Expression atlas of Dmrt genes across sex and development in the mouse brain: functional insights from the olfactory system

Abstract

The ancient DMRT family of transcription factors has been proposed as evolutionarily conserved effectors of sexual differentiation. While brain sexual differentiation has traditionally been attributed to a paracrine role of steroid hormones and intrinsic cues directly controlled by sex chromosomes, the downstream effector mechanisms remain elusive. To elucidate the role of Dmrts in mammalian brain sexual differentiation, we generated a comprehensive expression atlas for all family members in the mouse brain using in situ hybridization across sexes and development (E12.5, E14.5, E18.5, adulthood and, sometimes, P7). We found that all Dmrts, except Dmrt7, were expressed in the brain. This study expands understanding of the DMA-Dmrt subfamily beyond pallial structures and identifies their expression maintenance in adult neurogenic sites. For the first time, we described the neuronal expression of Dmrt2 and Dmrt6. Mouse Dmrts did not show clear sexually dimorphic patterns, but did show quantitative expression differences between sexes. Most Dmrts were maintained in postmitotic neurons during embryonic and postnatal stages, suggesting potential interactions with gonadal sex hormones during organizational (perinatally, with testosterone peaking at E18.5 in males) and activational phases (after puberty, with androgens in males and estrogens in females). Our analysis of Dmrt5 expression revealed its prominent presence in the mouse olfactory system. The olfactory system shows sex differences and is fundamental for controlling sex-specific innate behaviors in rodents. Thus, it represents an ideal scenario to test, as a proof of concept, the role of Dmrts in the sexual differentiation of the mammalian brain. The absence of Dmrt5 similarly affected the main olfactory epithelium, where sensory neurons reside, in both sexes; however, cell number-related defects in the olfactory bulb and piriform cortex differed between male and female embryos, revealing Dmrt5 interaction with sex in deeper layers of innate neural circuits. Our results provide a valuable resource for uncovering novel sites and mechanisms of sexual differentiation in the mammalian nervous system, potentially contributing to sex biases observed in psychiatric and neurological disorders.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13293-026-00836-6.

Plain Language Summary

This study looked at where and when certain genes called Dmrts are active in the mouse brain in both males and females, at different stages of life. The researchers studied many areas of the brain and found that these genes may affect how the brain develops differently in males and females. One gene, Dmrt5, was important for brain cells that respond to smells, and its effects depended on sex. These discoveries could help explain why some brain disorders affect men and women differently.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13293-026-00836-6.

Highlights

A comprehensive atlas mapped all Dmrt genes in mouse brain across sexes and development, finding all except Dmrt7 expressed, with novel neuronal expression of Dmrt2 and Dmrt6.

In the mouse brain, Dmrts lack clear binary sexual dimorphism but show quantitative sex differences, such as elevated Dmrt2 in male cerebral cortex.

Many Dmrts persist beyond embryogenesis in postmitotic neurons and adult neurogenic sites, suggesting roles in brain development and stem cell regulation.

Dmrt5 displays broad expression in the olfactory system; its absence causes defects in olfactory dopaminergic neuron specification that might affect innate behaviors.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13293-026-00836-6.

Introduction

Across phylogeny, the nervous system (NS) presents sex differences that are responsible for sex-specific behaviors. In humans, brain sexual variations might uncover sex-related susceptibilities to psychiatric and neurological disorders [1]. Sex hormones and genetic factors regulate the molecular mechanisms underlying the specification and maintenance of NS sex differences [2]. Among them, the ancient family of Dmrt genes (from doublesex and mab-3 related transcription factors) emerged as conserved effectors in the development of sex-specific traits of different animal species [3–5]. Within this gene family, Dmrt1 is the only one with a conserved role in gonadal sex determination across vertebrates. Moreover, in mice, Dmrt1 is essential for male somatic and germ cell differentiation, and in humans, deletions and point mutations in DMRT1 are associated with 46, XY complete gonadal dysgenesis [6].

The role of Dmrts in the sexual differentiation of the NS has been primarily studied in invertebrates such as Drosophila melanogaster and Caenorhabditis elegans. In these organisms, species-specific sex-determination cues control sexually dimorphic expression patterns of Dmrt genes in a binary mode (reviewed in [5]). The sex-specific expression of Dmrts in one sex versus the other generates dimorphic neuronal circuits from which sex-specific behaviors emerge. To accomplish this, Dmrts utilize distinct mechanisms, involving sex-specific control of neuron number, neuronal identity, and synaptic connectivity. Sex differences in neuron number arise through sex-specific dsx isoforms that promote proliferation in males or cell death in females in Drosophila [7–10]. Regulation of sex-specific neuronal identity is a common feature among Dmrts in C. elegans. For instance, mab-3 promotes the neurogenesis of male-specific sensory ray neurons [11–13], while in these neurons, dmd-3 and mab-23 determine their neurotransmitter identities [14, 15]. Regarding sexually dimorphic connectivity, the female-specific isoform of dsx represses midline crossing of axons from foreleg gustatory receptor neurons [16], and dmd-3 orchestrates the drastic male wiring remodeling of the sex-shared PHC neurons [17].

Sex biasing factors can generate but also suppress sex differences. At least in mammals, sexual differentiation results from the effects of sex chromosomes and sex hormones, which can be antagonistic, reducing sex differences between the two sexes [18–21].

Most animal genomes contain multiple Dmrts defined by a common DNA-binding domain known as the DM domain. In mammals, there are seven Dmrts, Dmrt1 to Dmrt7 [22–24], and up to three Dmrt-like genes (Dmrt8.1, Dmrt8.2, and Dmrt8.3) clustered in the X chromosome, which lack a functional DM domain [25]. However, the DM domain-encoding sequence of Dmrt8 is not conserved in the mouse genome and will therefore not be considered in this study. Dmrt family genes exhibit limited sequence homology outside the highly conserved DM domain, except for a subgroup that contains an additional domain, the DMA domain. This subfamily includes Dmrt3, Dmrt4 (also known as Dmrta1), and Dmrt5 (also known as Dmrta2) in mammals, dmrt93B and dmrt99B in D. melanogaster, and dmd-4 in C. elegans [26]. From this point forward, we will collectively refer to the Dmrt3-5 genes as DMA-Dmrts.

The DMA-Dmrts are the most studied Dmrts in the mouse NS, functioning in telencephalic patterning and corticogenesis [27]. DMA-Dmrts are crucial for determining the dorsoventral axis of the mouse telencephalon, defining the dorsal identity and the pallial-subpallial boundary [28–33]. On the other hand, the mouse Dmrt3 and Dmrt5 maintain the self-renewal ability of the neural stem and progenitor cells in the medial pallium [30, 33, 34], while in the lateral cortex, Dmrt5 has a neurogenic role [35]. Furthermore, in mice, other brain functions have been described for Dmrt3 and Dmrt5 beyond the telencephalon. Dmrt3 participates in the neuronal specification of the dI6 subdivision of spinal cord neurons and is critical for locomotor coordination [36]. Dmrt5 is crucial for the ventromedial neural fate specification of the midbrain according to in vitro experiments on neuralized mouse embryonic stem cells [37].

Behavioral defects are reported for Dmrt3 and Dmrt4 null mutant mice, while the Dmrt2 and Dmrt5 null mutations are perinatally lethal. Neonatal Dmrt3 null mutant mice present severe deficits in hindlimb coordination, and adult Dmrt3 null mice display gait disturbances and reduced running performance at high speeds [36]. A subset of Dmrt4 null mutant males displays sexual mounting on other males. However, Dmrt4 mutant males retain a preference for female-soiled bedding and do mate with females, suggesting lower selectivity rather than a reversed mate preference [38]. In Dmrt1 null mutants, postnatal transdifferentiation in the testis, from male Sertoli cells to their female equivalent granulosa cells, leads to severely reduced testosterone levels [39]. To our knowledge, behavioral experiments have not been performed in Dmrt1 mutants, but deficits in sexual and social behaviors (e.g., mounting and intermale aggression) are expected due to reduced testosterone levels [40]. In previous works, Dmrt6 and Dmrt7 single null mutants exhibit disrupted spermatogenesis [41–43], but impaired specific behaviors are not described. In humans, deletions or duplications of the short arm of chromosome 9, where DMRT1, DMRT2 and DMRT3 are located, have been associated with craniofacial malformations, intellectual disability and mental disorders, including autism spectrum disorder [44, 45] and obsessive‑compulsive disorder [45–47], in which DMRT2 has been proposed as a risk factor [48]. Moreover, homozygous mutations in DMRT5 cause microcephaly, underscoring the fundamental role of DMRT5 in cerebral cortex development, which is altered in multiple neurodevelopmental disorders such as attention-deficit/hyperactivity disorder, autism spectrum disorder, and schizophrenia [49].

A comprehensive characterization of Dmrt gene expression patterns in the mammalian NS is crucial for elucidating novel mechanisms of sexual differentiation. Among the regions of the NS, the olfactory system is an ideal context to interrogate the effect of the sex on Dmrt genes since it is key to elicit sex-specific behaviors in rodents. Previous work mainly focused on DMA-Dmrt expression at early developmental stages, mostly in the telencephalon, for which excellent reviews exist [23, 27]. However, no systematic comparisons between the sexes were undertaken in such studies. Furthermore, no studies have addressed Dmrt expression at late embryonic and postnatal stages, when novel functions or interactions with sex hormones might occur. Also, available gene expression atlases, such as the Allen Developing Mouse Brain Atlas or EURExpress, show poor or absent in situ hybridization (ISH) signals for Dmrts.

In this work, we generated a comprehensive map of the brain across sex and development for all Dmrts found in the mammalian genome. This study included embryonic and postnatal stages. The embryonic stages considered span the period between the peak of neurogenesis and the surge of testicular testosterone, encompassing embryonic (E) days 12.5, 14.5, and 18.5. The postnatal stages included adulthood and, for Dmrt5 and Dmrt6, the postnatal (P) day 7. The expression in the main olfactory epithelium and vomeronasal organ was also assessed for all Dmrts at E12.5, for all of them except Dmrt1 and Dmrt7 at E14.5, for Dmrt3 and Dmrt5 at P4, and for Dmrt5 in adulthood. Our study is primarily based on ISH; however, other techniques were also employed. Immunofluorescence (IF) was used to examine coexpression among DMRTs and between DMRTs and molecular markers. 3D brain reconstructions were used to precisely quantify phenotypes in Dmrt5 mutants. Quantitative RT-PCR (qRT-PCR) was also performed on Dmrt2 as a representative example to quantify how Dmrt family expression varies across sex and developmental stages, complementing the spatial resolution of ISH. We provide a molecular atlas of individual expression of Dmrt1, Dmrt2, Dmrt3, Dmrt4, Dmrt5, and Dmrt6 in male and female embryos and postnatal animals. Dmrt7 was not detected in the mouse NS. Our work is an invaluable resource for exploring the function of the DMRT transcription factors (TFs) in the mammalian brain and olfactory system, including the development of sex differences and their implications in the etiology of sex-biased mental disorders as protective or risk factors.

