Organization of the olfactory epithelium

The pseudostratified olfactory sensory epithelium is organized similarly in all vertebrates (reviewed in Eisthen and Polese, 2006). The most superficial somata belong to the sustentacular cells, a class of secretory supporting cells. The somata of the olfactory receptor cells lie below those of the sustentacular cells, and the deepest layer contains the basal cells, which serve as progenitors for the olfactory receptor and sustentacular cells. The surface of the olfactory epithelium is bathed in specialized mucus, which is largely produced by nasal glands adjacent to the olfactory epithelium, as well as by Bowman's glands, which lie deep to the olfactory epithelium. The secretory activity of the sustentacular cells also contributes to this fluid.

Olfactory receptor cells are unusual among vertebrate sensory cells in that they are primary neurons, rather than modified epithelial cells. The unmyelinated axons of these bipolar neurons generally form small bundles that coalesce into an olfactory nerve projecting to the olfactory bulb at the rostral pole of the telencephalon. At the other pole of the cell, the unbranched dendrite extends into the mucus layer lining the olfactory organ. The tip of the dendrite can contain cilia, microvilli, or both, depending on the species (reviewed in Eisthen, 2004). Transduction occurs when an odorant interacts with a receptor at the surface of the ciliary or microvillar membrane. The most prevalent receptors in all vertebrates are members of the large family of odorant receptor (OR) genes (reviewed in Mombaerts, 2004). A second class of olfactory receptors, the transient amino acid receptors (TAARs), was recently found to be expressed in a small proportion of olfactory receptor neurons, and appears to be widely distributed across vertebrates (Liberies and Buck, 2006). Although it is not yet clear how many TAARs may be expressed in a given cell, individual olfactory receptor cells appear to randomly express only one or only a small subset of OR genes, suggesting the odorant coding may be based in part on the identity of the individual cells that are activated (Ngai et al., 1993; Vassar et al., 1993; Chess et al., 1994; Buck, 2000; Sato et al., 2007).

Fig. 1.

GnRH-immunoreactive fibers in the olfactory epithelium and bulb of a goldfish (Carassius auratus). (A) Schematic illustration of TN-GnRH cell bodies and fibers in the olfactory system of the goldfish. For clarity, the position of the brain is illustrated dorsal to the eye and the sizes of the olfactory epithelium and bulb are magnified. In the goldfish, the TN ganglion cells are located in the transitional area between the olfactory nerve and the olfactory bulb, from which they extend fibers to the olfactory epithelium and bulb. Additional GnRH-immunoreactive fibers, not shown here, run through the medial olfactory tract and extend principally to areas of the telencephalon, optic tectum, dorsal thalamus, and spinal cord as well as through the optic nerve to the retina. LP, lamina propria; OB, olfactory bulb; OE, olfactory epithelium; ON, olfactory nerve; OT, olfactory tract. (B) Photomicrograph of GnRH immunoreactive fibers (arrowheads) underneath the olfactory epithelium of a goldfish. Scale bar, 50 µm. (C) Photomicrograph of a GnRH immunoreactive cell body (arrow) and fibers in the olfactory bulb of a goldfish. Scale bar, 50 µm.

Fig. 2.

Schematic illustration of the afferent sources to the TN ganglion in teleost fishes, as described by Yamamoto and Ito (2000). For clarity, the relevant regions are shown in their approximate positions in a side view of the brain of a dwarf gourami, Cotisa lalia, a teleost fish; anterior is to the left and dorsal is up. Large neurons of this ganglion are GnRH-immunoreactive. The TN ganglion receives input from olfactory areas in the forebrain (Vv and Vs) as well as from the nucleus tegmento-olfactorius (nTO, also called the nucleus tegmento-terminalis). Because the nucleus appears to relay somatosensory and visual information, the TN ganglion probably receives multimodal sensory inputs. BO, bulbus olfactorius; Dp, area dorsalis telencephali pars posterior; FR, formatio reticularis; NP, nucleus pretectalis; nVs, nucleus sensorius nervi trigemini; TN-ggl, ganglion of the nervus terminalis; VM, nucleus ventromedialis thalami; Vs, area ventralis telencephali pars supracommissuralis; Vv, area ventralis telencephali pars ventralis.

Fig. 3.

Possible mechanisms of modulatory effects of GnRH on the olfactory receptor neurons or olfactory bulbar circuits. (A) Schematic transduction mechanism in olfactory receptor neurons that express ORs. Odorant binding to ciliary ORs activates adenylyl cyclase via a G protein, opening cyclic nucleotide-gated cation channels. The calcium that flows in then secondarily activates calcium-activated chloride channels, depolarizing the ciliary membrane. In many cases, this stimulates voltage-activated sodium channels. GnRH may modulate these processes through phosphorylation via PKA-, PKG-, or PKC-mediated pathways. AC, adenylyl cyclase; CNG channel, cyclic nucleotide-gated cation channel; G_olf, olfactory-specific G protein; OR, odorant receptor. (B) Schematic illustration of the neural circuit in the olfactory bulb of a generalized teleost fish. In teleosts, unlike mammals, the somata of periglomerular cells are distributed among the mitral cells (Fuller et al., 2006). GnRH may act on one or more types of neurons, thereby modulating the processing of olfactory information. GL, glomerular layer; ICL, internal cell layer; MCL, mitral cell layer; ONL, olfactory nerve layer.

