The role played by the vomeronasal system (VNS) in the mammalian olfactory function is not completely clear (Brennan and Zufall, 2006; Spehr et al., 2006; Zufall and Leinders-Zufall, 2007; Broad and Keverne, 2012). Certainly much evidence shows how information captured by vomeronasal receptors (Vrs), once transformed, transported, and processed, causes changes in the behavior of the rodents of both sexes, including intermale aggression, sexual behaviors, and long term neuroendocrine alterations (Tirindelli et al., 2009; Brennan, 2010; Isogai et al., 2011; Koh and Carlson, 2011; Ma, 2012). It is true, however, that scientific data published over the past few years suggest that the functional independence of the main olfactory system (MOS) and the VNS is questionable, as both are in many ways complementary (Keverne, 2005; Levai et al., 2006; Baum and Kelliher, 2009; Slotnick et al., 2010; Baum, 2012). In contrast, it is worth noting that the vast majority of studies concerning the VNS of mammals have used mouse as the subject of study (Boehm, 2006), which is absolutely logical as it is the mammalian species most commonly used for neurobiological experimentation (Hedrick, 2012) and it is also a macrosmatic animal. Nevertheless, there are some disadvantages in using the mouse model exclusively, as it does not exactly match the pattern of other mammals, including macrosmatic species equipped with a well-developed VNS (Halpern and Martinez Marcos, 2003; Brennan and Zufall, 2006; Salazar et al., 2007; Shi and Zhang, 2007).
In our laboratory, we have studied aspects of the VNS in domestic mammals, including cows, sheep, goats, pigs, horses, dogs, and cats. We have noted several significant differences in the morphology of the VNS in these species compared with mice. It might be time to start considering the possibility of establishing different patterns concerning the morphological constitution and organization of the structures that make up the system, as well as its functional activity (Takami, 2002; Halpern and Martinez Marcos, 2003; Salazar and Sánchez-Quinteiro, 2009).
In the work reported here, we describe some of the most significant morphological characteristics of the vomeronasal organ (VNO) of the dog, labeling using the lectins Ulex europaeus agglutinin (UEA-I) and Lycopersicum esculentum (LEA), and the expression of Gαi2 and Gα0 proteins in both the VNO and the accessory olfactory bulb (AOB). The expression pattern of lectins and Gαi2 protein in the dog's VNO sensory epithelium (SE) is positive, although slightly different to the pattern in mice, whereas in the AOB the differences between dogs and mice in these types of staining are more striking. Gα0 protein staining is negative in dogs.
MATERIAL AND METHODS
A total of fourteen large adult mesaticephalic dogs of both sexes were studied. All animals came exclusively from the necropsy room and from the Department of Clinical Sciences of our School. The heads were intact and did not show clinical or postmortem evidence of neurological disease. All experiments were performed in accordance with the regulations and laws of the European Union (86/609/EEC) and Spain (RD 223/1998) for the care and handling of animals in research.
Dissections and microdissections of the nasal cavity and anterior parts of the cranial cavity were performed in three heads, and the VNOs and AOBs of the remaining heads were examined histologically. We have also employed samples of the nasal cavity (serial transverse sections) and the brain (serial horizontal and parasagittal sections of the olfactory bulbs) from our archival material. In addition, two adult singly housed female BALB/c mice bred in the Animal House of the University of Santiago de Compostela (Registry no. 15003AE) were also used, and their VNO and AOB tissue were studied as a comparison control.
Immediately after death, part of the occipital, temporal, parietal, maxillary, and nasal bones were removed, and then the heads were fixed by immersion in 10% buffered formalin. Olfactory bulbs and VNOs were dissected out and transferred to 4% buffered formalin (the samples of six heads), or transferred to Bouin's fixative, and after 24 hr transferred to alcohol 70% (samples of five heads). VNO: after paraffin embedding transversal sections 8–10 μm were serially cut and stained with hematoxylin-eosin. Olfactory bulb/AOB: after paraffin embedding horizontal sections 10 μm were serially cut and stained with Nissl cresyl-violet and Tolivia. Exactly the same procedure was followed with the heads of mice; one was fixed and stored in buffered formalin and one transferred to Bouin's fixative.