Results

All Dmrt genes, except Dmrt7, are detected in the embryonic NS

We initiated our analysis at E12.5, before testosterone can be secreted by the fetal testis (approximately at E13.0). We followed the E13.5 sagittal atlas of the Allen Developing Mouse Brain Atlas (ADMBA, available at: https://developingmouse.brain-map.org/static/atlas), which is based on the prosomeric model established by Luis Puelles [50]. To a lesser extent, we also used the Prenatal Mouse Brain Atlas at E12.5 [51], which includes coronal sections, while maintaining the nomenclature of the ADMBA atlas, except for the choroid plexus (ChPl), and the piriform cortex (Pir). This analysis relied solely on anatomical landmarks. When precision was uncertain, we deliberately avoided overly detailed annotations and instead assigned expression to broader topological regions. In Fig. 1, only a selection of images of one sex is depicted to complement Table 1, which compiles the expression of all Dmrts at E12.5. For full access to the expression data and sex comparisons, see Repository link: https://www.informatics.jax.org/reference/J:364871.

Dmrts are broadly expressed in early embryonic development. All panels depict ISH at E12.5 unless indicated. Only one sex is shown. Dmrt1 in the medial pallium (MPall) in sagittal (A) and coronal (A’) views. B-B’) Dmrt2 in the subpallial septum (SeSPall) and preoptic area (PO). B’) Higher magnification of Dmrt2 expression in the PO. C-E’) DMA-Dmrts extensively overlap in the pallium and the prethalamic eminence (PThE): Dmrt3 (C–C’), Dmrt4 (D-D’) and Dmrt5 (E-E’). Sagittal view in upper panels, and coronal view in lower panels. F-G) Dmrt6 is detected broadly in subpallial regions [central subpallium (CSPall), striatum (Str), pallidum (Pal), paraseptal subpallium (PaSe)] and other regions. f) Inset shows a magnification of the MPall region. G) Dmrt6 expression along the rostrocaudal (R–C) axis.H) Dmrt3, 4, 5 are expressed in the ventricular zone (VZ) of the PThE, and Dmrt5 is also expressed in the mantle zone (MZ). Dmrt6 is expressed in the MZ of the PThE, and excluded from VZ. I-I’) Dmrt4 is transiently expressed in the habenula (Hb) (I) and the isthmus (I’). J, K) Dmrt3 expression in the gigantocellular reticular area (Gi) of the hindbrain in a coronal plane (J), and in a sagittal plane (K). L-M) Dmrt6 in the VZ of the B1-B3 serotonergic (5-HT) group. Dmrt3 and Dmrt6 occupy distinct anatomical areas in the hindbrain, which is evident in coronal (J, L) and lateral views (K, M). O-P) Dmrt5 in the anterior (AL) and intermediate (IL) lobes of the pituitary gland (Pit) at E12.5 (O) and E14.5 (P). Q-U) Dmrt expression in the ventral midbrain: Q, Dmrt4 (ISH) and DMRT5 (IF) show the lack of overlap between the two TFs in the ventral midbrain. R-S) Dmrt4 (ISH) and TH (IF) in a sagittal view (R) and a coronal view (S). Dmrt4 is strongly expressed in the isthmus (R), and it is detected in mature DA-TH + neurons (S). T, U) Dmrt5 and Dmrt6 (ISH) and TH (IF). Dmrt5 occupies the VZ, while Dmrt6 is excluded from the VZ. None of them overlaps with TH. Dmrt6 fills the lateral ventral domain. V) Dmrt expression in the nasal epithelium. Dmrt2 is strongly detected in the respiratory epithelium (RE) and scattered cells in the vomeronasal organ (VNO); Dmrt3 and Dmrt5 show an overlapping pattern of expression in the main olfactory epithelium (MOE), lateral RE (latRE) and ventral RE (vRE), but not in the VNO, where only Dmrt3 is found. Dmrt4 is strongly detected in the latRE and, like Dmrt3, it fills up the entire VNO. For a complete series of expression, see: https://www.informatics.jax.org/reference/J:364871. 3V: third ventricle. 4V: fourth ventricle. Aq: cerebral aqueduct. ASPall: subpallial amygdala. ChPl: choroid plexus. LPall: lateral pallium. m1-2: mesomeres 1 and 2. Mam: mammillary area. p1-3: prosomeres 1, 2, and 3. MnR (B5): median raphe nucleus (serotonergic B5 cell group). NA group (A4-7): noradrenergic A4-7 cell group. PHy: peduncular hypothalamus. PPaA: peduncular paraventricular area of the hypothalamus. PSB: pallial-subpallial boundary. RM (B3): raphe magnus nucleus (serotonergic B3 cell group). RMa: retromammillary area. RN: red nucleus. SC: superior colliculus. SN (A9): substantia nigra (dopaminergic A9 cell group). Th: thalamus. THy: terminal hypothalamus. TPaA: terminal paraventricular area of the hypothalamus. VPall: ventral pallium. VTA (A10): ventral tegmental area (dopaminergic A10 cell group)

Table 1.

Dmrt expression in E12.5 embryos

(a) ventricular zone; (b) ventricular and mantle zones. When (a), (b) or a region between brackets is not indicated, the expression is detected only in the mantle zone, which occurs mainly for Dmrt6. Dark gray boxes: DMA-Dmrts; N: negative expression; Y: positive expression. A positive expression does not necessarily mean an ubiquitous expression in the nucleus. *The region of the pallial-subpallial boundary is not considered in the Allen Developing Mouse Brain Atlas. †The classic term of “posterior hypothalamic area”—although obsolete according to the prosomeric model—is included in the table

Our comprehensive analysis revealed that Dmrt7 was not detected in the NS (at any of the examined developmental stages). In contrast, Dmrt2, Dmrt3, Dmrt4, Dmrt5, and Dmrt6 showed robust expression in the NS. Dmrt1 was dimly detected in the medial pallium (MPall) at E12.5 (Dmrt1 was not detected in adult animals) (Fig. 1A). Dmrt2 expression was restricted to the subpallial septum (SeSPall) and the preoptic area (PO) (Fig. 1B). However, it was strongly expressed in the respiratory epithelium (RE) and scattered cells in the vomeronasal organ (VNO) (Fig. 1V). Previous reports indicate that Dmrt6 is expressed in the mouse embryonic brain by RT-PCR [42], but, to our knowledge, no spatial detection has been conducted in the brain before. Here, we show that Dmrt6 is one of the most broadly expressed Dmrts in the mouse NS (Table 1), encompassing the three major parts of the brain (forebrain, midbrain, hindbrain) and the spinal cord (Fig. 1F-H, L, U). Dmrt6 expression was absent from proliferative ventricular zones (VZ), except in the MPall (Fig. 1F’) and serotonergic cell groups in the hindbrain (Fig. 1G, L). This expression pattern suggests a potential postmitotic role for Dmrt6.

The expression of DMA-Dmrts has been extensively described before, with special attention to the pallium [27] (Fig. 1C-E). In addition to the three DMA-Dmrts, Dmrt1 and Dmrt6 were transiently expressed in the MPall (Fig. 1A, F’), where the hippocampus and dentate gyrus (DG) originate from.

For the ChPl, we confirmed that Dmrt3 and Dmrt5 were robustly detected [52], and lacked Dmrt2, Dmrt4, and Dmrt6 expression (Fig. 1B, D, F). All DMA-Dmrts were redundantly expressed in the VZ of the prethalamic eminence (PThE) (Fig. 1H), and Dmrt6 was only found in the mantle zone (MZ) of the PThE. In addition, Dmrt4, Dmrt5, and Dmrt6 were expressed in the terminal paraventricular area of the hypothalamus (TPaA). The greatest extent of Dmrt5 in this area overlapped with a small ventral Dmrt4 territory, and a dorsal Dmrt6 territory (Fig. 1H). Moreover, Dmrt5 and Dmrt6 were extensively expressed in the peduncular paraventricular area of the hypothalamus (PPaA). Dmrt5 expression was very broad and found in the ventral part (or tegmentum) of the three diencephalic (p3, p2, p1) and the two mesencephalic (m1, m2) neuromeres. This included the diencephalic portion of the former posterior hypothalamic area (diPHA), the ventral tegmental area (VTA), the substantia nigra (SN), and the red nucleus (RN) (Fig. 1E). Dmrt6 was found in the same tegmentum of p3, p2, p1, m1 and m2, but, in addition, expanded to the dorsal part (or tectum) (Fig. 1F). Dmrt6 continued anteriorly beyond p3 (relative to the anteroposterior axis) in both the mammillary (Mam, within the terminal hypothalamus) and retromammillary (RMa, within the peduncular hypothalamus) areas (Fig. 1F). From here, Dmrt6 was a continuum until the hindbrain. Dmrt5 overlapped with Dmrt4 in the pre-isthmic tegmentum of m2 (PIsTg), where a few dopaminergic (DA) ventral mesencephalic neurons originate. In this region, Dmrt4 was expressed in early-born postmitotic DA—as it colocalized with tyrosine hydroxylase (TH)—and Dmrt5 laid dorsal to the Dmrt4+ cells in the VZ and the subventricular zone (SVZ), corresponding to DA progenitors (Fig. 1Q-T). Dmrt6 was also expressed in the PIsTg, covering a more extensive area and partially overlapping with the domains of Dmrt4 and Dmrt5 (Fig. 1U).

