Ignacio Salazar, Department of Anatomy and Animal Production, Unit of Anatomy and Embryology, Veterinary Faculty, 27002 Lugo, Spain. E: email@example.com
The general morphology of the vomeronasal vessels in adult cows was studied following a classic protocol, including optical, confocal and ultrastructural approaches. This anatomical work was completed immunohistochemically. The vomeronasal organ in cows is well developed, and its vessels are considerable in size. This fact allowed some functional properties of the vomeronasal arteries to be evaluated and, for the first time, their isometric tension to be recorded.
Our functional studies were in agreement with the immunohistochemistry, and both corroborated the morphological data on the similarity between the vomeronasal vessels and those of the typical erectile tissue. In consequence, the vasoconstriction and vasodilation of the vomeronasal vessels would facilitate an influx and outflow of fluids in the vomeronasal organ, that is to say, this organ in cows would be able to work as a pump mechanism to send chemical signals to the vomeronasal receptor neurones.
In the majority of terrestrial mammals, odours are detected in two ways: by means of the main (MOS) and the accessory (AOS) olfactory systems. The MOS and the AOS – commonly called the vomeronasal system (VNS) – share a common pattern of organization, despite significant morphological and physiological differences, including the size of the structures of which they are formed (Halpern, 1987; Brennan, 2001).
An interesting and important contrast between these systems relates to the initial step on the normal pathway of olfactory information, that is to say the corresponding olfactory receptors, not only due to the obvious difference in their topography, but to the manner in which they are arranged. Whereas the main olfactory receptors (MOrs) are situated freely in the epithelium of the mucosa of the most posterior part of the nasal cavity, the vomeronasal receptors (VNrs) are hidden inside a tube, the vomeronasal duct (VNd), usually in the epithelium of its medial wall (Breer et al. 2006). This means that chemical signals which will be identified by the VNS need extra help or an ‘auxiliary apparatus’ to enable such substances to enter the VNd (Meredith et al. 1980).
Around the VNd there is a considerable amount of glands, vessels, nerves and connective tissue, which organize the soft tissue of the vomeronasal organ (VNO). Additionally, a thick capsule wraps all these structures (Vaccarezza et al. 1981; Keverne, 1999).
The presence of a pumping mechanism, powered by vasomotor movements, which can suck stimulus substances into the VNd, has been demonstrated in hamsters (Meredith & O’Connell, 1979). This physiological action may be similar in other rodents, as the characteristic innervation of the vomeronasal blood vessels suggests (Matsuda et al. 1996; Canto & Suburo, 1998).
With the information that we have today, the presence of such a pump mechanism is likely in species in which the VNd leads directly into the nasal cavity (Wysocki & Meredith, 1987). In other mammals in which the VNd leads into the incisive duct, there are no data concerning the characteristics of the vomeronasal vessels and their presumptive role in vomeronasal reception.
Here we report an anatomical, immunohistochemical and physiological study of the vomeronasal vessels in cows. This is a species representative of the ungulates order, which has a well developed VNO, and in which the diameter of the lumen of the vomeronasal arteries is large enough to measure some of their physiological characteristics. Our results suggest that the physiological pump mechanism may be capable of functioning perfectly in animals with the second alternative outlet of the VNd.
Materials and methods
Twenty three nasal cavities of adult cows (Bos taurus), males and females, were obtained from the slaughterhouse immediately after death. Eight were fixed by immersion in 10% buffered formalin. The vomeronasal organs of four others were removed and preserved in 2.5% buffered glutaraldehyde. Finally, the vomeronasal organs from the remaining 11 were transferred to cold physiological solution to carry out the corresponding functional studies from fresh material. The total number of samples was 46.
Processing of the vomeronasal organs
To obtain data about the size and topography of the VNO, as well as about the relationship between the vomeronasal cartilage and the cartilage associated with the incisive duct, in some of the formalin-fixed samples the part of the nasal cavity in which the VNO is located was decalcified in 5% ethylene-diamino-tetraacetic acid (EDTA), and after paraffin embedding, transversal sections 10–12 µm were serially cut and stained with haematoxylin-eosin.
