PAX2+ Mesenchymal Origin of Gonadal Supporting Cells Is Conserved in Birds

During embryonic gonadal development, the supporting cell lineage is the first cell type to differentiate, giving rise to Sertoli cells in the testis and pre-granulosa cells in the ovary. These cells are thought to direct other gonadal cell lineages down the testis or ovarian pathways, including the germline. Recent research has shown that, in contrast to mouse, chicken gonadal supporting cells derive from a PAX2/OSR1/DMRT1/WNT4 positive mesenchymal cell population. These cells colonize the undifferentiated genital ridge during early gonadogenesis, around the time that germ cells migrate into the gonad. During the process of somatic gonadal sex differentiation, PAX2 expression is down-regulated in embryonic chicken gonads just prior to up-regulation of testis- and ovary-specific markers and prior to germ cell differentiation. Most research on avian gonadal development has focused on the chicken model, and related species from the Galloanserae clade. There is a lack of knowledge on gonadal sex differentiation in other avian lineages. Comparative analysis in birds is required to fully understand the mechanisms of avian sex determination and gonadal differentiation. Here we report the first comparative molecular characterization of gonadal supporting cell differentiation in birds from each of the three main clades, Galloanserae (chicken and quail), Neoaves (zebra finch) and Palaeognathe (emu). Our analysis reveals conservation of PAX2+ expression and a mesenchymal origin of supporting cells in each clade. Moreover, down-regulation of PAX2 expression precisely defines the onset of gonadal sex differentiation in each species. Altogether, these results indicate that gonadal morphogenesis is conserved among the major bird clades.


INTRODUCTION
Gonadal sex differentiation during embryogenesis provides an excellent model for studying the genetic regulation of development. The somatic component of the vertebrate gonad arises from intermediate mesoderm, while the germ cells are of extra-gonadal origin, migrating into the gonad before somatic sex differentiation commences (Lawson, 1999;Nef et al., 2019). Among most vertebrates, the gonadal primordium, together with its germ cells, is initially morphically identical in both sexes. Subsequently, the somatic cells and the germ cells of the gonad are directed down the testicular or ovarian pathway via a cascade of sexually dimorphic gene expression that starts in the somatic compartment (Eggers et al., 2014;Liu et al., 2016;Yang et al., 2018). Two distinctive structures are initially distinguishable in the somatic compartment, an outer coelomic epithelium and underlying medulla (Smith and Sinclair, 2004). In the mouse, for which most data are available, proliferation of cells in the coelomic epithelium gives rise to so-called supporting cell progenitors, which enter the medulla. This key cell lineage generates Sertoli cells in the testis and pre-granulosa cells in the ovary. In both sexes, the supporting cells are thought to direct other uncommitted progenitor cells to the testicular or ovarian pathways, including the germline (Piprek et al., 2017;Rotgers et al., 2018). Germ cells are specified in the epiblast very early in development, and they migrate into the undifferentiated gonads (via the hindgut in mammals, via the bloodstream in birds) (Tagami et al., 2017). The germ cells of both sexes populate the gonad but are uncommitted to either the spermatogenesis or oogenesis pathway until somatic gonadal cells send inductive cues. Germ cell fate is therefore closely linked to somatic development of the gonad. In mouse, after the germ cells have settled in the gonads, the somatic supporting cell lineage begins to differentiate. In female mammals, somatic, and intrinsic signals induce germ cells to express Stra8 and enter meiosis during embryogenesis. A large body of evidence previously pointed to retinoic acid (RA) as the somatic indicative signal triggering meiosis in females (Bowles et al., 2006;Koubova et al., 2006). However, surprisingly, recent data has shown that compound mutant mouse ovaries lacking all retinoic acid receptors or all three RALDH2 enzymes that synthesize RA can still initiate meiosis (Chassot et al., 2020;Vernet et al., 2020). The exact gonadal somatic signal for female germ cell sexual development is therefore again open to investigation. In male mammal gonads, the germ cells do not enter meiosis during embryogenesis. Instead, they lose pluripotency and enter mitotic arrest (Spiller et al., 2017). These sexually dimorphic germ cell fates are intimately linked to the development of the gonadal somatic cells. In males, this lineage gives rise to pre-Sertoli cells and, in females, pre-granulosa cells . Signals such as Fgf9 sent from the Sertoli cells act with intrinsic factors, such as Nanos2, to antagonize meiosis and instead direct the germ cells down the male pathway, toward spermatogenesis (Suzuki and Saga, 2008;Bowles et al., 2010). The supporting cell lineage also sends inductive signals to the presumptive steroidogenic lineage, directing their differentiation into Leydig cells (in the testis) or thecal cells (in the ovary) (Yao et al., 2002;Rebourcet et al., 2014). In the developing mammalian ovary, proper follicle formation requires cross-talk between the female somatic and germ cell populations (Liu et al., 2010;Baillet and Mandon-Pepin, 2012). Hence, the sexual fate of the gonadal soma and the germ cells hinges upon the key supporting cell lineage.
