Brains were obtained from five specimens of three different species of the suborder Odontoceti, belonging to the family Delphinidae and stranded in the Canary Islands between 2005 and 2011 (Sacchini, 2015): three striped dolphins (Stenella coeruleoalba, Meyen 1833), one common dolphin (Delphinus delphis, Linnaeus 1758), and one Atlantic spotted dolphin (Stenella frontalis, Cuvier 1829). While cross-sections were used to study the histology of the AMY, cross, sagittal, and horizontal sections of whole brains were also made to provide a more accurate assessment of the AMY’s topography within the brain. For this reason, three Atlantic spotted dolphins and one Blainville’s beaked whale (Mesoplodon densirostris, De Blainville 1817, family Ziphiidae) were added to the research, and their brains were examined using sagittal and horizontal sections solely for macroscopical purposes, revealing the corresponding AMY. The AMY macroscopical localization and relationship to surrounding structures were all revealed by each cutting plane. A comprehensive protocol for sample collection, histochemical and immunohistochemical procedures, and brain handling is available from a previous published work (Sacchini et al., 2022). Briefly, at the time of necropsy, brains were extracted and coded for freshness (Decomposition Code ≤2, fresh). Both the cerebral and cerebellar hemispheres underwent some longitudinal incisions prior to immersion. Most of the cuts were superficial, but at least one of them went deep into each cerebral hemisphere to reveal the lateral ventricles and let the fixative enter the ventricular system. Brains were submerged in 4% formaldehyde in phosphate-buffered saline (PBS; pH 7.4), during at least 72 h. Subsequently, one or both entire AMY from each animal (Table 1) were isolated, postfixed during 24 h, and then cryoprotected by immersion in a 30% sucrose and 0.1% sodium azide solution in PBS (pH 7.4), at +4°C, to prevent freezing artifacts. Samples were cut rostrocaudally, using a freezing sliding microtome (Leica SM 2000 R, University of Bologna). Sections were stored in PBS containing sodium azide (0.01%). The high quality of the tissues was confirmed by thionine staining. Every fourth serial section was collected for Nissl staining (thionine) and the adjacent sections for calbindin-D28k (c-D28k) free-floating immunohistochemistry. Of all the sections obtained, 13 were chosen as main references and as equidistant as possible, in order to obtain common references to each AMY.

Immunoperoxidase staining procedure was carried out on free-floating 50-μm–thick coronal sections under moderate shaking conditions. Formalin fixed free-floating sections were treated with 3% H₂O₂ in PBS for 30 min at room temperature (RT), to quench the endogenous peroxidase activity, and rinsed in PBS (three times, 10 min each). To block non-specific binding, sections were incubated in a solution containing 10% normal horse serum (Vector, S-2000), containing 0.5% Triton X-100 (Merck, Darmstadt) to permeabilize the tissue, in PBS for 2 h at RT. Thereafter, a first set of sections was incubated in a mouse monoclonal anti c-D28k (Swant, Code N° 300), diluted 1:500, overnight at +4°C. The following day, the sections were rinsed in PBS (three times, 10 min each) and incubated for 45 min at RT with a secondary antibody biotinylated horse anti-mouse (Vector Laboratories, BA-2000, diluted 1:200) in a solution containing 1% normal horse serum in PBS. The sections were rinsed in PBS and incubated with an avidin-biotin complex, (ABC, Vector Laboratories; PK-4000) for 1 h at RT. Finally, peroxidase activity was revealed with a 3,3′-diaminobenzidine peroxidase kit (Vector Laboratories, SK-4100). The sections were mounted onto gelatine-coated slides and dried overnight at RT. Slides were then dehydrated in ethanol, cleared in xylene and coverslipped with Entellan (Merck, Darmstadt, Germany).