In the remaining samples the vomeronasal organs were dissected out just before their histological processing. Most samples were paraffin embedded, cut in transversal and sagittal 8-µm sections and transferred to slides, but, to perform the autofluorescence imaging study of the VNO, some were cryoprotected by immersion overnight in 30% sucrose in 0.1 m phosphate-buffered saline (PBS) at 4 °C, and cut in 40-µm sagittal sections using a freezing microtome.
Semithin and ultrastructural study
Vomeronasal vessels, arteries and veins, were carefully removed by microdissection and fixed by immersion in 2.5% buffered glutaraldehyde, postfixed in 1% aqueous osmium tetroxide, dehydrated in ethanol solutions, and embedded in epoxy resin (EMbed 812, Electron Microscopy Sciences), according to standard techniques.
Semithin sections 0.5 µm thick obtained from the epoxy blocks were collected on slides, stained with toluidine blue and photographed with a digital camera adapted to a photomicroscope (AxioCam MRc 5, Zeiss). Thin sections were counterstained with uranyl acetate and lead citrate and examined under a JEOL JEM-1011 transmission electron microscope.
The immunohistochemical study was as follows: The sections of the VNO were first incubated with hydrogen peroxide (DakoCytomation) for 30 min at room temperature to inactivate endogenous peroxidase activity. The following polyclonal rabbit antibodies were used: anti-neural nitric oxide synthase (n-NOS, Serotec) diluted 1/200, anti-neuropeptide Y (NPY, Oncogene Research Products) diluted 1/1000, and anti-calcitonin gene related peptide (CGRP, Chemicon) diluted 1/2000. All antibodies were incubated overnight at 4 °C. Antigen-antibody complexes were detected with an anti-rabbit peroxidase-labelled polymer (EnVision™+, DakoCytomation) applied for 30 min at room temperature, and colour development was achieved by incubating sections with diaminobenzidine (DakoCytomation). Finally, the slides were dehydrated through alcohols, cleared in xylene, and coverslipped.
Confocal laser scanning microscopy
Autofluorescence imaging of elastic fibers of the VNO was performed by Confocal Laser Scanning Microscopy. Transversal sections of 40 µm were collected in PBS 0.1 m pH 7.2 and stored at 4 °C until their free-floating processing. A Bio-Rad MRC 1024 ES set-up for dual-channel fluorescence using green autofluorescence and TO-PRO-3 iodide (far red fluorescence) filter settings (laser Ar 100 mW, wavelength 488 nm and laser He-Ne, wavelength 633 nm, respectively) was used. Images were taken using a 20× lens (NA 0.45) on a Nikon TE 2000 inverted microscope, with separated emission filtering for each fluorescence. For the staining of nuclear DNA the sections were incubated for 15 min in the nucleic acid-specific dye TO-PRO-3 iodide (1 µm; Molecular Probes). After three washes in PBS they were mounted in a drop of anti-bleaching mounting medium Slow-Fade solution (Molecular Probes Cat. # S7461). Some of the sections were transferred to slides and stained with haematoxylin-eosin.
Image acquisition and processing
Digital microscopy images were captured by using a Karl Zeiss Axiocam MRc5 digital camera. All images were processed using Adobe Photoshop 6.0 (Adobe Systems Inc., San Jose, CA). For the purpose of presentation all images were cropped, resized and rotated, and some of them were adjusted for contrast and brightness to equilibrate light levels. No additional digital image manipulation was performed.
Dissection and mounting
Vomeronasal arteries were carefully dissected from fresh tissues by removing the adhering connective and fatty tissue in cold physiological solution (PSS) of the following composition (mm): NaCl 119, KCl 4.7, KH2PO4 1.18, MgSO4 1.17, CaCl2 1.5, EDTA 0.027 and glucose 11. Arterial segments 2 mm long were mounted as ring preparations on 40-µm wires in double microvascular myographs (Danish Myotechnology, Denmark) for isometric tension recording. The vessels were equilibrated in PSS gassed with 5%/CO2/95%O2 at 37 °C for at least 30 min and then the relationship between passive wall tension and internal circumference was determined for each individual artery. From this, the internal circumference L100 corresponding to a transmural pressure of 100 mmHg for a relaxed vessel in situ was calculated (Mulvany & Halpern, 1977). The arteries were set to an internal circumference L1 equal to 0.9 times L100 (L1 = 0.9 × L100), at which tension development is maximal. The working internal diameter, l1, of the vessels used in this study ranged between 200 and 600 µm.
