The Shark Basal Hypothalamus: Molecular Prosomeric Subdivisions and Evolutionary Trends

The hypothalamus is a key integrative center of the vertebrate brain. To better understand its ancestral morphological organization and evolution, we previously analyzed the segmental organization of alar subdivisions in the catshark Scyliorhinus canicula, a cartilaginous fish and thus a basal representative of gnathostomes (jawed vertebrates). With the same aim, we deepen here in the segmental organization of the catshark basal hypothalamus by revisiting previous data on ScOtp, ScDlx2/5, ScNkx2.1, ScShh expression and Shh immunoreactivity jointly with new data on ScLhx5, ScEmx2, ScLmx1b, ScPitx2, ScPitx3a, ScFoxa1, ScFoxa2 and ScNeurog2 expression and proliferating cell nuclear antigen (PCNA) immunoreactivity. Our study reveals a complex genoarchitecture for chondrichthyan basal hypothalamus on which a total of 21 microdomains were identified. Six belong to the basal acroterminal region, the rostral-most point of the basal neural tube; seven are described in the tuberal region (Tu/RTu); four in the perimamillar region (PM/PRM) and four in the mamillar one (MM/RM). Interestingly, the same set of genes does not necessarily describe the same microdomains in mice, which in part contributes to explain how forebrain diversity is achieved. This study stresses the importance of analyzing data from basal vertebrates to better understand forebrain diversity and hypothalamic evolution.


INTRODUCTION
The hypothalamus is an important physiologic center of the brain. It integrates information from limbic, endocrine and autonomic sources to elaborate different kinds of homeostatic and behavioral responses such as feeding or reproduction. Its organization has been elusive for neuroanatomists as result of complex patterning processes converging at this point (Shimamura et al., 1995;Puelles and Rubenstein, 2003;Puelles et al., 2004;Medina, 2008;Szabó et al., 2009;Shimogori et al., 2010;Alvarez-Bolado et al., 2012;Beccari et al., 2013;Croizier et al., 2015).
Cartilaginous fishes, also known as Chondrichthyans, are a key group for evo-devo studies. They are among the most basal extant groups of gnathostomes (jawed vertebrates) being the closest out-group to osteichthyans (the other major phylum of gnathostomes, which includes bony fishes and tetrapods). Therefore, they are essential to address the ancestral condition of the vertebrate brain (Coolen et al., 2009). In previous work, we sketched prosomeric organization in the catshark Scyliorhinus canicula to better understand the ancestral condition of the vertebrate hypothalamus .
In this study, the expression of ScNkx2.1, ScDlx2/5, ScShh and ScOtp led to the identification of alar and basal domains, apparently homologous to the murine ones. However, this work also suggested that alar organization seems to be more conserved than basal one, what correlates with the development of conserved and divergent adult structures, respectively Rodríguez-Moldes et al., 2017). In subsequent work we deeply tested prosomeric assumptions in the alar hypothalamus on the light of additional makers. Our findings suggested conserved traits that can be traced back to the agnathan-gnathostome transition (Santos-Durán et al., 2016). Now, we revisit the organization of the basal hypothalamus with similar aims: (i) to look for further prosomeric molecular subdivisions; (ii) to test if new data on gene expression patterns support previous prosomeric interpretations; and (iii) to obtain some insights on the evolution of this region by comparative analysis. Noteworthy, conserved adult structures (i.e., tracts of the hypothalamic-hypophyseal system, a median eminence or the neurohypophysis) and divergent ones (i.e., inferior hypothalamic lobes and the saccus vasculosus) emerge from this territory offering an attractive scenario for evolutionary insights. To address these questions, previous data on ScNkx2.1, ScDlx2/5, ScOtp, ScShh expression and Shh immunoreactivity were revised jointly with new data on ScLhx5, ScEmx2, ScLmx1b, ScPitx2, ScPitx3a, ScFoxa1, ScFoxa2 and ScNeurog2 expression and proliferating cell nuclear antigen (PCNA) immunoreactivity patterns. Here we were able to identify a plethora of subdomains (microzones) in the catshark hypothalamus. A comparative analysis of microzone identity in catshark is made with mammals but not with other vertebrates due to the lack of detailed data. However gross comparisons among vertebrates prompt the idea that the caudal border of the hypothalamus, as it is currently defined, could be a derived character rather than a conserved one, a feature that deserves further investigation.