We also focused on the nasal and vomeronasal epithelia due to their relevance in rodent sex-typical behaviors [53]. Dmrt3 and Dmrt5 overlapped almost entirely in the main olfactory epithelium (MOE) and the RE (with Dmrt2). In the RE, Dmrt2 and Dmrt4 were expressed in the ventral and lateral parts, respectively. In the VNO, Dmrt2 presented a scattered pattern, while Dmrt3 and Dmrt4 filled up the whole structure (Fig. 1V). Previous studies have demonstrated that Dmrt can exhibit functional redundancy, often compensating for one another [31, 32, 35]. Therefore, exclusive sites of expression become significantly relevant in helping to identify the unique functions of each Dmrt. Dmrt3 was detected in the hindbrain (Fig. 1J), where it has been suggested to be expressed in the dB4 neurons of the hindbrain [54, 55]. However, to our knowledge, this assumption was based on the common developmental origin between dI6 and dB4 neuronal lineages and lacks a more in-depth analysis using molecular markers [55]. Dmrt6 was also present in the hindbrain but in the VZ of serotonergic nuclei (Fig. 1L) and did not occupy the same territory as Dmrt3. Lastly, Dmrt4 was expressed in the habenula (Fig. 1I) and the tegmentum of the pre-isthmus and isthmus (Fig. 1I’). However, similar to the other areas of the brain, Dmrt4 was only transiently expressed in these regions. No expression was detected after E14.5 (E14.5 expression is accessible at the MGI repository, https://www.informatics.jax.org/reference/J:364871).

Although this study did not explore tissues outside the NS, Dmrt5 expression in the developing pituitary gland was highlighted due to its relevance to hormonal secretion, including the control of sex hormone initiation during puberty. Dmrt5 was found from E12.5 until E18.5, as shown below (Fig. 1O, P, 5–7, Fig. S5). The postnatal expression of Dmrt5 in the pituitary gland was not explored in this work.

Fig. 5.

DMRT3 and DMRT5 in the postnatal neurogenic niches. A) Dmrt5 mRNA in the dorsal subventricular zone (SVZ) in adults (dashed line demarks the Dmrt5 area). B) DMRT3 and DMRT5 overlap in the DG—coronal section from a Light sheet scan obtained from an iDISCO-treated E18.5 brain. b) High magnification of the SGZ. Bright yellow dots are an artefact of the treatment. C, D) At P7, DMRT5 (IF) colocalizes with SOX2 in the hilus (i) and Reelin (RELN) in Cajal-Retzius (CR) cells (ii). In (C) DMRT5 IF, with nuclear Hoechst staining, shows the distribution of DMRT5 in the hippocampal formation. E) At P25, the inverted image shows that DMRT5 (cells shown in black, marked by arrowheads) persisted in the SGZ, overlapping with SOX2 in presumably NPCs (F). Insets (i) and (iii) show a high magnification of the SGZ. The white arrowheads depict DMRT5 +/SOX2 + colocalization. The orange arrowheads show SOX2 + cells outside the SGZ (potentially astrocytes) that do not express DMRT5. In ii and iv), a few CR cells, expressing both RELN and DMRT5, remain postnatally

Fig. 7.

Dmrt5 is required for the specification of the MOE in both sexes. A, B, and C) Dmrt5, Dmrt3, and Dmrt4 ISH in WT nasal epithelium at E14.5. A) Dmrt5 is not detected in Dmrt5^−/− nasal epithelium, and the MOE is reduced. B) In Dmrt5^−/−, Dmrt3 ISH does not show overt differences in expression, while Dmrt4 is ectopically expressed in the RE and upregulated in the MOE (C). (D, D’) SOX2 staining in basal and apical layers of the MOE in WT (D) and Dmrt5^−/− (D’). SOX2 cells are disorganized in the mutants. E, E’) TUJ1 expression in WT (E) and Dmrt5^−/− (E’). TUJ1 fills the entire MOE (a) and the VNO in WT conditions. In Dmrt5^−/−, the MOE is thinner (E’a’), although TUJ1 is maintained, and there is ectopic distribution in the RE (E’b). F, F’) OMP distribution is shown in inverted images. There is a significant reduction in the number of OMP neurons (quantification in (G)). In addition, there are ectopic OMP + neurons in the RE (red arrowheads). In (G) graphs represent the mean±SEM of the number of OMP neurons in one section of the unilateral MOE, area of the unilateral MOE (dashed red line), and the OMP + cell density. Each dot corresponds to a unilateral MOE from an individual embryo (n = 3 per experimental condition). Two-way ANOVA followed by Tukey multiple comparisons test: * p-value < 0.05; *** p-value < 0.001. B, basal; A, apical. Scale bars: 100 µm (A), 20 µm (D, a), 200 µm (E), and 100 µm (F).

The analysis of the whole family in the early embryo revealed a broad representation of these genes in the mouse NS. The expression of Dmrt1, Dmrt2, and Dmrt6 in the mouse NS was identified for the first time. At this early developmental stage, DMA-Dmrts were primarily found in progenitors in the VZ of distinct regions, including the pallium. Still, they also expanded to the ventral midbrain, where this early expression could be essential for its normal development.

Dmrt6 covers areas that control innate behaviors

Dmrt6 has been well characterized in the male gonad, particularly in spermatogonia, and, by RT-PCR, it has been detected in the adult ovary and the NS [41, 42]; however, the function of Dmrt6 in the brain remains unknown.

From the E12.5 analysis, we concluded that Dmrt6 was widely expressed in subcortical structures, including the limbic system, thalamic, and brainstem structures (https://www.informatics.jax.org/reference/J:364871). The limbic system comprises distinct neural circuits of different embryonic origins that control specific innate behaviors, which manifest without prior learning. Sensory stimuli significantly influence innate behaviors, and in rodents, olfactory cues are particularly crucial, modulated by intrinsic factors such as sex. Dmrt6 was present in the primary structures that comprise innate circuitry: the olfactory system, amygdala, bed nucleus of the stria terminalis (BST), and hypothalamus (Fig. 2A-J, pink-labeled nuclei). Table 2 complements Fig. 2, and compiles a complete list of the regions where Dmrt6 was maintained in late gestational embryos at E18.5 and postnatal P7 and P56 brains; Fig. S2).

Fig. 2.

Dmrt6 is maintained in the mouse nervous system, including postnatal brains. A-K) Dmrt6 ISH at E18.5. Regions in pink are associated with innate behaviors [56]; in green, with non-image visual circuits [59]; and in blue, with both innate behaviors and non-image visual circuits. H) Colocalization of Dmrt6 with TH in the VTA (H,h’) and SN (H,h’’). K) Dmrt6 expression in the ganglion cell layer (GCL) of the retina (higher magnification in k’). L) Dmrt6 ISH at P7 in the medial amygdala (Me) appears to occupy the Esr2 + region (shown in M). O-Q) Dmrt6 is restricted to the bed nucleus of the anterior commissure (BAC), bed nucleus of the stria terminalis (BST), triangular septum (TS), paraventricular nucleus of hypothalamus (PVH), supraoptic nucleus (SO),dorsomedial nucleus of hypothalamus (DMH), and zona incerta (ZI) in adult brains. ADT: anterodorsal nucleus of thalamus. AOB: accessory olfactory bulb. AOD: anterior olfactory area, dorsal part. AOV: anterior olfactory area, ventral part. APT: anterior pretectal nucleus. Co: cortical amygdaloid area. DLG: dorsal lateral geniculate nucleus. DMH: dorsomedial nucleus of hypothalamus. DR: dorsal raphe nucleus. DTg: dorsal tegmental nucleus. GCL: ganglion cell layer. GP: globus pallidus. HDB: horizontal nucleus of the diagonal band. IGL, intergeniculate leaflet. LPO: lateral preoptic nucleus. LS: lateral septal nucleus. MG: medial geniculate nucleus. MnR: median raphe nucleus. MPO: medial preoptic nucleus. MPT: medial pretectal nucleus. MS: medial septal nucleus. NBL: neuroblast cell layer. OPT: olivary pretectal nucleus. p2Tg: prosomere 2 tegmentum. p3Tg: prosomere 3 tegmentum. PAG: periaqueductal gray. Pir: piriform cortex. Pn: pontine reticular nucleus. PreG: pregeniculate nucleus. PVT: paraventricular nucleus of thalamus. Re: reuniens nucleus. RPE: retinal pigment epithelium. SC: superior colliculus. SIB: substantia innominate/basal nucleus. STh: subthalamic nucleus. VDB: vertical nucleus of the diagonal band. VL: ventral lateral nucleus of hypothalamus. VM: ventromedial nucleus of thalamus. VMH: ventromedial nucleus of hypothalamus

Table 2.

Dmrt6 expression across development (E18.5, P7 and adult)

If the Dmrt6 expression at E18.5 in a particular nucleus is maintained at P7 or adulthood, cells are colored with the same color as at E18.5 and without specifying the name of the nucleus. Empty white cells represent absence of Dmrt6 expression. Cells in white with “nd” (not determined) indicate that the particular nucleus has not been evaluated by ISH

At E18.5, as development progresses, Dmrt6 was detected in olfactory nuclei such as the accessory bulb (AOB), anterior olfactory nuclei (AOD, AOV), taenia tecta (TTe), and cortical amygdala (Co), one of the main limbic structures. Dmrt6 was located in the lateral (LS) and medial (MS) septal complexes, BST, and medial amygdala (Me), which integrates many rodent sex-typical behaviors and relays information from the olfactory system to the BST [56]. Postnatally, Dmrt6 coincided with Esr2 (estrogen receptor beta) in the same region, suggesting a potential interaction between them (Fig. 2L-M). Dmrt6 covered the PO (functionally included as part of the limbic system), and almost the whole hypothalamus, including the anterior (AH), dorsomedial (DMH), and ventromedial (VMH) hypothalamic nuclei. It overlapped with TH + neurons in the SN and the VTA (Fig. 2H), where it remained postnatally (Fig. S2). Lastly, the periaqueductal grey (PAG) is another critical region for innate behaviors where Dmrt6 was found.

Dmrt6 expression in areas of the central NS related to photosensation

At E18.5, Dmrt6 was located in the incipient ganglion cell layer (GCL) of the retina (Fig. 2K). From single-cell RNA sequencing data, Dmrt6 is probably found in the intrinsically photosensitive retinal ganglion cells (ipRGCs), which are rare mammalian photoreceptors essential for a wide array of visual functions, including circadian photoentrainment, pupillary light reflex, mood, and sleep, but not for image-forming visual functions [57–59]. Several articles have described ipRGCs in non-visual circuits, as demonstrated through tracing experiments in rodents. These retinal cells connect to several brain regions through the retino-hypothalamic tract [60], and Dmrt6 was present in many of them (with the critical exception of the suprachiasmatic nucleus, the central pacemaker of circadian rhythms) such as the lateral geniculate complex, including the dorsal lateral geniculate nucleus (DLG), intergeniculate leaflet (IGL), and pregeniculate nucleus (PreG) (also known as the ventral lateral geniculate nucleus); pretectal region, with the olivary pretectal nucleus (OPT), involved in pupillary light reflex; lateral habenula (LHb), involved in mood regulation; BST; LPO; hypothalamic nuclei, such as the supraoptic nucleus (SO); superior colliculus (SC), and PAG [59] (Fig. 2A-J, green- and blue-labeled nuclei).

