Abstract

Introduction

Most mammals and many reptiles and amphibians possess a secondary olfactory system, the vomeronasal system (VNS), which apparently detects particular classes of chemical signals (Keverne, 1983; Halpern, 1987; Brennan, 2001). It is composed of the vomeronasal organ (VNO), the accessory olfactory bulb (AOB), the vomeronasal amygdala (VNAg), and the nerves and tracts by which these components are connected (Wysocki & Meredith, 1987). The VNO is in turn composed of a number of distinct substructures: in domestic animals, the outermost part is the vomeronasal cartilage (cartilago vomeronasalis, VNCg), which in some species is in fact osseous, and the innermost the vomeronasal duct (ductus vomeronasalis, VNDc), an epithelial tissue that comprises areas of both sensory and non-sensory epithelium and forms a hollow tube that is closed caudally and communicates rostrally with the nasal and/or oral cavities (ICVGAN, 1994). Note that the term ‘vomeronasal organ’ should not be used (as it sometimes is) as a synonym for the sensory epithelium or the VNDc. The VNCg and VNDc are separated by soft tissue comprising blood vessels, nerves, glands and connective tissue. Since Meredith et al.'s (1980) study of the hamster VNO it has been known that in this species at least the large blood vessels that course along the lateral side do not merely support metabolism, but play a prominent role in the specific function of the VNO: their sudden constriction, induced by the autonomic nervous system (Eccles, 1982), causes a shrinkage towards the lateral VNCg that has the effect of sucking stimulus-bearing fluids into the lumen of the organ.

The VNCg has been studied in a variety of species, not only as a component of the VNS but because of the possibility of using its morphological characteristics for the purposes of phylogenetic classification (Wöhrmann-Repenning, 1984a, b; Wöhrmann-Repenning & Barth-Müller, 1994). The epithelium of the VNO has also been extensively studied (for that of the pig, see Salazar et al. 1998) because its sensory area is the tissue immediately responsible for obtaining secondary olfactory information and passing it to the brain for processing (Halpern, 1987). Less is known of the non-epithelial soft VNO tissues of mammalian species in general (Salazar et al. 1997; Salazar & Sánchez Quinteiro, 1998), perhaps because it has been thought unnecessary to confirm whether they all share the suction pump mechanism described above. Neither for the VNCg nor for the soft tissues is much known about prenatal development in mammalian species other than rodents. This paper reports the findings of a descriptive developmental study of the soft tissue and VNCg of the pig, a macrosmatic altricial species that has been extensively used in both clinical and basic research (for example, in research on xenotransplantation; see Bonenfant et al. 2003) and is one of the few mammals from which odorant receptor genes have been isolated (Matarazzo et al. 1998; Velten et al. 1998; Mombaerts, 1999).

Materials and methods

Pig fetuses at various developmental stages were obtained from the abattoir, and their ages were calculated from their crown-to-rump distance (Marrable, 1971) (Table 1).

For light microscopy, smaller specimens were fixed directly over 24 h in 10% formalin or Bouin's fluid, while older fetuses were washed and perfused before their heads were removed and the whole brain dissected out and immersed in the fixative. Some of the largest heads were decalcified with EDTA before dissection. After microdissection (when necessary), the tissue of interest was embedded in paraffin wax. Transverse sections 8 µm thick were cut and stained with haematoxylin and eosin, PAS, Alcian blue, or Weigert's resorcin-fuchsin.

For transmission electron microscopy (TEM), tissues were fixed either by perfusion of the fetus with 3.2% glutaraldehyde and 2.6% paraformaldehyde in cacodylate buffer (0.09 m, pH 7.35) followed by dissection of the tissue of interest and its immersion in the same fixative for several hours, or by direct immersion of dissected tissue. Both treatments were followed by immersion in 1% unbuffered osmium tetroxide, dehydration in graded solutions of ethanol, and embedding in Epon-Araldite. Thin sections were stained with uranyl acetate and lead citrate, and viewed with a JEOL 100SX TEM apparatus. As controls, semithin (0.5–3.0 µm) sections were stained with 1% toluidine blue in an equal volume of 2.5% sodium carbonate.