Experimental Animals
Some embryos of the catshark (lesser spotted dogfish; S. canicula) were supplied by the Marine Biological Model Supply Service of the CNRS UPMC Roscoff Biological Station (France). Additional embryos were kindly provided by the Aquaria of Gijón (Asturias, Spain), O Grove (Pontevedra, Spain) and Finisterrae (A Coruña, Spain). Embryos were staged by their external features according to Ballard et al. (1993). For more information about the relationship of embryonic stages with body size, gestation and birth, see Table 1 in Ferreiro-Galve et al. (2010). Sixty-nine embryos from stages 28 to 32 were used in this study. Eggs from different broods were raised in seawater tanks in standard conditions of temperature (15-16 • C), pH (7.5-8.5) and salinity (35 g/L). Adequate measures were taken to minimize animal pain or discomfort. All procedures conformed to the guidelines established by the European Communities Council Directive of 22 September 2010 (2010/63/UE) and by the Spanish Royal Decree 53/2013 for animal experimentation and were approved by the Ethics Committee of the University of Santiago de Compostela.

Tissue Processing
Embryos were deeply anesthetized with 0.5% tricaine methane sulfonate Sigma,St. Louis,MO,USA) in seawater and separated from the yolk before fixation in 4% paraformaldehyde (PFA) in elasmobranch's phosphate buffer [EPB: 0.1 M phosphate buffer (PB) containing 1.75% urea, pH 7.4] for 48-72 h depending on the stage of development. Subsequently, they were rinsed in phosphate buffer saline (PBS), cryoprotected with 30% sucrose in PB, embedded in OCT compound (Tissue Tek, Torrance, CA, USA), and frozen with liquid nitrogencooled isopentane. Parallel series of sections (12-20 µm thick) were obtained in transverse planes on a cryostat and mounted on Superfrost Plus (Menzel-Glasser, Madison, WI, USA) slides.
Whole mounts embryos were prepared for IHC as previously described in Santos-Durán et al. (2015). After fixation with 4% PFA in 0.01 M PBS at 4 • C for 2 days, embryos were washed in 0.9% NaCl in distilled water, dehydrated in graded series of methanol solutions (50%, 80%, 100%) and stored at −20 • C. Samples to be stained were placed on ice in 2 ml of dimethyl sulfoxide (DMSO)/methanol (1/1) until they sank. Then, 0.5 ml of 10% Triton X-100/distilled water was added, and the embryos were incubated for 30 min at RT. After washing in 0.05 M TBS with 0.1% Triton X-100 (TST, pH 7.4), samples were sequentially blocked using spin-clarified aqueous 1% periodic acid and 5% non-fat dried milk in TST (TSTM). Primary antibody (polyclonal rabbit anti-Sonic Hedgehog [anti-Shh], Santa Cruz Biotechnology, Santa Cruz, CA, USA, diluted 1:300) was diluted in TSTM containing 0.1% sodium azide for 2-4 days at RT with gently agitation on a shaking platform. The secondary antibody (horseradish peroxidase [HRP]-conjugated goat anti-rabbit, BIORAD, diluted 1:200 in TSTM) was incubated overnight. After a final washing in TST, the embryos were pre-incubated with 0.25 mg/mL diaminobenzidinetetrahydrochloride (DAB, Sigma) in TST with 2.5 mg/mL nickel ammonium sulfate for 1 h, and then allowed to react with DAB in TST containing 2.5 mg/mL nickel ammonium sulfate and 0.00075% H 2 O 2 for 20-40 min at RT. The reaction was stopped using Tris-HCl buffered saline and specimens were post-fixed with 4% PFA overnight at 4 • C. Epidermis and mesodermic derivatives were carefully removed and specimens were rinsed in graded series of glycerol (25%, 50%, 75% and 100%) and observed under the stereomicroscope.

