Six adults female Churra sheep (age = 16.60 ± 0.25 months); mean live weight [(LW) = 70 ± 6 kg] housed under natural light, temperature, and humidity conditions, fed ad libitum and provided with unrestricted access to clean drinking water, were used in this study over farm natural conditions. The experiments were performed in accordance with Spanish (RD 53/2013) and European (63/2010/EU) directives, and animal maintenance and care met all requirements of the agreement on the use of animals in scientific experimentation in Spain (Confederation of Scientific Societies of Spain - COSCE). Accordingly, the animals were sacrificed at the slaughterhouse MACRISA, in Medina de Rioseco, Valladolid, Spain, in April 2021, by neck cutting, severing the carotid and jugular veins, without prior drug administration. Heparinized 5-ml tubes were used to collect blood samples by jugular venipuncture; samples were immediately centrifugated at 3000 × g for 20 min, and plasma was stored at −20°C. Hormonal levels of progesterone was assayed by radioimmunoassay (RIA). All samples were run in a single assay. In all specimens, the progesterone levels were homogeneous, ranging from 15 to 20 mg/l.

After removing the skin, a horizontal cut was made in the skull from the external occipital protuberance to the upper edges of the orbits using an ultra-fast surgical saw with a 5 cm diameter radial disc. Using this approach, the dorsal superficial area of the cortex was cleanly cut, thus showing in depth the ventricles in the middle line maintaining intact the diencephalic area. This procedure provided access to the ventricles for a faster and more effective fixation. After tilting the skull backwards with a spatula and a pair of tweezers, the brain was gently pushed to ventrally cut the optic nerves to rostrally free the brain and to reveal and cut the pituitary stem. Caudally cranial nerves and meningeal flanges were severed. As a result, the brains shed and fell off under their own weight. The nerves were cut with a vise, and the meninges were removed from the pituitary stalk to facilitate infiltration of the fixative solution. The time elapsed from euthanasia to brain and pituitary fixation was approximately 5 min.

To set cytoarchitectonic boundaries in panoramic high-resolution images, nuclei were delimited by exploring differences in cell density, cellular shape and size, and Nissl staining intensity. In this study, the preoptic area (POA), the anterior and the posterior areas of the terminal hypothalamus (THy), were analyzed to more easily describe the localization of SFR (Figure 1). This approach helped us to locate, in a single sagittal section, the four nuclei of interest, ARC in the posterior (Figure 1 dot 1), SCh and SO (Figure 1 dots 2, 3), in the anterior and PVN (Figure 1 dot 4) close to the POA. By contrast, in horizontal and coronal sections, the nuclei must be analyzed in several serial sections. To estimate the related position of coronal and horizontal sections with respect to the sagittal plane, the coordinates were calculated from the first rostral (for the coronal plane) and ventral (for the horizontal plane) sections from carved blocks of tissue (sectioning plane), by multiplying the order number in the series by the averaged thickness of the sections (36 μm) (Red lines in Figure 2C in correspondence with Figures 3, 4).

The neurons of SFR nuclei showed Cb and Cr, but not Pv, immunoreactivity. All CBPs immunocytochemistry results from serial sections in the coronal plane were compared with the corresponding Nissl-stained alternate sections in the coronal plane.

No distinct features of astroglia and microglial cells were detected in IGL with respect to the remaining lateral geniculate body (not shown). In panoramic views, glial cell size and cell density were increased in extensive areas around ventricles in POA (rostral to the hypothalamus) and PHy, both for IBA 1 and GFAP immunocytochemistry (Figures 6, 7). In our material, medial (MPA) and lateral (LPO) regions of POA were distinguished based on differences in staining. IBA 1 immunoreaction labeled both areas, but GFAP immunoreaction only labeled MPA (compare Figures 6, 7).

Inside the POA and PHy, an intense immunoreaction in microglial and astroglial cells was observed inside SFR nuclei (Figures 5, 6 dotted lines). In PHy, a high concentration of IBA1-immunoreactive cells was identified along the border of the 3^rd ventricle to the ventral area around and inside ARC (Figures 6A, B dotted line); however, the density of immunoreactive cells decreased in the infundibular area, allowing us to define an ARC microglial-specific region separate from the periventricular zone (Figures 6A, B).

In IBA1 immunostained sections, from PVN and the 3^rd ventricle, to the lateral limit of LPO, cell density decreased gradually dorsally and laterally, as shown after comparing gray density values from selected square areas in Figure 5C (mean gray values: Square 1 – 191.48/ Square 3 – 104.9). MPA and LPO limits were also well defined by the laterally and dorsally slight decrease in immunoreactive microglial cells (Figure 6).

GFAP immunoreactivity also showed a high concentration of astrocytes around the ventricle and POA, but the densest reactive areas were more restricted to the paraventricular zone (Figure 6C). As in IBA1 sections, GFAP immunostaining density decreased laterally in POA, matching MPA and LPA borders (e.g., please note these differences in cell density in details 2, 3 and 4 in Figure 7). In GFAP-immunostained sections, PVN neurons appeared as negative dots framed by densely stained astrocytic feeds (Figure 7C square 3). In the most caudal, sections immunostained against GFP show tanycytes which cross the stalk and the median eminence forming a superficial well-stained limit and a superficial barrier (Figures 7A, B squares 1 and 2).

