The follicle‐sinus complex of the bottlenose dolphin (Tursiops truncatus). Functional anatomy and possible evolutional significance of its somato‐sensory innervation

Abstract

Vibrissae are tactile hairs found mainly on the rostrum of most mammals. The follicle, which is surrounded by a large venous sinus, is called "follicle‐sinus complex" (FSC). This complex is highly innervated by somatosensitive fibers and reached by visceromotor fibers that innervate the surrounding vessels. The surrounding striated muscles receive somatomotor fibers from the facial nerve. The bottlenose dolphin (Tursiops truncatus), a frequently described member of the delphinid family, possesses this organ only in the postnatal period. However, information on the function of the vibrissal complex in this latter species is scarce. Recently, psychophysical experiments on the river‐living Guiana dolphin (Sotalia guianensis) revealed that the FSC could work as an electroreceptor in murky waters. In the present study, we analyzed the morphology and innervation of the FSC of newborn (n = 8) and adult (n = 3) bottlenose dolphins. We used Masson's trichrome stain and antibodies against neurofilament 200 kDa (NF 200), protein gene product (PGP 9.5), substance P (SP), calcitonin gene‐related peptide, and tyrosine hydroxylase (TH) to characterize the FSC of the two age classes. Masson's trichrome staining revealed a structure almost identical to that of terrestrial mammals except for the fact that the FSC was occupied only by a venous sinus and that the vibrissal shaft lied within the follicle. Immunostaining for PGP 9.5 and NF 200 showed somatosensory fibers finishing high along the follicle with Merkel nerve endings and free nerve endings. We also found SP‐positive fibers mostly in the surrounding blood vessels and TH both in the vessels and in the mesenchymal sheath. The FSC of the bottlenose dolphin, therefore, possesses a rich somatomotor innervation and a set of peptidergic visceromotor fibers. This anatomical disposition suggests a mechanoreceptor function in the newborns, possibly finalized to search for the opening of the mother's nipples. In the adult, however, this structure could change into a proprioceptive function in which the vibrissal shaft could provide information on the degree of rotation of the head. In the absence of psychophysical experiments in this species, the hypothesis of electroreception cannot be rejected.

We describe the vibrissae of the newborn and adult bottlenose dolphin (Tursiops truncatus) using immunocytochemistry and histology. We found that even in adulthood, the follicle complex contains a vibrissal hair, albeit not protruding, and that the structure is likely mechanoreceptive in function.

1. INTRODUCTION

Vibrissae, also called whiskers, are modified tactile hairs that occur in most mammals except monotremes, anteaters, rhinoceroses, and humans (Cave, 1969; Chernova, 2006; Muchlinski, 2010; Van Horn, 1970). They are mainly located around the muzzle but can also be present in other parts of the head and under the carpus, depending on the species (Fundin et al., 1995; Sarko et al., 2011). Their main function is to convey mechanical (tactile) stimuli to the somatosensory cortex (Woolsey & Van der Loos, 1970). The hair follicle of each vibrissa is surrounded by a large venous sinus, together forming the "follicle‐sinus complex" (FSC) (Rice et al., 1986). The presence of vibrissae in Pinnipeds, Odobenids, Sirenids, and otters suggests that their somatosensory function is also functional in the water. However, vibrissae are present only in newborn cetaceans and generally disappear in adults. Therefore, the question arises whether the vibrissae of very young cetaceans perform a temporary function that is lost within a few weeks after birth, or whether they are just a remnant of a structure that evolution discarded in these mammals.

The morphology and innervation of the vibrissae have been studied extensively in rodents and cats (Ebara et al., 2002; Park et al., 2003; Rice, 1993; Rice et al., 1986), and thus our present knowledge of the structure and function of the FSC mostly derives from these species, although efforts have been developed in marsupials (Hollis & Lyne, 1974; Lyne, 1958; Marotte et al., 1992). In general, the FSC of terrestrial mammals consists of epidermal and dermal components. The epidermal parts include the hair bulb, the vibrissal shaft (VS), the inner and outer root sheaths, surrounded by a glassy membrane. The latter separates these components from the dermal parts, that is the mesenchymal sheath (MS) and the venous sinus. The sinus is horizontally divided into a proximal ring sinus (containing the ringwulst and the inner conical body), and a distal cavernous sinus (that contains a large number of trabeculae, filled with venous blood). The last dermal part is the connective tissue capsule that limits the follicle and caps it above the inner conical body with the outer conical body. Finally, the rete ridge collar is a thickening of the epidermis where the VS protrudes (Ebara et al., 2002; Rice et al., 1986).

