Topography and Development of TN
In mammals, olfactory placodes become evident as oval-shaped, thickened areas of the anterolateral ectoderm of the head in the late ontogenetic stage 2 of Štěrba's (1985, 1990) classification system. In the mouse-eared bat, Myotis myotis, this should happen at about gestational day (D) 25 and at a CRL of about 3.8 mm [calculated on the basis of Štěrba (1990)]. In early stage 3, the OPs show horseshoe-like ridges and begin to develop simple olfactory pits. In Myotis myotis, this presumably takes place at about D27 and a CRL of 4 mm. In recent studies, Whitlock et al. (2003, 2004) demonstrated the origin of at least the GnRH- immunopositive subset of TN cells from the anteriormost edge of the neural crest in D3 zebrafish [probably corresponds to stage 3 of Štěrba (1990)]. In human embryos, the neural crest was first observed by Müller and O'Rahilly (2004) in the Carnegie stage 13 [about D27, which corresponds to stage 3 of Štěrba (1990)]. Müller and O'Rahilly (2004) investigated over 300 serially sectioned human embryos and clearly demonstrated nasal crest material in several Štěrba stage 4 specimens directly underneath the basement membrane of the olfactory placode. This material originates from the nasal plate at the neurosomatic junction, as does the neural crest, and contains LHRH precursor cells (Verney et al., 2002). Müller and O'Rahilly (2004) observed that crest cells began to adhere to each other and to form cords, which a little later began to migrate toward the region of the olfactory tubercle. In this context, one can assume that the majority of TN neurons, particularly the endocrine cell populations of the TN in humans, originate from crest material of the neurosomatic junction, which is a major source of endocrine cells in general (von Bartheld and Baker, 2004; Whitlock, 2004). In contrast to this, a most recent investigation of the Indian major carp (Cirrhinus mrigala) by Biju et al. (2005) very clearly demonstrated GNRH peptide only in surprisingly well-differentiated bipolar epithelial cells of the olfactory placode and not in the neural crest on D1 [probably corresponding to stage 3 of Štěrba (1990)]. In this context, it should be noted that neural crest cells may also invade the placode (von Bartheld and Baker, 2004). Almost certainly the cells that were noted to leave the placode in the youngest specimens (6.5 and 9 mm CRL) of this investigation belong to another, probably nonendocrine cell population of the TN that probably is by far larger than the endocrine one. In bats, initial terminal nerve neuroblasts (TNn) and Schwann cells, ensheathing the olfactory and TN fibers later, leave the medial wall of the OP in stage 4. Some immature neurons penetrate the internal limiting membrane of the OP in the two youngest specimens investigated by Brown (1980). The smallest embryo of this study (stage 4; 6.5 mm CRL) already shows an aggregation of TNn near the surface of the telencephalic vesicle; thus, the onset of TNn migration must have happened earlier in ontogenesis, probably early in stage 4. This is confirmed by Brown (1980), who could not detect any TN material in smaller/younger embryos of other bats (Tadarida brasiliensis, Myotis velifer, and Myotis lucifugus) of 6 mm CRL (stage 3), whereas bats of more than 7 mm CRL (stage 4 and later) exhibit large numbers of TNn adjacent to the OP (Brown, 1980; Oelschläger, 1988). In case neural crest material would be an important source of the TN cells in bats, the relevant neuroblasts should have appeared before stage 4 and Brown should have encountered them; therefore, a placodal origin of the TN material seems more convincing for the moment.