The viability of the arterial preparations was tested at the beginning of each experiment by stimulating them twice with a high K+PSS (KPSS), similar to PSS except that NaCl was substituted for KCl on an equimolar basis. The presence of intact endothelium was assessed by the ability of acetylcholine (ACh) (10−5m) to induce relaxation in arteries precontracted with 1 µm noradrenaline. A relaxation greater than 50% was taken as evidence of endothelial integrity.
To evaluate the functional effects of the neurotransmitters contained in the periarterial nerves visualized by inmunohistochemistry, concentration-response curves were constructed by cumulative addition of noradrenaline (NA,10−8 – 3 × 10−5 m), NPY (10−9 – 3 × 10−7m), the nitric oxide (NO) donor S-nitrosoacetyl-d,l-penicillamine (SNAP) and CGRP. For the relaxation experiments (SNAP and CGRP), the agonists were added on the steady-state contractile response evoked by the α1-agonist phenylephrine (Phe).
Mechanical responses of the arteries were measured as force and expressed as active wall tension, ΔT, which is the increase in force, ΔF, divided by twice the segment length. Relaxant responses are given as percentage of the precontraction induced by Phe. Sensitivity to the agonists is expressed in terms of pEC50 = –log(EC50), EC50 being the concentration of the agonist required to give half-maximal relaxation. Results are expressed as mean ± se and n represents the number of arteries (2 from each animal). Statistical differences between means were analysed by Student's unpaired two-tailed t-test.
The VNO in cows is a paired structure attached to the vomer bone and the floor of the nasal cavity, and is entirely covered by the corresponding mucosa. The VNO has a considerable length, between 14 and 16 cm, and shows a different appearance according to the level of the section chosen (Fig. 1). The morphological modification of the organ concerns the vomeronasal cartilage, the soft tissue and indeed the duct. The VNd is moved slightly to the medial part of the organ, and longitudinally it finishes in the incisive duct (Fig. 2). The VNrs are located in the medial wall of the duct, that is to say, in the sensory epithelium (Fig. 3A,B). Between 6 and 10 large veins are located laterally to the VNd, (Figs 4, 5A), and other numerous smaller and thinner veins are irregularly distributed around the VNd (Fig. 4). Arteries are small and few and usually located close to the vomeronasal cartilage; in all cases their structural configuration is typical and quite uniform (Fig. 5B).
Semithin sections of both veins and vomeronasal arteries demonstrate clearly the composition of each vessel (Fig. 6), and transmission electron microscopy shows several details (Fig. 7).
Derivations of the caudal nasal nerve (Nnc) are responsible for innervating the soft tissue of the VNO. This myelinic nerve is one of the three branches of the pterygopalatine nerve, which initially conveys sensitive fibres in accordance to its origin, the maxillary nerve. Afterwards, it receives parasympathetic fibres from the pterygopalatine ganglion, and sympathetic fibres from the cranial cervical ganglion (Fig. 8). Once inside the VNO, the derivations of the Nnc are distributed by the soft tissue of the organ, including the vomeronasal vessels (Fig. 9).
Functional reactivity of VNO small arteries
The vomeronasal arteries investigated in the present study had an effective internal lumen diameter, l1, of 386 ± 42 µm (n = 10) and responded to a high K+ solution (KPSS) with a maximum increase in tension of 1.89 ± 0.22 Nm−1 (n = 10).
Noradrenaline-induced concentration-dependent contractions, pEC50 and maximum response for the adrenergic agonist were 6.33 ± 0.07 and 5.22 ± 0.89 Nm−1 (n = 8), respectively (Fig. 10). The sympathetic cotransmitter NPY evoked a potent small contraction (pEC50 7.80 ± 0.08, n = 5, P < 0.001 vs. NA) in five of the seven arteries tested, averaging 0.65 ± 0.14 Nm−1 (n = 5).