Measurements of the main central axis of the mammillary body in the three sectioning planes in histological Nissl-stained sections, compared with measures in macroscopic images, show small differences in retraction (<3%). As mentioned above, with measurements of the thickness of 20 randomly selected sections taken by three different observers (z-axis, and using a Leica X100 plan apo objective), we calculated a shrinkage of 4 (+/-) μm (final averaged thickness 36 μm). These data facilitated slight corrections in the translation of coordinates using the fixation and sectioning protocol applied here. Moreover, when coordinates are translated from frozen specimens, which is the most suitable approach to tissue extraction for molecular analysis, the measurements must be modified after evaluating changes in tissue contraction in each procedure. For micro-dissection, coordinates for extraction were defined in present paper in a caudal-to-rostral direction taking as referent point the anterior limit of the mammillary body (easily located macroscopically by its spherical shape close to the midline). From this reference point, ARC was localized in a sagittal view at 1.5 mm at the vertical axis, 2 mm at the horizontal axis (Figure 8A), and 1 mm laterally to midline (Figure 8B). By using as a reference, a parallel line to the ventral surface (tangential to the ventral most point of the mammillary body and the surface of the optic tracts – green line in Figure 8A), the periventricular region was located at its main axis oriented in an angle of 47.43^o. By using the same reference line, SCh specimens were localized 0.7 mm dorsally to the recess (Figure 8A, red arrows), and PVN was dissected by cutting a 4.2 × 1 mm rectangle (Figure 8A rectangle). Due to its localization, sandwiched between the optic tract (OT) and the preoptic area, SO coordinates cannot be accurately defined, but a tentative localization is suggested in Figure 8B (black dotted line). Furthermore, by its convoluted structure and oblique orientation, IGL coordinates are very difficult to precisely define in macroscopic views; nevertheless, a tentative localization is also provided in Figure 8 (red dotted rectangle).

After SFR nuclei activation by changes in photoperiod, peptide regulation induces neuronal activation and plasticity (11–13) leading to morphological changes in the hypothalamus. These changes in the estrous cycle, previously studied in rats (49) and as a function of seasonal fluctuations, mainly regulated by melatonin (33) may induce size and shape differences in SFR nuclei. In this study, ewes were euthanized under specific environmental, nutritional, and estrous cycle conditions. However, potential changes in morphology may be induced by plastic reorganization after changes in hormonal induction. Therefore, the anatomical data reported here must be revised under other estral conditions.

Relative SCh, ARC and PVN differences in shape and localization were identified in this study. SO showed substantial qualitative differences in both its layout and relative size in ewes, compared with other mammals (e.g., rodents, primates. pigs) (26, 31); however, the basic neuroanatomical plan of organization of ewe SFR nuclei, is similar to that of other mammals.

In line with differences in reproductive functions between rams (e.g., mounting) and Ewes (e.g., pregnancy and lactation), strong anatomical differences have been previously shown in the nuclei studied here (sexual dimorphism) (50, 51). Accordingly, in the future a specific anatomical systematization should be performed in rams for a complete neuroanatomical description of nuclei involved in seasonality in sheep (Ovies orientalis aries).

KNDy neurons form an interconnected network that fires synchronously to drive GnRH release during a pulse, inducing combined differences between nuclei closely positioned in the hypothalamus (23, 24). For future molecular analysis, functional and anatomical interconnections between hypothalamic nuclei involved in seasonal sexual activation should be studied using well-defined samples of tissue restricted to specific nuclei. A study using nuclei microdissection to analyze large areas of the hypothalamus quickly immersed in liquid nitrogen reported excellent results about changes in GnRH-R mRNA levels after stress in merino sheep (52). Thus, in the present study, we provide a cytoarchitectonic guide for a more precise microdissection of SFR nuclei, enabling researchers to collect more appropriate samples for conventional proteomic or genomic techniques, individual cell sequencing or chemical analysis of hormones, transcription factors, and peptides. Preparing serial sagittal sections after micro-dissection enables to confirm the appropriate extracted areas, as previously we reported in the rat (53).

IGL was first described in rats (54), and golden hamsters (55) as a well-defined band of cells located between the dorsal and ventral medial geniculate bodies. When comparing rodents and higher mammals, such as primates (26) or sheep (present results), the well-defined straight band between dorsal and ventral subdivisions of the visual thalamus were masked and hidden by the complex laminar organization of the dorsal geniculate bodies. However, in our Nissl-stained sheep sections (Figure 2A), large fusiform neurons, delimiting IGL, were more clearly observed in the coronal plane (Figure 3G). In rodents, nevertheless, its architecture is well defined, but both the localization and limits are difficult to distinguish in some gyrencephalic vertebrate species (26) because primates once had a generic wide pregeniculate nucleus (PGN) included both the ventral geniculate nucleus and IGL (56). In the visual thalamus of the rock cavy (Kerodon rupestris; rodent), IGL is identified as a well-defined and straight band between dorsal and ventral subdivisions, but in marmosets (primates), this subdivision is masked and hidden by the complex laminar organization of the dorsal part of the LG (57). Due to its intricate architectonic organization, the primate IGL should be neuroanatomically better defined by using additional neurochemical markers (e.g., VIP immunocytochemistry). However, in the ewe, positive Cb and Cr terminal fields, most likely coming from the retina„ allowed us to trace the borders of this subdivision. In summary, due to its larger fusiform neurons and Cb and Cr immunoreactivity (Figures 5B, C), IGL can be well differentiated from the remaining visual thalamus in the ewe (Figure 3A).