In the next stage (5), when the first fila olfactoria can be detected (Brown, 1980; Oelschläger, 1988; Jastrow, 1995; this study), outgrowing TN fibers gather in the mesenchyme rostral, ventral, and medial to the location where the primordial olfactory bulb will protrude from the telencephalic wall after being induced by olfactory fibers (Oelschläger and Buhl, 1985a, 1985b; Oelschläger, 1989). At small pit is found close to the rostral end of the OP (Fig. 3), when the majority of the TNn have left the OP. It indicates the formation of the first large lateral nasal gland. In the 9 mm CRL specimen (stage 5), most TN neurons seem to be present. Some of them are still on their way to the ganglion situated medial and rostroventral to the telencephalic vesicle. For the first time, the TN is traceable as a distinct fiber bundle in the 11 mm CRL embryo, whereas the formation of the main nasal ganglia begins at 10 mm CRL with the association of cell clusters (Brown, 1980; Oelschläger, 1988; this study). Additional smaller ganglia (S) are present on both sides of the nasal septum in the region where a vomeronasal organ normally would be expected. The central portion of the terminal nerve develops slightly later; in this study, it was first seen at 18.3 mm CRL (early stage 8): TN perikarya migrate and fibers grow out from the meningeal ganglion (M) in the direction of the future attachment site of the terminal nerve. Finally, TN neurons penetrate the leptomeninx. In part, the migration of neurons and the extension of the nerve are supplemented by a passive shift due to allometric phenomena since structures exceedingly growing next to the TN (e.g., the nasal septum) tend to stretch its fibers. Thus, the more or less spherical ganglion M of younger specimens may lengthen mainly by considerable proliferation of neighboring connective tissue in older specimens (Jastrow, 1995). Brown (1980) followed two central fiber bundles in embryos of 8 mm CRL (early stage 5) in the bat species mentioned above, one of the bundles entering the brain close to the primordial medial septal nucleus and the other in the area of the olfactory tubercle. In the material of this investigation, the first reliable central fibers are encountered at 18.3 mm CRL (early stage 8). In contrast to the findings of Brown (1980), who investigated Tadarida brasiliensis, Myotis velifer, and Myotis lucifugus, there is always one central TN fiber strand on each side in the mouse-eared bat.
From stages 5 to 8, the transformation of migrating young round cells into differentiated neurons seems to be achieved [cell processes much more prominent, increased volume of perikaryon, LHRH synthesis (Schwanzel-Fukuda and Pfaff, 1990)]. The ability of the cell somata to migrate is reduced and seems to come to an end by the development of cell processes, retaining the mostly multipolar TNn in their locations. It is surprising that Brown (1980) found Tadarida embryos of 13 mm CRL (stage 6) to be the oldest animals showing a TN in this species. Maybe this author did not investigate the anterior nasal roof in older bats where, in the present material, the majority of the TNn are found. There is considerable variation in the small septal nasal ganglia (S) and their fiber bundles (sfb), which may be lacking, even on both sides (18.3, 19, 22 mm CRL specimens), or can contain up to 1,357 TN neurons. Usually, one ganglion S is seen per side, but two ganglia of different size in combination with two sfbs are encountered as well. Most probably these ganglia persist even if a vomeronasal organ is not present in its vicinity because the neurons are involved in quick operation of the local (respiratory) nasal mucosa. For instance, they could be involved in local reflex mechanisms as discussed later in Functional Aspects. The meningeal ganglion (M) and its main fiber bundle (mfb) were found in all bats investigated in this study. In contrast to other species such as, e.g., the human or cat (Brookover, 1914, 1917; Larsell, 1918; Jastrow, 1995), there is no nasal fiber plexus with integrated small ganglia in adult Myotis myotis.
In the bat species investigated, the meningeal ganglia (M) seem to change in shape during ontogenesis; the initially irregular aggregation of cells (stages 4 and 5) becomes spherical (stage 6), develops processes of increasing length, and finally dissolves into two or more compact ganglia in the adult bat.
No Vomeronasal Organ But a Large Gland
The vomeronasal organ (VNO) is lacking in Myotis myotis (Frick, 1954), as is the case in other bats of the family Vespertilionidae (Mann, 1961; Baron et al., 1996; Wible and Bhatnagar, 1996) but not in Miniopterinae (Wible and Bhatnagar, 1996), such as, e.g., in Miniopterus schreibersi, where the organ exhibits a well-developed sensory epithelium (Cooper and Bhatnagar, 1976). In other mammals, TN fibers usually reach the VNO running next to vomeronasal nerve fibers, e.g., in the human (Brookover, 1917; Jastrow, 1995), cat, horse, and cattle (Larsell, 1918), rat (Bojsen-Møller, 1975), as well as rabbit (Huber and Guild, 1913). In the material scrutinized in this study, neither accessory olfactory bulbs (AOB) nor vomeronasal nerves are present. Conflicting data in the literature regarding the existence of a VNO in bats (Humphrey, 1936; Schneider, 1957; Mann, 1961; Cooper and Bhatnagar, 1976; Kämper and Schmidt, 1977; Bhatnagar et al., 1982) will not be discussed here. For reviews on the VNO and AOB, most of which are focused on bats, see Mann (1961), Wysocki (1979), Frahm and Bhatnagar (1980), Baron et al. (1996), Wible and Bhatnagar (1996), Bhatnagar and Meisami (1998), and Meisami and Bhatnagar (1998).
