Pheromones are chemosensory cues secreted by an individual, which are subsequently detected by a conspecific and elicit either behavioral or physiological responses to their mutual benefit (Karlson and Luscher, 1959; Meredith, 1983). In rodents, pheromones modulate the expression of a wide variety of social behaviors including mating and maternal behavior as well as intraspecies aggression (reviewed by Estes 1972; Keverne, 1983; Meredith, 1983). In the ferret, a carnivore, pheromones have been implicated in defining territories (Moors and Lavers, 1981; Clapperton et al., 1988) and in identifying perspective mates (Chang et al., 2000; Kindon et al., 1996). Although in some mammalian species (e.g., pig: Dorries et al., 1997; ewes: Cohen-Tannoudji et al., 1989) pheromones appear to be detected and processed via the main olfactory epithelium (MOE) and associated main olfactory bulb (MOB). The vomeronasal organ (VNO) and its associated accessory olfactory bulb (AOB) have more often been linked to pheromone detection and processing in rodent species.
The VNO is a paired organ located bilaterally at the base of the nasal septum where it is partially encased in a bony or cartilaginous capsule. Nonvolatile chemosignals gain access to the VNO either through the nasopalatine canal, which connects to the oral cavity through the incisive foramen (as in most ungulates and carnivores), or directly from the nasal cavity (as in rodents) (Wysocki, 1979). Many pheromones are probably detected by 7-trans-membrane domain receptors expressed in VNO receptor neurons, which differ from those in the main olfactory epithelium (Herrada and Dulac, 1997). The VNO sensory neurons send axonal projections to glomeruli of the AOB, which is typically located dorsally in the most caudal portion of the olfactory bulb and among eutherian mammals is best developed in rodents and lagomorphs (Estes, 1972). In the AOB, olfactory information is passed to mitral cells that project into a multisynaptic pathway comprising the medial amygdala, bed nucleus of the stria terminalis, medial preoptic area, and the lateral aspect of the ventral medial nucleus of the hypothalamus (Kevetter and Winans, 1981).
The morphology of the VNO was recently described in male ferrets (Weiler et al., 1999). These workers found no difference in VNO sensory epithelium volume in gonadally intact male ferrets killed during the reproductive (spring) and non-reproductive (fall) seasons. Furthermore, in preliminary studies (Chang et al., 2001; Wersinger and Baum, 1997b; Kelliher et al., 1998) we reported the presence of an AOB in a different location from that previously described for ferrets (Dennis and Kerr, 1969; Lockard, 1985). In our experiments we showed that either mating (Wersinger and Baum, 1997b), exposure to soiled bedding from male or female ferrets (Kelliher et al., 1998), or exposure to maternal odors (Chang et al., 2001) induced Fos-immunoreactivity (Fos-IR; an index of neuronal activation) in granule cells of the main olfactory bulb (MOB) but not in the AOB. These results suggest that the MOE-MOB system plays a central role in the detection and processing of reproductively significant pheromones in ferrets. In the rat, the VNO-AOB system is sexually differentiated and responsive to adult sex hormone manipulations (reviewed in Guillamon and Segovia, 1997), suggesting that it may have different functions in males and females. Although we cannot rule out a possible role of the VNO-AOB system in pheromonal communication in the ferret (a carnivore), we thought it would be useful to compare the morphology of this system in males and females, determine its responsiveness to hormone manipulation in adults, and contrast these features with those of other mammalian species (e.g., rat, mouse, hamster) in which the VNO-AOB system is known to mediate many pheromonal responses. Therefore, we compared the dimensions of the VNO sensory epithelium and AOB in adult gonadectomized ferrets of both sexes that had received either testosterone propionate (TP) or no steroid (oil vehicle). We studied the possible activational effects of TP treatment on the VNO-AOB morphology because our previous study (Kelliher et al., 1998) showed that this steroid dramatically enhanced neuronal Fos responses to estrous odors in several forebrain regions of both male and female ferrets. Since our recent characterization (e.g., Wersinger and Baum, 1997a,b; Kelliher et al., 1998) of the location of the ferret's AOB does not correspond with that of earlier studies (Dennis and Kerr, 1969; Lockard, 1985), we used several histochemical markers to confirm the location and size of this structure in the adult ferret. These included staining for soybean agglutinin (SBA), which binds specifically to glycoconjugates located on the vomeronasal nerve (Key and Giorgi, 1986). We also immunostained for luteinizing hormone-releasing hormone (LHRH), which has been found in the AOB but not the MOB of rodents (Zheng et al., 1988), and for tyrosine hydroxylase (TH), which is usually concentrated in the glomerular zone of the MOB but not of the AOB (Baker, 1986). Finally, we compared AOB and MOB volumes in Nissl-stained sections from adult ferrets of both sexes.