Controls and Specificity of the Antibodies
No immunostaining was detected when primary or secondary antibodies were omitted during incubations. The monoclonal anti-PCNA antibody specifically labels proliferating cells in the brain, retina and olfactory epithelium of this species (Rodríguez-Moldes et al., 2008;Ferrando et al., 2010;Ferreiro-Galve et al., 2010;Quintana-Urzainqui et al., 2014). The polyclonal anti-Shh antibody (Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA) was raised in rabbit against the amino acids 41-200 of the human Shh protein. We previously reported that the in situ hybridization (ISH) results were similar to those obtained by IHC, and therefore validate the specificity of the anti-Shh antibody used here .

Image Acquisition and Analysis
Light field images were obtained with an Olympus BX51 microscope equipped with an Olympus DP71 color digital camera. Photographs were adjusted for brightness and contrast and plates were prepared using Adobe Photoshop CS4 (Adobe, San Jose, CA, USA).

ScNkx2.1 and ScDlx2/ScDlx5 Expression. Comparison With ScShh-Expression/Shh Immunoreactivity
An overview of the expression of ScShh, ScNkx2.1, ScOtp and ScDlx2/ScDlx5 in the basal hypothalamus of S. canicula, mainly in early stages of development, has been previously described  Here we revisited these data to deepen in the genoarchitectonic profile of the basal hypothalamus and further characterize possible dorso-ventral and rostro-caudal subdomains of this territory. A detailed comparative analysis of the expression of such genes in sagittal and transverse sections is presented from stages 29 to 32, when the basic mature cytoarchitecture and organization of the adult hypothalamus are clearly recognized.

ScShh-Expression/Shh Immunoreactivity
Since ScShh detection by means of ISH at early developmental stages yields similar results to those obtained by IHC against anti-Shh (Santos-Durán et al., 2015), here we have used the antibody to analyze additional developmental stages and to ease pattern comparisons by means of double ISH-IHC staining. As described in Santos-Durán et al. (2015), from stage 29 onwards, Shh immunoreactivity is observed in part of the rostral and dorsal Tu domain and broadly detected within the RM domain, extending from here along the diencephalic basal plate (see Figures 1A-F; see also Santos-Durán et al., 2015). Shh immunoreactivity is not observed in the SPa domain (Figures 1B,E; see also Santos-Durán et al., 2016) or in the midline (acroterminal territory) just dorsal to the developing adenohypophysis (black arrowhead in Figure 1E). Caudally, Shh immunoreactivity is only observed in a portion of the RM domain but not at its dorsal-most and ventral-most portions ( Figure 1B and arrowhead in Figure 1F). At stage 30, Shh immunoreactivity is still detected in the hypothalamus and primary prosencephalon (arrowheads in Figure 1G; see also Santos-Durán et al., 2015) but it becomes reduced in the RM and basal plate of p3 (p3Tg) compared to previous stages.

ScNkx2.1 Expression
From stage 29 onwards, ScNkx2.1 is expressed ventral to the optic stalk through the whole basal hypothalamus except in the RM compartment (see Figures 1A-C,H-K). While ScNkx2.1 expression and Shh immunoreactivity co-distribute in part of the Tu domain (Figures 1C,I), Shh immunoreactivity does not match the dorsal border of ScNkx2.1 expression (black arrows in Figure 1I). In contrast to Shh, ScNkx2.1 is additionally expressed in the acroterminal territory dorsal to the adenohypophysis (arrowhead in Figure 1I). ScNkx2.1 expression in the MM abuts the RM, but it does not meet Shh immunoreactivity since Shh is absent from the ventral-most portion of RM, which creates a gap between both (Figures 1C,J; see also Santos-Durán et al., 2015). Of note, ScNkx2.1 expression in the ventricular zone forms a clear-cut border between the positive MM and the negative RM domain that is more evident on sagittal sections ( Figure 1K), though ScNkx2.1-expressing cells can be detected in the mantle of the RM domain (black arrowhead in Figures 1J,K).