In the mouse, the key supporting cell lineage derives from the coelomic epithelium via asymmetric cell division and egression into the underlying gonadal mesenchyme (Piprek et al., 2016;Lin et al., 2017;. Surprisingly, recent research has shown that, in contrast to mammals, the coelomic epithelium in the chicken embryo does not generate the supporting cell lineage (Sertoli or pre-granulosa cells). Rather, it gives rise to a non-steroidogenic interstitial cell population (Estermann et al., 2020). In chicken, the supporting cells develop from a mesenchymal source present in the gonad during early development (Sekido and Lovell-Badge, 2007;Estermann et al., 2020). These cells have a specific molecular signature, expressing the transcription factors PAX2, DMRT1 and OSR1, and the signaling molecule, WNT4. The finding that supporting cells in chicken derive form a different source to those in mouse was surprising, given the conservation of overall gonadal morphogenesis among vertebrate embryos (DeFalco and Capel, 2009). However, a major difference between birds and mammals is the genetic gonadal sex-determining trigger. In mouse and other mammals, the Y chromosomelinked SRY gene operates as the master sex switch, initiating Sertoli cell differentiation in male embryos (Koopman et al., 1990(Koopman et al., , 1991Sinclair et al., 1990;Goodfellow and Lovell-Badge, 1993). SRY is absent outside the mammalian clade, and in fact, birds have a different sex chromosome system. Birds have ZZ/ZW sex chromosomes, in which male (ZZ) is the homogametic sex and female (ZW) is heterogametic (Marshall Graves, 2008). The Z linked gene, DMRT1 operates as the testis determining factor via a dosage mechanisms (Smith et al., 2009;Lambeth et al., 2014;Ioannidis et al., 2021). Due to the lack of Z sex chromosome compensation, male supporting cells have double the dose of DMRT1 compared to females (Raymond et al., 1999;Smith et al., 2003;Ayers et al., 2015). DMRT1 knockdown or knock out results in feminization of the gonad. Moreover, over-expression of this gene causes gonadal masculinization, indicating that DMRT1 is the sex-determining gene in chicken, and presumably in all birds (Smith et al., 2009;Lambeth et al., 2014;Ioannidis et al., 2021). This would be consistent with the deep evolutionary conservation of the Z sex chromosome in birds, across some 60 million years (Handley et al., 2004;Zhou et al., 2014;Xu et al., 2019). In the male chicken embryo, DMRT1 is known to activate SOX9 expression, which is crucial in Sertoli differentiation, and AMH, which is important for Müllerian duct regression (Lambeth et al., 2014). In females (ZW), due to the lower levels of DMRT1 expression, supporting cells differentiate toward pre-granulosa cells by upregulating FOXL2 and Aromatase (Lambeth et al., 2013;Major et al., 2019). We previously characterized cell lineage specification during chicken gonadal sex differentiation and identified PAX2 as a novel marker of the early supporting cell lineage (Estermann et al., 2020). During the process of gonadal sex differentiation, PAX2 expression is down-regulated in chicken gonads (Estermann et al., 2020). This suggests that PAX2 down-regulation could be used to predict the onset of gonadal sex differentiation in chicken. However, the conservation of both the mesenchymal origin of gonadal cells and the role of PAX2 in birds beyond the chicken have not been previously explored.
