns of the trigeminal nerve in a variety of gnathostome species. With the exception of the diapsid species studied, we found that the maxillary nerve issues a branch (nasopalatine nerve in human) that innervates the medial nasal prominence derivatives. Because the trigeminal nerve in cyclostomes also possesses a similar branch, we conclude that the vertebrate maxillomandibular nerve primarily has had a premandibular branch as its dorsal element. The presence of this branch would thus represent the plesiomorphic condition for the gnathostomes, implying its secondary loss within some lineages. The branch for the maxillary process, more appropriately called the palatoquadrate component of the maxillary nerve (V2), represents the apomorphic gnathostome trait that has evolved in association with the acquisition of an upper jaw. J. Morphol. 275:17–38, 2014. © 2013 Wiley Periodicals, Inc.
- Top of page
- MATERIALS AND METHODS
- LITERATURE CITED
- Supporting Information
A major component of craniofacial primordia in vertebrates is the cephalic neural crest-derived ectomesenchyme (Le Douarin, 1982; Noden, 1983, 1988; Kuratani, 2012). Specified between the surface ectoderm and the neural plate, the neural crest produces delaminated, migratory cells with pluripotency, being able to differentiate into various cell types (Le Douarin, 1982; Gans and Northcutt, 1983; Noden, 1988). Of the neural crest, the cephalic component is characterized by its skeletogenic and mesenchymal differentiation, not found in the trunk part (Gans and Northcutt, 1983; Shimada et al., 2013). The migrating cephalic neural crest cells predominantly localize to the ventral part of the head, forming three massive ectomesenchymal cell populations. These cell populations are called, from anterior to posterior: the trigeminal crest cells that fill the mandibular arch and more rostral part of the head; hyoid crest cells in the second arch; and circumpharyngeal crest cells filling the postotic pharyngeal arches at the interface between the head and trunk portions of the embryonic body (Kuratani, 1997; Matsuoka et al., 2005). Of these, the trigeminal crest cells are so called because the distribution of these cells largely corresponds to the sensory innervation domains of the trigeminal nerves; this ectomesenchymal region is divided into mandibular and premandibular components (Johnston, 1966; Kuratani and Tanaka, 1990; Kuratani, 1997; Kuratani and Horigome, 2000). This domain is partly covered by the surface ectodermal region of trigeminal placodes. Therefore, the trigeminal ganglionic primordia also differentiate in this neural crest cell stream (Batten, 1957a; Noden, 1988; O'Neill et al., 2007; McCabe and Bronner-Fraser, 2008), similar to other sensory ganglia for branchiomeric nerves (VII–X) that develop both from cephalic crest cells and dorsal/epibranchial placodes in a segmental pattern, as seen in the pharyngeal arches (Fig. 1A). Thus, the peripheral nerve morphology is tightly associated with the craniofacial developmental programs of vertebrates.
Figure 1. Craniofacial development and anatomy of the trigeminal nerve. A: Left: The distribution of the placodes in an early pharyngula embryo of the chicken. Topographical relationships are obvious between the pharyngeal arch and epibranchial placodes. Right: The cranial ganglia and the sensory dorsal root ganglia (drg) at the later embryonic stage. the neuron are from both placode and neural crest. The trigeminal nerve (V) supplies the mandibular arch (MA) as well as the premandibular domain. Most of the facial nerve (VII) supplies the hyoid arch. Redrawn from Le Douarin (1986). B: Schematic drawing of a generalized branchial nerve. Somatosensory components are colored blue, visceral sensory components are green, and special visceral motor fibers are purple. Redrawn from Romer and Parsons (1986). C: Craniofacial primordia of the human embryo. Redrawn from Standring (2004). D: The Dlx code underlying the dorsoventral identity of the pharyngeal arch. The maxillary process (mx) develops as a dorsal growth of the mandibular arch and expresses only Dlx1/2 genes (light blue). Dlx gene expressions are dorsoventrally nested, constituting the Dlx code to specify the arch. Redrawn from Minoux and Rijli (2010). E: Developmental origins of the skeletal elements in the mouse cranium. The maxillary process derivatives are colored light blue and the mandibular process derivatives purple. The rostral components of the upper jaw, such as the premaxillary bone (opmx) and upper incisor (ui), are not mandibular arch derivatives that are patterned through the Dlx code. Redrawn from Depew et al. (2005). F: The anatomy of the trigeminal nerves of the mouse. The maxillary nerve (V) innervates most of the upper jaw and the ventral side of the nasal cavity. Redrawn from Greene (1935). For abbreviations, see the list in the text.
