Modern neuroanatomical, electrophysiological, and genomic studies have revealed that the concept of a single olfactory system is an oversimplification. Instead, the seemingly simple mammalian nasal cavity contains a significant number of olfactory systems, some of which have only been discovered relatively recently (Barrios et al., 2014a). The mammalian olfactory system can be divided into several subsystems based on the anatomical locations of the involved neuroreceptor cells, the receptors families that are expressed, the signaling transduction mechanisms employed, the chemosensory stimuli detected, and the targets of the sensory neuron axons within the rhinencephalon (Munger, 2009). However, growing evidence supports synergistic and cooperative interactions among these olfactory subsystems (Mucignat-Caretta et al., 2012; Pardo-Bellver et al., 2022).

The best-studied systems are the main olfactory system (MOS) and the vomeronasal system (VNS), which are both found in most mammalian groups, with a few exceptions (Trotier and Døving, 1998). The olfactory receptors (ORs) of the MOS are located in cilia within the mucosa lining the ethmoidal turbinates and in the caudal portion of the nasal septum (Salazar et al., 2019). Neuroepithelial axons project to the glomeruli of the main olfactory bulb (MOB; Scalia and Winans, 1975; Crespo et al., 2013). The MOS responds to thousands of volatile chemosignals, which carry information regarding food, pathogens, prey, predators, and conspecifics (Firestein, 2001).

The receptors of the VNS are located in the microvilli lining the neuroepithelium of the vomeronasal organ (VNO; Salazar et al., 1997, 2003). The vomeronasal neurons project to the accessory olfactory bulb (AOB; McCotter, 1912; Halpern et al., 1995). Among mammalian groups expressing the two vomeronasal receptors families, V1R and V2R (Herrada and Dulac, 1997), a morphofunctional antero-posterior subdivision is established within the AOB (Shinohara et al., 1992; Torres et al., 2022). The sensory neurons of the VNO detect a range of non-volatile natural ligands found in the exocrine secretions of conspecifics, which are involved in innate socio-sexual behaviors (Krieger et al., 1999; Villafranca-Faus et al., 2021).

In addition to the MOS and VNS, a range of sensory olfactory subsystems, including the Grüneberg ganglion (GG), the septal organ of Masera (SO), and guanylyl cyclase-D-expressing (GC D+) chemosensory neurons, have been characterized, primarily in laboratory rodents (Zimmerman and Munger, 2021). The GG (Grüneberg, 1973) consists of cells located in the dorsal region of the nasal vestibule that are immunopositive for olfactory-marker-protein (OMP) and are involved in the detection of highly membrane-permeant stimuli (Breer et al., 2006). GG cells project to the dorsal regions of the caudal MOB near the AOB (Fuss et al., 2005; Storan and Key, 2006). The SO is an isolated patch of sensory epithelium identified in Rodentia and located near the base of the nasal septum at the entrances to the nasopalatine ducts (Adams, 1972). The SO expresses 50–80 genes in the OR family (Kaluza et al., 2004). These neurons project to a small cluster of glomeruli located in the caudal, ventromedial aspect of the MOB (Levai et al., 2003).

The GC-D+ chemosensory neurons are a subpopulation of olfactory neurons that project to the well-defined necklace glomeruli, the most caudal glomeruli in the MOB. The necklace glomeruli, identified by Shinoda et al. (1993), were first defined in rats as a subset of OSNs immunoreactive to human placental antigen X-P2 (PAX) that converge on 7–9 glomeruli. These glomeruli overlap with a subset of “atypical” glomeruli with acetylcholinesterase (AChE) reactivity (Zheng et al., 1987; Weruaga et al., 2001). The projections from the GG and SO also overlap with the necklace complex. Although this receptor system remains minimally studied and poorly understood (Zimmerman and Munger, 2021), GC-D+ chemosensory neurons are not OMP+ and function as receptors for uroguanylin, guanylin, and urine components (Leinders-Zufall et al., 2007).

All known olfactory subsystems converge onto the transition zone located between the MOB and the AOB. This transition zone is a modified bulbar cortex, bounded anteriorly by the dorsal MOB and posteriorly by the anterior AOB (Larriva-Sahd, 2012) and has been named the olfactory limbus (OL). The OL contains specialized glomeruli that receive noncanonical sensory afferents and interact with the MOB and AOB, indicating that the OL may serve as an integration site for non-olfactory and atypical vomeronasal sensory inputs (Vargas-Barroso et al., 2017).