Dmrt6 was strongly expressed in other thalamic nuclei, such as the reuniens nucleus (Re), paraventricular nucleus (PVA), ventral lateral thalamic nucleus (VL), and zona incerta (ZI) (Fig. 2A-J. Table 2). Another structure that stands out because of the strong Dmrt6 signal is the RN (Fig. 2G). In the hindbrain, Dmrt6 was maintained in the dorsal raphe (DR) and median raphe (MnR) nuclei, and pontine reticular nucleus (Pn) at E18.5 and P7 (Fig. 2J; Fig. S2). However, additional markers are needed to disambiguate Dmrt6 + cell identities in the hindbrain.

Our results, together with previous RT-PCR data [42], confirm that Dmrt6 is broadly expressed in the NS, including regions associated with innate behaviors. Despite this, NS-related behavioral abnormalities have not been described in a Dmrt6 null mouse model [41]. The Dmrt6 locus is highly complex with multiple transcriptional variants, some of which may lack the DM domain (NCBI gene ID: 56296; Fig. S1). Thus, the allele used in Zhang et al. [41] may not have disrupted the variant expressed in the NS, or Dmrt6 may influence subtle neuronal functions detectable only through specific behavioral assays, such as the pupillary light reflex test to interrogate photosensitivity.

Dmrt2 is maintained in the mouse NS

At E12.5, Dmrt2 exhibited a restricted expression (Fig. 1B). However, it began to be strongly present in many regions within the cortex from E13.5 on [61] (https://www.informatics.jax.org/reference/J:364871). Figure 3 shows brains at P7, although the same areas were identified at E18.5 and adulthood (Fig. S3; data not shown). According to previous studies, Dmrt2 is expressed in corticothalamic projection neurons (CThPNs), which are specifically excluded from the somatosensory cortical area [62, 63], although no other regions have been specifically characterized. Dmrt2 was found in the neocortex, specifically in the orbital (ORB), prelimbic (PLC), and infralimbic (Ila) areas, and the cingulate cortex (CCx) [61] extending into the retrosplenial cortex (RSP) (Fig. 3A). In addition, Dmrt2 was found in other areas of the cerebral cortex, such as the claustrum (Cl), insular cortex (IA), endopiriform nucleus (EPd) in a few scattered cells, and the entorhinal cortex (ENT) (also forming a continuum with perirhinal-ectorhinal cortex, PER-ECT). Lastly, Dmrt2 was detected in the cortical amygdala and the basomedial amygdaloid nucleus (BMA). Apart from the prominent presence in cortical regions, Dmrt2 was found in the lateral septal nucleus in its caudal part (LSc), in a few scattered cells located in the PAG (more obvious at E18.5, Fig. 3B), and in the pyramidal layer of the subiculum (SUB).

Fig. 3.

Dmrt2 expression in postnatal brains. Dmrt2 was detected with ISH. A) Expression at P7. Only one sex is shown. B) Expression in the periaqueductal gray (PAG) at E18.5 is more robust than at P7. C) Expression at P56. Arrowheads point to Dmrt2 + cells. Colocalization of Dmrt2 with somatostatin (SST) or parvalbumin (PV) GABAergic markers by IF in the cingulate cortex (CCx) (I), the neocortex (II), the dentate gyrus (DG), and the hippocampus (IV) of adult (P56) brains. White arrowheads indicate colocalization of Dmrt2 with SST, and orange arrowheads with PV. D, E) Dmrt2 expression levels in the CCx (Adapted from [61]) and PO throughout development. The graph shows mean±SEM in arbitrary units (au). Each dot corresponds to one individual embryo. Two-tailed t-test: * p-value < 0.05 (n ≥ 5). Images show CCx at E14.5 (adapted from [61]) (D) and PO at P56 (E). R, rostral; C, caudal. AL: agranula insular area. BLA: basolateral amygdala nucleus. BMA: basomedial amygdalar nucleus. CCx: cingulate cortex. Cla: claustrum. COA: cortical amygdalar area. DG: dentate gyrus. EPd: endopiriform cortex. HPF: hippocampal formation. Ila: infralimbic area. LSc: lateral septal nucleus. ORB: orbital cortex. PAG: periaqueductal gray. Pir: piriform area. PLC: prelimbic cortex. RSP: retrosplenial area. SUB: subiculum

In adult brains, Dmrt2 expression was evident in the hippocampal formation (HPF), CA1 field, and DG. This expression was more prominent in adults, although it was already detected in E18.5 and P7 (Fig. S3; data not shown). Additionally, Dmrt2 was strongly detected in somatostatin (SST) GABAergic interneurons, confirmed by Dmrt2 ISH on adult SSTCre;TdT brains (Fig. 3C). Most Dmrt2 + cells colocalized with SST, especially in the medial part of the neocortex, where almost every Dmrt2 + neuron was SST +. We found less colocalization between SST and Dmrt2 in the lateral cortex (Fig. 3Ci-iv), and no colocalization with parvalbumin (PV) anywhere, except for some anecdotal neurons. A similar situation was found for Dmrt2 neurons in CA1, where a few colocalize with SST and none with PV. Dmrt2 did not colocalize with any of these markers in the DG. This finding aligns with the Allen Brain Single Cell Database and https://scheiffele-splice.scicore.unibas.ch/.

Dmrt2 expression levels differ between the sexes in the CCx and PO

We have previously shown that Dmrt2 expression levels in the CCx are higher in males than females during embryonic development [61]. In contrast, this differential expression is lost in postnatal and adult brains (Fig. 3D). The expression of Dmrt2 in the PO of early embryos (Fig. 1B) and adult (Fig. 3E) prompted us to focus on this region. The PO is one of the areas in rodents associated with innate functions, such as sexual and parental behaviors, and presents molecular sex differences in cell numbers and dendritic density, which are controlled by gonadal hormones [64]. Moreover, Dmrt2 is expressed in only a few cells of the PO, suggesting that this expression may correspond to a single, discrete cell population. We dissected the POs of male and female embryos at E14.5, P7, and P56, and conducted RT-qPCR. Dmrt2 exhibited higher expression levels in adult males, but not at other developmental stages (Fig. 3E). We did not conduct cellular quantifications or identification. With our analysis, we cannot conclude whether there was a difference in Dmrt2 + cell numbers or higher levels of expression per cell.

Our findings indicate that distinct mechanisms may regulate Dmrt2 expression levels. In the CCx, Dmrt2 levels vary between sexes when hormonal secretion is low [65]. In contrast, in the PO—a region under strong regulation by sex hormones—Dmrt2 expression increases in males following puberty, coinciding with the establishment of hormonal secretion in sexually mature animals. Further research is required to elucidate the origins of these differential Dmrt2 expression patterns and their physiological significance. Unfortunately, the study of Dmrt2 loss-of-function mutants is limited to embryonic stages due to their perinatal lethality [66].

Persistent expression of DMA-Dmrts beyond early embryonic stages

DMA-Dmrts expression in the NS has been described in detail at early embryonic stages [23, 27, 29, 30, 33, 52] (Table 1, Fig. 1C 2E, https://www.informatics.jax.org/reference/J:364871). To our knowledge, whether their expression persists in any brain region has not been documented before. We found that among the DMA-Dmrts, only Dmrt3 and Dmrt5 were maintained, while Dmrt4 was no longer detected anywhere in the mouse brain after E14.5 (https://www.informatics.jax.org/reference/J:364871).

We highlighted the expression of DMA-Dmrts in non-neural structures, which are closely related to brain function. In the ChPl, Dmrt3 and Dmrt5 were strongly detected across development from early embryos until adults (Fig. 1C’-E’, Fig. 4K, L). Both TFs were localized to epithelial cells of the lateral ventricle ChPl, but not in those of the third or fourth ventricles, which have different developmental origins. It has been previously described that Dmrt5 is necessary for the formation of the ChPl, although this might be a consequence of Dmrt5 expression in the cortical hem [30, 52, 67, 68].

Fig. 4.

Dmrt5 is maintained in late embryonic and postnatal brains. Dmrt5 was detected with ISH, except in a’ and b’, where double DMRT5 IF and Vglut2 ISH was performed. A, B, C, and D) Sagittal views of Dmrt5 at E18.5. a-f’’ coronal views corresponding to levels indicated in (A), (B), (C), and (D). Arrowheads in Aa, B, Bb, Bc, C, Ce and I showing Dmrt5 in potential neurogenic niches. a’ and b’) DMRT5 in the lateral hypothalamus (LH) and the VTA in glutamatergic (VGluT2 +) neurons. b) Dmrt5 does not colocalize with TH in the VTA. c’) Dmrt5 colocalizes with BRN3A in the RN. D) At E18.5, Dmrt5 fills up the whole RMS and other olfactory system regions: the olfactory tuberculum (Tu), MOE, and Grueneberg ganglion (GG) (D, g), including the MOB (j’ and j’’). Dmrt5 perdures in postnatal OBs (E, F). In adults, it is restricted to a bead-like pattern that resembles the necklace glomeruli (F, h). Dmrt5 expression is maintained in other regions in adulthood such as the LH and substantia innominata (SI) (G); VTA (H); posterior hypothalamic area (PHA) (I), including the VZ of the third ventricle (black arrowhead in I); and dorsal raphe (DR) (J) (K and L) Dmrt5 and Dmrt3 persist in the ChPl

Dmrt5, but not Dmrt3, was also found in the adenohypophysis, presumably expressed in progenitor cells at E12.5 and E14.5 (Fig. 1O, P). At E18.5, Dmrt5 was detected in the marginal cell layer (MCL) of the intermediate lobe (IL), and in a few cells of the anterior lobe (AL) (Fig. 4Bb”). The role of Dmrt5 has been described in the zebrafish hypophysis, where Dmrt5 is required for the differentiation of corticotropes and gonadotropes, as well as for the maintenance of lactotropes, ultimately affecting reproductive function [69]. Whether this function is conserved in mammals is unknown. In line with the phenotype in zebrafish, we did not observe an overt morphological pituitary phenotype in Dmrt5^−/− mice (data not shown).