Both the NO donor SNAP and CGRP maximally relaxed Phe-precontracted vomeronasal arteries (Fig. 11), the potency of the peptide being nearly two orders of magnitude higher than that of the NO donor. Thus, pEC50 values for SNAP and CGRP were 5.85 ± 0.14 and 7.61 ± 0.09 (P < 0.001, n = 8), respectively.
Firstly, it must be made clear that the researchers who work on the VNS usually employ the term VNO to refer to its sensory epithelium or in some cases to the VNd; in sensum latum, and in certain situations, this habit may be justified as the sensory epithelium is the part of the organ where the VNrs are located. However, it is evident that the VNO is more complicated than a simple epithelium, as we have pointed out in the Introduction. In fact, the remarkable location of the VNrs implies that the soft tissue of the organ must play an important role in its function (Salazar & Sánchez Quinteiro, 1998).
When a VNS is seen to exist in a specific mammal species, the VNO is certainly the most homogeneous of the three parts which make up the VNS, and follows an established pattern, whereas the variability of the accessory olfactory bulb (AOB) is enormous (Meisami & Bhatnagar, 1998). The so-called vomeronasal amygdala has not been explored in depth when it comes to comparison between species (Nieuwenhuys, 1998). If we take the mouse VNO as a universal model for a well-developed VNO in mammals, there are, nevertheless, substantial and interesting variations when we compare it with the VNO of other species (Takami, 2002), among which the following relate directly to the present issue.
The first point concerns the covering of the whole VNO, a lamina made of bone (rodents and lagomorphs) or cartilage (ungulates and carnivores), whose precise meaning is not clear, although it is likely to play a role of containment in the pump mechanism (Eccles, 1982; Doving & Trotier, 1998). Nevertheless, recently it has been demonstrated that the vomeronasal capsule is partially absent in some mammals (Bhatnagar & Smith, 2007). In any case, the presence of this capsule has been considered an interesting feature for the purpose of phylogenetic classification (Wöhrmann-Repenning, 1984).
The second difference is related to the modalities of termination of the VNd, which opens directly into the nasal cavity, for example in rodents, or into the incisive duct, in carnivores and ungulates, and through it communicates with both the nasal and oral cavities. The possibility of such a communication has led to speculations on the way that certain facial expressions may be involved in the stimulus access to the VNO. These include the flehmen (Estes, 1972; Ladewing & Hart, 1980; Melese-d’Hospital & Hart, 1985; Houpt et al. 1989), but also some specific tongue movements, for example when the animals lick their nostrils (Jacobs et al. 1981), when the animals chatter and the tongue presses the incisive papilla, or other behavioural mechanisms (Poran et al. 1993; Houpt, 1998).
With regard to the third point we would like to mention, the vomeronasal vessels, it must be recognised that, when establishing strict morphological differences between mammalian species with a well-developed VNO, the use of statistics is necessary (number and kind of vessels and differences between them), and there are no objective data on this issue. We have observed that the larger vomeronasal veins in cows are not exactly equivalent to the typical sinus venosus described in rodents (Breipohl et al. 1979; Taniguchi & Mochizuki, 1983; Salazar & Sánchez Quinteiro, 1998).
Although there are several interesting publications concerning the anatomy of the VNO in cows (Minett, 1925–26; Pearlman, 1934; Taniguchi & Mikami, 1985; Adams, 1986; Salazar et al. 1997), the morphological and functional characteristics of the vomeronasal vessels are not considered in depth. Nevertheless, the general idea is that the mentioned vessels seem to organize a sort of erectile tissue by a flow combination of the small but powerful arteries and the numerous veins. Careful study of the transversal series of the VNO, as we have done in the present work, corroborates that issue. The pattern of distribution of the vessels which reach the VNO is quite similar in cows to other mammals (Salazar et al. 1997).
Nevertheless, to have a more solid base on the nature of the vomeronasal vessels and their functional capability, it would be necessary to go further, and to bear in mind some functional peculiarities.