Finally, members of the two families of vomeronasal receptor genes, the V1Rs and V2Rs, are expressed in the olfactory epithelium in teleost fishes (Cao et al., 1998; Naito et al., 1998; Speca et al., 1999; Saraiva and Korsching, 2007) and in the receptor cells of the vomeronasal organ in amphibians, reptiles, and mammals. Interestingly, members of these two families have also been found to be expressed in small numbers in the olfactory epithelium of frogs (Xenopus laevis, Date-Ito et al., 2008), goats (Capra hircus, Wakabayashi et al., 2002; Wakabayashi et al., 2007), mice (Mus musculus, Karunadasa et al., 2006), and humans (Rodriguez et al., 2000). The phylogenetic diversity of these animals indicates that olfactory epithelial expression of V1R and V2R genes may be widespread among tetrapods, suggesting the existence of four classes of olfactory receptor neurons: those expressing ORs, TAARs, V1 Rs, and V2Rs. All four categories of receptors activate G proteins when stimulated. In the majority of olfactory receptor cells, i.e., those expressing ORs, the G protein activates adenylyl cyclase, opening cyclic nucleotidegated cation channels, which secondarily activate chloride channels. Receptor binding thus leads to an influx of cations and efflux of anions, depolarizing the receptor neuron (Fig. 3A) (reviewed in Schild and Restrepo, 1998, Firestein, 2001).

Sources of GnRH in the olfactory epithelium

The olfactory organ develops from the nasal placode, a thickened disk of tissue that lies anterior to the developing neural plate. The TN ganglion cells also develop from the nasal placode and then migrate centrally during early development, with some cells migrating as far as the hypothalamus (Schwanzel-Fukuda et al., 1985; Schwanzel-Fukuda and Pfaff, 1989; Wray et al., 1989; Hilal et al., 1996; Amano et al., 1998; Amano et al., 2004; Okubo et al., 2006). One or more TN ganglia usually form along the olfactory nerve or ventral olfactory bulb, with fibers that extend rostrally to the olfactory epithelium and caudally to the preoptic area, although collaterals can branch widely throughout the brain (Oka, 1997). The terminals of the rostral processes have proven difficult to locate precisely, but the fibers appear to end in or just underneath the lamina propria that surrounds the olfactory epithelium; additional fibers appear to end near the Bowman's glands (Wirsig-Wiechmann, 1993). Thus, GnRH and other compounds released from TN fibers may diffuse upward through the olfactory epithelium to stimulate cells, or may be released in the mucus overlying the olfactory epithelium, or both.

Although the TN is the only known source of GnRH-containing fibers that project to the olfactory epithelium, in some species other structures associated with the nose are also innervated by GnRH-containing fibers from other sources; for example, in tiger salamanders (Ambystoma tigrinum), the palatine ganglion of the trigeminal nerve contains GnRH-immunoreactive fibers that innervate the naris closure muscles (Wirsig-Wiechmann, 1993; Wirsig-Wiechmann and Ebadifar, 2002). In addition, the olfactory epithelium is also well vascularized. Thus, it seems possible that GnRH could reach the cells in the olfactory epithelium from sources other than the TN. Regardless of the mechanism of access or source, GnRH receptors are present in the olfactory epithelium, indicating that GnRH has direct effects on cells in the olfactory epithelium (Wirsig-Wiechmann and Jennes, 1993; Wirsig-Wiechmann and Wiechmann, 2001; Zhang and Delay, 2007). Unfortunately, we do not yet know which specific cell types express which subtypes of GnRH receptor genes.

GnRH immunoreactivity has also been described in primary olfactory receptor neurons in the walking catfish (Clarias batrachus; Subhedar and Rama Krishna, 1988) and in the Indian major carp (Cirrhinus mrigala; Biju et al., 2003; Biju et al., 2005). In the latter, expression occurs only in young larvae and in adult females, and is seasonal, peaking during the prespawning period (Biju et al., 2003; Biju et al., 2005). In both species, GnRH-immunoreactive olfactory receptor neurons project to glomeruli in the olfactory bulb (Subhedar and Rama Krishna, 1988; Biju et al., 2003; Biju et al., 2005), but it is not known whether these cells release GnRH into the olfactory epithelium, olfactory bulb, or both.