ScDlx2/ScDlx5 Expression
In the basal plate ScDlx2/5 is intensely expressed in a restricted subdomain of the Tu/RTu and the p3Tg domains ( Figure 1L). In the hypothalamus, it is expressed in a subdomain spreading from the RTu to the neurohypophysis (Figures 1L-N). Rostrally, ScDlx2/5 expression in the basal hypothalamus co-distributes with Shh immunoreactivity in a subdomain of the Tu (Figures 1C,M). Note that neither ScDlx2/5 expression nor Shh immunoreactivity can be observed in the acroterminal territory co-extensive with the adenohypophysis (arrowhead in Figure 1M) but it is expressed in the neurohypophysis ( Figure 1L). ScDlx2/5 expression in the caudal-most RTu almost abuts Shh immunoreactivity in the RM although a gap exists ( Figure 1C; arrow in Figure 1N). In the most caudo-ventral part of Tu, individual and dispersed ScDlx2/5-expressing cells can be recognized almost reaching the rostral and ventral-most part of the PM (arrowheads in Figures 1L,N) including the primordium of the saccus vasculosus. These cells are less intensely labeled but still observable at stage 31 (arrowhead in Figure 1O). At stage 32 the basic pattern described for ScDlx2/5 is maintained although its expression becomes reduced in intensity (see below).

ScLhx5 Expression
From stage 29 onwards, in the basal plate, ScLhx5 is observed in a subdomain of the dorsal-most and rostral-most Tu domain (Figures 2A,B,I,J). Dispersed ScLhx5-expressing cells are also observed in the most caudo-ventral part of Tu where individual and dispersed ScDlx2/5-expressing cells were observed (compare Figure 2D with Figure 1O). ScLhx5 expression can also be observed in the PM/PRM and MM domains (Figures 2A-D,I,J). Note that ScLhx5 expression in the MM domain describes a clear-cut border with the RM domain ( Figure 2D), though ScLhx5-expressing cells can be recognized in the mantle of the RM domain (arrowheads in Figures 2C,D).

ScOtp Expression
In the basal plate, ScOtp has been identified in Tu and PM/PRM domains . From stage 29 onwards ScOtp is expressed in the rostral-most part of the Tu domain, from the optic stalk to the primordial neurohypophysis ( Figure 2E). Specifically, it is restricted to the acroterminal territory of the Tu domain just dorsal to the adenohypophysis (Figures 2E,I; arrowhead in Figure 2F) codistributing with ScNkx2.8 (arrowhead in Figure 2F ). In the PRM domain, ScOtp expression abuts the RM but not the Shh immunoreactivity of this domain (Figures 2G,J). In the PM, ScOtp is also expressed in the acroterminal territory ( Figure 2E ). Note that the expression of ScOtp in the PM faces the MM (Figure 2H). Marginal ScOtp-expressing cells can be recognized in the RM (black arrowheads in Figures 2G,H) and p3Tg (not shown). This pattern is maintained until stage 32.