Download figure to PowerPoint
The trigeminal nerve is the rostralmost component of the branchiomeric nerves, the cranial nerves that innervate the pharynx. It predominantly serves cutaneous sensation in the craniofacial region and innervates the masticatory muscles arising from the first pharyngeal or mandibular arch (Goodrich, 1930; De Beer, 1937; Romer and Parsons, 1986; Standring, 2004; Kardong, 2008). Although the sensory ganglion of the trigeminal nerve appears as a single huge mass of sensory neurons in mammals, this nerve is generally understood to consist of two components: the ophthalmic and maxillomandibular nerves. These have been suggested to represent two successive cranial nerves in a hypothetical ancestor of vertebrates (Goodrich, 1930; De Beer, 1937; Romer and Parsons, 1986; see below). Although no vertebrate species exhibits an entirely separate ophthalmic nerve innervating an extra pharyngeal arch rostral to the mandibular arch, the trigeminal ganglion develops from two separate ganglionic primordia, the ophthalmic ganglion and maxillomandibular ganglion, in many vertebrate species including amniotes (Batten, 1957a,b,c; reviewed by Noden, 1993).
The name of the trigeminal nerve stems from its possession of three major branches. Of these, the ophthalmic nerve (V1) is purely sensory, associated with the ophthalmic ganglion and innervates the preoptic region of the head. This innervation domain does not contain any mandibular arch derivatives but is found rostral to the mandibular arch. In gnathostomes, the maxillomandibular nerve is further divided into two branches: the maxillary (V2) and mandibular (V3) nerves. The motor fibers innervating the mandibular arch muscles are included exclusively in the mandibular nerves, in gnathostomes. On the other hand, the maxillomandibular nerve of the lamprey, a cyclostome, has upper and lower branches, both of which contain motor fibers: these branches are not directly comparable to the V2 and V3 nerves of gnathostomes (see Johnston, 1905b; Song and Boord, 1993; Kuratani et al., 2001; see below).
The trigeminal nerve is classified as one of the branchiomeric nerves. This nerve, or at least its maxillomandibular moiety distributed predominantly in the mandibular arch, has been thought to represent a serial homologue of other members of this category, such as the facial, glossopharyngeal, and vagus nerves. As a general rule, branchiomeric nerves possess two sensory ganglia, the superior and inferior ones, as well as a certain number of branches called the pretrematic and post-trematic branches on both sides of the pharyngeal slit, and the pharyngeal branch that passes longitudinally in the pharyngeal roof (Starck, 1979; Romer and Parsons, 1986; Tanaka, 1987; Fig. 1B). Of those, it is the post-trematic branch that is generally treated as the main branch (Hauptast or Endast) of the branchiomeric nerve, carrying motor fibers to innervate the muscle in the pharyngeal arch, to which the nerve belongs. It is also the longest gustatory nerve innervating the taste buds, thus characterizing the segmental nature of branchiomeric nerves most conspicuously (Johnston, 1905b; Allis, 1920; Goodrich, 1930; Starck, 1979; Romer and Parsons, 1986; Tanaka, 1987; Kuratani and Tanaka, 1990). In this generalization, two major differences are recognized between the trigeminal and the more caudal branchiomeric nerves, highlighting the peculiarity of the trigeminal nerve among the cranial nerves. First, it lacks a long gustatory component innervating taste buds, as seen in the more caudal branchiomeric nerves. Second, the trigeminal ganglion does not develop from a typical epibranchial placode located dorsal to a pharyngeal pouch, as seen in the inferior ganglia of more caudal branchiomeric nerves (Kastschenko, 1887; Batten, 1957a,b,c; D'Amico-Martel and Noden, 1983; Webb and Noden, 1993; Fig. 1A).