Results

Early in its development, the VNDc differentiates as distinct lateral and medial epithelia, the latter of which is morphologically a typical sensory epithelium (Fig. 1). Even in its early stages, it communicates with the nasal and oral cavities via the incisive duct (Fig. 2). The non-epithelial soft tissue surrounding the VNDc comprises nerves, blood vessels, glands and connective tissue containing collagen and elastin, which is especially abundant and evident beside the VNCg (Fig. 3). The nerves, blood vessels and glands, and the VNCg, develop as follows.

Open in a separate window

Fig. 1

Transverse sections through the VNO of the prenatal pig. Centre, camara lucida drawing of a section through the whole VNO of a 12-cm specimen. Clockwise from top left: HE-stained sections showing the development of the vomeronasal epithelia; SE, sensory epithelium; VNCg, vomeronasal cartilage; VNDc, vomeronasal duct; d, dorsal; l, lateral. Scale bar = 100 µm.

Open in a separate window

Fig. 2

HE-stained transverse sections through the VNO of 3.1-cm and 22-cm pig fetuses, showing communication between the VNDc and the nasal (NC) and oral (OC) cavities via the incisive duct (asterisk). PI, papilla incisiva; VNDc, vomeronasal duct; d, dorsal; l, lateral. Scale bar = 100 µm.

Open in a separate window

Fig. 3

Transverse sections through the VNO of prenatal pig at two different levels and magnifications (A and B), showing the typical appearance and arrangement of connective tissue (arrows). VNCg, vomeronasal cartilage. Stain: Weigert's resorcin-fuchsin. Scale bar = 100 µm.

Nerves

The vomeronasal complex is innervated by the vomeronasal nerves (VNns) and branches of the nasocaudal nerve (NCn). Throughout development, the former run dorsal, medial and ventral to the VNDc and the latter mainly lateral and ventral (Fig. 4A–C). From the 25-cm stage onwards, the NCn branches can be seen to be myelinated (Fig. 4D,E) and the VNns unmyelinated (Fig. 4F,G).

Open in a separate window

Fig. 4

Transverse sections through the VNO of several prenatal pig specimens stained with haematoxylin-eosin (A–C) or toluidine blue (D–G). The general topography of the soft tissue is illustrated in the main picture (C), which shows a section taken caudally. The other pictures show the myelinization status of the nasocaudal nerves (D, E, and arrowhead in C) and the vomeronasal nerves (F,G, and arrows in A–C). Scale bars = 20 µm (A), 50 µm (C1,D,F), 100 µm (B,C2,E,G) and 500 µm (C).

Blood vessels

Arteries can be recognized by their well-developed three-layer wall structure (tunica adventitia, tunica media and intima) at quite early stages (Fig. 5A,B). Subepithelial capillaries are generally related to the sensory epithelium (Fig. 5C) rather than the non-sensory epithelium. Veins and venules are recognizable by their large lumina and thin walls composed of a thin endothelium and loose surrounding connective tissue, sometimes with a poorly developed tunica media (Fig. 5D–G). Under the light microscope both kinds of vessel are seen around the VNDc throughout its length, but veins are not only larger but also more numerous than arteries, and the veins (usually singly, but occasionally in pairs) form venous sinuses (Fig. 5D,E) very close to the lateral epithelium. Once constituted, both kinds of blood vessel retain their main characteristics throughout development.

Open in a separate window

Fig. 5

Transverse sections through the VNO of several prenatal pig specimens stained with Weigert's resorcin-fuchsin (A,B) or haematoxylin-eosin (the others), showing arteries (A,B, arrows) with a clear tunica intima, subepithelial capillaries (C, arrows), veins (V) and venous sinuses (asterisks). LE, lateral epithelium; SE, sensory epithelium. Scale bar = 20 µm (C,D), 50 µm (A,B) and 100 µm(E–G).