Calbindin-D28k is a calcium-binding protein of the EF-hand (helix-loop-helix motif) family related to calmodulin and troponin-C. Through their roles as Ca²⁺ transporters across cell membranes, Ca²⁺-modulated sensors, or decoding Ca²⁺ signals, calcium-binding proteins, and thus c-D28k, contribute to the regulation of Ca²⁺ concentration in the cytosol and take part in a variety of biological functions (Yáñez et al., 2012). Monoclonal c-D28k is a mouse IgG1 generated through hybridization between mouse myeloma cells and mice’s spleen cells that were immunized with purified c-D28k from chicken. The amino acid sequence of the c-D28k was compared with the mouse. The Ensembl genome browser¹ shares comparative genomics and sequence variation. The sequence of c-D28k is shared with the mouse (Mus musculus) for over 96% in the following cetaceans: bottlenose dolphin, beluga whale (Delphinapterus leucas; Pallas 1776; family Monodontidae), sperm whale (Physeter macrocephalus; Linnaeus 1758; family Physeteridae), and even in the blue whale (Balaenoptera musculus; Linnaeus, 1758; family Balaenopteridae). Specifically, the protein alignment between the mouse and the bottlenose dolphin is: 261 amino acids length, 96% protein identity, and 100% coverage. To evaluate the specificity of the secondary antibody, negative controls were conducted using only PBS and the secondary antibody instead of the primary antibody.

In order to conduct the neuronal morphometric study, neurons were photographed at 40× magnification, using the Olympus XC50 digital camera. The measures were conducted using a digital image software (CellSens Standard, Olympus). In the thionine-stained sections, neurons that exhibited a distinct border, without any blurring, and a discernible nucleolus were identified for measurement. Neural populations were classified using indicators such as dendritic count, neuron size and shape and grouped into four groups: polygonal, spherical, fusiform, and large polygonal. A grade of affinity to thionine (weak, moderate, or strong) was also assigned to each cell, and the perikaryal area and perimeter were measured. The data were presented as mean ± standard deviation (SD).

By identifying neighboring structures, all types of sections—transverse, sagittal, and horizontal—provided information on the macroscopic morphology of the AMY and made it possible to establish the rostro-caudal, dorso-ventral, and medio-lateral limits. The AMY appeared as a well-defined, almond-shaped structure of a considerable relative size if compared to the total brain volume (Figures 1, 2). Its approximate size was measured only in the Atlantic spotted dolphin and had a mean of 1.3 cm (length) for 2 cm (wide).

In serial cross sections of the brain, just at the level of the hypothalamus (Figure 2A), the AMY was lateral to the diencephalon (thalamus and hypothalamus), separated from them by a thick internal capsule. Medially and ventrally, the optic tract was evident. The neocortex of the temporal lobe and the sylvian fissure were observed laterally to the AMY. A very thin external capsule separated the AMY from a thin, sheet-like neuronal structure, the claustrum, observed only rostrally and just at the beginning of the AMY. On the other side, a thin extreme capsule separated the claustrum and the AMY from the neocortex of the temporal lobe (Figures 2A, A’). Dorsally, the fundus striati (FStr), the most ventral part of the striatum, was present and this was quite evident through parasagittal sections (Figures 2B, B’). The AMY extended from the caudal limit of the claustrum to the rostral limit of the hippocampal formation and the lateral ventricle. Therefore, the hippocampal formation and the onset of the lateral ventricle were observed, marking the caudal limit of the AMY (Figures 2B, B’, C, C’). Figure 2B’ will be used to suggestively sequence the development of the CeA.

Microscopically, the AMY appeared as a heterogeneous grouping of brain nuclei, some of which were located rostrally, while others caudally. Twelve nuclei divided into three main groups were observed. The basolateral complex (deep nuclei or basolateral amygdala) was the largest and composed by:

Pyramidal and non-pyramidal neurons in all deep nuclei presented a disorganized distribution.

The cortical or superficial area were made up of:

With the exception of the medial nucleus, the pyramidal and non-pyramidal neurons of the superficial nuclei presented a tri-stratified arrangement.

The remaining areas were composed by:

The neurons of the remaining areas were similar to those of the striatum or FStr, without organization in layers, except for the AHA in which a stratified distribution was observed, with the presence of pyramidal and non-pyramidal neurons.

In the examined dolphins, neither the nucleus of the lateral olfactory tract (NLOT) nor the bed nucleus of the accessory olfactory tract (BAOT) were identified.