ScEmx2 Expression
The expression of ScEmx2 has been analyzed by Derobert et al. (2002) in the brain and related tissues from early stages of development (stage 19) until mid-gestation stages (stages 28-30).
Here we analyze in detail the expression of ScEmx2 in the basal hypothalamus from stage 29 until stage 31. From stage 29 onwards, ScEmx2 is expressed in the basal hypothalamus in a well-defined domain spreading into part of rostral and ventral-most Tu domain, the PM/PRM and the MM domains (Figures 3A-D). Of note, its expression lacks in midline domains of the Tu (acroterminal territory) such as the neurohypophysis ( Figure 3A) and saccus vasculosus ( Figure 3B) but is present immediately caudal to the last (Figures 3C,D).
A comparison of ScEmx2 expression with other genes reveals several correlations. We compared ScEmx2 expression patterns with the presence of PCNA-immunoreactive (-ir) cells (Figures 3E-H). PCNA-ir cells define proliferative zones that are separated by non-proliferative (PCNA-immunonegative) ventricular regions, which are believed to define important segmental boundaries (reviewed in Candal et al., 2005). Of note, the caudal border of ScEmx2 expression in the MM domain (arrowheads in Figures 3A,E) correlates with a domain of reduced PCNA immunoreactivity in the ventricular zone (arrowheads in Figures 3E-G). The caudal border of ScEmx2 expression in the PRM domain also correspond with a domain of restricted PCNA immunoreactivity (compare arrowheads in Figures 3D,H). Thus, a band of reduced or negative proliferation seems to spread from the rostral and dorsal border of the RM (Figures 3F-H), p3Tg and zona limitans intrathalamica (zli; not shown).
Finally, we compared ScEmx2 expression with that of other genes usually expressed in the basal hypothalamus to better understand its organization. A comparison with Shh immunoreactivity (Figure 3I) revealed that the ScEmx2expressing domain in the PRM is fairly complementary to Shh in the RM domain (compare Figures 3A,I). Besides, this ScEmx2expressing domain includes the caudal domain expressing ScLhx5 in the PM/PRM and MM (compare Figures 3A,J) and ScOtp in the PM/PRM (compare Figures 3B,L). Of note, ScEmx2, ScLhx5 and ScOtp define consecutively more restricted domains (compare Figures 3B,K,L). Moreover, the expression of ScEmx2 abuts that of ScDlx2/5 in dorsal and caudal positions (RTu domain) while they co-distribute in more rostral and ventral positions (not shown).
ScLmx1b, ScPitx2, ScPitx3a, ScFoxa1, ScFoxa2 and ScNeurog2 as Markers of the RM ScFoxa1 and ScFoxa2 are expressed in fairly the same spatial and temporal patters in the regions and stages considered in this study and thus are conjointly referred as ScFoxa1/2.
At stage 29 ScLmx1b, ScPitx2, ScPitx3a, ScFoxa1/2 and ScNeurog2 are expressed in a similar pattern spreading caudally from RM into the diencephalon including the zli in the case of ScPitx2, ScPitx3a, ScFoxa1/2 and ScNeurog2 (Figures 4A-D,B ,C ,O). On transverse sections the expression of these genes dorsally abuts the PRM (Figures 4E-H,F ). Rostrally these genes also abut the MM (arrowheads in Figures 4E-H). From stage 29 onwards this general pattern persists although some differences emerge. At stage 31, ScLmx1b becomes downregulated being restricted to the floor plate ( Figure 4I) while Shh immunoreactivity is still found in the basal plate (arrowheads in Figure 4I). At stage 31 ScPitx2 ( Figure 4J) is still expressed in the pattern observed at stage 29, while ScPitx3a is restrictedly expressed in the caudal-most diencephalon (not shown). In the case of ScFoxa1/2, there is slight dorsal and ventral downregulation but the main pattern persists through RM and diencephalon ( Figure 4K). From stage 30 onwards, ScNeurog2 becomes downregulated in the mentioned territories although it can be recognized in the zli and habenulae (data not shown). Finally, ScLmx1b, ScPitx2 and ScFoxa1/2 still present a sharp border of expression that abuts the MM domain at this developmental stage (arrowheads in Figures 4L-N).

DISCUSSION
In a previous work, we identified the shark basal hypothalamus harboring three domains (Tu/RTu, PRM/PM and MM/RM) based on the basal expression of ScNkx2.1, ScShh, ScOtp and ScDlx2/5 (Santos- . In the present work we revisit such analysis on the light of ScLhx5, ScEmx2, ScLmx1b, ScPitx2, ScPitx3a, ScFoxa1, ScFoxa2 and ScNeurog2 expression, besides Shh and PCNA immunoreactivity. Different subdomains were identified within the aforementioned domains and within the basal acroterminal region, the basal rostral-most neural tube (Figures 5, 6; Table 1). These genes present a robust and conserved expression across vertebrates. However, their roles and functions are still poorly understood and their expressions do not necessarily define domains per se. Therefore, caution has to be borne in mind during the following analysis.