Physiological effects of GnRH in the olfactory epithelium

In axolotls (Ambystoma mexicanum), a type of aquatic salamander, GnRH has been shown to modulate responses evoked by amino acids, which act as food cues for aquatic vertebrates (Park and Eisthen, 2003). Specifically, GnRH reduces the magnitude of the odorant response measured using a technique called an electro-olfactogram (EOG) recording that largely reflects a summed generator potential recorded from many olfactory receptor neurons (Scott and Scott-Johnson, 2002). Interestingly, the magnitude of the EOG response rebounds quickly once the GnRH is washed off, and in some cases becomes significantly larger than the initial baseline response (Park and Eisthen, 2003). Although the biological significance of this effect is not yet clear, perhaps GnRH release suppresses responses to food odorants during courtship or breeding, and later promotes an enhanced response to the same odorants to compensate.

The modulatory effect of GnRH on odorant responses has also been measured at the level of individual olfactory receptor neurons in another aquatic salamander, the mudpuppy (Necturus maculosus). Zhang and Delay (2007) used either a cocktail of volatile odorants with no inherent behavioral significance for mudpuppies or a mixture of 3-isobutyl-1-methylxanthine (IBMX) and forskolin to stimulate the cAMP-gated odorant transduction pathway. Both stimuli generally suppressed outward currents in olfactory receptor neurons, and GnRH reduced, or counteracted, this suppression. The IBMX+forskolin mixture also elicited an inward current, and the effects of GnRH on this current were variable. The mechanisms by which GnRH alters odorant responses are not known, but the available data suggest that GnRH alters odorant signal transduction. One possible category of mechanisms, phosphorylation of one or more elements of the odorant signal transduction pathway, is illustrated in Fig. 3A.

GnRH does not appear to activate ionic currents directly (Zhang and Delay, 2007), but does modulate the voltage-activated currents in olfactory receptor neurons. Specifically, in isolated mudpuppy olfactory receptor neurons, brief (2- to 3-min) application of GnRH suppresses the tetrodotoxin-sensitive, voltage-activated sodium current in the majority of olfactory receptor neurons; in about half the cells that respond to GnRH, the current is only suppressed for about a minute and then is enhanced for a longer period. These modulatory effects appear to be mediated by PKA and/or PKG, which are activated by one or more cyclic nucleotides (Zhang and Delay, 2007). In separate experiments involving mudpuppy olfactory neurons in epithelial slices, application of GnRH enhanced the magnitude of the same sodium current over a longer time course, 5–40 min (Eisthen et al., 2000). Both studies found varying effects on voltageactivated outward currents, but did not identify the currents involved nor systematically investigate the details of the effects (Eisthen et al., 2000; Zhang and Delay, 2007). Given that the sodium current affected by GnRH likely underlies the action potential in olfactory receptor neurons, these data suggest that GnRH might initially make the cells less excitable, but then in a subset of cells or in the intact epithelium it might make the cell more excitable or more likely to respond to weak stimuli. The effects on voltage-activated and odorant-stimulated outward currents are more complex, and current-clamp recordings during odorant stimulation should be used to measure directly the effects of GnRH on cell excitability and odorant sensitivity.

Both studies that examined effects of GnRH at the single-cell level in mudpuppies found that about twice as many olfactory receptor neurons responded to GnRH during the breeding season compared with the non-breeding season (Eisthen et al., 2000; Zhang and Delay, 2007). Park and Eisthen (2003) only used axolotls in breeding condition, but their data suggest that in such animals GnRH may briefly suppress responses to food odorants. Finally, Propper and Moore (1991) showed that GnRH levels in TN fibers increase during courtship in another salamander, female rough-skinned newts (Taricha granulosa). In the central nervous system, GnRH plays a key role in coordinating endocrine activity during vertebrate reproduction; taken together, these results suggest that the GnRH-containing fibers of the TN may also help coordinate sensory responses to olfactory stimuli that are important for reproduction.

The results described above are intriguing, but many questions remain concerning the mechanisms and function of GnRH modulation in the olfactory epithelium. For example, the study by Zhang and Delay (2007) demonstrates that GnRH directly affects at least some olfactory receptor neurons, as the authors recorded from isolated cells. However, we do not know whether GnRH might also influence the other types of cells in the olfactory epithelium, the sustentacular or basal cells, or other nearby tissues such as secretory glands. In addition, it seems possible that GnRH may affect the activity of some cells directly, but may also cause the release of secondary compounds that modulate activity in other cells. Finally, we do not know why GnRH affects the activity of some olfactory receptor neurons but not others. Do the two groups of cells differ in some way? Perhaps GnRH affects the activity of neurons that respond to certain odorants, or classes of odorants, while having no effect on neurons that respond to other odorants. Similarly, it seems possible that GnRH receptors are co-expressed with some types of odorant receptor, or some subcategories of the OR family, but not others. These questions should be resolved by studies using imaging techniques or single-cell RT-PCR.