As mentioned above, marine mammals also develop vibrissae, and a description of their morphology and dimensions in seals and otter has been recently reported in comparison with several terrestrial species (Dougill et al., 2020). Walruses have the highest number of vibrissae (up to 350 on each side), while pinnipeds possess large and richly innervated FSCs, divided into three parts, with up to 1600 axons reaching it (Hyvärinen, 1989, 1995; Hyvärinen et al., 2010; Ling, 1966, 1977; Marshall et al., 2006). In manatees, extensive studies have described the vibrissae, which are spread out on the muzzle and the body (Reep et al., 1998, 2001; Sarko et al., 2007). Mysticetes have vibrissae in large quantity caudally to the blowhole and on the rostro‐lateral sides of the upper and lower jaws with numbers up to 250 in the bowhead whale (Balaena mysticetus) (Slijper, 1962; Yablokov & Klevezal, 1964). On the contrary, most adult toothed whales have no facial hair and show 2–10 bilateral rows of vibrissae only during fetal life and the early postnatal period (Ling, 1977; Reidenberg & Laitman, 2009; Yablokov et al., 1972). Toothed whales show fully developed vibrissae only in the early phases of their postnatal life (Cozzi et al., 2017; Czech‐Damal et al., 2013; Dehnhardt & Hanke, 2017). From morphological comparisons among odontocetes, a classification divided them into four groups based on the development of the FSC (Yablokov et al., 1972). Following this classification, the bottlenose dolphin falls into a group comprising species in which the VS is still present in the early postnatal period but disappears in the majority of adult individuals. This is not the case in river dolphins such as the Guiana dolphin (Sotalia guianensis), of which a recent study described the FSC (Czech‐Damal et al., 2012). The FSC of this species was renamed vibrissal crypt because of its different anatomical structure, characterized by the absence of the VS, hair papilla, clear root sheaths, blood sinus, and capsule (Czech‐Damal et al., 2012). The FSC lumen is filled with desquamated corneocytes and keratinous fibers, that together may be considered a highly electrically conductive biogel (Czech‐Damal et al., 2012), part of an electrosensory system that facilitates the hunt of small bottom‐living prey in turbid water, where echolocation is not possible or potentially not efficient enough, by detection of their electric field (Czech‐Damal et al., 2012).

The somatosensory innervation of mystacial vibrissae is provided by three subdivisions of the maxillary branch of the trigeminal nerve. The deep vibrissal nerve originates directly from the infraorbital nerve, supplies a single FSC, penetrates the capsule, and arborizes dorsally at various levels. The superficial vibrissal nerves (SVNs) come from superficial cutaneous nerves and supply several FSCs. Small‐ to fine‐caliber nerve fiber branches reach the FSC from the base and supply the hair papilla and hair bulb (Ebara et al., 2002; Rice et al., 1986). The deep vibrissal nerve ends in mechanoreceptors such as Merkel nerve endings (MNEs), lanceolate endings, and free nerve endings (FNEs) along the follicle. The SVNs, instead, provide innervation to lanceolate endings at the level of inner conical body and MNEs at the level of the rete ridge collar (Ebara et al., 2002; Fundin et al., 1997a). The somatomotor innervation is provided by motoneurons placed in the lateral part of the facial nucleus and innervates the extrinsic (mimic) and intrinsic musculature of the mystacial pad (Haidarliu et al., 2010; Herfst & Brecht, 2008). The visceral innervation (sympathetic and parasympathetic) regulates blood flow in the FSC, supplied by the deep vibrissal artery, and consequently regulates blood pressure, which is essential for the activation of receptors that respond to specific stimulation thresholds. (Fundin et al., 1997b; Maklad et al., 2004).

Here, we describe the FSC in a series of postnatal and adult bottlenose dolphins, aiming at characterizing the changes in the anatomy and morphology of this structure at different life stages by histochemical and immunohistochemical techniques. Special attention was dedicated to the innervation of the FSC, the nature of the nerve fibers, and its functional potential.

2. MATERIAL AND METHODS

2.1. Animals

The samples of vibrissae from 11 bottlenose dolphins (Tursiops truncatus, Montagu 1821) were obtained from the Mediterranean Marine Mammal Tissue Bank (MMMTB, http://www.marinemammals.eu), housed in the Department of Comparative Biomedicine and Food Science (BCA) of the University of Padova. The MMMTB is a CITES recognized institution (IT 020) that collaborates with the Italian Ministry of the Environment. The MMMTB collects, processes, and stores samples of tissues of various cetacean species that have stranded along the Italian coastline since 2000. Additional samples derived from marine mammals that died at marine theme parks and aquaria, and whose bodies were delivered to BCA for diagnostic post‐mortem. More details of the specimens used in this study can be seen in Table 1.

TABLE 1.

Origin of specimens

ID	Species	Sex	Age class	Origin
# 83	Tursiops truncatus	M	Newborn	Died in a marine theme park
# 114	T. truncatus	M	Newborn	Died in a marine theme park
# 123	T. truncatus	F	Newborn	Died in a marine theme park
# 124	T. truncatus	M	Newborn	Died in a marine theme park
# 144	T. truncatus	M	Newborn	Died in a marine theme park
# 145	T. truncatus	M	Newborn	Died in a marine theme park
# 162	T. truncatus	M	Newborn	Wild
# 229	T. truncatus	M	Newborn	Died in a marine theme park
# 146	T. truncatus	M	Adult	Died in a marine theme park
# 159	T. truncatus	M	Adult	Died in a marine theme park
# 444	T. truncatus	M	Adult	Wild

2.2. Sample processing

Each sample was obtained by carving out around the VS on both sides of the rostrum in the newborn and around the dimple containing the orifice in the adult (Figure 1a,b).