The peculiarity of the trigeminal nerve is extended into its topography. As seen in many textbooks, the innervation pattern of the trigeminal nerve branches has been explained by the embryonic composition of the craniofacial primordia (Fig. 1C), just like the other branchiomeric nerves, which develop in a segmental fashion in register with pharyngeal arch segmentation (Goodrich, 1930; Jarvik, 1980; Le Douarin, 1982). In this interpretation, the ophthalmic nerve fibers grow into the medial and lateral nasal prominences and its derivatives (premandibular region; Kuratani, 1990), the maxillary nerve in the maxillary process; and the mandibular nerve in the mandibular process (Kuratani, 1990). Because the jaw has traditionally been regarded as arising from a dorsoventral division of the mandibular arch (Rathke, 1827; Romer and Parsons, 1986; also see Kuratani, 2012), it would thus be reasonable to assume the entire upper jaw to be generally innervated by the maxillary nerve. However, in mammals, recent studies have revealed that the upper jaw is a composite structure derived from both the maxillary process and from the premandibular component. Their rostralmost region, including the incisors that are innervated by the maxillary nerve, is derived from the medial nasal prominence, which is presumed to be innervated by the ophthalmic nerve (Depew et al., 2002; Lee et al., 2004; Ozeki et al., 2004; Minoux and Rijli, 2010; Talbot et al., 2010; Wada et al., 2011; Gillis et al., 2013; Fig. 1D,E). The above concept of jaw development leads to the conclusion that the innervation domain of the maxillomandibular component of the trigeminal nerve does not strictly correspond to the mandibular arch and its derivatives, even if the maxillary nerve is perfectly associated with the upper jaw as described in human anatomy (Standring, 2004; Fig. 1F).
Further insights into the evolution and embryonic morphogenesis of the trigeminal nerve will be fundamental to our understanding of the evolutionary origin of the vertebrate head, the establishment of the oral apparatus and the evolutionary acquisition of the jaw. This study aims at understanding and reconciling the apparent discrepancy between the trigeminal nerve branching patterns, in particular the maxillary nerve, and the developmental composition of the craniofacial primordia with special focus on the innervation pattern of the so-called maxillary nerve in the gnathostomes. Here, we compared embryonic development of the trigeminal nerve in several gnathostome species according to phylogeny, and found that the maxillary nerve in gnathostomes is not primarily associated with the dorsal half of the mandibular arch. Instead, it primarily possesses branches that also innervate the premandibular domain. Together with the morphological pattern of the trigeminal nerve in the cyclostomes, the maxillomandibular nerve must have extended into the premandibular domain in the common ancestor of the gnathostomes.
- Top of page
- MATERIALS AND METHODS
- LITERATURE CITED
- Supporting Information
In the present study, we described the developing trigeminal nerves in the embryos of a variety of gnathostome species, namely, M. musculus, G. gallus, P. pictus, the Bester, and S. torazame, to understand evolutionary changes in the peripheral morphological patterns of this nerve, especially in their relationships to craniofacial development. Simultaneously, we rationalized the nomenclature of the trigeminal nerve branches, in particular those arising from the maxillomandibular component of the nerve. As a result, we came to realize that the branch of the maxillary nerve called the nasopalatine nerve potentially represents a highly enigmatic component of the maxillary nerve, and is a key to understanding the craniofacial evolution of vertebrates.
Comparative and experimental embryologists have long recognized the parallel relationships between the morphology of the peripheral nerves and mesenchymal composition in developing embryos (Kuratani, 1990, 1997; Kuratani and Tanaka, 1990; Kuratani and Kirby, 1991, 1992; Kuratani et al., 1997; Kuratani and Eichele, 1993; Noden, 1993; Schlosser and Roth, 1997a,b). Thus, the peripheral morphology of the nerves reflects the body plan of embryos. As for the segmental pattern of the spinal nerves, which is widespread among vertebrates (Goodrich, 1930; Ota and Kuratani, 2008), segmentation of the mesoderm has primary roles in patterning the peripheral axonogenesis. Namely, the growth patterns of motor fibers as well as the distribution of dorsal root ganglion-forming crest cells depend on the presence of somites (Detwiler, 1934; Keynes and Stern, 1984; Tosney, 1988). As for the morphology of the branchiomeric nerves, nerve root formation and neuronal specification have been shown to be prefigured by the selective adhesion of the neural crest cell populations, especially onto the even-numbered rhombomeres of the hindbrain (Moody and Heaton, 1983a,b,c; Kuratani and Eichele, 1993), and not by the pattern of paraxial mesoderm. More peripherally, inferior ganglia arise from the epibranchial placodes that are induced by endodermal pharyngeal pouches (Begbie et al., 1999). Therefore, for the peripheral morphology of the branchiomeric nerve, pharyngeal pouch segmentation imposes the development of serially homologous patterns among the branchiomeric nerves; in that respect, the trigeminal nerve appears rather exceptional.