Other modulatory Influences on the olfactory epithelium

In addition to GnRH, the TN contains other potentially modulatory compounds; for example, the TN in most vertebrates is FMRFamide-immunoreactive. Although FMRF-amide is a molluscan peptide that is not present in vertebrate brains (Chartrel et al., 2006; Tsutsui and Ukena, 2006), structurally similar molecules (RFamides) are likely to be present. In goldfish, chromatography and radioimmunoassay data suggest that two RFamides are present in the TN (Fischer et al., 1996). One of these peptides may be an LPXRFamide, a member of a family of peptides that terminate in the sequence LPXRF-NH₂ (where X=L or Q; see Tsutsui and Ukena, 2006, for review). In goldfish, a novel precursor gene encodes three different LPXRFamides, although only one peptide (gfLPXRFa-3) may be produced in the brain and TN (Sawada et al., 2002). The other RFamide may be Carassius RFamide (C-RFa, Fujimoto et al., 1998). Although its presence in the goldfish TN has not been demonstrated unequivocally (Satake et al., 1999), the olfactory tract of Atlantic salmon (Salmo salar) contains C-RFa immunoreactive fibers, suggesting that the peptide could be produced by TN cells (Montefusco-Siegmund et al., 2006). In addition, some anti-FMRFamide antisera cross-react with NPY, which has a similar C terminal (Chiba et al., 1996; Chiba, 2000), and the TN in some species almost certainly contains NPY (Chiba, 2005; Mousley et al., 2006). Finally, the TN has been reported to contain acetylcholine (ACh), a compound often present in modulatory systems: histochemical and pharmacological data suggest its presence in the TN of tiger salamanders and bonnethead sharks (Sphyrna tiburo), as well as that of at least young rats (Rattus norvegicus) and chicks (Gallus domesticus) (Schwanzel-Fukuda et al., 1986; Wirsig and Leonard, 1986a; Wirsig-Wiechmann, 1990; White and Meredith, 1995). Some reports also suggest that glutamate may serve as a co-transmitter in GnRH-containing neurons (Dumalska et al., 2008).

The relative distribution of TN compounds appears to vary considerably among species. For example, ACh histochemistry and GnRH immunoreactivity distinguish separate populations of fibers in tiger salamanders (Wirsig and Leonard, 1986b), but acetylcholine appears to be present in both RFamide- and GnRH-immunoreactive fibers in bonnetheads (White and Meredith, 1995). Immunoreactivity data from tiger salamanders, bonnetheads, and cloudy dogfish (Scyliorhinus torazame) indicate that GnRH and RFamides or NPY are found in separate populations of cells (White and Meredith, 1995; Chiba, 2000; Wirsig-Wiechmann et al., 2002), suggesting that release of these peptides could be regulated independently. In contrast, these peptides appear to co-localize in the TN in teleosts and perhaps in all actinopterygians (Batten et al., 1990; Chiba et al., 1996; Chiba, 1997; Wirsig-Wiechmann and Oka, 2002; Chiba, 2005), suggesting that they are co-released. If so, the effects of GnRH may be modified by simultaneous release of ACh, NPY, or RFamides.

The effects of some of these TN-derived compounds have been studied in the olfactory epithelium. In contrast to GnRH, which suppresses EOG responses evoked by amino acids, NPY enhances the magnitude of responses to the same odorants in axolotls (Mousley et al., 2006). Like GnRH, though, NPY enhances the magnitude, but does not alter the kinetics, of the tetrodotoxin-sensitive sodium current in olfactory receptor neurons in slices of olfactory epithelium from adult axolotls (Mousley et al., 2006). As with GnRH, these effects depend on the physiological state of the animal: the EOG response is enhanced only in fooddeprived animals, and the effect on the sodium current occurs in more than three times as many cells in fooddeprived animals compared with well-fed animals (Mousley et al., 2006). These results are interesting given that NPY plays an important role in regulating hunger and appetite (see Michel, 2004, for review). Unfortunately, we do not yet know whether these physiological variables interact; for example, perhaps GnRH suppresses responses to food odorants in well-fed animals during the breeding season, but enhances the same responses during the non-breeding season to support later reproduction.

Park and colleagues (2003) examined the effects of FMRFamide on the olfactory epithelium in axolotls, and found that it enhanced the magnitude of the sodium current in olfactory receptor neurons but had no effect on EOG responses elicited by amino acids. A more recent study (Ni et al., 2008) using isolated olfactory receptor neurons from mice demonstrated that FMRFamide enhanced the magnitude of the delayed rectifier potassium current (I_K), but had no effect on the kinetics of the current nor on the magnitude or kinetics of the fast transient potassium current (I_A). The authors conclude that FMRFamide could cause olfactory receptor neurons to repolarize more quickly following an action potential, facilitating odorant responses (Ni et al., 2008). The results of both studies are difficult to interpret, however, in part because FMRFamide is not an endogenous compound in vertebrates, and in part because the authors of these studies did not manipulate the physiological state of the subjects. RFamides may serve a variety of functions (e.g., Chartrel et al., 2006) but seem to play a role in feeding behavior (Dockray, 2004); perhaps FMRFamide would have different effects in food-deprived mice or axolotls compared with the well-fed animals that were used in these experiments. Finally, ACh has been reported to evoke direct excitatory responses from olfactory receptor neurons in frogs (Rana ridibunda, Bouvet et al., 1984; Bouvet et al., 1988), but it is not clear whether it also exerts modulatory effects.