The examined dolphins possessed a well-developed AMY that comprised the basolateral and superficial nuclear groups as well as the remaining amygdaloid nuclei, as seen in other mammals. Contrarily to Patzke et al. (2015), and in common with what described by other authors in different species of cetaceans (Jansen, 1953; Schwerdtfeger et al., 1984; Buhl and Oelschläger, 1988; Marino et al., 2003, 2004a; Marino, 2004; Tartarelli and Bisconti, 2006; Oelschläger and Oelschläger, 2009; Rambaldi et al., 2017), the AMY had a considerable size, compared to the total brain volume. The inferior horn, located in the temporal lobe of the hemisphere, is reached by the caudate nucleus, which travels along the lateral ventricle from the anterior horn. Here, the AMY and the tip of the sharp caudate nucleus can occasionally be continuous (fundus striati) (Cozzi et al., 2017). It is remarkable how large the AMY is in dolphins considering the overall diminished state of their olfactory system. In contrast to NLOT, which is totally dependent on olfactory input, the remaining AMY is highly developed in anosmatic (odontocetes) and microsmatic (mysticetes and human) species. This implies that the AMY receives information inputs of different kind, such as auditory, in addition to olfactory stimuli. Compared to primates, dolphins appear to have an even larger AMY overall (Schwerdtfeger et al., 1984). The nuclear group with the largest dimensions was the basolateral nuclear group, which was made up of the paralaminar, lateral, basal, and accessory basal nuclei. In the examined dolphins, it was possible to identify the same parcellation of the amygdala present in other mammalian species (McDonald, 2003; Sah et al., 2003; Amunts et al., 2005; Morgan and Amaral, 2014), according to the nomenclature introduced by Price et al. (1987). Previous studies have shown that although the amygdaloid complex of different mammal species shares common characteristics, there is a clear difference in the organization, parcellation, and in the total and relative size of each amygdaloid body. Hence, this work represents the first classification of the AMY in the species striped dolphin, common dolphin, and Atlantic spotted dolphin, according to the currently accepted classification. Previous classification differs significantly from the one we have today. In addition, a discrepancy has been detected in the divisions commonly used in literature and the one proposed in the Nomina Anatomica Veterinaria (International Committee on Veterinary Gross Anatomical Nomenclature) and the International (human) Neuroanatomical Terminology (Table 3). Twelve nuclei have been found, arranged into three groups: the deep nuclei, the superficial area, and the remaining areas.