The traditional view of comparative morphologists assigned the ophthalmic nerve to the hypothetical ancestral pharyngeal arch once believed to have been present in front of the mandibular arch (premandibular arch). That concept was highly influenced by the idea of head segmentation. Whether the paraxial mesoderm in the head contains segmented blocks of compartments serially homologous with somites in the trunk, it was almost unanimously accepted that the pharyngeal arches are serially homologous among themselves, each of which is accompanied by branchiomeric cranial nerves (V1, V23, VII, IX and X; Gegenbaur, 1872; Balfour, 1878; Marshall, 1881; van Wijhe, 1882; Goodrich, 1930; De Beer, 1937). Based on the dual nature of the trigeminal nerve, the existence of one or more premandibular “arches” was assumed by some morphologists, and prespiracular gill slits were sought in structures such as eyes, mouth, and nostrils (Goodrich, 1930; Sewertzoff, 1931; De Beer, 1937; reviewed by Bjerring, 1977; Jarvik, 1980; and by Kuratani et al., 1997). It is true that the premandibular domain of the vertebrate head contains extensive neural crest-derived ectomesenchyme that develops into cartilages resembling pharyngeal arch skeletal elements (trabecula cranii: Huxley, 1874; Couly et al., 1993; Wada et al., 2011; Kuratani, 2012). Nevertheless, it lacks some fundamental embryonic structures that define pharyngeal arches such as a pharyngeal endodermal pouch, pharyngeal arch muscles and their primordia (see Kastschenko, 1887; Graham et al., 2005). In both extant and fossil cyclostomes, no evidence was found for the presence of any muscularized pharyngeal arches in front of the mandibular arch (Janvier, 1996; see Johnels, 1948 and Kuratani et al., 2004 for the nature of the lamprey trabecula). Absence of Dlx transcripts in the premandibular ectomesenchyme also emphasizes the different developmental pattern of this ectomesenchyme from the rest of the cephalic ectomesenchyme that resides in the pharyngeal arch (see below). Based on developmental and recent comparative morphological evidence, we will not discuss the morphology of the trigeminal nerve in the context of head segmentation in the present paper. Rather than regarding the ophthalmic nerve as a degenerated branchiomeric nerve and its innervation domain as a highly modified pharyngeal arch, hereafter we will simply call this domain filled with the Dlx-negative ectomesenchyme the premandibular “region” or “domain.”
Morphology of the Trigeminal Nerve
Contrary to the general understanding among developmental biologists, the innervation domain of the so-called maxillary nerve does not coincide with the pattern of upper jaw development. It is probably the most curious problem associated with the morphology of the gnathostome trigeminal nerve. Namely, the upper jaw in jawed vertebrates is generally composed of at least two different primordia: the maxillary process derived from the mandibular arch, and derivatives of a more rostrally located primordium, the medial nasal prominence. The latter represents a premandibular structure, found rostral to the mandibular arch (Lee et al., 2004; Standring, 2004; Wada et al., 2011). Thus, the neural crest's contribution and apparent morphological innervation pattern in the upper jaw do not correspond to each other in terms of crest cell contribution and embryonic composition of the craniofacial region. In mammals, the rostral part of the upper jaw, containing the premaxillary bone and incisors, is innervated by a single component of the trigeminal nerve: the maxillary nerve (Fig. 1F). Exceptions are encountered in the elasmobranchs and sturgeons, where the upper jaw skeleton is solely formed from the palatoquadrate. This suggests that their upper jaw is purely of mandibular arch origin (Figs. 7D,E and 8C,C′). Nevertheless, the situation of the upper jaw skeleton such as these examples appears to represent a derived condition, because more basal lineages of chondrichthyes, a single ethmoidal internasal septum exists in the anterior midline and it separates palatoquadrates left and right, as found in the fossil records. Moreover, the mesial part of the upper tooth row is on the ethmoidal internasal septum in some animals (e.g., Doliodus), that condition is considered to be plesiomorphic for chondrichthyes (Brazeau, 2009; Maisey et al., 2009; Khonsari et al., 2013). Thus, the above noted inconsistency regarding the maxillary nerve appears to hold generally true for living gnathostomes (as for jawless vertebrates; see below).