Anatomically, six types of neurons have been described in previous studies using the Golgi method in the rat ARC nucleus (58). Different neurochemical phenotypes have also been reported, particularly a neuronal type in which kisspeptin, neurokinin A and dynorphin coexist, known as KNDy neurons (59). These key neurons regulate seasonality outputs for reproduction (kisspeptin, neurokinin, dynorphin) by activating GnRH pulses in sheep (32) Because these neurons are mainly located in ARC (23, 60), techniques demonstrating kisspeptin, neurokinin or dynorphin (not available in our material) should be used as complementary markers to be able to better correlate the morphology and topography of this nuclei with present results. In ARC, several chemical neuronal types have also been described; thus, an accurate analysis in the ewe will require a peptide immunocytochemistry analysis, which was not performed in our study. Nevertheless, in our material, the morphology (Nissl) and immunoreactivity (CBPs) of dorsal ARC neurons showed that they were roughly similar in size and location to KNDy neurons.

PVN shows the highest peptidergic and glial activity in the hypothalamus and is located at the border of the third ventricle. Surrounding this structure, some regions of special interest for their association with seasonal reproduction in sheep, such as the set of A15 dopaminergic cells, regulate the negative feedback of estradiol, which in turn is another compound that regulates both GnRH production and pulses (32, 61, 62). In ewes, the columnar distribution at the sagittal plane allowed us to distinguish strongly Nissl-stained neurons into three potential subdivisions (Figure 2E).

Light cycle-sensitive pacemaker SCh neurons are directly regulated by retinal and indirectly by IGL projections through the geniculate thalamic tract (63). SCh, a highly phylogenetically conserved structure with a similar location in many species of vertebrates, is placed between the 3^rd ventricle and the optic chiasm and contains small and packed neurons (64). In ewes, we also observed small neurons (~430 μm²) and strong immunoreactivity for fibers containing Cr or Cb presumably belonging to retinal projections (Figures 5L, M).

SO interconnects several of the seasonality nuclei (PVN and ARC), directly projecting to the pituitary gland and to the region surrounding the third ventricle (44, 45) Because SO is involved in glial activation, we will analyze its morphological features more extensively in the next section (51). Despite differences in animal size, SO nucleus and its neurons are much larger in ewes than in rodents, but they have the same topographical distribution (26, 41).

To our best knowledge, this is the first analysis of glial cell architecture in the ewe hypothalamus and diencephalon. Glial cells are widely known to play a key role in the synthesis, recycling and delivery of hypothalamic hormones and factors (65) and thus GnRH in the hypothalamus of different mammals, including sheep (51, 66, 67). Increases in cell density and cell size around the 3^rd ventricle shown in our material after IBA1 and GFAP immunoreactivity, indeed expresses glial overactivation (increase cell size, number of profiles, immunostaining and cell density) which can be related to potential areas of higher metabolic regulation of sexual hormones (GnRH) (68). Accordingly, the densest areas of microglial and astroglia immunoreaction around PVN and SO, in the anterior hypothalamus, and around the ARC, in the THy, may be related with a heightened endocrine or paracrine regulation.

Notwithstanding the above, both types of reactive glial cells (as shown by larger and denser population of microgliocytes and astrocytes) are found around neurons. These findings suggest the potential role of denser immunoreactive areas with overactivated paracrine secretion. When comparing IBA1 and GFAP immunoreactivity, microglial reactivity was more extended in all section levels, suggesting a microglial more widespread in areas potentially implicated in secretory regulation (please compare Figures 6, 7). Tanycytes formed a dense ventricular barrier of the 3^rd ventricle in the stalk-median eminence (Figures 7A, B squares 1, 2). Due to the absence of any other reactive glial cells, this finding suggests that these areas, as in other species, regulate the access of metabolic signals to the hypothalamus (69).

Our results establish a guide for future histological, physiological and molecular experiments related to hypothalamic effects of interventions in farm animals aimed at improving fertility regulation (e.g., GnRH neurons, KNDy cells, and melatonin receptors). The data provided here may also have translational relevance because the sheep brain, considering its similarity in anatomy and organization, is a good experimental model for human neuroanatomical research, as recently suggested (70). Finally, by establishing a procedure for the section-to-brain translation of coordinates, this research may also help us to develop suitable procedures for nuclei/tissue extraction from specific nuclei of interest toward assessing changes in mRNA levels by RT-PCR or in protein levels by Western blot.