Among the three DMA-Dmrts, Dmrt5 exhibited the broadest expression pattern in early embryos (Table 1). It was maintained in many more nuclei in late embryonic and postnatal brains than Dmrt3 (Figs. 4, 5, 6, 7). Dmrt3 was also detected in adult DG and hindbrain (data not shown). Dmrt5 presented a scattered expression pattern in the PO and subpallial structures, such as the olfactory tuberculum (Tu) (Fig. 4C, D), the vertical nucleus of the diagonal band (VDB) (Fig. S4), and the SI (Fig. 4A, G). Furthermore, Dmrt5 was expressed in the island of Calleja major (ICjM), a postnatally born structure (Fig. S4), and the lateral hypothalamic area (LHA) (Fig. 4a’,C).

Fig. 6.

Dmrt3 and Dmrt5 in the MOE across development. A, B) Dmrt3 and Dmrt5 (ISH) are expressed in the MOE and VNO of P4 animals. C) Identification of DMRT5 cellular subtypes in the MOE (i, ii, and iii) and the VNO (iv, v, and vi) at E14.5. In the MOE, DMRT5 colocalized with SOX2 in the apical layer (magenta arrowhead) but to a lesser extent in the basal layer (blue arrowhead indicates a DMRT5 +/SOX2 + cell, and white arrowhead indicates a DMRT5 +/SOX2- cell) (i). In the intermediate zone, DMRT5 colocalized with TUJ1 (immature) and OMP (mature) sensory neurons (red arrowheads in ii and iii). In the VNO, DMRT5 does not overlap with SOX2 (white arrowhead), although it does with TUJ1 (v) and OMP (vi) (red arrowheads). D, E) DMRT5 is maintained in adult MOE. The Hoechst staining in (D) depicts the turbinates in an adult animal. E) Inset of one turbinate with DMRT5 staining robustly located in the sustentacular layer (magenta arrowhead). DMRT5 remains in a few cells in the basal layer where olfactory neural stem cells are located (e’), and in the “cul-de-sac”, there is a prominent concentration of DMRT5 cells (e’’). A: apical layer of sVNO; B: basal layer of sVNO; L: lumen. Scale bars: 200 µm (A), 20 µm (a’, e', i), 10 µm (a’’), 100 µm (C), 500 µm (D), and 50 µm (E).

Dmrt5 encompassed a plurisegmental region from the diencephalon to the midbrain tegmentum (Fig. 1E) [30, 37]. In early embryos, Dmrt5 was expressed in midbrain LMX1A + DA progenitors and extended laterally to the glutamatergic m6 domain (Fig. 1T) [37]. The continuous Dmrt5 + domain is largely glutamatergic (colocalizes with Vglut2) and persists through development until adulthood along the anteroposterior axis covering the diPHA (Fig. 4Bb, b’, H, I) and several nuclei such as the VTA, where Dmrt5 showed minimal overlap with DA (TH +) neurons (Fig. 4Bb); RN, overlapping with BRN3A + (POU4F1) neurons (Fig. 4Bcc’); and Edinger-Westphal nucleus (EW) (Fig. 4Bc).

In the ventral midbrain, Dmrt5 plays an early role in patterning, in addition to regulating DA identity at the expense of GABAergic neurons. Dmrt5 suppresses the transcriptional program controlling the ventral-lateral midbrain and promotes ventral-medial progenitor fate (Msx1, Lmx1a, Foxa2) [37]. It would be interesting to test whether Dmrt5 regulates DA neurons in vivo and their postmitotic function.

Dmrt5 is enriched in olfactory structures across development

Apart from MOE, VNO, and Tu, Dmrt5 was maintained in other olfactory structures, such as the Grueneberg ganglion (GG) (Fig. 4A, D), and the main olfactory bulb (MOB) (Fig. 4Df’, Df’’, E, Fh). In the MOB of the late embryo, Dmrt5 appeared in the SVZ, and in the granular (Gr) and glomerular (Glom) layers (Fig. 4Df’, Df’’). P7, MOB layers are fully formed, and Dmrt5 expression was maintained in the Glom. However, the expression is more restricted than at E18.5 (Fig. 4E). In the adult, expression of Dmrt5 in the Glom is limited to a few glomeruli of the ventrolateral zone of the MOB (Fig. 4Fh). The bead-like expression pattern resembles the necklace glomeruli. However, specific markers are needed to confirm this type of glomeruli, such as Gucy2d or Pde2a [70]. The embryonic function of Dmrt5 in the olfactory system will be approached later in this manuscript.

Dmrt3 and Dmrt5 persist in postnatal neurogenic niches

By analyzing late embryos and postnatal animals, we observed that Dmrt5, and Dmrt3 to a lesser degree, persisted in postnatal neurogenic niches, including the SVZ, the DG, and the MOE. Furthermore, Dmrt5 expression was detected in three niches lining the VZ of brain ventricles. Two of them were in the third ventricle: one rostral (Fig. 4Aa, B, black arrowheads), and other caudal (Fig. 4Bb, white arrowhead). The other niche was on the cerebral aqueduct (Fig. 4Bc, C, Ce, black arrowheads). In adults, the expression of Dmrt5 was only detected in the caudal niche of the third ventricle (Fig. 4I).

At E12.5, Dmrt3 and Dmrt5 are prominently expressed in neocortical progenitor cells in the pallial VZ (Fig. 1C, E) [27]. At E18.5, those progenitors are depleted; therefore, DMA-Dmrts expression was not detected in the cortex. However, the SVZ persists after VZ depletion [71]. Concomitantly, Dmrt5 expression remained in the dorsal and septal SVZ (not the striatal) in its entire anteroposterior extension (Fig. 4A, D), following the rostral migratory stream (RMS) (Fig. 4D). In the OB, Dmrt5 filled the olfactory ventricle and the SVZ. We could still detect Dmrt5 mRNA (but not the protein) in adult animals in the dorsal SVZ (Fig. 5A, dashed line demarked by black arrowheads).

A second prominent site of postnatal neurogenesis is the subgranular zone (SGZ) of the DG, where progenitor cells persist into adulthood in the mouse brain [72]. DMRT3 and DMRT5 proteins overlapped in the E18.5 DG (Fig. 5B). At P7, DMRT5 colocalized with SOX2, a marker for neural stem cells and progenitor cells in the hilus [73], corresponding to the tertiary dentate matrix (Fig. 5C, D). At P25, SOX2 +/DMRT5 + cells are absent from the hilus and confined in the SGZ (Fig. 5D, E, black arrowheads). In addition, at P7 and P25, DMRT5 colocalized in the DG with reelin (RELN), a marker for Cajal-Retzius cells (CR) (Fig. 5ii, iv). Dmrt5 is crucial for the generation of CR in the medial pallium [30]. In Dmrt5^−/− mutants, there is a complete absence of cortical hem-derived CRs (the ones that eventually populate the postnatal hippocampus) and those from the pallia-subpallial boundary (PSB) [29, 30, 52]. We observed a decline in Dmrt5 expression from P7 to P25 in the SGZ, which might correlate with the reported reduction of postnatal neural progenitors over time [74] (Fig. 5C, E, respectively).

Another neural structure that can regenerate throughout life is the MOE [75, 76]. Dmrt3 and Dmrt5 were postnatally maintained there, showing a strong signal in the basal layer, where olfactory neural stem cells (oNSCs) are lifelong maintained, and in the sustentacular cells of the apical layer [77] (Fig. 6A, B). In addition, Dmrt3 was found in the apical layer of the sensory epithelium of the VNO (sVNO), the marginal zone, and a few cells of the basal layer of the sVNO (Fig. 6Aa’’). The marginal zone is located at the outer edges of the sVNO, where it interfaces with the non-sensory epithelium of the VNO (nsVNO), constituting the lifelong stem cell niche of the VNO [78, 79].

We characterized DMRT5 + cell types in the MOE through immunostaining (Fig. 6C, D). DMRT5 colocalized with SOX2 + oNSCs, with a few DMRT5 +/SOX2- cells likely corresponding to differentiating intermediate progenitors (Fig. 6Ci, white arrow). DMRT5 and SOX2 co-localization was also observed in the sustentacular cells (Fig. 6Ci, magenta arrow). Additionally, scattered DMRT5 + cells in the intermediate layer corresponded to immature (TUJ1 +/OMP-) and mature (OMP +) olfactory sensory neurons (iOSNs and mOSNs, respectively; Fig. 6Cii, Ciii, red arrows). This expression pattern persisted into adulthood (Fig. 6D), where DMRT5 continued to mark the sustentacular cells (Fig. 6E, indicated by the magenta arrow). In addition, a few scattered cells laid within the basal layer (Fig. 6Ee’, white arrow). DMRT5 was also found in the curved portions of the regions called “cul-de-sacs” (Fig. 6Ee’’), which are blind-ended recesses within the posterior end of the nasal cavity that contain non-canonical OSNs defined by the expression of the transmembrane receptor guanylate cyclase D and often exhibit a low level of OMP [80]. Their axons project to the necklace glomeruli of the olfactory bulb [81, 82], where Dmrt5 has also been detected in this study (Fig. 4g’). They detect social cues, such as the peptide uroguanylin and the gas carbon disulfide, to mediate food-related social learning [83–85].

Similar to the MOE, the embryonic VNO exhibited scattered DMRT5 + immature and mature olfactory sensory neurons in the sensory epithelium (Fig. 6Bb’’, Civ-vi). However, unlike the MOE, DMRT5 was absent from the sustentacular cells of the VNO. The non-sensory VNO epithelium contains DMRT5 + cells that colocalized with TUJ1 and OMP (Fig. 6Cv and Cvi, respectively), though their function remains undetermined.

We showed that DMRT5 protein was postnatally maintained in well-defined neurogenic sites: the basal progenitors of the MOE, SVZ, and SGZ of the DG, and also in potential neurogenic niches lining the VZ of the third and fourth ventricles. The colocalization of DMRT5 and SOX2 in the SGZ places DMRT5 as a novel marker of neural progenitor cells (NPCs) [73].

Previous studies have shown that Dmrt5 sustains progenitor proliferation by regulating Hes1 and cell cycle genes, and its loss strongly impairs cerebral cortex development, leading to a total agenesis of the hippocampus and a reduced caudomedial neocortex [33, 34]. Similar phenotypes have also been observed in zebrafish, indicating a strong conservation across different species [86, 87]. Double-knockout mice for Dmrt3 and Dmrt5 exhibit a more severe reduction of the caudomedial neocortex than single knockouts [32]. If similar mechanisms were to be employed in postnatal NPCs, Dmrt3 and Dmrt5 might be maintaining postnatal progenitors in a quiescent state. It would be interesting to investigate in future experiments how they contribute to adult neurogenesis and whether they might integrate sexual cues, such as Dmrts in Drosophila and C. elegans [88].