The dual origin of the gnathostome upper jaw in a functional sense (the tooth-bearing region of the upper jaw) was inferred from several labeling experiments (Shigetani et al., 2000; Cerny et al., 2004; Lee et al., 2004; Wada et al., 2011). Labeling the mandibular arch ectomesenchyme resulted in the distribution of the labeled cells only in the posterior part of the upper jaw as a functional unit, corresponding to the domain of the palatoquadrate: the dorsal half of the mandibular arch skeleton (Goodrich, 1930; De Beer, 1937). The more rostral moiety of the upper jaw is derived from the premandibular part of the trigeminal crest cells, especially those called the preoptic crest cells that localize in the lateral and medial nasal prominences (Johnston, 1966; Osumi-Yamashita et al., 1997; Wada et al., 2011). Thus, most probably, the enigmatic nature of the trigeminal nerve is to be ascribed to an atypical pattern of embryonic environment where the nerve develops, and especially to the fact that the mandibular arch represents one of the most functionally specialized pharyngeal arches in the differentiation of the oral apparatus. In the following discussion, we will focus on the comparative morphology and development of the maxillary nerve in gnathostomes. To understand better the evolution of the mandibular arch-related and premandibular domains in vertebrates, we also discuss the cyclostome embryonic patterns to explore further the evolutionary origin of the gnathostome jaws.
Composition of the Upper Jaw
Understanding the morphology of the trigeminal nerve is relevant to understanding the development and evolution of the mandibular arch and its derivative, the jaw. Several genetic experiments support the dual origin of the upper jaw as presented above. For example, homeobox-containing Dlx genes are upregulated in the pharyngeal arch ectomesenchyme of vertebrate embryos and, especially in the gnathostomes, the Dlx gene members are expressed in a nested fashion along the dorsoventral axis of the arches. Dlx1 and Dlx2 are expressed in the entire pharyngeal ectoderm, Dlx5 and Dlx6 in the ventral half of the arches (including the dorsoventral hinge regions) and Dlx3 and Dlx7 in the ventralmost domain of the arches (Depew et al., 2002, 2005; Fig. 1D). This pattern of Dlx family expressions, generally known as the Dlx code, is understood to underlie the dorsoventral specification of the gnathostome pharyngeal arches (Gillis et al., 2013). In the mouse, double knockout of Dlx5 and Dlx6, the genes specifying the ventral moiety of the mandibular arch, results in transformation of the identity of the lower jaw into that of the upper jaw (Depew et al., 2002). In contrast, gain-of-function experiments affecting the same genes lead to the duplication of the upper jaw identity in the upper jaw domain (Depew et al., 2005; Sato et al., 2008). Curiously, the upper jaw region of the latter mutant possesses two pairs of incisors, one pair arising from the rostral tip of the duplicated dentary and the other on the endogenous premaxillary bone. Thus, the rostral domain of the upper jaw is independent of the Dlx code, as is expected from the expression pattern of the Dlx genes and only the posterior part of the upper jaw was transformed into the entire lower jaw as the mandibular arch derivative in the double knockout mutant mouse (Sato et al., 2008).
Given the dual origin of the gnathostome upper jaw, its rostral part should be innervated by the branches of ophthalmic nerves, not by those of the maxillomandibular. That pattern corresponds exactly to our findings in the two diapsid species, G. gallus and P. pictus (Haller and Hallerstein, 1934; Watanabe and Yasuda, 1970; Figs. 4 and 5). Clearly, this prediction will only be justified if the maxillomandibular innervation is restricted to the mandibular arch derivatives. However, it is also true that the other branchiomeric nerves do not follow such a rigid pattern as typically seen in the pretrematic nerves (Alcock, 1898; Johnston, 1905a; Kuratani et al., 1997). Even the chorda tympani, which tends to be regarded as the post-trematic, major branch of the facial nerve, takes a pretrematic course in some diapsids (Kuratani et al., 1988 and references therein). In the maxillary nerve as well, many gnathostomes develop a branch, the nasopalatine nerve, that grows beyond the limit of the maxillary process recognized by its Dlx transcripts, and extends into the premaxillary bone and nasal regions (Fig. 2C′, D′). This branch is universally seen in mammals (Greene, 1935; Standring, 2004), but not in diapsids (Haller and Hallerstein, 1934; Watanabe and Yasuda, 1970).
It appears that the condition in diapsids accords well with the hitherto expected relationship between innervation patterns and developmental primordia, namely, the maxillomandibular nerves for the mandibular arch derivatives, and the ophthalmic nerve for the premandibular domain (Figs. 4 and 5). This makes the mammalian configuration somewhat exceptional, and it was apparently acquired as an autapomorphy. By observing basal lineages of gnathostomes, we performed an out-group comparison to determine the nature of the premaxillary innervation (see below).