Activity in the olfactory epithelium is subject to modulation by compounds from other sources in addition to the TN. in rodents, leptin appears to be synthesized in nasal glands and released into the mucus overlying the olfactory epithelium, where both the receptor neurons and the supporting cells express leptin receptors (Getchell et al., 2006; Baly et al., 2007). In rats, both cell types also express orexins and orexin receptors (Caillol et al., 2003). While these data suggest that orexins and leptin could modulate odorant responses in the olfactory epithelium through paracrine mechanisms, direct evidence of this phenomenon has not yet been published. Finally, a recent paper (Czesnik et al., 2007) demonstrates that endocannabinoids modulate odorant responses in African clawed frogs (Xenopus laevis), indicating another mechanism by which hunger could modulate activity in the olfactory epithelium. A physiological source of endocannabinoid release into the olfactory epithelium has not yet been identified. Similarly, both adrenaline and serotonin have been shown to modulate activity in the olfactory epithelium, although a plausible local source for these compounds has not been identified (Arechiga and Alcocer, 1969; Kawai et al., 1999; Wetzel et al., 2001).

In addition to the TN, the trigeminal nerve also modulates the activity of olfactory receptor neurons. Dopamine decreases odorant sensitivity and excitability of olfactory receptor neurons (Hegg and Lucero, 2004). The olfactory epithelium is innervated by dopamine-containing fibers from the superior cervical ganglion (Kawano and Margolis, 1985), and dopamine levels in the olfactory mucus are elevated following stimulation of the trigeminal nerve (Lucero and Squires, 1998). Thus, dopamine may be released as part of a feedback loop to protect olfactory receptor neurons from the potentially damaging effects of noxious chemicals that stimulate trigeminal fibers (Hegg and Lucero, 2004). The modulation of olfactory receptor cell activity caused by release of substance P and acetylcholine from trigeminal fibers (Bouvet et al., 1987a; Bouvet et al., 1987b, 1988) may serve a similar function. Damaged olfactory receptor neurons release ATP, which reduces odorant sensitivity in adjacent cells, again serving a neuroprotective function (Hegg et al., 2003).

Taken together, the available data clearly indicate that neural activity in the olfactory epithelium is subject to strong modulation, both to protect the neurons from potentially damaging chemicals and to alter responding with respect to the animal's reproductive and nutritional state. We do not yet know how the sources of modulatory influence in the olfactory epithelium interact with each other.

Organization of the olfactory bulb

The olfactory bulb is the primary olfactory center, as it is the first relay in the central nervous system to receive direct projections from the olfactory epithelium. It is organized similarly across all vertebrates. In general, the olfactory bulb is composed of morphologically and functionally distinct neuronal types arranged in distinct layers, and these neurons contribute to the processing of olfactory information that is conveyed from the olfactory receptor cells (Fig. 3B) (Shepherd, 2004). First, the olfactory nerve, which consists of bundles of axons of the olfactory receptor neurons, enters the olfactory bulb. These bundles then defasciculate profusely in the superficial part of the olfactory bulb to form synapses in the olfactory “glomeruli”, round bundles of neuropil in which the axons of the olfactory receptor neurons interact with dendrites of bulbar neurons. In the glomeruli, axons of olfactory receptor neurons expressing the same odorant receptor genes converge onto one or a few glomeruli in mice and zebrafish (Mombaerts et al., 1996; Sato et al., 2007). Each odorant activates distinct subsets of glomeruli, and thus the olfactory information may be coded in part using patterns of spatial activity (Shepherd, 1994; Friedrich and Korsching, 1997).

In a glomerulus, the axon terminals of olfactory receptor neurons are in contact with the dendrites of mitral cells, the axons of which project out of the olfactory bulb into other olfactory regions in the central nervous system. The mitral cells also form reciprocal dendro-dendritic synapses with granule cells, with cell bodies in the deepest layer of the olfactory bulb and dendrites that extend radially into the mitral cell layer (Rall and Shepherd, 1968). In these reciprocal synapses, the mitral cells form excitatory glutamatergic synapses with the dendrites of granule cells, and the granule cells form GABAergic synapses with the dendrites of the mitral cell. These microcircuits may enhance the tuning specificity of odor responses by lateral inhibition (Rall and Shepherd, 1968; Shepherd and Brayton, 1979; Yokoi et al., 1995).