Dmrt5 is required for the development of several components of the olfactory system

The olfactory system plays a crucial role in regulating rodent innate behaviors, which often exhibit pronounced sex differences [89]. Given the robust expression of Dmrt5 in multiple olfactory nuclei across various levels of information processing, we were prompted to investigate its function in specifying the olfactory system in both males and females at distinct circuit levels.

The olfactory system is the only sensory system in which the information reaches the primary sensory cortex without passing through the thalamus. Sensory neurons in the MOE and VNO detect pheromones and other social cues. Then, these neurons pass the information on to OB neurons that project to the olfactory cortex, which includes the Pir and the Tu. Previous data in Xenopus laevis indicate that Dmrt5 is necessary for neurogenesis in the nasal placode [90], but this has never been assessed in mammals. For our analysis, we used Dmrt5 null mutants kindly provided by Prof. Bellefroid (Dmrt5^−/−) (Fig. S1) [30]. The perinatal lethality of Dmrt5^−/− homozygous mice limited our study to embryonic stages.

Dmrt5 is required for the general olfactory epithelium specification

DMA-Dmrts could compensate for each other functionally [32, 90]. DMRT3 and DMRT5 have similar DNA-binding properties, and in cortical development, they have a synergic effect on their target genes [31, 32]. Therefore, we first established the epistatic relationship between the three DMA-Dmrts in the nasal epithelium by analyzing Dmrt3 and Dmrt4 in E14.5 Dmrt5^−/− mutants. At E14.5, Dmrt3 and Dmrt5 overlapped completely in the nasal epithelium (including the MOE and the RE), while Dmrt4 was only dimly detected in the lateral part of the RE. Dmrt3 and Dmrt4 showed a similar expression pattern in the VNO: filling up the entire apical layer and the marginal zone. In contrast, their expression was more restricted in the basal layer, being localized to the most basal cells (Fig. 1V, 8A-C). In Dmrt5^−/− mutants, Dmrt4 expression was ectopically expressed in the MOE and upregulated in the RE in both sexes (Fig. 8C’), while Dmrt3 showed an expression similar to that of the wildtype (WT), both in male and female (Fig. 8B’). No gross defects were found in the VNO for Dmrt3 or Dmrt4. These results suggest that Dmrt5 is repressing Dmrt4 expression in WT conditions in both sexes, following a similar logic to that of the pallium [31], although the sex was not evaluated as a variable.

Fig. 8.

Density of OB-TH neurons and number of Pir-TH neurons increase in Dmrt5^−/− female embryos. A, B) TH antibody staining in OB slices of WTs and Dmrt5^−/− in both sexes at E18.5. The Glom layer is thicker in the mutants. C-E) Quantification of OB-TH neurons. The graphs represent quantifications expressed as mean±SEM from whole OBs in 3D-claired brain reconstructions. C) Total number of TH + neurons, D) OB volume, E) Density of TH neurons (number of TH/Volume). G-J) TH IF in E18.5 brain slices in the Pir and neocortex of WTs and Dmrt5^−/− in both sexes. Ectopic TH neurons are detected in Dmrt5^−/− mutant neocortices in both sexes (g’-j’). F) There is a higher increase in Pir-TH neurons in Dmrt5^−/− in female embryos than in males (g’’-j’’). The graph represents mean±SEM from whole 3D-claired brain reconstructions. At least 3 animals were used per experimental condition. Each dot corresponds to a hemibrain except in panel F, where each dot represents the average of the two hemibrains. Two-way ANOVA followed by Tukey multiple comparisons test: * p-value < 0.05, ** p-value < 0.01; *** p-value < 0.001; **** p-value < 0.0001. Scale bars: 100 µm (A), 200 µm (G), and 10 µm (g’)

In Dmrt5^−/− mutants, the MOE area was reduced, likely due to a decrease in the total number of cells (Fig. 7F, G). In WTs, SOX2 + cells were arranged in basal and apical layers (in neural progenitors and sustentacular cells, respectively). In Dmrt5^−/− mutants, SOX2 + cells were randomly distributed across the epithelium, indicating that the MOE lost its organized layer stratification. SOX2 protein distribution was not abolished; however, the apical layer presented fewer SOX2 + cells (Fig. 7D’). Furthermore, in Dmrt5^−/− mutants, the neuronal markers TUJ1 + (for both iOSNs and mOSNs) and OMP + (specific for mOSNs) were ectopically located in the RE, in its lateral and ventral parts (Fig. 7E’b, F’). Next, we measured the mOSN density in both sexes (number of mOSN divided by the MOE area depicted in dashed red lines in Fig. 7F, F’). Both the total number of mOSNs and the MOE area were reduced in mutants of both sexes. As a result, the mOSN density was not different between genotypes or sexes, indicating that Dmrt5 affected the overall MOE specification similarly in males and females, and did not specifically regulate sensory neurons in the MOE (Fig. 7G).

These results suggest that Dmrt5 is required for the mouse MOE specification, with a similar effect in both sexes. As a result, neuronal and non-neuronal (such as sustentacular cells) cell types are reduced in Dmrt5^−/− mutants. Previous studies in Xenopus have shown that XDmrt5 and XDmrt4 may function redundantly upstream of the pro-neural genes XNgnr1 and XEbf2 in olfactory placode neurogenesis [90]. The ectopic expression of Dmrt4 and the continued presence of Dmrt3 in the MOE may together compensate for the absence of Dmrt5. To further elucidate the role of Dmrt5, future studies employing specific molecular markers and epistasis analyses with established TFs involved in MOE-specific processes are essential.

Sex and Dmrt5 interact to regulate the number of DA neurons in the OB and Pir

In early embryos, Dmrt5 (together with Dmrt3 and Dmrt4) determines the position and size of the PSB [33] from where OB interneurons (OB-i, TH +) and OB projection neurons (OB-p, Calretinin +) derive. In fact, Dmrt5^−/− embryos show an expansion of the OB-i population [52]. To investigate the potential interaction between Dmrt5 and sex in the OB, we performed 3D brain reconstructions in clarified brains (by iDISCO +, see Experimental Procedures) stained with an anti-TH antibody. This technique allowed us to precisely measure the number of TH interneurons normalized by the OB volume (Fig. 8A-E). The number of OB-TH neurons was similar across all experimental conditions (WTs and KOs of both sexes) (Fig. 8C). KOs presented a significant reduction in OB volume in both males (adj p < 0.0001) and females (adj p = 0.0001). However, the OB volume did not show sex differences either in WTs (adj p = 0.6430) or KOs (adj p = 0.4609) (Fig. 8D). The decrease in OB volume in KOs is consistent with previous studies [30, 52]. The density of OB-TH neurons did not present sex differences in WTs (adj p = 0.9995). However, when comparing KOs vs. WTs, the density of OB-TH neurons increased significantly in females (adj p = 0.0016) but not in males (adj p = 0.5037). Consequently, the density of OB-TH neurons exhibited sex differences in KOs (adj p = 0.0172), with females displaying 1.97-fold more density than males (Fig. 8E), meaning that sex differences emerge in the absence of Dmrt5. Therefore, Dmrt5 suppresses the sex differences in the density of OB-TH neurons. Regarding the density of OB-TH neurons, there was a significant interaction between Dmrt5 and sex (accounting for 11.17% of the total variance, p = 0.0261). Dmrt5 explained a notable part of the variance (34.35%, p = 0.0003), and sex contributed to a lesser extent to the variance (12.8%, p = 0.0189).

Taking advantage of the TH staining in the clarified brains, we quantified the effect of Dmrt5 in a group of TH + neurons located within the Pir, at the boundary between Pir and the Tu (Fig. 8G-J). These DA neurons, of unknown function, were PAX6 +, which suggests their PSB origin (Fig. S5). They have been previously reported to be derived from the PSB, where Dmrt5 and Dmrt3 are strongly expressed within the Pax6 territory [33, 91]. The number of Pir-TH in WT embryos did not show sex differences (p = 0.9702). Dmrt5^−/− embryos presented a significant increase in Pir-TH neurons in both male (adj p = 0.0018) and female (adj p < 0.0001) embryos compared to WT embryos. Moreover, Dmrt5^−/− females presented significantly 1.53-fold more Pir-TH neurons than Dmrt5^−/− males (adj p = 0.0181) (Fig. 8F). This indicates that Dmrt5 is suppressing the sex differences in the number of Pir-TH neurons. There was a significant interaction between Dmrt5 and sex (accounting for 4.45% of the total variance in Pir-TH neuron numbers, p = 0.0384). Dmrt5 explained the majority of the variance in (82.78%, p < 0.0001), whereas sex contributed modestly to the variance (6.96%, p = 0.0148). In addition to the Pir-TH neurons, we found that in Dmrt5^−/− embryos of both sexes, TH/PAX6 + cells appeared ectopically at the border of the neocortex and the Pir (Fig. 8H, J). These neurons were never detected in WT brains and appeared in both sexes.

These results indicate an interaction between Dmrt5 and sex in DA neurons from olfactory structures. These neurons originate from the PSB domain, where all DMA-Dmrts are expressed in early embryos (Fig. 1C’−1E’). Depletion of Dmrt5 leads to the expansion of the PSB more dorsally by converting the identity of cortical progenitors to that of the PSB [32, 33]. As a result, the persistent expression of PAX6 and an increase in OB-i were observed as they migrate to form the olfactory cortex. No effect was observed for OB-p neurons.

It would be interesting to check if there are any sex differences in DMA-Dmrt expression levels in the PSB. The precise control of DMA-Dmrts dosage is crucial for cortical patterning, specifically in the PSB. Therefore, intrinsic differences in their expression between the sexes could have a significant effect.

Concluding remarks

The comprehensive analysis of Dmrt gene expression establishes the framework for understanding how this TF family contributes to mammalian brain development. We provide the first evidence of Dmrt1 in the mouse brain and confirm robust expression of Dmrt2 and Dmrt6, expanding their known roles beyond the gonadal and peripheral organs. The spatial and temporal distribution observed here suggests that Dmrts act in both unique domains to specify discrete neuronal populations and in overlapping regions where multiple family members may interact cooperatively.

Several regions emerged as particularly relevant. Multiple Dmrts colocalize in the PThE, a source of migrating neurons targeting brain regions such as the PO and hypothalamus, essential for innate behaviors [92]. In the MPall, most Dmrts, except Dmrt2, are coexpressed in progenitors that give rise to the hippocampus. Synergistic functions of Dmrt3 and Dmrt5 are well established here [33], but the presence of Dmrt1 and Dmrt6 suggests additional roles yet to be defined. Interestingly, Dmrt2, undetected in hippocampal progenitors, appears in DG interneurons, involved in hippocampal plasticity [93]. The ventral midbrain also shows enriched Dmrt expression: Dmrt5 in progenitors, Dmrt4 in postmitotic DA neurons, and Dmrt6 in distinct mantle regions. Such a spatial configuration raises the possibility of epistatic interactions regulating DA identity specification.