MATERIALS AND METHODS
Eight European (Fitch) ferrets of each sex were purchased from Marshall Farms, North Rose, NY, at 14 weeks of age and were housed individually under a long-day photoperiod (16L:8D). Ferrets were fed moistened ferret chow (Purina) once daily, and water was available ad libitum. Eight male and 8 female animals were gonadectomized under ketamine (95 mg/kg) and xylazine (12 mg/ kg) anesthesia and subsequently were injected s.c. daily with TP (5 mg/kg) in sesame oil. When administered to castrated male ferrets, this dose of TP elicits the full range of masculine sexual behaviors and sustains plasma levels of testosterone that are characteristic of gonadally intact male ferrets in breeding condition (Sisk and Desjardins, 1986; Lambert and Baum, 1991). In conjunction with another experiment (Kelliher et al., 1998), all ferrets were treated for 4 weeks with TP whereupon 4 males and 4 females continued to receive daily TP injections for five additional weeks while the other 8 ferrets received only oil vehicle. A single additional male and female ferret, both gonadally intact and in breeding condition, were purchased to provide optimal material for VNO histological photographs (see Histology section). For LHRH and TH immunocytochemistry, some additional archival material was also used from gonadally intact breeding male and estrous female ferrets whose olfactory bulb tissues were stored at −20°C in a sucrose, ethylene glycol, and PVP-40 antifreeze solution. These experiments were conducted in accordance with Guiding Principles for the Care and Use of Research Animals and were approved by Boston University IACUC (protocol 97-017).
Animals were killed with an overdose (100 mg/kg) of sodium pentobarbital and transcardially perfused with 0.1 M phosphate buffered saline (PBS; pH = 7.4) followed by 4% paraformaldehyde. The olfactory bulbs were immediately removed and postfixed in 4% paraformaldehyde for 2 hr. After postfixation, the bulbs were cryoprotected in 30% sucrose/PBS solution until they sank. Olfactory bulbs were then frozen on dry ice and sectioned either coronally, sagitally, or horizontally at 30 μm on a Reichert-Jung SM2000R table-top sliding microtome. Sections were either processed immediately for cresyl violet staining or were stored in antifreeze (500 ml 0.1 M Tris-buffered saline [TBS], 300 ml ethylene glycol, 300 g sucrose, 10 g polyvinylpyrrolidone [Mw 40,000], and distilled H2O to a volume of 1.0 L) at −20°C until processing for immunocytochemistry.