Basal Acroterminal Domains
The acroterminal domain involves the alar and basal plate spreading from the rostral-most roof plate to the rostral-most floor plate. It has been suggested that specialized structures like the lamina terminalis, the optic chiasm and the neurohypophysis emerge here under particular signaling events (Puelles et al., 2012;Puelles and Rubenstein, 2015). We have identified at least 6 subdomains inside the basal acroterminal region named 1-6 from dorsal to ventral (BAt1-6; Figure 6B). Noteworthy the acroterminal territory (medial) is easily distinguishable from the remaining hypothalamus (lateral) due to genes differentially expressed at these locations (compare Figures 6A,B).
The two dorsal-most domains (BAt1-2) are positive for ScNkx2.1 ( Figure 1I) and ScOtp (Figure 2F) but only BAt1 shows expression of ScLhx5 ( Figure 3J). Furthermore, ScOtp co-distributes with ScNkx2.8 (Figures 2F,F ). BAt1-2 subdomains are negative for other genes broadly expressed in the Tu such as ScDlx2/5 and Shh (compare Figure 1M with Figure 2F). Of note, at later stages ScLhx5 is absent from the midline (arrowhead in Figure 3K). Noteworthy, BAt1-2 are almost co-extensive with the developing adenohypophysis (Figures 1I, 6B), which in part is co-extensive with negative subdomains for ScDlx2/5-expression and Shh immunoreactivity ( Figure 1M). Of note, these gaps are as wide as the adenohypophysis, which has been noted in other vertebrates even for different adenohypophysis sizes (see Figure 2N in Manning et al., 2006), suggesting a role for the adenohypophysis in the local patterning of the hypothalamus. In shark, both the gaps of ScDlx2/5-expression and Shh immunoreactivity and the expression of ScNkx2.8 are wider than the medio-lateral extension of ScOtp-expression (Figures 2F,F ), which suggests the existence of additional medio-lateral subdomains. BAt1 Asterisks indicate that cells expressing this gene are not located in the ventricular zone. For abbreviations, see list.
Ventrally to BAt3, we identified BAt4 as a subdomain that corresponds to the primourdium of the saccus vasculosus ( Figure 6B; see also Van de Kamer and Shuurmans, 1953;Sueiro et al., 2007). The initial tiny domain expands becoming morphologically distinguishable (stage 29, Figures 1H, 2E; stage 30, Figure 1I). This domain is characterized by the expression of ScNkx2.1 and dispersed ScDlx2/5-expressing cells (Figure 1L). Since ScDlx2/5 is involved in the development of a GABAergic phenotype (Anderson et al., 1999), its expression in the saccus vasculosus could explain the existence of GABAergic cells at this point (Sueiro et al., 2007). Besides, GFAP-immunoreactivity has been described to be restricted to BAt4 (the developing saccus vasculosus), and it is not observable in more caudal subdomains (Sueiro et al., 2007). Finally, BAt4 is also characterized by lack of ScEmx2 (Figure 3B) which, however, is present in more caudal acroterminal subdomains (Figures 3C,D; see also BAt5 in Figure 6B) and in lateral (non-acroterminal) domains (Figures 6A,D). Noteworthy, in S. canicula, the tip of the notochord has been described to reach the primordium of the saccus vasculosus (BAt4; Figure 1 in Van de Kamer and Shuurmans, 1953) suggesting a causal relationship to saccus vasculosus development.
BAt4 shares ScNkx2.1 expression and dispersed ScDlx2/5expressing cells with the domain ventral to it (BAt5; Figure 6B). However, as commented above, BAt5 differentially presents a lack GFAP-immunoreactivity and the presence of ScEmx2 expression (Figures 3C,D; see also Figure 6B) at late stages of development.
The ventral-most acroterminal domain is BAt6 which express ScNkx2.1, ScEmx2, ScOtp and ScLhx5, but not ScDlx2/5 (see Figure 6B). However, we cannot discard that this territory could be interpreted as the floor plate of the MM (Figure 6D

Perimamillar (PM/PRM) Subdomains
Ventral to Tu6, we identified PM1 a domain where ScNkx2.1/ScEmx2/ScLhx5 are co-expressed (compare Figures 1K, 3C,K; see Figure 6D). We term PM2 (PM-like in Santos-Durán et al., 2015) the subdomain expressing ScNkx2.1/ScEmx2/ScLhx5/ScOtp (compare Figures 1K, 3C,K,L; see Figure 6D). The caudal continuation of PM1 and PM2 are referred as PRM1 and PRM2 (PRM-like in Santos-Durán et al., 2015) and express the same genes ( Figure 6D). We have identified three dorso-ventral subdomains in the ventral and caudal-most point of the basal hypothalamus here referred as RM1, RM2 and RM3 (Figure 6E), which together fairly correspond to the previously defined RM-like territory ; see also Figure 5). The dorsal-most domain, RM1, may be defined based on lack of Shh immunoreactivity at stage 29 (as does not reaches the ABB; see Santos-Durán et al., 2016) and the expression of ScLmx1b/ScPi tx2/ScPitx3a/ScFoxa1/ScFoxa2/ScNeurog2 (compare Figure 1A with Figures 3A-D; see Figure 6E). In the RM2, these genes co-distribute with Shh immunoreactivity while the ventral-most domain, RM3, is again characterized by the lack of Shh immunoreactivity (compare Figure 1F with Figures 1E-H; see Figure 6E). Of note, from stage 30 onwards, ScPitx2 is expressed in the whole RM (Figures 4J,M) but the downregulation of other genes suggest the existence of even more dorso-ventral subdomains. Finally, RM3 also presents ScLhx5and ScOtpexpressing cells likely born in neighbor domains (Figures 2D,H,  6E; Table 1).