FIGURE 1.

Macroscopic images of the rostrum of some specimens of (a, left four) newborn and (b, right two) adult bottlenose dolphin. The arrows indicate where the vibrissae emerge from the skin, as can be seen in the newborns (a) or the concavity found in the adults (b)

The samples were fixed by immersion in 4% neutral buffered paraformaldehyde and stored at 4°C. Tissues for Masson's trichrome were then included in paraffin and cut in 5 µm‐ and 10 µm‐thick sections either longitudinal or transversal to the main axis of the FSC by use of a rotatory microtome (Leica). Sections were mounted on gelatinized slides and air‐dried. Samples bound for immunocytochemistry were washed in standard phosphate buffer solution (PBS) overnight at 4°C, stored in PBS containing 0.1% Na‐azide and sucrose at 30%, immersed in OCT Compound (Tissue Tek, Sakura Finetek Europe, NL), and frozen at −80°C in isopentane cooled with liquid nitrogen. 25 µm‐thick sections of the longitudinal and transversal planes were subsequently taken with a cryostat (Leica).

2.3. Histological techniques

The morphology of the FSC was stained using a Masson's trichrome staining protocol. Briefly, the sections were immersed in three baths of xylene for 5 min each and subsequently hydrated with a descending series of graded alcohol solutions (100%, 95%, 90%, 80%, 70%, 50%). Then, they were stained with Mayer's Emallume for 5–10 min and rinsed with tap water. Later, the sections were colored for 5 min in a solution of distilled water (300 ml) containing Ponceau 2R (0.2 g), acid fuchsin (0.1 g), and acetic acid (0.6 ml). After rinsing with a 1% acetic acid solution, the sections were put in a solution of distilled water (100 ml), phosphomolybdic acid (3–5 g), and orange G (2 g) for 5 min, and rinsed again in an acetic acid solution. The sections were then colored for 5 min in light green (0.1–0.2 g in 100 ml distilled water) and acetic acid (0.2 ml). After the last rinsing in a 1% acetic acid solution, the slides were dehydrated directly with absolute alcohol, and passed in xylene (3 × 3 min) and coverslipped with Entellan (Merck).

The innervation of the FSC was characterized by immunocytochemistry, either via immunoperoxidase (IP) or immunofluorescence (IF), using the neuronal markers shown in Tables 2 and 3.

TABLE 2.

List of the primary antibodies used for immunoperoxidase (IP) or immunofluorescence (IF)

Primary antibody	Used for	Immunogen/host	Supplier	Dilution	Antibody RRID	Validation
Protein gene product 9.5	IP	Polyclonal rabbit	Millipore	1:500	AB_91019	PMID:19296476
	IF			1:1000
Substance P	IP	Polyclonal rabbit	Immunostar	1:1000	AB_572266	PMID:10087030 PMID:10196365
	IF	Monoclonal rat	Fitzgerald Industries International	1:200	AB_2313816	PMID:22740069 PMID:26713509
Calcitonin gene related peptide	IP/IF	Monoclonal mouse	Santa Cruz Biotechnology Inc.	1:200	AB_2259462	PMID:30971286 PMID:29943954
	IF	Polyclonal rabbit	Peninsula Laboratories Inc.	1:1000	AB_2313775	PMID:18186028 PMID:28680400
Human tyrosine hydroxylase	IP/IF	Monoclonal mouse	Monosan	1:50	ID: MONX10786 a	PMID:29615733
Neurofilament 200 kDa	IP/IF	Monoclonal rabbit	Sigma	1:1000	AB_477272	PMID:18022951 PMID:19937712

^aAntibody RRID are universally identified codes and were taken from the website the antibody registry (https://antibodyregistry.org/) which integrated the antibody database of the Journal of Comparatve Neurology. For each antibody, there is at least one publication correlated to a unique PMID (PubMed Identifier). For the antibodies whose lot number are MONX10786 and 401314, there are still no current RRID available but the validation appears in one publication (Bombardi et al., 2010).

TABLE 3.

List of the secondary antibodies used for immunoperoxidase (IP) or immunofluorescence (IF)

Secondary antibody	Used for	Immunogen/host	Supplier	Dilution	Antibody RRID	Validation
Biotinylated anti‐rabbit	IP	Goat	Vector Laboratories	10 μg/ml	AB_2313606	PMID:19127523 PMID:23766132
Anti‐mouse	IP	Goat	Vector Laboratories	10 μg/ml	AB_2336171	PMID:23766132 PMID:25057794
Anti‐mouse Alexa 594	IF	Goat	Thermo Fisher Scientific	1:200	AB_141372	PMID:23913443 PMID:25057190
Anti‐rat Alexa 594	IF	Donkey	Thermo Fisher Scientific	1:200	AB_2535795	PMID:25933105 PMID:28089909
Anti‐rabbit‐FITC	IF	Goat	Calbiochem	1:100	ID: 401314 a	PMID:29615733

^aAntibody RRID are universally identified codes and were taken from the website the antibody registry (https://antibodyregistry.org/) which integrated the antibody database of the Journal of Comparatve Neurology. For each antibody, there is at least one publication correlated to a unique PMID (PubMed Identifier). For the antibodies whose lot number are MONX10786 and 401314, there are still no current RRID available but the validation appears in one publication (Bombardi et al., 2010).