VNO Morphology: Effects of Sex and Steroid Treatment
After brains were removed, the heads from adult ferrets were placed in 4% paraformaldehyde and shipped to Florida State University where they were stored at 4°C until VNO histology and measurements were performed. Four percent paraformaldehyde was used because it was the optimal fixative for immunocytochemistry of the olfactory bulbs. To obtain VNOs that were optimally fixed for the illustration of fine histological detail, one additional male and female (both gonadally intact and in breeding condition) were perfused with Bouin's solution. The photomicrographs shown in Figure 1 are from these ferrets. Heads were first decalcified with RDO® (Apex Engineering Products, Plainer, IL). Snouts were imbedded in paraffin and cut coronally at 15 μm using a microtome. Every tenth section was saved and mounted onto gelatin-coated slides. Sections were then stained using cresyl-violet and luxol-fast-blue. Several measurements of the VNO were made using an image analysis system (Image-pro®, Media Cybernetics, Silver Springs MD). Length of the VNO was calculated by multiplying the number of VNO sections by the distance between sections. Surface area of the VNO sensory epithelium was calculated by summing the lengths of the sensory epithelium surface facing the lumen in each coronal section and multiplying by the distance between these sections. The VNO neuroepithelium volumes were calculated by outlining the area of the sensory epithelium visible in each individual section (i.e., for the full thickness of the epithelium) and then multiplying by the distance between that section and the next measured section. These inter-section volumes were then summed to provide the volume of the entire sensory epithelium. Data for surface area and volume of the sensory epithelium were analyzed using a two-way analysis of variance (ANOVA).
AOB Identification Using Histochemical Criteria
All histochemistry was performed on tissue from a minimum of eight different adult animals including at least two subjects from each treatment group (males and females treated with either TP or oil). Olfactory bulbs were sectioned in either the coronal or horizontal plane. For soybean agglutinin (SBA) histochemistry, sections were first incubated in 3% normal goat serum/1% H202/PBS for 120 min at room temperature (RT). Between each step, the sections were washed four times for 10 min in 0.1 M PBS. The sections were then incubated in biotinylated SBA 15 μg/ml (Sigma) for 40 min at RT. Sections were then incubated for 120 min with ABC (1:100; Vector Labs). After washing, sections were reacted with nickel-DAB (prepared according to the manufacturer's recommendation; Vector Labs) for 7 min. Sections were then mounted onto gelatin-coated slides and coverslipped using Permount®.
Both coronal and horizontal olfactory bulb sections were used for LHRH-ICC. Sections were first preincubated in 3% normal goat serum/1% H202/PBS for 120 min at 4°C. Sections were then incubated with LHRH primary antiserum, LR-1 (1:100,000 in 0.4 % Triton X-100/0.1 sodium azide/PBS; LR-1, gift of Dr. Robert Benoit) for 16 hr at RT. Following every incubation, sections were washed four times in 0.1 M PBS, pH 7.4, for 10 min. Sections were incubated with biotinylated goat anti-rabbit IgG (1:200; Vector Labs) for 2 hr followed by ABC (1:100; Vector Labs) for 2 hr. The sections were reacted for 7 min with nickel DAB and were mounted onto gelatin-coated slides and coverslipped using Permount®.
TH-ICC was performed on coronal and horizontal sections using the same protocol as for LHRH ICC. The primary antiserum used was a monoclonal mouse anti-tyrosine hydroxylase (1:1,000; DiaSorin, Stillwater, MN), and the secondary antibody used was biotinylated goat anti-mouse IgG (1:200; Vector Labs). Omission of primary antibody for TH or LHRH eliminated immunoreactive staining. LR-1 and TH antibody specificity has been demonstrated previously in the ferret (Lambert et al., 1992; Wersinger and Baum, 1997a).
Olfactory Bulb Morphology: Effects of Sex and Steroid Treatment
Histological analysis was performed on the MOB and AOB of adults. Coronal sections (30 m) were cut through the olfactory bulb of 1 hemisphere from each ferret. Every third section was saved and mounted onto gelatin-coated slides. Sections were then stained for Nissl substance using cresyl violet. An image analysis system (Image-pro®, Media Cybernetics, Silver Springs, MD) was used to digitize and measure the areas of the cell layer in every AOB section and the granule cell layer in every tenth MOB section. There is no apparent lamination into individual mitral and granule cell layers in the ferret AOB. Therefore, the term “cell layer” was used to refer to a layer of mitral + granule cells adjacent to the AOB glomeruli. The areas of the AOB and MOB in each section were plotted as a function of the distance from the rostral tip of the olfactory bulb. Sigma Plot software (SPSS, Chicago IL) was then used to calculate volumes with a transformation algorithm for determining the area beneath a curve. Volumes of the entire AOB (glomerular+cell layer) and of the entire olfactory bulb were also obtained by this method. The percentage of the whole olfactory bulb devoted to AOB (AOB % of OB) was calculated by dividing the entire AOB volume by the volume of the entire olfactory bulb and multiplying by 100. Data were analyzed using a two-way ANOVA.