Using electrophysiological and pharmacological methods, Meredith & O’Connell (1979) demonstrated the presence of a pumping mechanism in hamsters, powered by vasomotor movements, which can suck stimulus substances into the VNO. This mechanism is activated by autonomic nerve fibres running in the nasopalatine nerve, a terminal branch of the Nnc, and operates by local changes in the vascular resistance that alter the degree of engorgement of the vomeronasal cavernous tissue. Thus, sympathetic-mediated vasoconstriction and parasympathetic-mediated vasodilatation result in an influx and outflow of the fluid from the VNd, respectively (Meredith & O’Connell, 1979).
Likewise, Eccles, in his studies on the innervation of the nasal blood vessels in cats, dedicated a publication to the autonomic innervation of the VNO (Eccles, 1982), in which a model of possible vasomotor pumping mechanism was described and the pathways taken by the autonomic fibres to the organ were also considered.
In the present study, two experimental approaches have been applied to analyse the innervation of the vomeronasal vessels in cows. The first was using immunohistochemical methods, and the second by functional evaluation of the vessels according to the data obtained after mounting them in the corresponding microvascular myographs.
It is worth remembering that blood vessels may be innervated by up to three major classes of neurones: sympathetic vasoconstrictor, sympathetic or parasympathetic vasodilator, and peripheral fibres of sensory neurones which can mediate vasodilation. The Nnc in cows, whose termination enters the VNO, conveys the three kinds of fibres.
Nowadays, there is an increasing body of evidence for the neurotransmitter roles of such peptides, for example the substance P and CGRP in sensible neurones, vasointestinal peptide (VIP), NOS and choline acetyltransferase (ChAT) in autonomic vasodilator neurones, and NPY, galanin (GAL) and tyrosine hydroxylase (TH) in sympathetic vasoconstrictor neurones (Morris et al. 1995). Some of these peptides have been used to demonstrate immunoreactivity in vomeronasal vessels, especially in mice (Kishimoto et al. 1993; Nagahara et al. 1995; Matsuda et al. 1996) and rats (Canto & Suburo, 1998).
Following a similar protocol to the previous authors (Nagahara et al. 1995; Matsuda et al. 1996), although using different peptides which covered the three possibilities of innervated vessels (Morris et al. 1995), we have obtained results in cows similar to those in rodents. The dilation and constriction of arteries and veins have been demonstrated, showing that the vomeronasal vessels in cows could be involved in the pump mechanism regulating influx and efflux from the VNd.
The inmunohistochemical findings of the present study suggest that NO, NPY and CGRP derived from nitrergic, sympathetic and sensory perivascular nerves, respectively, play a role in the regulation of arterial tone and in the distribution of blood flow in the soft tissues of the VNO. The location of such nerve fibres around veins likewise suggests an active regulation of penile venous tone. These observations, along with the demonstration that the neurotransmitters contained in the perivascular nerve fibres elicited both contractile and relaxant responses in the vomeronasal arteries, further support a role for the pumping mechanism that draws stimuli from the nasal cavity in and out of the VNS through the VNd (Meredith et al. 1980; Meredith, 1994).
Sympathetic vasoconstrictor nerves usually utilize two or more cotransmitters, the most common being noradrenaline, adenosine-5′-triphosphate (ATP) and NPY (Lundberg, 1996). Noradrenaline acted as a powerful vasoconstrictor in bovine vomeronasal arteries, as depicted by the large contractions evoked by this agonist, which represented about 300% of the standard response evoked by a high K+ solution. This in vitro finding would be consistent with the earlier physiological findings of Meredith & O’Connell (1979) showing that stimulation of the superior cervical sympathetic ganglion or application of adrenaline produced contraction of the VNO with inflow at the duct, this contractile effect turning into a small dilation after cervical sympathectomy. Thus, blood vessels surrounding the vomeronasal lumen would be constricted repetitively by bursts of activity in the vasomotor sympathetic nerves, and contribute to the pumping mechanism regulating influx from the VNd (Meredith, 1994). NPY had only a moderate contractile activity in isolated bovine vomeronasal arteries, despite the presence of peptidergic-NPY perivascular nerve fibres. However, this peptide usually acts as a neuromodulator by regulating the effects of other sympathetic transmitters such as noradrenaline at both presynaptic and postsynaptic sites (Prieto et al. 1997, 2004).