A key question in this study was whether these genes are sexually dimorphically expressed in the NS, given the evident dimorphism of Dmrt1 in the gonad and multiple Dmrts in the invertebrate NS [3, 106]. At the whole mR

NA level, determined by ISH, we observed broadly similar expression between sexes, with only quantitative differences, such as higher Dmrt2 expression levels in male CCx primordium [61] or adult PO. Nonetheless, we did not delve into the intricate mechanisms of gene regulation, such as alternative splicing and post-translational regulation, which can also generate sex bias. The presence of splice variants predicted for Dmrt2, Dmrt4, and Dmrt6, and the precedent of sex-specific degradation through the DMA domain in C. elegans [94], suggests additional layers of complexity. Functional analyses already suggest that Dmrts can buffer latent sex differences: knockdown of Dmrt2 reveals male-biased effects on progenitor proliferation in the CCx [61]. Here, we showed the absence of Dmrt5 unmasks sex differences in the density of OB-TH neurons and in the number of Pir-TH neurons. Thus, Dmrts may serve as regulators of hidden sexual programs, only revealed under perturbation. Our results also highlight roles that extend beyond embryogenesis. Most Dmrts, except Dmrt1 and Dmrt4, persist into postnatal stages. The postnatal function of Dmrts has received limited attention, in part because some Dmrt null mutants do not survive past the perinatal period. Moreover, sex, as a biological variable, is frequently overlooked in neuroscience research [95]. Dmrt5 expression in SOX2 + progenitors in postnatal neurogenic niches strongly suggests that Dmrt5 participates in the regulation of neural stem cell activity. Given known sex differences in hippocampal neurogenesis [96], including estrogen-driven proliferation in females [97] and testosterone-sensitive responses in males [98], Dmrts could mediate intersections between hormonal signals and intrinsic cell-cycle regulation. The translational implications are notable. Human deletions at the DMRT1–3 locus are associated with intellectual disability [44, 45, 99], while DMRTA2/DMRT5 mutations cause microcephaly in both humans and mice [68, 100], demonstrating evolutionary conservation. Similar mechanisms to those observed in the increase of OB-TH neuronal density and Pir-TH neuron number in Dmrt5 mutants, where the interaction between Dmrt5 and the sex of the organism modulates development, may operate in the hippocampus, hypothalamus, amygdala, and midbrain dopaminergic zones. This raises the possibility that dysregulated DMRT function contributes to sex differences in psychiatric etiology and symptomatology.

Looking ahead, several avenues stand out. First, advancing from transcriptional mapping to multi-omics integration—including proteomics and splicing analysis—will uncover hidden biases in Dmrt regulation. Second, conditional knockouts will be key for elucidating postnatal roles, given the perinatal lethality of some null mutants. Finally, dissecting how Dmrts intersect with sex hormones and their receptors may unravel molecular links between developmental gene networks and sexual cues.

Despite limited overt transcriptional dimorphism, Dmrts exhibit complex spatial and temporal expression patterns, positioning them as critical regulators of brain development. Alternative splicing, protein regulation, and gene–gene interactions likely encode latent sex differences that shape neurogenesis, circuit formation, and disease susceptibility. By extending our understanding of DMRT biology into postnatal and human contexts, future research can uncover their contributions to cognitive functions and psychiatric disorders, with important implications for sex-specific vulnerabilities in brain disease.

Experimental methods

Animals and tissue preparation

Animals were housed and bred at the CBM animal facility and treated in accordance with European Communities Council Directive 86/609/EEC. All procedures were approved by the Bioethics Subcommittee of CSIC (Madrid, Spain) and the Comunidad de Madrid under the PROEX 286/19; RD 53/2013.

Animals used in this study include male and female C57BL/6JRccHsd (ENVIGO) and C57BL/6J (The Jackson Laboratory). Heterozygous mice B6J.Cg-Sst^{tm2.1(cre)Zjh}/MwarJ (028864, The Jackson Laboratory) and B6.129P2-Pval^{btm1(cre)Arbr}/J (017320, The Jackson Laboratory) [101] were crossed with the reporter floxed mice tdTomato B6;129S6-Gt(ROSA)26Sor^{tm14(CAG−tdTomato)Hze}/J (007914, The Jackson Laboratory).

For embryonic sex genotyping, tail tissue was collected and digested in 50 μl of MgCl₂-free PCR reaction buffer (10013–4104, Biotools) supplemented with 400 μg/ml proteinase K (3115879001, Roche) at 55°C with shaking for 1–3 h (h), followed by heat-inactivation of the proteinase K at 95°C for 5 min. The protocol of genomic DNA amplification was adapted from McFarlane et al. [102]. The primers used were: Sx-Forward 5’-GATGATTTGAGTGGAAATGTGAGGTA−3’ and Sx-Reverse 5’-CTTATGTTTATAGGCATGCACCATGTA−3’.

The presence of a vaginal plug estimated embryonic day (E) 0.5 of pregnancy after overnight male mating. Embryos were collected from independent litters. For staining on slices, heads from embryos (E11.5, E12.5, E13.5, E14.5, E15.5, and E18.5) were fixed in 4% paraformaldehyde (PFA) (1.04005.1000, Merck) diluted in phosphate-buffered saline (PBS) for 2 h at room temperature (RT). Postnatal day (P) 4, P6 or P7 heads were fixed in 4% PFA overnight (o/n) at 4°C. Adult brains (P56) and adult noses (P71) were fixed in 4% PFA o/n at 4°C from animals previously perfused intracardially with sterile PBS followed by 4% PFA. After fixation, samples were washed several times in PBS. Adult noses were decalcified in 0.25 M EDTA for 3 days at 4 °C and, after that, washed several times in PBS All samples were cryoprotected in 15%−30% sucrose (1.07651.1000, Merck) in PBS. Lastly, they were embedded in 7.5% gelatin (G2625, Sigma)/15% sucrose (1.07651.1000, Merck)/PBS and frozen at −80 °C. Coronal and sagittal cryostat sections were produced at 16 μm thickness.

In situ hybridization (ISH)

Tissue sections were hybridized with digoxigenin-labeled riboprobes as described in a previous work with minor modifications [103]. In the case of E12.5 embryos, the analysis was performed in coronal and sagittal planes on consecutive sections for at least three embryos of both sexes. In the case of postnatal tissue from P25 onward, sections were firstly treated with proteinase K (10 μg/ml; Merck, 3115879001) for 5 min, washed in PBS (pH 7.4), and post-fixed for 10 min in 4% PFA. For embryonic and neonatal tissues, the procedure began directly with post-fixation for 10 min in 4% PFA. After fixation, sections were washed in PBS (pH 7.4), acetylated, and washed in PBS with 0.1% Triton X-100. Then, sections were incubated for 1 h at RT with hybridization buffer [50% deionized formamide (Millipore, S4117), 1X salts (pH 7.5, 10 mM Tris, 200 mM NaCl, 5 mM NaH₂PO₄, 5 mM Na₂HPO₄, 5 mM EDTA), 1X Denhardt’s (Merck, D2532), 10% dextran sulphate (Merck, 4911), tRNA 1 mg/ml (Merck, R6625)] in a humified chamber with 5X SSC (pH 4.5, 750 mM sodium citrate, 75 mM NaCl) and 50% deionized formamide (Millipore, S4117). Then, sections were hybridized o/n at 72°C with riboprobes (1:1000 in hybridization buffer). After hybridization, sections were washed for 90 min in 0.2X SSC at 72 °C, and blocked in 2% blocking solution (Roche, 11096176001) in 1X MABT [pH 7.5, 100 mM maleic acid, 150 mM NaCl, 0.1% Tween-20 (Merck, P1379)] for 1 h at RT. After blocking, sections were incubated o/n at RT with anti-DIG antibody conjugated with the alkaline phosphatase (1:5000, Roche, 11093274910) in MABT. After extensive washing in TBST (100 mM Tris, 150 mM NaCl, 0.1% Tween-20) at pH 7.5 and then in TBST at pH 9.5, the alkaline phosphatase activity was detected at RT by using 100 μg/ml 4-nitro blue tetrazolium chloride (Roche, 11383213001) and 175 μg/ml 5-bromo-4-chloro-3-indolyl phosphate p-toluidine salt (Roche, 11383221001) in NTMT (pH 9.5, 100 mM Tris, 100 mM NaCl, 50 mM MgCl₂, 0.1% Tween-20). Sections were mounted in Mowiol (Merck, 81381). At least three biological replicates per sex were used for ISH performed on embryos, and one biological replicate was used for ISH on postnatal mice. Riboprobes were generated in this study de novo for Dmrt6, Dmrt7, and GlyT2 genes (Fig. S1; Suppl. Table 1). The oligos used to isolate the cDNA sequences were the following: for Dmrt6 (5’-ATGCTTCGCGCCCCCAAGTG−3’ and 5’-TGCTCCTGGGGAGAGTCCTG−3’); for Dmrt7 (5’-TGCTCTCCACCACTGTTCTG−3’ and 5’-GCTACAAGAATAGCCACCCTCAG−3’); for Esr2 (5’-GCCGACTTCGCAAGTGTTAC−3’ and 5’-GCTTCGAGGGTACAAGTCCTCA−3’), and for GlyT2 (5’-ACAGCGCACAGGCCAATCC−3’ and 5’-GTTCGGATAGGCAGTCATGCAG−3’). The rest of the riboprobes have been published before and kindly donated by Prof. Eric Bellefroid for Dmrt1, Dmrt3, and Dmrt5; Dr. Nuria Flames for Vglut2 [104] and Gad65, Dr. Takahiko Sato for Dmrt2, and Dr. Dajiro Konno for Dmrt4 (Fig. S1; Suppl. Table 1).

Immunofluorescent staining (IF)

IF was performed following standard protocols. Brain sections were washed with PBS with 0.1% Triton X-100 (T9284, Sigma-Aldrich) (PBST) and blocked for 1 h in blocking solution [PBST and 1% of bovine serum albumin (BSA) (A2153-100G, Sigma)]. Primary antibodies were diluted in a 1/10 blocking solution and incubated o/n at 4 °C (see Supp. Table 2); the corresponding secondary antibodies were incubated at 1:1000 for 2 h at RT. Hoechst 33,342 (1/1000) (H1399, Invitrogen™) was used for cell nuclei staining. For double ISH or FISH and IF staining, ISH was performed before IF, as previously described [103].