VNO Morphology: Effects of Sex and Steroid Treatment
Coronal sections through the snouts of male and female ferrets revealed the presence of a paired VNO that is located at the base of the nasal septum above the hard palate (Fig. 1A). The general morphology of the VNO in both male and female ferrets corresponds with what was described in the male ferret (Weiler et al., 1999). Briefly, the rostral opening to the lumen of the organ (the VNO duct) is located at the nasal end of the nasopalatine canal. The VNO-epithelium is surrounded by large blood vessels, some glandular tissue, and unmyelinated vomeronasal nerve bundles (Fig. 1B). The VNO capsule is 3.5–4.0 mm in length; however, only slightly more than half of the cylindrical organ contains a sensory epithelium. The pseudostratified columnar sensory epithelium, which is 4–5 cells thick, contains bipolar receptor neurons and nonsensory supporting cells whose nuclei form a single layer adjacent to the lumen (Fig. 1C). The sensory epithelium is located predominantly on the medial side of the VNO, although it also extends onto the lateral side in some individuals (Fig. 1B). We observed no morphological differences in the VNOs of males and females. We also found no sex differences or effects of adult steroid treatment on the volume or surface area of the VNO sensory epithelium (Table 1).
Table 1. Vomeronasal organ (VNO) dimensions in adult gonadectomized male and female ferrets treated with either testosterone propionate (TP) or oil vehicle*
VNO length (mm)
VNO neuroepithelium surface area (mm2)
VNO neuroepithelium volume (mm3)
Data are expressed as mean ± SEM.
3.58 ± 0.31
1.64 ± 0.14
0.08 ± 0.01
3.22 ± 0.08
1.60 ± 0.13
0.07 ± 0.01
3.77 ± 0.16
1.60 ± 0.08
0.08 ± 0.01
3.72 ± 0.25
1.47 ± 0.16
0.08 ± 0.01
AOB Identification Using Histochemical Criteria
In horizontal sections (Fig. 2B), the vomeronasal nerve (VN) was observed running caudally along the medial side of the bulb and arborizing into glomeruli of the AOB. SBA consistently bound heavily to the VN and AOB glomerular layer (AOB GL) (Fig. 2A) of all animals examined. Some light SBA binding was also seen in glomeruli of the MOB (not shown); however, this staining was very sparse, with no consistent pattern being seen in different subjects.
LHRH-IR fibers (Fig. 3A) were seen aggregated within and around the AOB in association with the VN and AOB GL. Fibers were also observed within the cell layer and lateral olfactory tract nearest to the AOB (not shown). A few LHRH-IR fibers were seen scattered superficially in the nerve layer of the MOB (not shown). Numerous TH-IR cell bodies and fibers were seen associated with the glomeruli of the MOB (Fig. 3B). By contrast, TH-IR was completely absent in the glomerular and cell layer of the AOB (Fig. 3B).
AOB Morphology: Effects of Sex and Steroid Treatment
The morphology of the ferret AOB is shown in Nissl-stained sections cut in the coronal (Fig. 4A), sagittal (Fig. 4B), and horizontal (Fig. 4C) planes. Although the AOB could be seen in Nissl-stained sections cut in all three planes, it was most easily recognized in the coronal and horizontal sections. The AOB GL is located on the medial side of the bulb and, unlike the MOB, has only a few visible periglomerular cells. Individual mitral and granule cell layers could not be identified in the AOB. Instead, an AOB cell layer is located just lateral and somewhat caudal to the glomeruli. This cell layer contains a few putative mitral/tufted cells based on their morphology. In addition, there appear to be numerous granule cells; however, they were difficult to distinguish from glia (Fig. 5). No gross differences in AOB morphology or cell layer volume were evident between gonadectomized males and females, regardless of whether they received TP or oil vehicle in adulthood. A significant sex difference was observed in the volume of the MOB granule cell layer (F[2,35] = 9.207, P < 0.001) (Table 2). However, this difference was not reflected in a statistically significant sex difference in the percent of olfactory bulb volume that was devoted to AOB.