Lectin Histochemistry Protocol
Two lectins—Ulex europaeus agglutinin (UEA-I) and Lycopersicum esculentum agglutinin (LEA)—were obtained as biotin conjugates from Sigma (St. Louis, MO) and detected as follows: (1) incubation for 30 min at room temperature with 2% bovine serum albumin in 0.1 M Tris buffer (pH 7.2); (2) incubation for 24 hr at 4°C with lectin at various dilutions in 0.1 M Tris buffer containing 0.5% bovine serum albumin; (3) washing for 2 × 10 min in PBS; (4) incubation for 90 min at room temperature with Vectastain ABC reagent (1:250 in PBS), and visualization of peroxidase activity by incubation in a solution containing 0.05% 3,3-diaminobenzidine and 0.003% H2O2 in 0.2 M Tris–HCl buffer (pH 7.6). Controls were run without lectin and by preabsorption of lectin by an excess amount of the corresponding sugar. As an additional control, mice sections were stained following identical procedure. Usually, the sections were alternatively stained using the Nissl and Tolivia methods.
The immunohistochemical detection of the following specific markers was done in paraffin-embedded sections: G-protein Gαi2(Santa Cruz Biotechnology, 1:100) and G-protein Gα0 (DML, 1:100). The samples were embedded in paraffin wax, cut on a microtome (6–8 μm thickness), and finally transferred to slides.
The procedure was as follows: after dewaxing in xylene and hydration the slides were transferred to PBS. Then all the sections were incubated for 30 min at room temperature in 5% horse normal serum containing 2% bovine serum albumin in PB. Sections were sequentially incubated in primary antibody for 24 hr at 4°C, biotinylated secondary antibody for 1 hr, avidin-biotin-horseradish peroxidase complex (ABC Vectastain reagent) for 2 hr and visualized with 3,3-diaminobenzidine following standard procedures for the visualization of horseradish peroxidase complex. Finally, the slides were dehydrated through alcohols, cleared in xylene, and coverslipped. As an additional control to the routine immunohistochemical controls mice sections were stained following identical procedure. Usually, the sections were alternatively stained by Nissl and Tolivia methods.
Image Acquisition and Processing
Digital images were captured using a Karl Zeiss Axiocam MRc5 digital camera. All images were processed using Adobe Photoshop 6.0 (Adobe Systems, San Jose, CA). When necessary, images were adjusted for contrast and brightness to equilibrate light levels, and cropped, resized, and rotated for purpose of presentation. No additional digital image manipulation was performed.
The paired VNO is a symmetrical structure located in the floor of the nasal cavity on each side of the vomer bone. It comprises the vomeronasal duct (VNd), which is surrounded by soft tissue integrated by connective tissue and a sizeable number of glands, vessels, and nerves. Those elements are enveloped by an incomplete cartilaginous lamina, which is externally covered by the mucosa of the nasal cavity (Figs. 1 and 2). By dissection and microdissection it is possible to isolate the VNd, which has an approximate length of 80 mm; it is blind-ended at its posterior end, and its anterior extremity opens into the incisive duct. From a morphological point of view the incisive duct establishes a communication between the nasal and oral cavities. Its communication with the oral cavity is clear because it opens in a specific anatomical prominence, the incisive papilla. In contrast, in the nasal cavity the incisive duct is diffuse, and it can be considered as the projection of the mucosa, which covers the edges of the palatine fissure. Moreover, the anterior part of the vomeronasal cartilage is associated to the same fissure, and its prolongation through the incisive papilla becomes an incomplete reinforcement of the incisive duct (Figs. 3 and 4). The area of the SE is limited to the central portion of the VNd, on its medial wall, which means that the Vrs are concentrated in a relatively small area of the VNO, and at the same time are hidden and very well protected. The different elements that organize the strata of the SE are the processes of receptor cells, supporting cells and receptor cells. A basal lamina separates the SE from the underlying tissue (Fig. 5).
Only by microdissection it is possible to identify the AOB as a small ovoid asymmetrical protuberance of variable development, usually located on the posterior third of the medial edge of the MOB. In order to facilitate the description of the stratified structure of the AOB, reference will be made to the conventional arrangement of strata as seen in the main olfactory bulb (MOB). Classical nomenclature includes the following layers: nervous, glomerular, external plexiform, mitral, internal plexiform, and granular. Both the nervous or vomeronasal layer and the glomerular layer are easily distinguishable, whereas the rest of strata are difficult to identify due to the lack of definition and organization of AOB cells. Beneath the glomeruli scattered neurons with large pale-staining nuclei and intensely stained nucleoli, are present. Because of the difficulty in distinguishing mitral from tufted cells, both groups of neurons form a putative stratum defined as mitral/tufted cells. In agreement, the plexiform strata are undistinguishable. Likewise, an AOB granular layer could not be identified with certainty (Fig. 5).