Imaging

Chromogenic ISH images were acquired on a Leica DM5000B vertical microscope equipped with a Leica DFC500 color camera. IF images were acquired on the previous Olympus SpinSR10 confocal, a Leica DM5000B vertical microscope fitted with a Leica DFC350 FX monochrome camera, or a Zeiss LSM900 confocal coupled to a vertical Axio Imager 2 microscope.

Immunolabeling-enabled three-dimensional imaging of solvent-cleared organ plus (iDISCO+)

Tissue preparation and whole-mount immunolabeling were performed using the iDISCO+ protocol, optimized for embryonic tissues [105]. E18.5 embryos were collected in ice-cold Leibovitz L-15 medium or PBS, and allowed to drain blood for 5 min on ice before fixation in 4% PFA at 4°C o/n with agitation.

Fixed samples were pretreated with methanol for tissue dehydration and bleaching. This involved a graded methanol series (20–100%, 1 h each at RT with shaking) followed by o/n incubation in a 66% dichloromethane (DCM) (5.89581, Sigma-Aldrich)/33% methanol (2:1 v/v; 1.06009, Supelco) solution with shaking. After additional washes in 100% methanol, samples were bleached in freshly prepared 5% hydrogen peroxide (H₂O₂)/methanol o/n at 4 °C. Following rehydration through a reverse methanol gradient to PBS, samples were washed twice in PTx.2 buffer (2% Triton X-100/PBS) for 1 h each at RT with shaking.

Pre-treated samples are permeabilized using a solution containing 0.3 g/mL Glycine and 20% DMSO (472301, Sigma-Aldrich) in PTx.2 at 37 °C with shaking for up to two days. Samples were then blocked using a 6% Donkey Serum (017-000-121, Jackson Immunoresearch) and 10% DMSO/PTx.2 blocking buffer at 37 °C with shaking for up to two days. Samples were incubated with the primary antibody in PTwH (10–50 µg/ml Heparin sodium salt in 0.2% Tween20/PBS) supplemented with 5% DMSO and 3% donkey serum for three (E18.5 brains out of skull) or four (E18.5 embryos head) days at 37°C with shaking. Rabbit anti-TH at 1:1000 (Merck; AB152) was used; For DMRT3, a guinea pig anti-DMRT3 at 1:400 (a gift from Irene Miguel-Aliaga) was applied. Subsequently, samples were extensively washed in PTwH and left in PTwH o/n at 4°C before proceeding with secondary antibody incubation. The secondary antibodies used were Donkey anti-Rabbit IgG (H + L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor 647 (at 1:500; Jackson ImmunoResearch; 711-605-152. Secondary antibodies were applied under similar conditions to primary antibodies for three or four days. Antibody solutions were filtered through a 0.2 µm membrane or centrifuged at 20,000 × g for 10 min before use to minimize the formation of precipitate. After several washes in PTwH, clearing was achieved by dehydrating samples through a methanol gradient followed by incubation in DCM/Methanol (2:1 v/v) for 3 h, and two additional 15-min incubations in 100% DCM. Final optical clearing was performed in dibenzyl ether (DBE) (108014, Sigma-Aldrich), with samples fully immersed in DBE-filled tubes to minimize oxidative damage.

Cleared samples were placed in an imaging reservoir made of 100% quartz filled with ethyl cinnamate (ECi) (W243000, Sigma-Aldrich) and illuminated from the side by a laser light using a LaVision Biotec Ultramicroscope II (Miltenyi Biotec, Germany)/ImspectorPro (LaVision BioTec) software. For DMRT3, ECi was used to fill the sample chamber, and cleared samples were fixed to the sample holder using UV adhesive (Bondic, BC5000) for 3D imaging using a ZEISS Lightsheet 7 (Carl Zeiss, Germany)/Zeiss Zen Black (Carl Zeiss, Germany) software.

3D images and their subsequent processing were carried out using Imaris × 64 software (version 9.3.1, Bitplane). Zeiss Arivis Pro (version 4.3.0, Carl Zeiss) for DMRT3 DG stainings. The different volume reconstructions (OB and Pir) were generated manually by creating a mask using the “surface tool” and they were pseudo-colored. The TH + cell number was obtained using the “spots tool”, which was applied to each of the regions under study, and was conducted on both brain hemispheres to enable the detection of potential interhemispheric differences in marker expression, structural organization, or cellular distribution. This approach enhances the sensitivity of the analysis to subtle anatomical or molecular differences that may be obscured in unilateral assessments. It ensures that any hemispheric bias in staining, labeling, or tissue handling does not compromise the accuracy of the results.

RNA isolation and qRT-PCR

Total RNA was isolated using NZY Total RNA Isolation kit (MB13402, Nzytech) following the manufacturer’s instructions. cDNA was obtained from 1 μg of total RNA with First-Strand cDNA Synthesis kit (27-9261-01, GE) in a 15 μl reaction volume. qPCR was performed using GoTaq® qPCR Master Mix (A6002, Promega) according to the manufacturer’s protocol in a CFX-384 Touch Real-Time PCR Detection System (Bio-Rad). The following gene-specific primer pairs were used to detect Dmrt2: Dmrt2-Forward 5’-CCTGAGCCTGGTTCTTG−3’, Dmrt2-Reverse 5’- CAGTGGAGTCCCACAGCTATC−3’. For housekeeping, Gus-Forward 5’-AGCCGCTACGGGAGTCG−3’, and Gus-Reverse 5’-GCTGCTTCTTGGGTGATGTCA-3’. These primers were previously validated to be invariable between sexes and developmental stages. Dmrt2 and gfp expression levels were quantified in triplicate and normalized to Gus levels. The data were analyzed using the comparative Ct method.

Analysis of the MOE in Dmrt5^−/− null mutants

16-micron coronal cryostat sections of the MOE were prepared from Dmrt5^−/− null mutant and WT E14.5 embryos of both sexes. Sections were immunostained with primary anti-OMP antibody, followed by incubation with Alexa Fluor 555-conjugated secondary antibody (1:1000). Nuclei were counterstained with Hoechst 3334 (1:1000). For each sample, a single section representing the central region along the anteroposterior axis of the unilateral MOE was selected for analysis. Quantification of OMP + cells and measurement of MOE area were performed on these selected sections.

To determine the number of OMP + cells, immunostained sections were imaged using an Olympus SpinSR10 spinning disk confocal system mounted on an IX83 inverted microscope. For complete coverage of the unilateral MOE, 1–2 images were acquired per sample using a 20× objective. Each image consisted of a Z-stack comprising 51 optical sections, which were subsequently Z-projected to generate a single image. OMP + cells were manually counted on the Z-projected images using the Cell Counter plugin in Fiji (ImageJ). OMP + cells located outside the thickness of the epithelium were not counted.

To measure MOE area, each sample was imaged with Hoechst nuclear staining. Images were acquired using a Leica DM5000B upright microscope equipped with a Leica DFC350 FX monochrome camera and a 10× objective. The unilateral MOE area was delineated and measured using the freehand selection tool in Fiji.

OMP + cell density for each unilateral MOE was calculated by dividing the number of OMP + cells by the corresponding MOE area.

Statistical analyses

The researchers were blinded to the sex and genotype at the time of analysis. The number of quantified animals is indicated in each plot; a minimum of three mice were included. The normality of the distribution was tested using the Shapiro–Wilk test in all cases. Each individual test is indicated in the figure legend of the corresponding experiments. Statistically significant differences were considered and represented by p-value: *p < 0.05, **p < 0.01, and *** p < 0.001. Statistical analyses were conducted using GraphPad Prism 9 and 8.0.2.

Author contributions

ES-S conceived and designed the study and supervised the project. ES-S, RC-N, AB-S, and RTC developed the theoretical framework. All authors discussed the results, ES-S, RC-N contributed to the writing of the manuscript. RC-N performed the ISH for Dmrt3-Dmrt7, as well as their anatomical identification. Additionally, RC-N conducted functional experiments to investigate the role of Dmrt5 in the olfactory system. AB-S performed all ISH for Dmrt2 . RTC performed iDISCO+ stainings for DMRT3 in the DG. V–O did the DMRT5 double stainings in the DG of postnatal animals. LC-C did Dmrt1 and Dmrt7 ISH. SA-B, MPM-V y SJ contributed with TH iDISCO staining in wildtype and Dmrt5^−/− brains, including quantifications between sexes. All authors approved the final version.

Funding

Open Access funding provided thanks to the CRUE-CSIC agreement with Springer Nature. Grant PGC2018-101751-A-100 and PID2021-127235NB-I00 funded by MCIN/AEI/10.13039/501100011033, and ERDF, “A way of making Europe”. RC-N holds an FPU from the Spanish MCIN, FPU19/02352; AB-S holds an FPI fellowship from the Spanish MCIN, PRE19-089366; RTC holds an FPI from the CAM, PIPF-2022/SAL-GL-24909, cofounded by ESF. LC-C holds a JAE from CSIC, JAEINT24_EX_0840 SJ was supported by Grant PID2020-113878RB-I00/AEI/10. ES-S. was supported by Grant RYC-2016–20537 funded by MCIN/AEI/10.13039/501100011033 and ESF, “Investing in your future”.

Data availability

This published article and its supplementary information files include all data generated or analyzed during this study. In situ raw data are available at [https://www.informatics.jax.org/reference/J:364871] (https:/www.informatics.jax.org/reference/J:364871). MGI Direct Data Submission to Mouse Genome Database (MGD). The Jackson Laboratory, Bar Harbor, Maine. World Wide Web (URL: [http://www.informatics.jax.org] (http:/www.informatics.jax.org)).

Declarations

Ethics approval and consent to participate

All experimental procedures involving animals were conducted in accordance with the European Directive 2010/63/EU on the protection of animals used for scientific purposes and the Spanish Royal Decree 53/2013 (BOE 34/08.02.2013). Experiments were approved by the Comité de Ética de Experimentación Animal (CEEA) of the Consejo Superior de Investigaciones Científicas (CSIC) and authorized by the Comunidad de Madrid under project license number PROEX 286/19. All efforts were made to minimize animal suffering and reduce the number of animals used, following the principles of Replacement, Reduction, and Refinement (3Rs) prescribed by Spanish and EU legislation. Consent to participate is not applicable.

Consent for publication

Written informed consent for publication was obtained from all participants included in the study.

Competing interests

The authors declare no competing interests.