Little is known regarding the olfactory subsystems present in carnivores. No sensory systems equivalent to the SO, GG, or necklace glomeruli have been described in either dogs or cats (Salazar and Sánchez-Quinteiro, 2011; Barrios et al., 2014b). However, studies characterizing the AOB in minks, meerkats, and dogs have indicated a more complex glomerular organization in the OB than has been described for other mammalian orders. The presumptive AOB in minks comprises not only the main dorsocaudal protuberance but also smaller lateral and medial regions (Salazar et al., 1998), whereas a recent study of the meerkat OB found a subpopulation of atypical glomeruli with a strong calretinin (CR)-positive neuropil in the vicinity of the AOB (Torres et al., 2021). A study by Miodonski in the dog (1968) described the AOB as consisting of distinct glomerular aggregations, both “in the dorsomedial side of the OB, at the posterior margin of the MOB and downward along the caudal edge of the MOB, descending to its base both on the medial and on the lateral side.” Miodonski’s observations may represent the erroneous attribution of atypical glomerular structures to the accessory olfactory system instead of alternative olfactory subsystems; however, the multiform characterization of the dog AOB was not confirmed by subsequent lectin and immunohistochemical studies (Nakajima et al., 1998; Salazar et al., 2013). Only Nakajima et al. (1998) identified a small group of glomeruli with NADPH-diaphorase reactivity “between the glomeruli at the most caudal portion of the MOB, ” which they attributed to the specific projections from subsets of neurochemically distinct OR cells.

Recent studies of the AOB in wild canids, including the African wild dog (Lycaon pictus; Chengetanai et al., 2020) and the red fox (Vulpes vulpes; Ortiz-Leal et al., 2022a) indicate significant differences in the AOB between domestic and wild canids. In the red fox, a small atypical glomerular formation was identified in the proximity of the AOB. In the present study, we focused on characterizing the OL in the red fox by performing general and specific histological stainings on serial sections, using both single and double immunohistochemical and lectin-histochemical labeling techniques. We describe an OL in the red fox that displays an uncommonly high degree of development and complexity, suggesting that the red fox may represent a novel mammalian model, the study of which could improve our understanding of the noncanonical pathways involved in the processing of chemosensory cues.

Both macroscopic and microscopic studies were performed on the OB to describe the topographical anatomy of the fox OL (Figure 1). The OL is mainly located along the caudomedial margin of the OB, rostrally to the olfactory peduncle (Figure 1A). A minor portion projects along the caudolateral margin of the OB (Figure 1B). The entire length of the OL and AOB can be better visualized on a dorsolateral view of the OB after removing the telencephalic frontal lobe (Figure 1C). The AOB is located caudoventrally to the OL, on the OB medial surface. A transverse section of the OB at the level of the OL (Figure 1B) shows from a caudal viewpoint the topographic relationship between the OL, the MOB, and the olfactory peduncle (OP; Figure 1D).

A microscopic study of the fox OL was performed on serial, horizontal sections obtained from the anterior region of the telencephalon across an area that includes the AOB, MOB, and frontal lobe (FL). The OL was defined as the transition zone between the MOB and AOB, consisting of a large area located in the medial OB, delimited by the dorsocaudal edge of the MOB and the anterior end of the AOB (Figure 2). The OL is not a sharp frontier but is instead a heterotypical bulbar cortex with distinct cytological characteristics relative to the MOB and AOB. A laminar pattern can be observed throughout the OB, whereas the OL features varying degrees of lamination, which is its most obvious characteristic. The widespread presence of atypical glomeruli in the fox OL is remarkable (Figures 2B,C). In contrast to the linear arrangement and uniform size of glomeruli in the MOB (100–150 μm maximum diameter), glomeruli in the OL are characterized by an atypical, disarrayed arrangement, irregular shapes, and a wide range of sizes (50–500 μm maximum diameter).