After odorant information is transmitted from olfactory epithelium to the olfactory bulb, the information is processed through these olfactory bulbar neural circuits, evoking oscillatory activity in the olfactory bulb (Adrian, 1950; Hasegawa et al., 1994; Kashiwadani et al., 1999; Friedrich et al., 2004). Specifically, the mitral cells undergo synchronized subthreshold oscillations in the membrane potential, which are mainly driven by inhibitory inputs of granule cells (Schoppa, 2006). The mitral cells tend to produce action potentials near the peaks of these oscillations in membrane potential, contributing to synchronization of mitral cell action potentials (Friedrich et al., 2004). Although the functional significance of the oscillatory activity in the olfactory system is not thoroughly understood, studies in insects indicate that it is important for odorant discrimination (Stopfer et al., 1997; Laurent et al., 2001).

Through processing in the olfactory bulb, the representation of each odorant by firing in each mitral cell changes continuously. The temporal firing patterning of each mitral cell progressively reduces the similarity between related odors, making the representation of each odorant more specific (Friedrich and Laurent, 2001; Friedrich et al., 2004). Thus, although olfactory information is partially coded by patterns of spatial activity in glomeruli, the temporal patterning of the mitral cell activity is also important in olfactory information processing.

Localization of GnRH fibers and receptors in the olfactory bulb

GnRH-immunoreactive cell bodies and fibers located near or within the olfactory bulb have been described in various vertebrate species, including teleosts (Oka and Ichikawa, 1990; Amano et al., 1991; Kim et al., 1995; Yamamoto et al., 1995; Gonzalez-Martinez et al., 2001; Gonzalez-Martinez et al., 2002), amphibians (D'Aniello et al., 1995), birds (Teruyama and Beck, 2000), and mammals (Kim et al., 1999). In dwarf gourami (Cotisa lalia), a dense projection of GnRH-immunoreactive fibers extends not only ipsilaterally but also bilaterally from the TN ganglion into the olfactory bulb (Oka and Matsushima, 1993).

The function of GnRH in the olfactory bulb can be inferred from the pattern of projections of GnRH-immunoreactive fibers. GnRH-immunoreactive fibers project to both the mitral and granule cell layers in teleosts (Oka and Matsushima, 1993; Kim et al., 1995). Thus, GnRH probably acts on multiple types of olfactory bulbar neurons, and may modulate the olfactory information processing in a complex fashion. In goldfish, the olfactory bulb is functionally subdivided into a few clearly defined areas. Amino acid odorants, which indicate the presence of food, strongly activate neurons in the lateral region of the olfactory bulb; in contrast, the neurons of the medial region in the olfactory bulb respond to sex pheromones (Hanson et al., 1998). In spite of this possible functional segregation, projections of GnRH-immunoreactive fibers in the olfactory bulb in goldfish appear to be extensive, and not limited to a certain region (Kim et al., 1995). Thus, GnRH seems to be involved in the modulation on the olfactory bulbar neurons that process different categories of olfactory information. Perhaps GnRH has different effects in different regions, depending on the type of olfactory information being processed.

The mechanisms by which the activity of neural circuits in the olfactory bulb is modulated by GnRH would be clearer if we knew more about the types of olfactory bulbar neurons that express GnRH receptors (Illing et al., 1999). In addition, different subtypes of GnRH receptor may be differentially distributed in the olfactory bulb, as has been reported for various regions of the nervous system in goldfish (Peter et al., 2003). In the retina of a cichlid fish, Astatotilapia (Haplochromis) burtoni, for example, differential distribution of two GnRH receptors has been reported: the type-I receptor is expressed in cells in the amacrine cell layer, while the type-II receptor is expressed in ganglion cells (Grens et al., 2005). Because amacrine cells are involved in processing visual information and ganglion cells relay the processed information to the brain, GnRH would influence each step of this process. Similar information concerning the distribution of GnRH receptors in the olfactory bulb would provide insight concerning the functional role of GnRH in odorant information processing in the olfactory bulb.

Unfortunately, however, few reports describe the distribution of GnRH receptors in the olfactory bulb in detail (Peter et al., 2003; Soga et al., 2005; Albertson et al., 2008). To date, one report indicates that in cichlid fish expression of the GnRHRIII subtype is restricted to the granule cell layer (Soga et al., 2005), and another demonstrates that in mice mitral cells express GnRHRI receptors (Albertson et al., 2008). Clearly, more information concerning the regions and types of neurons that express GnRH receptors, as well as the subtypes of receptors expressed, is needed.

Other modulatory Influences within the olfactory bulb

The neuromodulatory effects of some classical neurotransmitters and neuromodulators other than GnRH have been demonstrated in olfactory bulb. These data may provide clues as to the possible neuromodulatory effects of GnRH, and therefore will be summarized here; Fig. 3B illustrates the potential mechanisms discussed below.

First, it is possible that GnRH alters the membrane excitability of mitral cells, thereby modulating their firing activities. In the rat olfactory bulb, for example, orexin A causes some mitral cells to depolarize and others to hyperpolarize (Apelbaum et al., 2005; Hardy et al., 2005). Thus, if GnRH modulates the olfactory bulbar neural circuit via similar pathway, it may be capable of enhancing sensitivity for certain odorants that are critical for the animal to detect at that time.