Histochemistry and Immunohistochemistry
The results obtained in this section are shown considering both the organ SE and the AOB together. This allows comparisons between both structures. Likewise, it is possible to compare the data obtained from dogs and mice.
Figure 6 shows the staining pattern of lectins UEA-I and LEA, and proteins Gαi2 and Gα0, in the SE and in the AOB of dogs. Both lectins labeled the mucomicrovillar complex of the SE, and the vomeronasal and glomerular layers of the AOB in a similar way. The main difference could be established inside the SE as LEA labeled the basal epithelium more specifically than other parts. The Gαi2 staining was found in the apical zone of the SE more strongly than other parts, but was not found in the mucomicrovillar complex. This protein was also expressed in the vomeronasal and glomerular layer of the AOB. No sign of immunostaining for Gα0 was seen in either the VNO or AOB.
In mice (Fig. 7) staining for both the lectin UEA-I and the protein Gαi2 was seen in the apical SE. Staining for the protein Gα0 was seen in the basal SE, and for the lectin LEA was seen in both zones. This pattern of labeling is in agreement with labeling seen in the AOB targets of VNO input: thus UEA-I and Gαi2 labeling was seen in the anterior part of the AOB, Gα0 staining was seen in the posterior part of the AOB, and LEA staining was seen in both segments of the AOB, although different intensities of staining were observed.
A species comparison of the morphology of VNS may provide insights into species differences in the function of this system. In order to do this, it seems sensible to begin with the VNS model of mice, which has been widely studied and is usually considered to represent a universal pattern. The three parts that make up the VNS of mice—VNO, AOB, and vomeronasal amygdala—are highly developed, and the connections between them, the vomeronasal nerves (VNns) and the accessory olfactory tract, are clearly distinguishable (Mucignat-Caretta, 2010).
Our observations, which exclude the vomeronasal amygdala, show that there are several morphological differences between the VNS of dogs and mice. Amongst those differences, we highlight the following: in contrast to the dog, in mice the VNd ends directly in the nasal cavity, there is a clearcut definition and development of the sensory epithelial cells, VNns are easily identifiable both in the VNO transversal sections and along the nasal septum, and the whole VNO covering is a virtually complete osseous lamina (Wysocki and Meredith, 1987; Døving and Trotier, 1998; Keverne, 1999; Salazar and Sánchez-Quinteiro, 2003).
In mice the right and left AOBs are symmetrical; they have an invariable topography and a similar size. The AOB strata are more clearly defined in mice than in dog, although the mitral and tufted cells distinction in the AOB is a problem not yet resolved (Mori, 1987; Salazar et al., 2006). The lack of definition in the stratification of the AOB is widened by the morphological diversity of such structure, and different points of view have been sustained by different authors. For example, in the ferret it has been shown that the mitral/tufted and granule cells layers are poorly defined, and for that reason it has been used the term of “AOB cell layer” to include this group of cells, but even so it is difficult to establish the distinction between granule and glia cells (Kelliher et al., 2001).
Every morphological species difference mentioned earlier is easily seen; however, drawing definitive conclusions about the functional consequences of these differences is far more complicated. There are other striking features of the dog's VNS morphology, including the exact place where the anterior part of VNd links with the incisive duct. The topographic connection between both ducts in dogs is more complex that has been lately considered, mainly due to the complicated disposition of the cartilaginous lamina associated to them (Ramser, 1935; Adams and Wiekamp, 1983; Salazar et al., 1984). As was previously mentioned, in mice the VNd opens directly in the nasal cavity. This species difference in the termination of the VNd could be linked to a difference in the ease with which pheromonal stimuli reach the VNO sensory epithelium. In any case, in the hamster it has been demonstrated that exists a pumping mechanism to regulate the stimulus access to the VNO (Meredith and O'Connell, 1979; Meredith et al., 1980), mechanism later corroborated by us in cows following a different strategy (Salazar et al., 2008). The characteristics of the enveloping part of the whole system, osseous or cartilaginous, are supposed to be related in some way to this mechanism, although this link has so far been taken into account only to establish phylogenetic classifications (Wöhrmann-Repenning, 1984).