On both sides, the investigated Myotis specimens measuring more than 10 mm CRL show a small duct in the anterior third of the nasal roof, which opens into the rostromedial nasal cavity, extends caudalward and lateralward, widens slightly, and has a blind end (Figs. 3–6, 8, and 10). This duct persists up to the adult and belongs to the large lateral nasal gland I (Frick, 1954), a fact confirmed in this investigation. In an earlier study (Jastrow, 1995), the ending of the duct was not scrutinized in later fetal stages up to the adult and erroneously suggested to be a vomeronasal organ in an unusual location. The nasal opening of this duct is situated some micrometers away from the main nasal ganglion (N) of the TN, but there is no obvious connection to the latter. Later the distance between the duct and N increases to some hundred micrometers.
Criteria for defining what neural components can be included as constituents of the TN were discussed by von Bartheld (2004) with respect to the multitude of potential markers for its cells. When considering the ontogenetic development, topographical relationships, cytological criteria, and immunohistochemical data, it is obvious that the terminal nerve must have different functional implications, which in part may overlap with those of the main olfactory and accessory olfactory systems. Moreover, regarding these aspects, there are differences between species and groups of animals and there is even considerable inter- and intraindividual variation; thus, e.g., the complete TN may be absent on one side (Jastrow, 1995; Jastrow et al., 1998; this study).
It is interesting to note that TN neurons precede those of the telencephalon as to developmental criteria: up to a CRL of about 19 mm, these peripheral neurons have larger perikarya, develop processes much earlier, and stain more intensely than central neurons (Jastrow, 1995; Müller and O'Rahilly, 2004). It was only for this reason that in the mouse-eared bat of 18.3 mm CRL, some TN neurons that were continuous with the central TN fiber tract could be detected inside the telencephalic wall with routine stain.
As to their morphology, terminal nerve neurons comprise unipolar cells with pericellular baskets of gliocytes or fibrocytes, bipolar or multipolar perikarya, which may have two nuclei (Larsell, 1918; Filogamo and Robecchi, 1969; Bojsen-Møller, 1975; Mendoza et al., 1982; Wirsig and Leonard, 1986b; Zheng et al., 1988, 1990; Wirsig-Wiechmann, 1990; Oelschläger and Northcutt, 1992). In the material scrutinized in this investigation, apart from some spindle-shaped bipolar cells, multipolar TNn clearly predominated. This situation may be typical for bats in general (Brown, 1980, 1987; Oelschläger, 1988; this study), and such neurons may have a sensory or vegetative function. In their study on the TN of different mammals, Huber and Guild (1913) suggest that TN neurons mostly resemble those of cervical sympathetic ganglia. Accordingly, Larsell (1918) mentions a morphological similarity of the TNn to sympathetic neurons in that they also have pericellular baskets. In contrast to this, Brown (1980), who compared TN neurons to those of other ganglia in bats, found the best correspondence with small somata in the ganglia of cranial nerves V and IX. As a consequence, he favors a sensory function of the TN, which is supported by their placodal origin and fiber terminations between epithelial cells of the nasal mucosa. Ridgway et al. (1987) depicted TNn in the adult bottlenose dolphin with large diameters, sparse processes, and a dense sheath of potential gliocytes around their perikarya. These features are indicative of pseudounipolar neurons and may have a (somato?) sensory function (Oelschläger et al., 1987; Oelschläger and Northcutt, 1992). Interestingly, between these numerous large and rounded perikarya, a few bipolar LHRH-immunoreactive (ir) cells exist (Ridgway et al., 1987; Oelschläger and Northcutt, 1992).