Glands

Vomeronasal glands (VNGl) appear early in development near the VNDc. Most open into the lumen in areas of the lateral wall, especially close to its theoretical boundary with the medial epithelium (Fig. 6A–C), but they are abundant throughout the parenchyma (Fig. 6D,E). In some 20–25-cm specimens they begin to react with PAS and Alcian blue, and from the 29-cm stage on they consistently stain strongly positive for neutral mucopolysaccharides with PAS (Fig. 6F,H) and weakly positive for acid mucopolysaccharides with Alcian blue (Fig. 6G,I). TEM shows that the VNGl are active during prenatal development: there are no secretory granules in the cytoplasm of 15-cm stage gland cells, which also exhibit other ultrastructural signs of immaturity, but by the 21-cm stage electrodense granules have appeared in the supranuclear cytoplasm, and their numbers increase until by the 37-cm stage they almost fill this region; cytoplasm and granule-containing apical processes can frequently be seen protruding into the acinar lumen (Fig. 7A). The granule contents are released by fusion of the granule membrane with the apical membrane of the gland cell; although this was first seen in a 33-cm specimen (Fig. 7B), electrodense material that is morphologically identical to that seen in the secretory granules was consistently observed in the lumina of acini and their excretory ducts from the 21-cm stage on.

Open in a separate window

Fig. 6

Transverse sections through the VNO of several prenatal pig specimens stained with haematoxylin-eosin (A–C), PAS (D,F,H) or PAS and Alcian blue (E,G,I), showing vomeronasal glands opening into the lumen of the VNDc (A–C) and reactivity of vomeronasal glands (D–I, arrows). Scale bar = 20 µm (A) and 100 µm (others).

Open in a separate window

Fig. 7

TEM micrographs of the apical cytoplasm of VNGl cells showing the release of the contents of secretory granules by fusion of granule and cell membrane (arrows). L, lumen. Scale bar = 1 µm.

Cartilage

Series of transverse sections of fetuses of various stages show rostrally the morphological relationship between the VNCg and the cartilaginous layer associated with the incisive duct (Fig. 8I–III). In sections posterior to the separation of the VNDc from the incisive duct (Fig. 8III), the VNDc and its associated soft tissue are progressively surrounded by the VNCg until it closes completely (Fig. 8III–V), but more caudally the VNCg opens dorsally, allowing renewed contact between the soft tissue of the VNO and nasal mucosa (Fig. 8VI). The above general layout holds throughout development, with just minor variations. Under the light microscope, however, differences between early and later fetuses are apparent: (1) early fetuses fail to show clear differentiation between the outer, fibrous layer of the VNCg perichondrium and surrounding mesenchymal tissue (Fig. 9A), whereas older fetuses show a perichondrium with clearly differentiated outer fibrous and inner cambial layers (Fig. 9B, arrows); (2) mitotic figures are common in the cartilage of young fetuses (Fig. 8C,D) but rare in older fetuses; (3) by contrast, the number of isogenous groups of chondrocytes (Fig. 9E) increases with fetus age; (4) the numerically predominant VNCg cells in older specimens are round, oval or spindle-shaped chondrocytes (Fig. 9F).

Open in a separate window

Fig. 8

Selected members of series of HE-stained transverse sections through the VNO of several prenatal pig specimens, showing the vomeronasal cartilage (in black) at various levels (I–VI) as described in the text.

Open in a separate window

Fig. 9

Cellular components of the vomeronasal cartilage at various stages of prenatal development, showing the fibrous layer of the perichondrium (A,B, double arrow), mitotic figures (C,D, arrows), cell nests (E, arrows) and round and oval chondrocytes (F, arrows). Scale bar = 50 µm.