We have handled jointly the immnunohistochemical results concerning the SE and the AOB so they are commented in this section. In previous studies, we verified that all of the lectins chosen for this work excel at staining the olfactory tissue, and that they are complementary, as LEA stains the main and accessory olfactory tissue, whereas UEA-I stains only the vomeronasal olfactory tissue (Salazar et al., 1992, 2001). Our observations corroborate previous results, showing that there is a perfect morpho-functional synchrony between the sensory transduction site and the first relay stage of the VNS in mice (Halpern et al., 1998; Mori et al., 2000; Knöll et al., 2003), whereas this was not the case in dogs. In that species, the AOB vomeronasal and glomerular strata—those that are stained in mice—are uniform, hence there seems to be no segregation or zones when lectins are used as markers.
In contrast, Gαi2 and Gαo proteins are associated with Vr1 and Vr2, respectively, and their expression should be therefore related to the VNS (Berghard and Buck, 1996; Halpern and Martínez-Marcos, 2003). In some mammals such as mouse (Jia and Halpern, 1996; Wekesa and Anholt, 1999), rat (Jia and Halpern, 1996), guinea pig (Sugai et al., 1997), and opossum (Halpern et al., 1995) it has been demonstrated that the apical SE of the VNO and the anterior part of the AOB were Gαi2 positive, whereas the basal SE of the VNO and the posterior part of the AOB were Gαo positive, as we had corroborated in this work in mice. On the contrary, in other species such as horse (Takigami et al., 2004), goat (Takigami et al., 2000), sheep (Salazar et al., 2007; Salazar and Sánchez-Quinteiro, 2011), and cat (Salazar and Sánchez-Quinteiro, 2011), all of them belonging to the group of domestic mammals, the immunoreactivity of Gαo was always negative, both in the SE and in the AOB. These findings reinforce the notion of the existence of differences in the segregation of the VNS according to species.
At the moment to evaluate our result concerning the expression of G proteins, we could state that the expression of the Gαo is negative in the dog, based on the fact that we had the opportunity to follow exactly and simultaneously the same protocol with vomeronasal and olfactory bulb tissue of dogs and mice. Each time that our methodology has been repeated the results were the same. Fortunately, another solid argument to support our results is that although the genome of the dog has not yet been completely sequenced, to date Vr2 receptor genes have not been identified in the dog (Young and Trask, 2007; Grus and Zhang, 2008). Nevertheless, the group of Morrison found resemblances in the SE immunoreactivity of Gαi2 and Gαo proteins in dog, which were positive in both cases (Dennis et al., 2003). The antigen retrieval procedure employed by these authors could explain this finding. Actually they affirm themselves that the possibility of amplifying a signal from a cross-reaction with a nonspecific epitope remains. From our point of view those results must be taken with caution until the SE immunoreactivity is confirmed by its projection to the AOB, if that were the case.
The results obtained in our study, together with additional published data on the VNS of domestic mammals (Halpern and Martínez-Marcos, 2003; Salazar and Sánchez-Quinteiro, 2009), provide strong evidence for the existence of VNS morphological models in addition to mice. This possibility is strengthened by recent findings obtained in other disciplines, for instance those related to the genome in different animal species. Thus, several groups reported that the olfactory system as a whole (but also the VNO) express significant species differences (Grus et al., 2005; Young et al., 2005; Niimura and Nei, 2007; Nei et al., 2008; Shi and Zhang, 2009; Young et al., 2010), even between similar species (Kurzweil et al., 2009). Although there are some small differences in the number of genes reported by authors, we have selected some data in order to show an example of the olfactory diversity. The functional olfactory receptors, Vr1 and Vr2 genes are 1063/239/121, 1198/93/86, 822/9/0, and 1152/40/0, respectively, for mouse, opossum, dog, and cow.
In the next few years, we will probably have accurate information about genes related to the VNS in bats, an extraordinary example of morphological diversity amongst mammal species (Cooper and Bhatnagar, 1976; Frahm and Bhatnagar, 1980). This task has already been undertaken by Jianzhi Zhang's group (Zhao et al., 2011). We anatomists have the responsibility of giving specific descriptions of each part of the VNS, in as many species as possible, in order to identify potential relationships among morphology, physiology, and genetics (Liao et al., 2010).
The authors thank J. Rodríguez Fernández and J. Castiñeira for the technical assistance. The comments of anonymous reviewers improved the clarity of the manuscript.