For IP staining, contiguous sections were initially immersed in a 0.4% solution of Triton X‐100 (Merck) in PBS at 4°C for 24 h. They were then rinsed in PBS baths for 3 × 10 min. Next, sections were treated with 1% H₂O₂ in PBS for 30 min. After three 10‐min washes in PBS, a 3% solution of normal goat serum (NGS, Sigma G‐9023) was applied for 2 h at room temperature. Thus, sections of each sample were incubated in a wet chamber for 48 h, at 4°C with the primary antibodies (Table 2) in antibody diluent (1.8% NaCl in a 0.01 M sodium phosphate solution containing 0.1% Na‐azide). After primary incubation, the sections were washed with PBS and incubated with the specific secondary antibodies (Table 3) diluted in PBS in a wet chamber for 2 h at room temperature. After further three washes in PBS, they were transferred for 30 min in an avidin‐biotin complex solution (ABC Standard, ABC kit Vectastain, Vector Laboratories, PK 6100) and washed again in PBS. Finally, IP was developed using 3.3′‐diaminobenzidine (DAB kit Vector Laboratories, BA‐9200). The sections were dehydrated in ethanol, passed in xylene, and covered with a coverslip using Entellan.

The slides of both Masson's trichrome and IP were observed with an optic microscope (Zeiss Axioplan, Carl Zeiss), captured with the microscope Nikon Coolscope (Nikon) and subsequently elaborated with the programs Elipsenet 1.20.0 (Nikon) and GIMP 2 (GNU Image Manipulation Program 2.10).

For the IF procedure, slides were placed in a wet chamber. A first PBS wash was performed to rehydrate the sections. A Blocking Serum solution (0.5% Triton X‐100, 10% NGS, Vector, in PBS) or 10% Normal Donkey Serum (NDS; Jackson) was used at room temperature for 2 h. Then, the sections of each sample were incubated in a wet chamber for 48 h, at 4°C, with the primary antibodies (Table 2) in antibody diluent. After 48 h, the sections were washed with PBS and either pure NGS or 10% NDS, (5 × 10 min on a stirrer). Next, the sections were incubated for 3 h, at room temperature, with specific secondary antibodies (Table 3), diluted in PBS. After further five 10‐min washes in PBS, the slides were air‐dried and prepared with glycerol buffered with 0.5 M sodium carbonate (pH 8.6) to be finally sealed with nail polish.

The slides obtained were observed under an epifluorescence optical microscope (Axioplan, Carl Zeiss), equipped with a system of filters that allowed the distinction of the fluorescence FITC (given by fluorescein) from Alexa 594 fluorescence. The images were acquired using a digital camera and DMC 2 software (Polaroid Corporation). The images were processed using Adobe Photoshop (Adobe Systems).

3. RESULTS

3.1. Morphology

3.1.1. Newborn dolphins

In the newborns, the external part of the VS was approx. 10 mm long. The FSC of all specimens consisted of an epidermal and dermal part. The epidermal part comprised the hair with its sheaths, overlying the dermal venous sinus. The VS originated from the bulb and consisted of three concentric layers, which, from inside to outside, were identified as the medulla, the cortex, both made of keratinized cells, and the cuticle, which consisted of a simple squamous keratinized epithelium. The MS and the capsule were fused near the follicle outlet. At the base of the FSC, the bulb resembled a highly innervated and vascularized dermal papilla (Figure 2a). The VS was wrapped by the inner root sheath, attached to the cuticle, and the outer root sheath, surrounded by the glassy membrane (Figure 2b). The hair shaft was surrounded by a venous sinus and delimited by a connective tissue capsule. The sinus comprised the MS internally, in contact with the glassy membrane and externally by a capsule (Figure 2c).

FIGURE 2.

Longitudinal section of a typical follicle‐sinus complex in newborn bottlenose dolphin. (a) The vibrissa is surrounded by a venous sinus (s). A cp envelops the complex. Several nerves (arrows) reach the root of the vibrissa. The arrowhead indicates the fusion between the cp and ms. (b) Detail at higher magnification of the epidermal components. (c) Detail at higher magnification of the dermal components. b, bulb; cp, capsule; gm, glassy membrane; irs, inner root sheath; ms, mesenchymal sheath; ors, outern root sheath; p, papilla; s, venous sinus; vs, vibrissal shaft. Masson's Trichrome stain. Scale bars: a = 1 mm; b, c = 100 μm

In the slides analyzed, it was never possible to observe either a ringwulst or a ring sinus. Furthermore, no muscle fiber or gland was present around the follicle.

3.1.2. Adult dolphins

In the adults, the VS was present but did not reach the skin surface. Apart from this feature, the FSC of the adult dolphins showed the same structure as those of the newborns.