During ontogenesis, the TN is an important source of LHRH (= mGnRH) neurons that innervate/invade the hypothalamus to establish the hypothalamo-hypophysial-gonadal axis (Schwanzel-Fukuda et al., 1981, 1985; Oelschläger and Northcutt, 1992; Jastrow, 1995; Schwanzel-Fukuda, 1999). This outstanding developmental phenomenon seems to be obligatory for all mammals; disturbances in the formation of the TN result in sterile individuals with the olfacto-genital (Kallmann) syndrome (Kallmann et al., 1944; Gauthier, 1960; Schwanzel-Fukuda et al., 1989). The claim that GnRH neurons of the terminal nerve arise from the neural crest, however, should be viewed with caution as it is impossible to discriminate neural crest from placodal ectoderm in teleosts and because this is in contrast to the findings of Biju et al. (2005), who report that, in the Indian major carp, GnRH first and only appears in cells of the olfactory placode.
It was shown in zebrafish that only the neuromodulatory mGnRH-positive cells are part of the terminal nerve (Whitlock et al., 2003). As mentioned above, these cells originate from the rostralmost edge of the neural crest (Whitlock, 2004). In early embryonal stages they associate with the material of the olfactory placode before it is morphologically distinct, and they migrate along TN fibers into the hypothalamus during early fetal development. In contrast, the immediately endocrine hypothalamic GnRH neurons that express species-dependent forms of GnRH (Dubois et al., 2002) constitute another cell population originating from the anteriormost neural crest (adenohypophysial placode) and reach the hypothalamus independent of the TN (Kouki et al., 2001; Whitlock et al., 2003; Whitlock, 2004). Among the prenatal bats under scrutiny, there were no specimens from ontogenetic stages early enough to check what happens to cranial neural crest material.
In teleost GnRH- and FMRFamide immunoreactive cells of the TN form a plexus in the upper inner plexiform layer of the retina before terminating on dopaminergic interplexiform amacrine cells or ganglion cells. By this and the fact that locally secreted GnRH causes horizontal processes to invaginate deeper into terminals (of cones only), the TN exerts a modulatory effect on the retina reducing its sensitivity during light adaptation processes (Behrens and Wagner, 2004). In this context, the topography of the TN fibers can explain why odorants of food extracts influence the b-wave of electroretinograms, demonstrating an olfactory influence on the visually guided behavior in fish (Weiss and Meyer, 1988). Other implications of the terminal nerve may come via colocalized GnRH/FMRFamide fibers, which innervate the basal forebrain and hypothalamus, including the hypothalamo-hypophyseal-gonadal axis, and influence the mating behavior (Oelschläger et al., 1998; Helpert, 2006). The GnRH component of the TN alone is not sufficient to explain how this system manages peripheral and central sensory and motor functions related to behavior and reproduction (Wirsig-Wiechmann, 2004). In fact, the GnRH-immunoreactive cells comprise only a minor population (∼10%) of all TN neurons while other TN cell populations are labeled by a large variety of other markers as reviewed by von Bartheld (2004). This is not surprising since the neural crest is known to generate several types of endocrine/neuroendocrine cells (von Bartheld and Baker, 2004; Whitlock, 2004).
Due to the fact that the present investigation could not involve any immunohistochemistry, the discussion was not extended beyond the essential issue of LHRH neurons and their migration.