Double immunohistochemical labeling against different markers allowed us to characterize the topographical relationships between the atypical glomeruli in the OL (Figure 2A), the AOB, and the axonal endings of the vomeronasal nerve (NVN; Figure 2B). Double immunolabeling against Ulex europaeus agglutinin (UEA) lectin and the G protein subunit Gαo (Figure 2A) identified two superficial layers in the AOB, the NVN, and glomerular layers, both of which were intensely stained by UEA. Conversely, anti-Gαo staining was observed in the deep layers of the AOB and throughout the MOB, with more intense staining observed in both the olfactory nerve and glomerular layers. Strikingly, the OL, delimited by arrowheads in Figure 2A, was comprised of irregularly shaped glomeruli with no homogeneous immunostaining pattern.

Double immunolabeling against the G protein subunits, Gαo and Gαi2 (Figure 2B), is equally useful to double staining with UEA and Gαo. Gαi2 is expressed exclusively in the VNS, at both the AOB surface and in the NVN (Figure 2C). By contrast, the nerve endings projecting to the OL glomeruli were Gαo-immunopositive (Figure 2C).

Using both hematoxylin-eosin (Figure 3A) and Nissl staining (Figure 3B) revealed the OL texture, which is characterized by high structural complexity. The two most striking features are dense, compact clusters of neuronal somata (Figures 3A.1,B.1) and a broad nervous formation consisting of less densely aggregated neuronal somas distributed within a neuropil that is distinct from the underlying glomeruli (Figures 3A.2,B.2). We have termed this previously undescribed nervous formation the macroglomerular complex (MGC). Both features are characterized by a remarkably high concentration of neuronal somata located in the most superficial bulb layer, proximal to the pial surface. Small, irregularly shaped glomeruli deep into both structures complete the organization of the OL.

To exclude the possibility that these atypical formations are the result of interindividual anatomical variability, we performed Nissl staining in four additional animals, as shown in Figures 4 and 5. In the first specimen (Figure 4), horizontal OL sections were stained at three levels dorsal to the AOB, which allowed for the detailed characterization of the appearance and dimensions of both neuronal clusters (Figures 4A,A.1) and the MGC (Figures 4A,A.2,B,B.1,C). Somata in the clusters were similarly sized, with polyhedral, ellipsoid, or oval morphologies (Figures 4A.1,A.2,B.1,C). The MGC was always located on the bulbar surface, close to glomeruli populations located in the immediately deeper plane, and the magnitude of the MGC was notable, extending over several millimeters in some sections (Figures 4A,B). Most of the glomeruli associated with the MGC featured irregular, atypical shapes (Figures 4A,B), although some glomeruli with typical spherical shapes and well-defined boundaries were also observed (Figure 4A). Nissl staining of the other three individuals (Figures 5A–C) verified these findings.

Figure 6 summarizes our interpretation of the histological data in a schematic drawing of a histological MOB section (Figure 6A) and an enlargement of the caudodorsal area, showing the irregular arrangement and dimensions of atypical glomeruli and the structure of the MGC (Figure 6B). To further characterize the fox OL neurochemically and identify potentially distinctive features relative to the AOB and MOB, we conducted an immunostaining study using various antibodies and the lectin UEA.

Immunolabeling against the G protein subunit Gαo resulted in the well-characterized pattern of immunopositivity throughout the neuropil of the FL and OB, with Gαo immunonegativity observed only in the nerve and glomerular layers of the AOB (Figure 7A). However, the atypical glomeruli proximal to the AOB showed more intense immunolabeling than the typical glomeruli of the MOB. The MGC showed a labeling intensity similar to that observed for typical glomeruli. When the MGC was observed at higher magnification (Figure 7B), two clearly differentiated areas could be identified, one containing numerous Gαo-immunopositive neuronal somata and the other devoid of them, in contrast with the Gαo-immunonegative mitral cell somata of the MOB (Figure 7C).

We returned to the original double immunolabeled sections shown in Figure 2 to more closely examine the structural organization of the MGC. Figure 8 shows higher magnifications of the double immunolabeled OL sections presented in Figure 2. In Figures 8A,B, the nerve and glomerular layers of the AOB appear intensely labeled by the lectin UEA, whereas the MOB glomeruli are not labeled. In the OL, some glomeruli are UEA-negative, whereas the remaining, primarily smaller glomeruli, are strongly labeled with UEA (Figure 8A). Remarkably, Gαo-positive somata belonging to the nervous formation are embedded in a dense network of UEA-positive fibers (Figures 8B,C). The NVN fibers segregate into UEA-positive (Figure 8B, black asterisk) and Gαo-positive (Figure 8B, white asterisk) components.