3.2. Innervation

Anti‐PGP 9.5 immunoreactive (‐ir), anti‐neurofilament 200 kDa (NF 200)‐ir, anti‐tyrosine hydroxylase (TH)‐ir, and anti‐SP‐ir nerve fibers were evident in all the samples. No anti‐calcitonin gene‐related peptide (CGRP)‐ir fibers were observed.

3.2.1. Newborn dolphins

PGP 9.5‐ir fibers penetrated the FSC at the level of the hair bulb, and yielded an intricate arborized network of ramifications (Figure 3a,b). The nerve fibers protruded at various levels in the MS, giving rise to button‐like terminations characterizing MNEs (Figure 3c). They derived from large and medium‐sized fibers that ran to form clusters of button‐like endings with a smooth and regular surface, between which fine spiral‐like fibers were present (Figure 3d).

FIGURE 3.

Innervation of the follicle‐sinus complex (FSC) in a newborn bottlenose dolphin. The nerve fibers were immunolabeled for protein gene product 9.5 (PGP 9.5) (a–d) and neurofilament 200 kDa (e,f). (a) Several PGP 9.5‐ir nerve bundles (asterisks) reach the root of the vibrissa. (b) Few thin‐calibre fibers (arrowhead) enter the papilla (p) and terminate as free nerve endings (arrows). (c) In the mesenchymal sheath, some nerve fibers (arrow) give rise to Merkel nerve endings (MNEs) (asterisks). (d) High magnification showing MNEs. Note the characteristic button‐like endings. (e) The dense network of nerve fibers around the bulb. (f) A nerve bundle penetrate the FSC laterally. vs, vibrissal shaft. Scale bars: a,e = 100 μm; b,c,f = 200 μm; d = 50 μm

NF 200‐ir fibers were also detected penetrating the bulb (Figure 3e), first running parallel to the VS and then entering at different levels along the follicle (Figure 3f). Nerve fibers of different calibers were distributed along the VS, progressing either in a straight line or along a winding path until they reached the top of the FSC (Figure 3f).

Numerous nerve fibers were observed in transversal sections of the FSC, from the hair bulb to the apex (Figure 4). These fibers innervated the hair bulb (Figure 4b,c) and sent small groups of axons to surround the follicle (Figure 4d,e), ending into MNEs (Figure 4f). Some of these axons penetrated the venous sinus, ran along the trabeculae (Figure 4g), and ended at the MS that wrapped the VS with MNEs (Figure 4h). This rich innervation was evident in all transverse sections up to the outlet of the vibrissa (Figure 4i). We did not identify other receptors with certainty and, as mentioned above, most fibers seemed to end as FNEs.

FIGURE 4.

Longitudinal (a) and transverse (b–i) sections (b) of follicle‐sinus complex (FSC) in newborn bottlenose dolphin showing the innervation at different levels from the basal (b) to the apical (i). The nerve fibers were immunolabeled with antibodies to protein gene product 9.5. b. Some nerves (arrows) reach the root of the vibrissa. (c) The nerves break into several fascicles (arrows) that ascend close to the papilla (p). (d) The nerve fibers (arrows) surround the follicle. (e) Some nerve fibers (arrows) penetrate the venous sinus (s) and branch in the mesenchymal sheath (arrowhead). (f) Some fibers terminate on Merkel nerve endings (MNEs) (arrowhead), while others continue along the FSC (arrow). (g) A nerve fiber (arrow) passes through one of numerous trabeculae of the venous sinus. (h) At the level of dermo‐epidermal border, the nerve fibers disappear but the MNEs are still present (arrow). (i) Section through the skin and the dermal papilla. Scale bars: a = 1 mm; b–h = same magnification of i; i = 350 μm. vs, vibrissal shaft

SP‐ir fibers ran either grouped in bundles or alone close to blood vessels (Figure 5a,b). Double IF for PGP 9.5 and TH showed that PGP 9.5‐ir fibers were qualitatively four‐fold the TH‐ir fibers (Figure 5c,d). TH‐ir fibers were mainly located around the blood vessels, and sometimes presented a tortuous pattern. They contained thin‐caliber axons that ran first on the surface of the adventitia and then penetrated the wall (Figure 5e). Few TH‐ir fibers were found at the base of the bulb and in the MS (Figure 5f).

FIGURE 5.

Substance P‐ (a,b) and tyrosine hydroxylase (TH)‐ (c–f) immunoreactive fibers in newborn bottlenose dolphin. (a) Transverse section of a nerve bundle showing many immunoreactive fibers. (b) A nerve fiber (arrow) runs parallel to a blood vessel (asterisk). (c,d) Double immunofluorescence protein gene product 9.5‐FITC (c)/TH‐Alexa 594 (d) of a nerve bundle in transverse section. Note the TH immunoreactivity of some fibers. (e) Several nerve fibers (arrows) reach the tunica adventitia of a vessel (asterisk). (f) A thin fiber run within the mesenchymal sheath. Scale bars = 100 μm

3.2.2. Adult dolphins

Immunohistochemical results in adult dolphins showed the same general pattern of that of newborns, with some notable exceptions. Masson's trichrome revealed structures, likely lamellar corpuscles (LCs) (Figure 6a), located in the MS (Figure 6b) upper half, relatively close to the VS.