A TN function related to olfaction is highly probable since the areas where the nerve enters the brain, i.e., the medial septal nucleus or its primordium, olfactory tubercle (Brown, 1980; Buhl and Oelschläger, 1986; Oelschläger et al., 1987; Oelschläger, 1988; Oelschläger and Northcutt, 1992; this study) are part of or related to the olfactory system. Free TN endings in and immediately beneath the mucosa of the nasal septum and in the vomeronasal organ were interpreted as part of a general visceroafferent system consisting of uni- and bipolar neurons (Huber and Guild, 1913; Larsell, 1918, 1950; Stewart, 1920; Simonetta, 1932; Simonetta and Magnoni, 1939; Pearson, 1941). Allen (1936) assumed that TN neurons are receptive to stimuli other than those adequate for trigeminal nerve terminals. Most probably, TN-specific odorants do not bind to superficial cell surface receptors reaching into the nasal mucus as is the case in olfactory sensory cells but, in this case, agents must be able to penetrate somewhat deeper to reach free nerve endings between epithelial cells. It is possible that in Myotis myotis and other bat species that lack a vomeronasal organ, TN terminals can perceive pheromones so that a vomeronasal system is not required. However, it seems premature to suggest that the TN takes over the vomeronasal function when the main olfactory system is another candidate and by far more similar to the VN system. All primates have a functional main olfactory system, and this appears to be true for bats as well. Thus, the main olfactory system may well take over for the missing VNO in both Myotis (and many other bats) as well as in Old World primates (Smith et al., 2001). In this context, Biju et al. (2005) note that the occurrence of GnRH in olfactory receptor neuron-like cells of Indian major carp larvae coincides with the onset of feeding and suggest a food intake-related chemosensory perception of this TN cell population.
Apart from the suggested shift of postnatal VNO function to the TN, the latter could also take over the prenatal functions of the VNO system, e.g., the generation of LHRH neurons usually located in the VNO. One could consider this since this investigation revealed that Myotis myotis not only lacks a VNO, but moreover not even forms a VNO primordium or an accessory olfactory bulb (at least in the investigated microslide series) as this is the case in many other mammals. There is a parallel provided by Catarrhine primates: adult Old World monkeys, including apes, and some humans lack a VNO, and lower Old World monkeys lose the VNO earlier in development than apes or humans, which appear to retain a vestige (Smith et al., 2001; for a comprehensive summary of VNO system regression in humans, cf. Trotier et al. 2000). Some studies have reported LHRH neurons in the human VNO or entangled with vomeronasal nerves during embryonic development (Boehm et al., 1994; Kjaer and Fisher-Hansen, 1996; Schwanzel-Fukuda, 1999). These neurons most probably belong to the terminal nerve. Further, LHRH neurons are associated with the TN of macaques (Quanbeck et al., 1997). These authors even describe two migratory “waves” of GnRH neurons in the macaque and related both of them to the TN. Since the TN provides a documented source of LHRH migration and a pathway for these neurons, it is certainly reasonable to suggest that it may make up for the absence of the VNO in Myotis. The scenario presented above relies on comparison to other taxa with similar characteristics. In a study on adult humans, no neurons or vomeronasal nerve fiber bundles were observed by Trotier et al. (2000). This suggests that in this special case, the VNO may not function as sensory organ. In contrast, all Old World monkeys studied to date do not develop a VNO vestige (Smith et al., 2001) and yet still have a TN.
On the other hand, blood pressure monitoring and modulation as a function of afferent TN fibers was inferred from potential sensory terminals that were encountered between smooth muscle cells in the Tunica media of adjacent cerebral arteries (Larsell, 1918; Pearson, 1941). In this context, the regulation of blood flow to the olfactory bulb and the nasal submucosa could be realized by TN branches along the supplying blood vessels. Apart from this, it is most probable that there are anastomoses between the TN and autonomic nerve fibers along these vessels. In the present study, nasal TN fiber bundles were observed in close vicinity of or directly attached to blood vessels. Therefore, it can be assumed that thin nerve terminals enter their walls. In the adult mouse-eared bat, the immediate contact of meningeal ganglia with blood vessels on both sides also implies an innervation of the latter (in the case of M. myotis: posterior cerebral arteries) (Grosser, 1904). Unfortunately, investigated sections were too thick to follow potential TN fiber endings to their target structures.