In the present study, NOS-IR nerve fibres were localized around bovine vomeronasal blood vessels, which confirms previous findings in the VNS from rodents, where nitrergic nerves are distributed in the receptor-free epithelium and around blood vessels and glands of the cavernous tissue (Matsuda et al. 1996; Canto & Suburo, 1998). NO plays an important role in the regulation of blood flow, particularly at arteriovenous anastomoses, in the brain and in the genital organs (Simonsen et al. 2002; Toda & Okamura, 2003). In penile erectile tissue, NO is the main nonadrenergic-noncholinergic inhibitory neurotransmitter, released upon stimulation of parasympathetic nerves and responsible for the relaxation of arterial and cavernous tissue leading to penile erection (Simonsen et al. 2002; Prieto, 2007). The fact that exogenous NO induces arterial vasodilation of vomeronasal arteries further supports a role for nitrergic nerves in the regulation of the vomeronasal vascular tone by relaxing the muscular walls of the cavernous vessels and thus contributing to the function of the pump. In fact, in the early in vivo experiments of Meredith & O’Connell (1979), stimulation of the caudal nasal nerve after sympathectomy elicited vasodilation which was atropine-resistant, suggesting that neurotransmitters other than acetylcholine were responsible for the vasodilation of the VNO upon parasympathetic nerve fibre stimulation.
The finding that CGRP is present in perivascular fibres and potently relaxes isolated vomeronasal arteries is consistent with the universal vasodilator role of this peptide, usually located in capsaicin-sensitive sensory periarterial nerve fibres and released upon antidromic stimulation and also by chemical stimulants such as low pH, which occurs with ischaemia, inflammation and tissue damage (Lundberg, 1996).
During the last three decades the concept of olfaction in general and of the VNS in particular has evolved (Halpern, 1987; Breer et al. 2006). In this context, it is interesting to consider the key role that may be played by the differences in the general organization of the VNS in different species, which, as a consequence, may contribute to the clarification of some important issues of the secondary olfactory system.
With regard to the VNO, the main difference between rodents and ungulates concerns the outlet of the VNd. Nevertheless, the morphological and functional characteristics of the vomeronasal vessels of cows in the present study provide evidence that the VNO, in this representative species of Ungulata, is perfectly capable of working as it does in rodents with regard to the pump mechanism, that is to say, in the reception of chemical signals, as could be expected in some ways.
However, the AOB shows significant variation among mammals (Meisami & Bhatnagar, 1998). Whereas in rodents, mice are the best pattern (Salazar et al. 2006), the AOB is well developed, in ungulates this structure shows some peculiarities, mainly because of the poor population of the mitral/tufted cell layer (Salazar et al. 2007). This would probably imply that in ungulates the first relay station of the VNS – the AOB – would not be able to process the vomeronasal information appropriately. One could speculate that in some species, ungulates for example, the VNS may be in regression or involution (for more information see Salazar et al. 2007). The literature informs us that some mammals have a typical VNO, whereas an AOB was not detected (Halpern, 1987; Meisami & Bhatnagar, 1998).
It is evident that further studies will be necessary to confirm the supposition of the involution of the VNS; in this way, in the future specialists from fields ranging from comparative neurobiology to molecular biology will be able to concur. Nevertheless, nowadays some recent publications (Boehm et al. 2005; Mandiyan et al. 2005; Yoon et al. 2005) have demonstrated that the main olfactory system is able to assume functions previously assigned exclusively to the VNS; therefore the precise role of the VNS is probably more uncertain than ever (Baxi et al. 2005); however, it is necessary once more to bear in mind differences between species (Keverne, 2005).
The authors wish to thank the staff of the veterinarian service of the slaughterhouse for kindly providing the material employed in this study, to J. Castiñeira for his technical help, and to J. Curtin and I. Hughes for the revision of the English text. This work was supported in part by the research grant number BFU2004-01004 from the Spanish Ministry of Education and Science. The comments of anonymous reviewers greatly improved the clarity of the manuscript.