The Nasopalatine Nerve in Gnathostomes
In the present study, we found that the nasopalatine nerve (V2) does not exist in nonmammalian amniotes (Figs. 4 and 5). This situation raises two possibilities: either the nerve is generally absent in amniotes and only mammals have acquired it; or it was lost in the lineage leading toward diapsids. However, data on out-groups are rather poor for determining the evolutionary origin of this nerve. For example, various patterns exist for the trigeminal nerve's pathway in amphibians. In the axolotl, Ambystoma mexicanum (Supporting Information Fig. S1), development of the maxillary nerve is relatively later and smaller than other branches of the trigeminal nerve. It was seen initially as fine fibers along the infraorbital lateral line nerve and possessed no branches that innervated the primary palate. Instead, the ophthalmic nerve was well developed, issuing a branch that innervated the rostral end of the palate. Similar traits could be found in urodeles, which are generally thought to represent a derived condition among chordates (Francis, 1934; Paterson, 1939).
In the African clawed frog, Xenopus laevis, the maxillary nerve tends to be diminished and a well-developed ophthalmic nerve innervates the rostral end of the upper jaw, as seen in urodeles (Paterson, 1939). The condition as such, in which the ophthalmic nerve takes over the maxillary nerve's innervation pattern, is specific to the Pipidae including Xenopus. This phenomenon is regarded as an autapomorphy of anurans (Paterson, 1939). However, in other frog genera such as Rana and Leiopelma, the maxillary nerve is divided into two rami mediolaterally, of which the medial one makes communication with the palatine branch of VII (Stephenson, 1951). The latter pattern is also observable in Discoglossus pictus (Schlosser and Roth, 1997a) and Eleutherodactylus coqui (Schlosser and Roth, 1997b).
In the Gymnophiona, the sister group of anurans and urodeles, part of the maxillary nerve passes along the palatine branch of VII (Maddin, 2011), although its exact innervation domain has not been examined (Wiedersheim, 1879; Waldschmidt, 1887).
In dipnoans (e.g., Neoceratodus forsteri), the maxillary nerve divides mediolaterally into two branches during development, of which the medial one communicates with the facial nerve (Bartsch, 1993). Distally, this nerve reaches the ventral aspect of the ethmoidal cartilage, resembling the mammalian nasopalatine nerve. However, the presence of the nasopalatine nerve was uncertain in anatomical analyses of Protopterus annectens (Pinkus, 1895). Similarly, anatomical observations of the coelacanth (Latimeria chalumnae) have also not identified the nasopalatine nerve (Northcutt and Bemis, 1993).
As noted above, the origin of the mammalian nasopalatine nerve remains uncertain. However, the rostral part of the so-called maxillary nerve in the shark and a major part of the upper branch of the maxillomandibular nerve in the sturgeon that does not innervate the upper jaw exhibits conspicuous similarity to the nasopalatine nerve in mammals (Fig. 9A). All these nerves leave the maxillary process and supply the postoptic region of the premandibular domain. Thus, the nasopalatine nerve is more likely to represent a plesiomorphic feature among gnathostomes. Also in holocephalans, the sister group of the Elasmobranchs, the maxillary nerve divides into two branches, of which the medial issues fine nerves to innervate the nasal sac. Of those, a few further extend rostrally to reach the region around the nostril to innervate the skin and muscles in that region (Cole, 1896). The latter nerve resembles the rostral component of the maxillary nerve in sharks, especially with the lack of upper jaw innervation. As was implied by the developmental pattern in the sturgeon, the nasopalatine nerve might even develop as a major component of the upper branch of the maxillomandibular nerve (Fig. 9A). Namely, the part of V2 that is restricted to the maxillary process develops to innervate this gnathostome's autapomorphic domain. Thus, the condition in some amphibians and diapsids might represent derived conditions among the entire gnathostomes.
Figure 9. Proposed scheme of the evolution of the trigeminal nerve. A: The maxillomandibular ganglion and the maxillary nerves are colored orange and the mandibular arch derivatives light blue. During evolution, the maxillomandibular nerve ancestrally supplied not only the mandibular arch derivatives but also the postoptic region of the premandibular domain. The chicken and gecko do not have a nasopalatine nerve, but this condition would have been established secondarily in the diapsid lineage. B: The schematic development of the maxillomandibular nerve in the gnathostome embryo. C: The trigeminal nerve innervation patterns in lamprey (stage 28 embryo). Redrawn from Kuratani et al. (1997). D: The schematic developmental pattern of the maxillomandibular nerve in the lamprey. For abbreviations, see the list in the text.