FIGURE 6.

Longitudinal section of the follicle‐sinus complex (FSC) in adult bottlenose dolphin. (a) Example of the location of the LCs along the FSC. (b) Higher magnification of the corpuscles in the mesenchymal sheath. LC, lamellar corpuscles. Scale bars: a = 200 µm; b = 100 µm

PGP 9.5‐ir and NF 200‐ir nerve fibers were clear (Figure 7a–c) and the MNEs bound to the MS were smaller in adults (Figure 7d). SP reactivity was found in large‐caliber fibers near the dermo‐epidermal junction (Figure 8a), where they ran parallel to the skin before bending toward the FSC and ending as FNEs (Figure 8b). TH‐ir fibers were rarer. Very thin TH‐ir fibers were present in the trabeculae of the venous sinus and the MS ending with isolated oval corpuscles (Figure 8c–e).

FIGURE 7.

Longitudinal sections of follicle‐sinus complex (FSC) in adult bottlenose dophin. The nerve fibers were immunolabeled for protein gene product 9.5 (a,b) and neurofilament 200 kDa (c,d). (a) Note the six nerve bundles (arrows) reaching the FSC. (b) A vibrissa is clearly visible inside the follicle. Some nerve fibers (arrows) reach the bulb, others (arrowheads) run within the mesenchymal sheath (MS). (c) High magnification showing the rich innervation (arrows) of the MS. (d) Detail of a merkell nerve ending (arrow). Scale bars: a, b = 1 mm; c = 200 μm; d = 100 μm. b, bulb; p, papilla; vs, vibrissal shaft

FIGURE 8.

Substance P‐ (a,b) and tyrosine hydroxylase (TH)‐ (c–e) immunoreactive fibers in an adult bottlenose dolphin. (a) Two positive fibers reach the follicle‐sinus complex laterally. (b) Few nerve fibers in the mesenchymal sheath (MS). (c,d) Double immunofluorescence protein gene product 9.5‐FITC (c)/TH‐Alexa 594 (d) of a large nerve bundle in longitudinal section. Note the few TH‐ir fibers. (e) Few fibers (arrow) in the MS. Scale bars: a,b,e = 200 μm; c,d = 100 μm

4. DISCUSSION

In the present work, we describe the morphology and the innervation of FSC of newborn and adult bottlenose dolphins. Our findings reveal that newborn specimens possess a complete structure divided in an epidermal and a dermal part, hence the term "FSC", whereas, in adults, the VS lies within the follicle (Figures 2 and 7). The absence of a ring sinus (ringwulst) of the erector pili muscle and of any kind of associated gland constitutes differences with the vibrissae in terrestrial mammals. The blood sinus of the bottlenose dolphin consists of just one cavity forming a trabecular net just like the tammar wallaby (Macropus eugenii) (Marotte et al., 1992) and is not divided into two parts as in terrestrial mammals (Ebara et al., 2002; Rice et al., 1986) or three parts as in pinnipeds (Marshall et al., 2006) (Figure 9).

FIGURE 9.

Schematic drawn representing the main differences between the follicle‐sinus complex (FSC) of a terrestrial mammal (left), pinniped (center) and dolphin (right). On the right, the follicle represented is that of the adult as the dotted lines and transparent areas (SVNs and MNEs in the dermo‐epidermal junction) are of the newborn. In terrestrial mammals the FSC is divided into two halves, the inferior cavernous sinus, and the superior ring sinus. The receptors are of various nature, are positioned at various heights depending on their receptor (sensory) nature and come mainly from the deep vibrissal nerve. In the pinniped it is instead divided into three portions, including an upper cavernous sinus. In this case, however, the fibers, which derive only from the deep vibrissal nerve, innervate up to the inner conical body, without reaching the epidermis. Finally, in the dolphin there is a trabecular component that forms a single venous sinus in which the receptors, deriving mainly from the deep vibrissal nerve, are distributed along the follicle until they reach the epidermis in the newborn. Also note the tyrosine hydroxylase‐ir (green) and substance P‐ir (light blue) fibers which accompany the blood vessels. a, artery; b, bulb; c, capsule; CEs, circular endings; cs, cavernous sinus; DVN, deep vibrissal nerve; e, epidermis; FNEs, free nerve endings; icb, inner conical body; irs, inner root sheath; LC, lamellar corpuscles; lcs, lower cavernous sinus; LEs, lanceolate endings; MNEs, Merkel nerve endings; ms, mesenchymal sheath; ocb, outern conical body; ors, outern root sheath; p, papilla; REs, reticular endings; rs, ring sinus; rw, ringwulst; SVNs, superficial vibrissal nerves; ucs, upper cavernous sinus; v, vein; vs, venous sinus; VS, vibrissal shaft (modified from Rice, 1993 and Sarko et al., 2007)