The majority of research on the CeA focuses on its interactions with other brain areas (Price and Amaral, 1981; Fung et al., 2011). These studies have primarily focused on rats and non-human primates in addition to humans. The CeA receives sensory information from the basolateral complex and projects it to the brainstem and hypothalamus, where it mediates fear behavior with autonomic and behavioral response signals (Sah et al., 2003). Its function is not limited to expressing fear; it also plays a role in learning fear and creating memories (Wilensky et al., 2006). It was also noted in this work that the presence of CeA was consistently associated with the NBmc, in accordance with the description in the bottlenose dolphin (Morgane and Jacobs, 1972). There are some disparities, though, regarding the CeA’s divisions. In fact, different authors have given different descriptions to different CeA subdivisions within a single species. There have been reports of two (Davis and Whalen, 2001), three (Sah et al., 2003), and up to four (Pitkänen and Kemppainen, 2002) subdivisions in rats (Rattus norvegicus). The CeA of primates and humans, on the other hand, clearly shows two divisions (Pitkänen and Kemppainen, 2002). Two distinct subdivisions of the CeA were identified in our dolphins: one was lateral and dorsal to the NL, and the other medial and dorsal to the NBmc. The Atlantic spotted dolphin, however, was devoid of the medial subdivision (this finding will be later examined in more detail). Neither the bottlenose dolphin nor the porpoise have been reported to show these two subdivisions (Morgane and Jacobs, 1972). In terms of neuronal distribution and morphotypes, the two subdivisions were identical. Depending on the animal species, the medial and lateral parts of the CeA have different neuropeptides and projections to other amygdaloid nuclei and brain structures (Pitkanen and Kemppainen, 2002). Therefore, further investigation is needed to fully understand the roles played by each of the two CeA divisions in dolphins. We have acquired quantitative parameters of the CeA of an adult male common dolphin, which served as a rough estimate of its size. Nevertheless, measuring and comparing areas and perimeters across the various species under study was not the goal of our research. It is believed that within a species, CeA is the same size in both sexes. In fact, previous studies suggest that the medial nucleus (Cooke and Woolley, 2005) and the NLOT are the only amygdaloid nuclei that exhibit sexual dimorphism (Guillamon and Segovia, 1997). Additionally, the size of the human CeA is used to predict the development of various mental disorders. A study on males found that individuals with disorders affecting emotional processing (psychopathy) had an increased size of CeA ranging from 10% to 30% (Boccardi et al., 2011). The size and presence of the CeA have been determined by the presence and size of the NBmc. In particular, the maximum size of the NBmc has implied the maximum development of the CeA, especially its medial division. The adult male Atlantic spotted dolphin under examination was the only one whose AMY did not exhibit the medial portion of the CeA, making this data impractical. The CeA was smaller than that of the other species even after the NBmc reached its largest size. We have linked these data to the reported behavior of this species, which is less elusive and more gregarious when interacting with humans, than other species, such as the striped dolphin (Herzing, 1997). The authors have also firsthand experienced with this conduct during whale-watching boat excursions. This kind of behavior would be less common in adult males and more common in juveniles and mothers with their calf (Herzing, 1997). To support this theory and investigate the relationship between the size of the AMY and CeA, as well as the extent to which different toothed whales display gregarious behaviors and possible stress responses resulting from human interaction, more research is required. Since studies have shown that the AMY in humans decreases with age, another theory would be to relate the animal’s age to the size of the CeA (Bickart et al., 2011). The CeA is distinct from other amygdaloid nuclei because it originates from the striatum (McDonald, 1982; Schiess et al., 1999). This was also evident in the CeA neurons, strikingly similar to those of the FStr as previously reported in the bottlenose dolphin and porpoise (Breathnach and Goldby, 1954; Morgane and Jacobs, 1972). Large polygonal neurons were dispersed throughout the CeA or through the white matter space between the CeA and the NL. Similarities between LP neurons and those in the adjacent substantia innominata were observed. This is probably due to the origin of some CeA neurons from the substantia innominata (McDonald, 2003).

Cetaceans in general, and dolphins in particular, are wild marine animals and neuroanatomical studies in these mammals are often very challenging. The postmortem period degrades the quality of the tissues, and stereological studies are difficult to carry out because some of the samples from various brain regions need to be designated for neuropathological examinations. In addition, fixation by immersion has been often deemed inadequate, particularly in large brains where the fixative is slow to penetrate. Inadequate or late fixation of subcortical structures led to low tissue quality, sections more prone to rupturing, poor c-D28k immunoreactivity and a very low affinity for thionine. Actually, because of the poor quality of their AMY, we had to exclude four toothed whales, namely two bottlenose dolphin, one short-finned pilot whale (Globicephala macrorhynchus, Gray 1846, family Delphinidae), and one Blainville’s beaked whale (M. densirostris, De Blainville 1817, family Ziphiidae) (Sacchini, 2015). In the study of the AMY, one of the main difficulties has been its deep location inside the brain, compared to cortical areas. Its position represents a limit for the penetration of the fixative. In some cases, immunoreactivity against c-D28k has not been homogeneous, with only small areas of positivity being detected. The AMY which provided the most comprehensive description were those with the best tissue quality, which resulted in improved thionine staining, and a strong and uniform immunoreaction to c-D28k. Further complicating the investigation is the small size of CeA’s neurons and their poor affinity to thionine. The morphometric study was then carried out on the animals with the best thionine staining: the common dolphin and the two striped dolphins. Furthermore, CeA areas and perimeters were reported only in one common dolphin. Any estimation of the CeA’s size was only approximative because its limits were frequently not well defined, according to other authors (Breathnach and Goldby, 1954). Future studies should employ additional calcium binding proteins and a larger range and number of cetacean species in order to deepen our understanding of the AMY in these marine mammals.