Table 2. Dimensions of the main and accessory olfactory bulbs in adult gonadectomized male and female ferrets treated with either testosterone propionate (TP) or oil vehicle
Effects of Sex and Adult Hormone Treatment on the VNO-AOB
The dimensions of the ferret's VNO and AOB were similar in the two sexes, and there was no effect of adult testosterone treatment on either structure. These results confirm and extend those of Weiler et al. (1999) who found no differences in VNO volume between male ferrets killed in the Spring (when circulating testosterone levels are high) or Fall (when testosterone levels are low). In contrast, the volume of peripheral segments of the chemosensory pathway including both the VNO and AOB are reportedly sexually dimorphic in rats. Segovia and Guillamon (1982) reported that the total volume of the VNO was significantly larger in males than in females, with males having a larger VNO neuroepithelial volume and more large bipolar neurons. Either neonatal castration of males or neonatal testosterone treatment of females abolished this sex difference. While Segovia and Guillamon did not control for possible steroid effects during adulthood, the fact that early steroid manipulation abolished the observed sex differences in VNO morphology suggests that it resulted from an organizational action of testosterone in the developing male.
The volume of the AOB is also reportedly larger in male than in female rats (Segovia et al., 1984) due to a greater number of mitral and granule cells in males (Segovia et al., 1986). As with the VNO, male rats castrated on postnatal day 1 (P1) had AOB volumes comparable to those of control female rats whereas females given TP on P1 had AOB volumes comparable to those of gonadally intact males (Segovia et al., 1984, 1986). Again, in ferrets we found no sex difference in the volumes of either the AOB GL or cell layer. By contrast, we did find that the volume of the MOB granule cell layer was significantly larger in male than in female ferrets. This may not be surprising considering that overall brain size is greater in male than in female ferrets (Tobet and Hanna, 1997). Even so, there was no statistically significant difference between males and females in the percentage of the olfactory bulb devoted to AOB. While the lack of a sexually differentiated VNO-AOB in the ferret is not necessarily evidence against the existence of a functional VNO-AOB in this carnivore, it does highlight another difference with rodent species whose VNO-AOB systems are known to respond to pheromones.
Species Comparison of the VNO-AOB
Our results together with those of Weiler et al. (1999) for males show that the ferret's VNO and associated AOB are relatively small compared with their dimensions in other mammals, including rodents, marsupials, and ungulates (McCotter, 1912; Crosby and Humphrey, 1938). The size of the ferret's VNO and AOB is comparable to those of two other carnivores, the dog (Salazar et al., 1992, 1994) and the cat (Salazar et al., 1996). We found that the volume of the ferret's VNO neuroepithelium is approximately 50% of that reported in the rat (Segovia and Guillamon, 1982) in which pheromones are known to act via the VNO-AOB. In some ferrets, the sensory neuroepithelium extended onto the lateral wall of the VNO. This contrasts with many species in which the VNO sensory neuroepithelium is restricted to the medial wall. Interestingly, this extension of the neuroepithelium onto the lateral wall of the VNO has also been observed in some species of bat (Cooper and Bhatnagar, 1976) that, like ferrets, have a relatively underdeveloped VNO-AOB. Perhaps the increased neuroepithelium surface area compensates for the small overall size of the organ.