Immunolabeling with the anti-microtubule-associated protein 2 (MAP-2) antibody, a consistent marker for the dendritic branching of neuronal cells in all species, results in the uniform staining of glomeruli and the external plexiform layer of the MOB (Figure 9A). In the OL, anti–MAP-2 also stains atypical glomerular formations, but a patch of the MGC remains unstained (Figure 9B, asterisk). Counterstaining of an immunolabeled section with hematoxylin (Figure 9C) confirmed that the unlabeled area corresponds to the superficial area rich in neuronal bodies receiving UEA-positive innervations, shown in Figure 8B.

Calcium-binding proteins are commonly used as markers in the study of the OB. We selected three markers to characterize the OL: anti-calretinin (CR, Figure 10), anti-calbindin (CB, Figure 10), and anti-secretagogin (SG, Figure 11).

Anti-CR immunostaining resulted in intense labeling in the MGC of the OL and in the AOB (Figure 10A). However, the glomeruli of the MOB did not show anti-CR immunolabeling in the neuropil, although the periglomerular cells were clearly labeled (insets A.1 and B in Figure 10A). In the MGC, both the neuronal somata and the neuropil (Figure 10B) were anti-CR immunopositive. Remarkably, anti-CR did not stain the mitral somata of the MOB. In sections performed at the level of the NVN, where the MGC is elongated and the neuronal somata are dispersed, the pattern of anti-CR immunopositivity was maintained (Figure 10C), although not all atypical glomeruli were CR-positive at this level.

The immunolabeling pattern obtained with anti-CB immunostaining was similar to that observed with anti-CR immunostaining. Somata belonging to the MGC were intensely stained, and atypical glomeruli were variably labeled, differentiating CB-positive and CB-negative subpopulations of atypical glomeruli (Figures 10D–F).

Anti-SG immunostaining produces a different immunolabeling pattern from anti-CR and anti-CB immunostaining. Anti-SG immunostaining labeled the entire glomeruli population in the OL and MOB, and periglomerular cells were intensely labeled (Figure 11A). Both the neuropil and the neurons scattered in the MGC were anti-SG immunopositive, although with varying degrees of labeling (Figure 11B).

As a summary of the results obtained, Table 2 shows an overview of the main immunohistochemical and structural features of the olfactory limbus in comparison to those of the glomeruli in both the main and accessory olfactory bulb.

Prior OB investigations have primarily focused on the MOS and the VNS. Although we have made substantial advancements in our understanding of other subsystems, particularly the necklace region, which consists of a ring of interconnected glomeruli encircling the caudal end of the MOB and the anterior AOB, most research has been performed in rats and mice, and the presence and organization of olfactory subsystems in other mammalian groups remain very poorly studied. However, neuroanatomical studies of the MOB and AOB comparing laboratory rodents with other mammals, such as lagomorphs (Villamayor et al., 2020), bats (Frahm and Bhatnagar, 1980), canids (Choi et al., 2010; Ortiz-Leal et al., 2022b), artiodactyls (Park et al., 2014; Kondoh et al., 2017a), and primates (Alonso et al., 1998) have identified substantial differences in their organization between orders. Therefore, undescribed integrated olfactory subsystems may exist in other species, characterized by noncanonical morphological patterns, and these should be thoroughly investigated to comprehend interspecies differences in olfactory physiology from a solid and reliable morphological basis. The present study presents neuroanatomical, immunohistochemical, and lectin-histochemical evidence supporting the presence of an OL region characterized by a complex anatomical and morphofunctional organization that can consistently be detected in the fox, a wild Canidae model species. This finding supports the need for additional research exploring the organization of potential olfactory subsystems such as the OL in other mammalian species, including humans, as these have been poorly defined in most mammals thus far.

The original contributions presented in the study are included in the article, further inquiries can be directed to the corresponding author.

Ethical review and approval was not required for the animal study because all the animals employed in this study died by natural causes. Accordingly, the Bioethics Committee of the University of Santiago de Compostela has determined that ethics approval was not required for this study. Their report has been added to the submission.

PS-Q, AL-B, and LF collected the tissues. IO-L, MT, and PS-Q processed the tissue. IO-L, MT, PS-Q, AL-B, LF, VV-B, and JL-S analyzed and discussed the results. IO-L, MT, PS-Q, VV-B, and JL-S wrote the manuscript. All authors contributed to the article and approved the submitted version.