Anti‐neuronal antibodies (anti‐PGP 9.5, ‐NF 200, ‐TH, and ‐SP) helped to characterize the innervation of the FSC together with the Masson's trichrome (Figure 6). Antibodies directed against proteins of the neurofilaments (anti‐PGP 9.5, ‐NF 200, Figures 3, 4, and 7) identified several fiber bundles which bifurcated from the deep vibrissal nerve at the bottom of the FSC to reach various heights of the follicle, similarly to what was previously described in manatees (Sarko et al., 2007). Few fibers were also found at the dermo‐epidermal junction (Figure 4i). MNEs and LCs were evenly distributed along the MS of the FSC, although, in the newborn MNEs were also found at the dermo‐epidermal junction (Figures 3 and 4). Since no striated muscle fibers and no glands were present in the FSC, the nature of the present innervation is likely somatosensory and derived from the maxillary nerve (V2) of the trigeminal nerve. The V2 runs—from its exit of the skull base—rostrally to enter the dolphin equivalent of the maxillary foramen and subsequent canal. The rostralmost fibers of the V2 run as the infraorbital nerve in maxillary canals. Some fibers of the latter penetrate the maxilla in dorsal direction to provide sensorial innervation to the skin of the rostrum (Rauschmann, 1992). Dolphins possess no movable lips and have virtually no snout fascia or mimic muscles beyond those, more caudal, that act on the melon, thus implying a virtual absence of a somato‐motor component in the facial nerve this far forward on the face. Comparisons with other mammals are difficult. Whisking rodents possess the mystacial pad, a thickening of the snout fascia where mimic muscles, sensory receptors, and collagen structures form a highly developed motor‐sensory organ (Haidarliu et al., 2020). In the rat and mouse, a column of neocortical neurons in the whisker somatosensory cortex (wS1, or barrel cortex) corresponds to each FSC, with a highly developed layer IV receiving the thalamic afferent (Bosman et al., 2011; Jeanmonod et al., 1981; Pearson et al., 2006; Rice & Van Der Loos, 1977; Schröder et al., 2020; Van der Loos, 1976; Van der Loos & Woolsey, 1973; Woolsey & Van der Loos, 1970). The barrel cortex is usually associated to most rodent species (although it is notably absent in the beaver), despite having been detected in selective other taxa such as in the ferret (order Carnivora), and suspected in the rabbit (order Lagomorph) (Fox & Woolsey, 2008). It has, however, not been found in other species (i.e. mammals of the genus Felis or Panthera), even ones which present whisking behavior such as the short‐tailed opossum (Ramamurthy & Krubitzer, 2016; Waite et al., 1991). Furthermore, the neocortex of dolphins and whales lacks a layer IV and it is currently hypothesized that thalamic projections reach layer II instead of IV (for a general description, see Cozzi et al., 2017), thus making any comparison with the highly specialized barrel cortex of rodents difficult. In particular, the study by van Kann et al. (2017) pointed out the main differences in the primary neocortical areas layering between the common dolphin, the wild boar, and humans.

The use of antibodies against peptides (SP, CGRP) and against a key enzyme in catecholamine synthesis (TH) allowed further characterization of the innervation. SP is involved in nociception, and a subpopulation of sensory neurons in the mammalian trigeminal ganglion contains SP, colocalized with CGRP (Alvarez et al., 1988; Fundin et al., 1997b; Waite & Ashwell, 2012). SP‐ir neurons have also been described in the dorsal root ganglia of bottlenose dolphins (Bombardi et al., 2010).

Several SP‐ir were located in the dermo‐epidermal junction and, deeper lateral to the follicle. Other SP‐ir nerves were also found around the sinus’ blood vessels. Both SP and CGRP are vasodilators. However, the presence of SP‐ir fibers within big nerve bundles suggests a nociceptive function, while the presence of SP in thin fibers around blood vessels might indicate a parasympathetic activity on the vessels of the FSC (Fundin et al., 1997b). No CGRP‐ir fiber was detected in the vibrissae of the dolphins in this study. This absence is difficult to explain from a functional point of view. Indeed, CGRP has been found in the gastrointestinal tract of the striped dolphin (Stenella coeruleoalba) (Domeneghini et al., 1997), and has been proposed to be present in the CNS of the bottlenose dolphin (Rambaldi et al., 2016). Additionally, the presence of this molecule was demonstrated in the FSC of manatees (Sarko et al., 2007). Therefore, the apparent absence of CGRP in the FSC of the bottlenose dolphin could be due to a loss of nociceptive function or different trigeminal organization.

TH‐ir fibers consisted of thin branches located around the surface of the blood vessels connecting to the sinus, even though some elements were also found in the MS and trabeculae. Considering their peripheral location, TH‐ir fibers are indicative of a noradrenergic sympathetic innervation derived from the cranial [superior] cervical ganglion. In mammals, sympathetic fibers follow both the external and internal carotid arteries. From the former, fibers then run along the infraorbital artery. However, the internal carotid artery obliterates early in postnatal life in dolphins (Boenninghaus, 1903; Cozzi et al., 2017), as in many Cetartiodactyls. Therefore, the precise route of the TH‐ir that was observed in the FSC of dolphins remains to be ascertained. It may be possible that the internal carotid artery develops in the odontocete embryo to guide the growing sympathetic fibers from the cervical ganglion since it is known that the axonal outgrowth of mammalian sympathetic precursors proceeds in dependence and thus in parallel to the development of the internal carotid artery (reviewed in Kameda, 2020). Thus, possibly only after the sympathetic fibers have found their route, the internal carotid artery obliterates in odontocetes. Alternatively, these fibers in cetaceans may rely exclusively on the external carotid path.