Download figure to PowerPoint
It has long been known that in sturgeons, whose jaws are released from the neurocranium and are derived solely from the mandibular arch, there is an extensive innervation of the rostrum and associated structures by the so-called “maxillary nerve” (Norris and Hughes, 1920; Allis, 1923; Norris, 1925; Sewertzoff, 1928; Marinelli and Strenger, 1959, 1973; Takahashi and Kobayashi, 1989; Kardong, 2008). As shown in Figures 6C,D′ and 7D,E, the nerve innervating the upper jaw in these animals arises as a branch of the mandibular nerve, as the best example to exhibit decoupling of this nerve from the rest of the maxillary nerve (although the latter no longer innervates the upper jaw). Furthermore, the nasopalatine nerve in these basal jawed vertebrates appears to show its nature as a principal component of the trigeminal nerve, because decoupling of the maxillary process from the premandibular domain does not appear to have arrested the axonal growth and innervation of the nasopalatine nerve. The dual nature of the maxillary nerve is apparently shared by other vertebrate species, as seen in the bifurcation of the maxillary nerve into medial (nasopalatine) and lateral components (maxillary process-related maxillary nerves). In the above scheme, that part of the maxillary nerve restricted to the maxillary process derivatives should more appropriately be termed the “palatoquadrate” nerve to distinguish it from the rest of the maxillary nerve, the nasopalatine nerve. Thus, the previously defined maxillary nerve consists of nasopalatine and palatoquadrate nerves anteroposteriorly, and the latter can occasionally merge with the mandibular nerve, as seen in the sturgeon.
A characteristic pattern of the palatoquadrate nerve in the sturgeon is that, unlike in other gnathostome species, some fibers take the course of the nasopalatine nerve for their proximal part. These fibers branch off from the nasopalatine nerve ventrally to enter the upper jaw (Fig. 6D,D′). This is reminiscent of some communicating fibers observed between the maxillary process and premandibular palatal domains in other embryos (Fig. 2D′). As for the formation of the common nerve trunk for the mandibular and palatoquadrate or maxillary nerves, similar morphological patterns have been observed among some osteichthyan stem groups (Allis, 1897; Norris, 1925). The flexibility or violation of the boundary between embryonic compartments in the proximal growth of the palatoquadrate nerve as described above would be most similar to the pattern observed in the avian chorda tympani (internal mandibular nerve of VII; Kuratani et al., 1988 and references therein); in the chicken this nerve mainly takes a pretrematic course. For most of the peripheral nerve branching patterns, growth of the nerve fibers is primarily confined to the craniofacial primordia, as typically seen between the medial and lateral nasal prominences and the maxillary process. However, such a restriction does not appear to be rigid along the longitudinal boundary to limit the maxillary process and the premandibular ectomesenchyme mediolaterally, where a number of fibers form communications between different nerves (Figs. 4B′–D′ and 5C′,D′). The ophthalmic nerve branches also grow secondarily across a similar boundary from the lateral aspect and innervate the upper incisors in diapsids. It is conceivable that this type of developmental variability would also have led to evolutionary variations, which is actually the case in the morphological patterns of the nerves innervating the palate.
As described in the result, in shark, the palatine branch of the VII does not communicate with trigeminal nerve branches whereas in other gnathostomes the communications are established between those nerves. This trait also has been reported that the palatine branch of VII in other shark species such as Chlamydoselachus anguineus, Cephaloscyllium umbratile, or Squalus acanthias, or in rays such as Dasyatis akajei, Raja kwangtungensis, or Mobula diabolus, resembles that of S. torazame, especially in their innervation of oral roof mucosa and upper jaw teeth (Allis, 1923; Takahashi and Kobayashi, 1989). This trait of the palatine branch appears to be a common trait among elasmobranchs. On the other hand, the palatine branch of VII in the holocephalans passes the oral roof and ventral aspect of the nasal sac and then turns caudally to innervate the upper jaw teeth, unlike in the shark (Cole, 1896). It remains unknown if this nerve is truly analogous to the palatine branch in other gnathostome groups.