Evo-Devo Considerations Concerning the Basal Hypothalamus
The prosomeric model offers a key tool to study homologies and brain evolution (Puelles andRubenstein, 2003, 2015). Counterparts of the genes here considered have been also studied in other vertebrates. Nevertheless, the lack of detailed data in prosomeric terms makes difficult to perform comparisons with most groups, except for mammals, at the level of microdomains. Below, detailed comparisons are made with mammals and we assume they are mostly transferable to other amniotes (Figure 7) due to the similarity of several patterns observed between mice and birds (Manning et al., 2006;Bardet et al., 2008Bardet et al., , 2010García-Calero et al., 2008;Abellán et al., 2010;also reviewed in Domínguez et al., 2014). Gross comparisons are also made with anamniotes that are, however, still informative (Figure 8).
In both, mouse and shark, the Tu/RTu is characterized by the expression of Nkx2.1, Dlx2/5 and Otp (Figures 7A,B; see also Santos-Durán et al., 2015). However, differences in sub-compartmentation emerge while considering additional genes. In mouse, Lhx5 is broadly expressed in the basal hypothalamus before stage 13.5 (Szabó et al., 2009;Abellán et al., 2010; Developing Mouse Brain Atlas). Nevertheless, after this stage, it becomes downregulated and restrictedly expressed in non-Tu/RTu domains ( Figure 7A). In shark, ScLhx5 is expressed in the dorsal-most Tu from stage 29 to 31 ( Figure 7B). Besides, in the mouse, the expression of Shh is restricted to the dorsal and caudal-most Tu and RTu ( Figure 7A) while in S. canicula it is fairly expressed in two subdomains of Tu but absent in the RTu ( Figure 7B). Noteworthy, the differential spatial distribution of Shh between both models leads to the emergence of a subdomain in the mouse Tu (asterisk in Figure 7A) containing Dlx2/5 alone, apparently not present in shark ( Figure 7B). However, a careful view suggests that this domain could correspond to the reduced Tu5 domain found in shark (see Figure 6C), so that in mouse this Dlx2/5-expressing domain could have been expanded at expenses of Shh-expressing domains due to changes in ventro-caudal signaling. Finally, in mouse, Emx2 is absent from the Tu/RTu while in shark it is expressed in Tu6 (Figures 6C, 7A,B). The PM/PRM is characterized in both models by Otp and Lhx5 expression (Figures 7C,D). Of note, in shark, this compartment can be subdivided in a rostral PM1/PRM1 that lacks ScOtp and a caudal PM2/PRM2 that does express this gene. In shark, both compartments also express ScEmx2 (Figures 7C,D). Noteworthy, the PRM of shark lacks Shh expression in contrast to mouse (Figures 7C,D).
In the MM/RM, in both species genes expressed in the RM, like Lmx1b or Foxa1, appear to form a clear-cut border of expression with respect to those expressed in the MM (and even in the PRM) as Nkx2.1, Emx2 and Lhx5 (see Figures 7E,F). This seems to be a conserved feature across vertebrates (see below). However, in mouse, genes like Pitx2 and Neurog2 are expressed in both domains (Figures 7E,F) though Neurog2 is restricted to RM at earlier stages (termed as p3Tg in Figures 2B,B in Osório et al., 2010).
Together, this analysis reveals at least two things. First, one to one comparisons are useful to understand how interspecific variability emerges. Interestingly, the same set of homologous genes defines new or different microdomains among different species. Second, the number of microdomains identified increase with the number of genes analyzed, though their significance in terms of homology (common ancestry) becomes elusive. This raises non-trivial questions concerning homology establishment (Abouheif, 1997;Puelles and Medina, 2002) and compel us to consider other possible interpretations in the context of the prosomeric model.