Our ability to document the existence of an AOB in the ferret may not seem surprising in so far as the ferret possesses a VNO. However, in mammals the presence of a VNO is not always associated with the existence of an obvious AOB. Several species of bat (Frahm and Bhatnagar, 1980) reportedly lack an AOB in adulthood even though a VNO has been identified. A similar situation exists in humans and in some other old-world primate species (see discussion in Wysocki, 1979). Adult humans have been identified as having a VNO-like structure (Johnson et al., 1985; Trotier et al., 2000), although it is not clear whether it contains functional sensory receptor neurons. Adult humans reportedly lack an AOB (Crosby and Humphrey, 1938; Stephan, 1965; Meisami et al., 1998; Meisami and Bahatnagar, 1998). It became clear to us in the course of establishing the existence of the ferret's AOB that a structure that is so small and ill-defined could easily be overlooked or incorrectly identified using only standard histological methods. It is possible that the poorly developed AOB that has been described in the human fetus (Humphrey, 1940; Pearson, 1942; Boehm et al., 1994; Meisami et al., 1998) may persist into adulthood, but be difficult to discern. Further histochemical analysis, using several different markers and looking explicitly for an AOB, might establish definitively whether this structure exists in the adult human.
There are at least two shortcomings in many previous anatomical studies describing the presence or absence of an AOB in various species. First, researchers often relied solely on conventional histological techniques, such as hematoxylin staining, to identify this subregion of the olfactory bulb. Second, there was a false expectation about the size and location of the AOB based on comparisons with the large and well-defined structures present in rodent and marsupial species. Both of these shortcomings are evident in previous studies that specified the location of an AOB in the carnivore subfamily, Mustelidae. Thus, an AOB was reportedly either absent (weasel, polecat, and marten: Jawlowski, 1956) or poorly delineated while occupying the entire dorsal-caudal portion of the olfactory bulb (weasel: Crosby and Humphrey 1939; mink: Jeserich, 1945; ferret: Dennis and Kerr, 1969; Lockard, 1985). The suggestion that the AOB of these Mustelid species encompasses the entire dorsal-caudal portion of the olfactory bulb was likely based on expectations derived from rodent studies and on the exclusive use of hematoxylin-stained coronal sections to assess AOB structure. In the ferret, our three-dimensional analysis combined with histochemical markers revealed a small AOB that is smaller and more medially located in the olfactory bulb than was previously reported. Using only Nissl staining, this small structure would have been extremely difficult to identify for the first time.
Is the Ferret's VNO-AOB Functional?
As explained in the Introduction, we previously obtained no evidence that the VNO and its associated AOB are functional in either adult or neonatal ferrets. Increased neuronal Fos-IR has been used as a marker of neuronal activation to show that AOB mitral and granule cells are activated by pheromonal stimuli in several species, including the hamster (Fernandez-Fewell and Meredith, 1994), rat (Paredes et al., 1998), shrew (Gill et al. 1998), and mouse (Halem et al., 1999). In the adult ferret, neither exposure to soiled bedding from opposite- or same-sex ferrets (Kelliher et al., 1998) nor mating stimulation (Wersinger and Baum, 1997b) induced Fos-IR in the AOB. In contrast, either stimulus induced neuronal Fos-IR in the MOB granule cell layer and in those limbic regions that receive olfactory inputs including the medial amygdala and bed nucleus of the stria terminalis of ferrets of both sexes. In females, the receipt of an intromission (Wersinger and Baum, 1997b) or exposure to male bedding also induced neuronal Fos-IR in the medial preoptic area and the ventrolateral portion of the ventromedial hypothalamic nucleus. While the lack of stimulus-induced increments in neuronal Fos-IR in the adult ferret's AOB tells us little about the potential responsiveness of neurons in this structure to pheromones, these results suggest that odor cues may activate amygdaloid and hypothalamic neurons after their detection by the MOE and processing via the MOB. Future studies will need to assess the possible contribution of the VNO-AOB system to all aspects of chemosensory communication in this carnivore.
This work was supported by National Institutes of Health grants HD 21094 (M.J.B.) and DC 00906 (M.M.). We thank Witten Hall for his assistance in measuring and calculating VNO parameters, Gay Howard for her excellent technical assistance with VNO data management and analysis, and the staff at the Boston University animal care facility for caring for our ferrets.