Contrarily with what was reported by Yablokov et al. (1972), we were able to confirm the presence of the VS in adult specimens, as was previously described by Palmer and Weddell (1964). In all the adult animals analyzed in the present study, the FSC was complete, i.e. the epidermal and dermal components were discernable, and the VS was still present, albeit only in the follicle. In fact, the hair papillae maintained the same morphology found in the newborns. Moreover, the VS did not have the aspect of a formless cluster due to the epithelial regeneration. Since the papilla is responsible for the production of the VS's components, a lack of this structure can justify the absence of the VS in the adult Guiana dolphin, in which it had transformed into an agglomeration of fat cells (Czech‐Damal et al., 2012). Interestingly, we found MNEs, FNEs and putative LCs but no other receptor, which is relatively coherent with what was described in the Guiana dolphin (Czech‐Damal et al., 2012). Nevertheless, LCs have been described in the neonate bottlenose dolphin, and in the harbor porpoise (Phocoena phocoena) throughout life stages (Czech, 2007). The present study would be the first to mention LCs in the vibrissae of adult bottlenose dolphins. The few we observed were present in the upper half of the MS, but further studies could confirm this finding.

The Guiana dolphin is in fact very similar to the ubiquitous bottlenose dolphin in general body morphology and proportions, although somewhat smaller. Yet, the preferred habitat of the Guiana dolphin is estuarine and coastal, while bottlenose dolphins are distributed worldwide and live both along coastlines and in the high seas. The FSC complex in Guiana dolphins has been linked to the sensitivity to electric fields generated by prey burrowed in sand or hidden by murky waters (Czech‐Damal et al., 2012, 2013). A progressive morphological adaptation to different specific environments over time may be the explanation to the differences in the structure of the FSC in those two similarly sized species. Whether the evolution was toward a loss or a gain of function remains unclear.

Based on the absence in dolphins of morphological features typical of terrestrial mammals with vibrissae, i.e. the absence of a wide range of receptors, the mystacial pad or the barrel cortex, we can infer that the FSC of dolphins are relatively rigid structures, with a sensitivity potentially reduced which would not allow the perception of dynamic changes the way fully formed whiskers do. Yet, since the vibrissae persist in newborn dolphins, contrarily with other structures lost during fetal growth (pelvic limbs, to name the most striking difference with land mammals), it is possible that the vibrissae play a role in the early postnatal life, in the days immediately following parturition.

Pinniped mothers recognize their newborn mostly through olfaction while the newborn, from the first minutes of life, vocalizes for most of the first day, a behavior that diminishes gradually afterwards (Trillmich, 1981). Underwater, this kind of recognition is almost impossible for dolphins since they cannot smell (Cozzi et al., 2017). As proposed by Cozzi et al. (2012, 2015), an evolutionary adaptation in cetaceans to immediate recognition is the early ossification of the tympanic bulla, which would allow the newborn to locate the mother's vocalizations and help it in the postpartum period. However, full echolocation capacities are likely not to fully develop until month 1–3 after birth (Harder et al., 2016); and while it is difficult to think that vibrissae may act for this recognition, it is more plausible that it could allow the newborn to find the opening of the nipples immediately after birth, thanks to the rich SP innervation and vasomodulation regulated by the TH‐ir plexus (Fundin et al., 1997b; Maklad et al., 2004).

A VS is still present within the FSC in the adult dolphin and the innervation remains relatively developed. The original tactile function of the FSC may lose value in the growing dolphins that, later in life, may rely on different sensory modalities to interact with their mother and the rest of the pod. This organ could act as a proprioceptor as suggested by Yablokov et al. (1972). The VS could be sensitive to low‐frequency oscillations and water movements caused by head rotation and would consequently activate the receptors that provide this information to the central nervous system. This would allow the dolphin to always have a perception of the angular position of the head. Nervous fibers could modify the pressure inside the venous system and help maintain thermoregulation and modify the threshold of the receptors (Fundin et al., 1997b). This idea does not preclude that the structure found by Czech‐Damal et al. (2012) in the Guiana dolphin could function as an electroreceptor. However, to confirm or reject this hypothesis, further psychophysical experiments should be conducted in the bottlenose dolphin.

AUTHOR CONTRIBUTIONS

TG, AG, JMG, and BC designed the study, TG, AG, CT, and MDS acquired and analyzed the data. TG, AG, JMG, and BC wrote the draft. AG, SDV, and SH critically revised the manuscript. All authors approved the article.

Gerussi T, Gra\xEFc J-M et al The follicle‐sinus complex of the bottlenose dolphin (Tursiops truncatus). Functional anatomy and possible evolutional significance of its somato‐sensory innervation J. Anat.2021;238:942–955. 10.1111/joa.13345

DATA AVAILABILITY STATEMENT

Data are available on request from the authors.