Comparisons With Anamniotes
Many of the genes here considered have been already studied in agnathans and teleosts (Figure 8). Though it is difficult to establish one to one comparisons at the level of subdomains, common traits do exist between these groups (including chondrichthyans). Therefore, such characters are assumed to be transferable to other anamniotes.
In amniotes, genes expressed in the Nkx2.1-expressing hypothalamus (Emx2, Otp, Lhx5) abut those expressed more ventro-caudally (Shh, ScPitx2, ScPitx3, ScNeurog2 and ScLmx1b). Furthermore, the last group of genes describes a continuous line from the floor plate of the terminal hypothalamus and extends into the zli (Figures 6A,C and Figure 8). Such abutted expression is typical of segmental boundaries (for definition of segmental boundaries see Dahmann et al., 2011;Cavodeassi and Houart, 2012;Kiecker and Lumsden, 2012; see also Larsen et al., 2001;Puelles et al., 2012) and has not been observed at other points of the caudal secondary prosencephalon. Having this in mind we decided to look for other evidences for segmental boundaries at this point as reduced cell proliferation and the presence of signaling centers. Noteworthy, at least in sharks, reduced cell proliferation (PCNA-negative cells at the ventricular zone in Figures 3E-H) can be also detected bordering the domain where Shh and other markers are expressed in the caudal hypothalamus. Moreover, transverse Wnt signals seem to describe such border from the zli to the MM/RM boundary in different vertebrates ( Figure 8A ; see also Guérin et al., 2009;Quinlan et al., 2009) while the situation in mice remain unclear. Finally, such border also seems to be the same as that delineated by Figdor and Stern (1993) between segment D1 and D2. Noteworthy, in the lamprey (Figure 8A), the expression of Wnt signals resemble that of other Wnt genes between rhombomers (Riley et al., 2004) an idea already suggested in Santos-Durán (2015). Though our results suggest that such border could get deformed on the course of evolution as shown in Figure 8A , these evidences are not necessarily supported in mice where the expression of genes like Neurog2 and Pitx2 is continuous through the MM/RM rather than restricted to the RM (Figure 7). Of note, a recent review on vertebrate forebrain development also suggests the existence of a novel secondary organizer at this point (Puelles, 2017) as previously suggested in Santos-Durán (2015).
However, the idea that the currently identified as MM/RM border could indeed represent the hypothalamic-diencephalic border implies a new model of the hypothalamus that would require re-examination of other postulated limits and thus deserves further investigation.

CONCLUSION
This work belongs to a series of articles  addressing the development and evolution of the chondrichthyan hypothalamus and also the new proposals of the prosomeric model (Puelles et al., 2012;Puelles and Rubenstein, 2015) on the mentioned region.
Here, our combinatorial analysis revealed the existence of many different microdomains within the main subdomains of the prosomeric basal hypothalamus of the shark (Tu/RTu; PM/PRM; MM/RM). The genes considered in this study (ScOtp, ScDlx2/5, ScNkx2.1, ScShh, ScLhx5, ScEmx2, ScLmx1b, ScPitx2, ScPitx3a, ScFoxa1, ScFoxa2 and ScNeurog2) are well conserved in vertebrates. However, detailed comparisons at the level of microdomains under the prosomeric framework can be only performed with mammals, on which abundant data are available. Such analysis reveals a number of microzones that do not exactly fit those described in mice (Ferrán et al., 2015). Understanding the homology and evolution of such microdomains results daunting and can be misleading. However, these results illustrate, at least in part, how organisms became different in spite of expressing similar set of homologous genes.

AUTHOR CONTRIBUTIONS
GNS-D, IR-M and EC designed the study and analyzed the data. SM contributed to data acquisition. GNS-D, AM and SF-G performed the experiments. GNS-D wrote the manuscript with inputs from all authors.