The swelling of the nasal mucosa in correlation with air flow and composition is certainly influenced by free endings of the TN in the submucosa that may detect odorants or irritating agents and quickly mediate a reaction via efferent TN fibers (perhaps of the same perikaryon in a nasal ganglion) to relevant glands and vessels (Holmgren, 1920; Larsell, 1950; Graziadei, 1976). Potential vasomotor fibers terminating at smooth muscle cells of the tunica media of blood vessels were demonstrated by Larsell (1950) and are also seen along vessels of the septal mucosa and the anterior brain (Oelschläger, 1988). Haymaker et al. (1982) therefore suggest the involvement of the TN also in the cyclic swelling of the nasal mucosa. Although quite probable, with the small nasal ganglia and their fibers located close to larger glands of the nasal septum (Fig. 6), a direct innervation of Bowman's glands by terminals of multipolar TNn could not be traced in the sections of the Myotis myotis presently examined. Perhaps the TN houses the parasympathetic (efferent?) component of innervation leading to vasoconstriction.
Free TN fiber endings were also considered to serve thermoreception (Grüneberg, 1973; Bojsen-Møller, 1975). Such an additional afferent innervation of nasal glands would allow the moistening of air by a reflex via the TN and thus serve thermoregulation. In vertebrates, thin fiber bundles of the TN were reported to be in close association with the nasopalatine nerve via anastomoses (Brookover, 1917). By this, the TN may modulate the activity of structures as different as glands and blood vessels in the nasal septum as well as the naris constrictor muscle (Wirsig-Wiechmann and Ebadifar, 2002). From the fact that, in some of the bats investigated in the present work, thin TN fibers run from the large ganglion M into the meninges, it may be assumed that there is a contribution to their innervation with possible anastomoses to those branches of the trigeminal nerve being responsible for the meninges.
In addition to the TN, autonomic nerve fibers carried by other nerves (mainly trigeminal branches) reach the nasal submucosa and influence its swelling, nasal air flow, and secretion. The persistence of the TN into the adult stage could be a sign that this innervation alone is not sufficient for the required protective mechanisms.
Despite its connections particularly with the olfactory nerve, the TN has to be regarded an independent cranial nerve (Jastrow 1995; von Bartheld, 2004; Wirsig-Wiechmann, 2004). On the one hand, this is obvious from its origin, development, course, and attachment to the brain; on the other hand, from its presence in toothed whales. These animals completely lack the vomeronasal system and lose the anterior olfactory system (olfactory nerves, bulbs, and peduncles) during early fetal development and the concomitant total reconstruction of the nasal region as a consequence of their phylogenetic adaptation to aquatic habitats (Johnston, 1914; Sinclair, 1951a, 1951c, 1966; Oelschläger and Buhl, 1985a, 1985b; Buhl and Oelschläger, 1986; Oelschläger et al., 1987; Ridgway et al., 1987; Oelschläger, 1992; Cranford et al., 1996; Klima, 1999; Cranford, 2000; Huggenberger, 2003). In toothed whales, which echolocate by means of their highly modified upper respiratory tract (epicranial complex, nasal complex) (Cranford et al., 1996, Huggenberger, 2003), the terminal nerve attains a maximal number of neurons among the Mammalia (Buhl and Oelschläger, 1986; Oelschläger et al., 1987; Ridgway et al., 1987). The coincidence of this hypertrophy with the loss of the nasal chemoreceptor systems clearly demonstrates the considerable degree of independence of the TN from these systems.
As is the case in the adult bottlenose dolphin (Ridgway et al., 1987), the terminal neurons in the adult Big Brown Bat (Eptesicus fuscus) for the most part do not express LHRH but may have to be attributed to other functional systems rather in microchiropterans (Oelschläger and Northcutt, 1992). It was suggested earlier that in toothed whales the TN innervates parts of the upper respiratory tract (Buhl and Oelschläger, 1986; Oelschläger et al., 1987) involved in the generation of sonar signals by means of a secondary pneumatically driven vocalization mechanism (Cranford and Amundin, 2004; Goodson et al., 2004). At the moment, however, it is unclear whether the TN can possibly play a similar role in the vocalization of bats. Microchiropterans use ultrasound signals of comparable intensity, directivity, and frequency range (Cranford and Amundin 2004). However, as the mouse-eared bat produces ultrasound signals with the larynx and emits them through the mouth (open-mouth vocalizer) and not the nose, it appears rather unlikely that the TN should be involved here in sound emission.