Second, it is possible that GnRH modulates the activity of reciprocal synapses between mitral and granule cells. In fact, neuropeptide Y (NPY) alters the efficiency of synapse transmission in the rat olfactory bulb by reducing the amplitude of calcium currents in mitral cells, thereby reducing the probability of glutamate release. Transmission at the excitatory synapses from mitral cells to granule cells is thus inhibited (Blakemore et al., 2006), which could result in a decrease in odorant contrast in the olfactory bulb. GnRH has been shown to affect synaptic transmission in other regions of the nervous system: for example, in the optic tectum of rainbow trout, GnRH modulates the efficiency of synaptic transmission between retinal fibers and periventricular neurons (Kinoshita et al., 2007). In this study, the excitatory postsynaptic currents that are activated by ionotropic glutamate receptors were enhanced by application of GnRH. If GnRH has similar effects in the olfactory bulb, it could affect odorant discrimination or other aspects of olfactory information processing.

We should also consider the possibility that GnRH may act on olfactory bulbar circuits via other neuromodulatory systems. For example, the action of GnRH on retinal neural circuits is mediated by dopaminergic neurons. In teleost retinas, TN-GnRH neurons make synapses on dopaminergic interplexiform cells (Zucker and Dowling, 1987). Application of GnRH causes horizontal cells to depolarize and enhances their responses to small spots while also diminishing responses to full-field light (Umino and Dowling, 1991). The effects of GnRH on horizontal cells are similar to those of dopamine, and are blocked by the application of haloperidol, a dopamine antagonist (Umino and Dowling, 1991). These results indicate that GnRH acts on horizontal cells indirectly, by stimulating the release of dopamine from interplexiform cells. The olfactory bulb also contains dopaminergic interneurons, and the activity of the olfactory bulbar circuits are affected by the application of dopamine (Duchamp-Viret et al., 1997; Hsia et al., 1999; Wachowiak and Cohen, 1999; Ennis et al., 2001; Davison et al., 2004). In rats and box turtles (Terrapene Carolina), for example, dopamine mediates presynaptic inhibition of olfactory receptor neurons by reducing calcium influx via D2 receptors (Hsia et al., 1999; Wachowiak and Cohen, 1999; Ennis et al., 2001). In northern leopard frogs (Lithobates [Rana] pipiens), dopamine similarly affects the firing activity of mitral cells and inhibits synapse transmission from mitral cells to granule cells (Duchamp-Viret et al., 1997; Davison et al., 2004). Thus, dopamine is broadly involved in the processing of odorant information. If GnRH acts on the dopaminergic interneurons in the olfactory bulb, it could indirectly affect the responsiveness of olfactory bulbar neural circuits as it does in the retina.

Role of GnRH in the olfactory bulb

We have discussed the presence of GnRH-immunoreactive fibers and receptors in the olfactory bulb, as well as the possible modulatory effects of GnRH in the olfactory bulb. What is the functional role of GnRH in the olfactory bulb? If GnRH is released in the olfactory bulb in a physiologically adaptive manner, this question is closely associated with the question of when the concentration of GnRH in the olfactory bulb changes.

Electrophysiological recordings from single TN-GnRH neurons in dwarf gouramis reveal that the majority of these neurons show regular pacemaker activities (Oka and Matsushima, 1993; Abe and Oka, 2007). Furthermore, GnRH release is evoked by a high-potassium depolarizing stimulus that increases the firing frequency of TN-GnRH neurons (Ishizaki et al., 2004). The firing frequencies of TN-GnRH neurons should determine the concentration of GnRH in the regions they innervate, and the activity of neurons that express GnRH receptors should be regulated by this concentration (Abe and Oka, 2007).

The multimodal sensory inputs to the terminal nerve should affect the release of GnRH into the olfactory bulbar neural circuits. Interestingly, some reports indicate that GnRH concentration in the olfactory bulb does change depending on sensory input. For example, one hour after female prairie voles are exposed to male urine, the GnRH concentration in the posterior olfactory bulb is significantly elevated (Dluzen et al., 1981). Furthermore, the GnRH concentration in the olfactory bulb is also elevated when male mice are exposed not only to the ovariectomized females but also to other males (Dluzen and Ramirez, 1983). Similarly, when a male goldfish is exposed to a female that displays spawning behavior, GnRH concentrations in the olfactory bulb are significantly elevated 1–2 hrs after the exposure (Yu and Peter, 1990). In addition, this increase in GnRH concentration is abolished when the medial olfactory tract, which carries information on female sex pheromones (Stacey and Kyle, 1983), is cut in males (Yu and Peter, 1990). In contrast, when males are exposed to other males, no increase in GnRH concentration occurs. The alteration of GnRH concentrations in the male olfactory bulb could be induced either by pheromonal activation of the GnRH neurons or behavioral interactions with a female fish. Overall, these studies suggest that conspecific chemical signals could affect the activity of neurons in the olfactory bulb via GnRH release.