Craniofacial Development and Evolution of the Trigeminal Nerve
Based on these comparisons, our proposed evolutionary sequence of trigeminal nerve morphology is summarized in Figure 9A. The gnathostome craniofacial region develops from the mandibular arch-derived maxillary and mandibular processes as well as from more rostral primordia that are divided by a pair of nasal pits into the medial and lateral nasal prominences (Johnston, 1966; Kuratani and Tanaka, 1990; Fig. 9B). These processes are formed by the neural crest cell stream that migrates rostral to the eye (preoptic crest cells; Kuratani et al., 2001). The rostral part of the upper jaw in many gnathostomes, such as the intertrabecula (the rostral part of the trabecula), the premaxillary bone that carries the incisors in mammals, the nasal capsules and nasal septum are derived from this cell population (Depew et al., 2002; Lee et al., 2004; Ozeki et al., 2004; Minoux and Rijli, 2010; Talbot et al., 2010; Wada et al., 2011; Gillis et al., 2013). There is another cell population that tends to be neglected among the premandibular crest cells populating the postoptic domain, which later differentiates into the caudal part of the trabecula (Cerny et al., 2004; Lee et al., 2004; Wada et al., 2005, 2011; reviewed by Kuratani et al., 2012).
Of the trigeminal nerve branches, the ophthalmic nerve predominantly innervates the premandibular components, and the maxillomandibular nerve innervates the mandibular arch derivatives. However, the preoptic domain is primarily a doubly innervated area whose ventral side is primarily innervated by the upper branch, the nasopalatine nerve, of the maxillomandibular nerve. Compartment-dependent restriction is only clear between the derivatives of the nasal prominences and maxillary process. Given that the maxillary prominence apparently represents a secondary formation utilized to acquire dorsoventrally biting jaws, the nasopalatine nerve might rather represent an ancestral component of the maxillomandibular nerves (Fig. 9B).
In this regard, it would be worthwhile considering the morphological pattern of the trigeminal nerves in cyclostomes (lampreys and hagfishes), a sister group of the gnathostomes. Anatomical comparisons between cyclostomes and gnathostomes have been attempted by many: the basic morphological patterns of cranial nerves are shared between the cyclostomes and gnathostomes, and the cyclostome trigeminal nerves are also divided into ophthalmic and maxillomandibular components (Alcock, 1898; Sewertzoff, 1931; De Beer, 1937; Marinelli and Strenger, 1954, 1956; Ronan, 1988; Mallatt, 1996, 2008; Oisi et al., 2013). The trigeminal nerve innervation domain in the lamprey is embryonically filled with the rostralmost cephalic crest cell population, also called the trigeminal crest cells in this animal (Horigome et al., 1999; McCauley and Bronner-Fraser, 2003; reviewed by Kuratani et al., 2001). The mandibular arch is again identified and molecularly specified as the Hox gene code-default domain among the pharyngeal arches (Takio et al., 2004). The maxillomandibular nerve of the lamprey and hagfish also possesses upper and lower branches, often called the maxillary and mandibular nerves (Alcock, 1898; Johnston, 1905a; Sewertzoff, 1931; De Beer, 1937; Marinelli and Strenger, 1954, 1956; Ronan, 1988; Mallatt, 1996, 2008; Kuratani et al., 1997; Fig. 9C). However, given that the gnathostome jaw was obtained through systematic reorganization and repatterning of the craniofacial ectomesenchyme, the latter terminology is inappropriate: the upper and lower lips of the ammocoete larva do not correspond to the upper and lower jaws of gnathostomes (Shigetani et al., 2002; reviewed by Kuratani, 2012 and Kuratani et al., 2012). In particular, the upper branch of the maxillomandibular nerve (called “V23a” by Oisi et al., 2013) contains not only sensory fibers, but also motor fibers to innervate the mandibular arch-derived upper lip muscles (Koyama et al., 1987; Kuratani et al., 2004; Murakami and Kuratani, 2008; Oisi et al., 2013).
In the lamprey oral apparatus, the upper lip contains postoptic neural crest cells that are not utilized as a component of the oral region in gnathostomes (reviewed by Kuratani, 2012 and Kuratani et al., 2012; Fig. 9D). Nevertheless, it is worth mentioning that the upper lip environment is similar to the region where the proximal part of the gnathostome nasopalatine nerve passes and communications are established between this nerve, and the palatoquadrate and palatine nerves. In other words, in the cyclostomes, the boundary is violated by the maxillomandibular nerve fibers between the mandibular arch derivatives and the premandibular domain, if the exact homologue of the nasopalatine nerve is absent from the lamprey. These lines of evidence suggest that the maxillomandibular nerve in vertebrates has always possessed both upper and lower trajectories (i.e., a shared feature of vertebrates), of which the upper primarily extended into the premandibular domain (preoptic or postoptic regions) in agnathan vertebrates. Thus, in a narrow sense, the maxillary nerve evolved when the maxillary process emerged as a new structure in the acquisition of the jaw in the lineage of late gnathostomes.