Discussion

Recent years have seen remarkable progress in our knowledge of the functioning of the vomeronasal system (Keverne, 1999; Buck, 2000; Dulac, 2000; Pantages & Dulac, 2000; Zufall & Munger, 2001). Most of those working in this field understandably direct their attention to the mechanisms and structures directly involved in the detection and processing of chemical signals, and therefore concentrate on the sensory epithelium (SE), the AOB and the VNAg. However, an unwritten morphological rule states that no tissue works in isolation. VNS structures other than the SE, AOB and VNAg may therefore also exhibit characteristics and behaviour that are specifically adapted to VNS function – the pumping action achieved by constriction of the large blood vessels is in fact a case in point – and their examination may therefore also clarify certain issues or suggest new perspectives (Meredith et al. 1980; Eccles, 1982). This possibility should particularly be borne in mind given the considerable and far from fully elucidated interspecies variation in VNO morphology and functions (Bertmar, 1981; Dawley, 1998; Døving & Trotier, 1998; Meisami & Bhatnagar, 1998), gross examples of which are the three different arrangements for stimulus entry (from the nasal cavity in rodents, from the oral cavity in snakes, and from both – via the incisive duct – in carnivores and some ungulates) and the two different kinds of outer capsule surrounding the VNDc and its associated soft tissue (bone in most rodents, cartilage in carnivores and ungulates) (Wysocki & Meredith, 1987). This variation makes it unwise to extrapolate from one species to another (Salazar et al. 1985, 1994, 1996, 2000), and highlights the importance of general morphological studies that provide an adequate context for the orientation and interpretation of specifically functional research. This is the purpose of the present work on one of the animals in which certain other aspects of the vomeronasal system have been most thoroughly studied (Mombaerts, 1999).

Pre-1970 studies of the VNS of the adult pig from various points of view have been reviewed by Matthay (1968). Since then, Frewein (1972) has shown the topographical projection of the VNO into the nasal cavity, Kratzing (1980) has examined the system by light microscopy and TEM, Adams (1992) has described the fine structure of the VNGl, and Salazar and co-workers have described the morphology of the VNCg (Salazar et al. 1995) and the soft tissues (Salazar et al. 1997). Partly because of different objectives, and also no doubt because of the use of different species of pig, the findings of these studies differ slightly not only in the nomenclature used but also as regards the rostrocaudal extent of the VNCg, the distribution and location of glands and blood vessels, and the degree of myelination of the nerve fibres; but certain ultrastructural features apart, the general morphology of the VNCg and soft VNO tissues of the adult pig is very clear when whole series of transverse sections are examined. In particular, the VNCg is hyaline, is continuous rostrally with the cartilage associated with the incisive duct, and envelops and protects the VNDc and its associated soft tissue (but leaving a dorsal gap). It has been suggested that the morphology of the VNCg is specific for major taxa, and therefore of value for the elucidation of phylogeny (Wöhrmann-Repenning, 1984a, 1984b; Wöhrmann-Repenning & Barth-Müller, 1994).

In this study we have seen that the morphology of the VNCg and the soft VNO tissues in prenatal pig is very similar to that of adults from the 112-cm stage on, and that the VNGl are active during this period. This suggests that during the later prenatal stage the VNO is functional, a hypothesis supported by the finding that in these post 31-cm fetuses the sensory epithelium of the VNDc (Salazar et al. 1998), the VNns and the AOB are also well developed and are labelled by Ulex europaeus agglutinin I and Lycopersicon esculentum agglutinin (results not shown). We have recently described a similar situation in prenatal sheep (I. Salazar et al. unpublished observations); in both cases, sheep and pig, a salient observation has been that the glomerular layer of the AOB, to which the VNns project, appears relatively early in development, whereas in rodents it is not present until birth (Hinds, 1968; I. Salazar & P. Sánchez-Quinteiro, unpublished observation).

The hypothesis that the VNS may be functional in utero is not new (Pedersen et al. 1983; for a review, see also Hudson, 1993). The findings reported here support the need for this hypothesis to be explored more thoroughly by various means. In this study we have found that the prenatal VNO of the pig seems morphologically capable of supporting function. In future work we shall extend the investigation of this issue to the whole porcine VNS.

Acknowledgments

References