Furthermore, GnRH concentrations in the olfactory bulb may be related to the physiological condition of the animal. In female Indian major carp, GnRH immunoreactivity in the olfactory bulb changes seasonally, with the density of GnRH-immunoreactive fibers peaking during the prespawning season (Biju et al., 2003). In chum salmon (Oncorhynchus keta), GnRH gene expression in the olfactory bulb is elevated when prespawning salmon migrate upstream (Onuma et al., 2005). These lines of evidence strongly suggest that the function of GnRH in the olfactory bulb may be closely associated with the reproductive status of the fish.

How does the effect of GnRH vary seasonally?

The concentration of GnRH in the olfactory system changes in accordance with reproductive state in some species. Perhaps seasonal changes in hormone levels or sensory inputs act on TN-GnRH cells so that the amount of GnRH peptide in the olfactory system is increased, regulating olfactory responsiveness in concert with the physiological condition of the animal. However, it is possible that over the course of the breeding season, the expression of GnRH receptors changes in addition to or instead of changes in the amount of GnRH itself. As described above, the modulatory effect of GnRH on the voltage-activated currents in olfactory receptor neurons changes seasonally (Eisthen et al., 2000; Zhang and Delay, 2007). Perhaps this result is due not to changes in the availability of the endogenous ligand, but to changes in the expression or function of GnRH receptors. In European sea bass and masu salmon, it has been reported that the expression level of GnRH receptors in the brain increases or decreases with the transition of seasons (Gonzalez-Martinez et al., 2004; Jodo et al., 2005). Additionally, alternative splice variants of GnRH receptors may be also involved in the control of the sensitivity of neurons to GnRH: in the bullfrog brain, four types of splice variants are generated from the primary transcript of bf GnRHR-3, a subtype of GnRH receptor, and these splice variants inhibit wild-type GnRH receptor-mediated signaling (Wang et al., 2001). Furthermore, the expression levels of the splice variants vary seasonally in many regions of the brain (Wang et al., 2001). Thus, it is possible to control the consequences of GnRH receptor binding by regulating the splicing process of the GnRH receptor primary transcript seasonally.

What is the function of GnRH in the olfactory system?

Because the concentration of GnRH in the olfactory system is related to reproductive state, it seems logical to surmise that the function of GnRH in the olfactory system is to modulate the processing of olfactory information that is important for reproduction. The simplest hypothesis is that GnRH modulates activity in the olfactory system such that the animal can detect chemical cues from prospective partners efficiently, promoting sexual behavior. Nevertheless, it is also possible that GnRH is involved in other interactions with conspecifics (Dluzen et al., 1981; Dluzen and Ramirez, 1983). For example, the modulatory effect of GnRH on the olfactory system may enhance detection of chemosignals from any conspecific animal, facilitating interactions that could be important for reproduction.

Interestingly, in axolotls GnRH suppresses odorant responses elicited by feeding cues (Park and Eisthen, 2003). Perhaps in this context GnRH acts to inhibit feeding behavior so that the salamanders can instead apply their energies to sexual behavior; at the same time, GnRH may also enhance processing of sex pheromones. Such a scenario would explain the observation that GnRH fibers are distributed widely throughout the olfactory bulb, not simply in areas involved in processing odorant cues directly related to sexual behavior (Kim et al., 1995).

We should also consider the possibility that GnRH simply enhances odorant contrast for many categories of odorants in the olfactory system. The TN-GnRH system has been suggested to affect the motivational or arousal state of the animal (Oka and Matsushima, 1993; Yamamoto et al., 1997; Abe and Oka, 2007). If GnRH enhances odorant contrast in general, a motivated animal can respond more appropriately to the external environment. In this case, it is possible that the close relationship between GnRH concentration in the olfactory system and reproductive state may simply reflect motivational state. For example, GnRH gene expression in the olfactory bulb of chum salmon is elevated when they are highly motivated to migrate upriver (Onuma et al., 2005). During this time, it is important for salmon to discriminate the chemical characteristics of rivers via their olfactory system so that they can home precisely to the natal habitat (Dittman and Quinn, 1996). If GnRH enhances odorant contrast in the olfactory system, it would facilitate homing to the natal habitat.

Overall, the TN-GnRH system receives sensory inputs from the somatosensory, visual, and olfactory systems, and such modalities are thought to influence the activity of the neurons involved. Furthermore, some hormones (thyroid hormone, testosterone) are involved in the regulation of GnRH gene expression in the TN-GnRH system in tilapia (Oreochromis niloticus) (Soga et al., 1998; Parhar et al., 2000). Therefore, the release of GnRH from the TN-GnRH system into the olfactory system would be influenced by changes in environmental factors or physiological status, and thus the olfactory responsiveness of the animal would be regulated.