Modern birds are classified into two main groups, the Palaeognathe, (the flightless ratites and volant tinamous) and Neognathae (all other birds). The Neognathae is divided into two clades, the Galloanserae (chickens, quails, and ducks et al.) and Neoaves, the perching birds (around 95% of all extant avian species) (Cracraft, 2001;Hackett et al., 2008). Most research on avian gonadal development has focused on the chicken, or related members of the Galloanserae clade, such as quail and duck (Takada et al., 2006;Smith et al., 2009;Ayers et al., 2015;Bai et al., 2020;Okuno et al., 2020). Additionally, these studies have focused mainly on the conservation of mammalian genes involved in gonadal sex differentiation. There is very little information regarding gonadal sex differentiation among the other major avian clades (Hirst et al., 2017a). Comparative analysis in birds is required to fully understand the mechanism of avian sex determination and gonadal sex differentiation.
Historically, gonadal sex differentiation has been characterized on the basis of morphology, whereby the condensation of Sertoli cells marks the onset of testis formation and organization of pre-follicular cells marks the onset of ovary formation (Wilhelm et al., 2007). In the chicken embryo, the first morphological sign of testis formation, as in mammals, is the appearance of Sertoli cells and their coalescence in the medullary cords of the gonad. In the female chicken embryo, the first overt morphological sign of sex differentiation is a thickening of the outer coelomic epithelium into a cortex and fragmentation of the underling medulla (Carlon and Stahl, 1985). However, histology alone has proven to be inconsistent and inaccurate to determine the precise time of gonadal sex differentiation onset. A clear example is studies on Japanese quail embryos (Coturnix japonica). Several histological hematoxylin-eosin based analyses determined that quail gonads were sexually differentiated at embryonic day 6 (E6) (stage 30), E7 (stage 32) or E8 (stage 35) (Kannankeril and Domm, 1968;Intarapat and Satayalai, 2014;Mohamed et al., 2017). However, sexual differentiation is likely to be triggered at the genetic level prior to overt histological differentiation. In the quail, the Sertoli cell marker SOX9 is detectable at E5 (stage 27) (Takada et al., 2006), indicating that sexual differentiation begins at the molecular level distinctly prior to morphological differentiation. The development of more accurate molecular methods to determine sexual differentiation is required to improve knowledge of avian sex determination and for informing methodologies targeting species conservation.
Here we report the first comparative molecular characterization of gonadal sex differentiation in birds from each of the three main clades, Galloanserae (chicken and quail), Palaeognathe (emu), and Neoaves (zebra finch). Our analysis demonstrates a conservation of the PAX2 + mesenchymal origin of supporting cells in all analyzed birds. In addition, PAX2 down-regulation immediately precedes up-regulation of male and female supporting cell markers, and the morphological onset of sexual differentiation. PAX2 gonadal down-regulation precisely predicted the onset of sex differentiation in each avian clade, more accurately than previous histological analysis. Altogether these results indicate that the process of gonadal sex differentiation is conserved among the major bird clades. This research proposes PAX2 immunodetection as a new methodology to evaluate gonadal differentiation in birds.

Eggs
Fertilized HyLine Brown chicken eggs (Gallus gallus domesticus) were obtained from Research Poultry Farm (Victoria, Australia). Wild type Japanese quail eggs (Coturnix japonica) were provided by the Monash transgenic quail facility. Fertilized emu (Dromaius novaehollandiae) eggs were purchased from Emu Logic (Toorahweenah, NSW). Zebra finch (Taeniopygia guttata) embryos were obtained from wild-derived birds. The zebra finch colony is a captive population derived from wild caught birds under Deakin University Animal Ethics #G23-2018. The birds used in this study were several generationcaptive birds derived from this initial population. Fresh eggs were collected in nests in outdoor aviaries and artificially incubated at Deakin University (Geelong, Australia). Eggs were incubated under humid conditions at 37.5 • C until collection and staged (Hamburger and Hamilton, 1951;Ainsworth et al., 2010;Nagai et al., 2011;Murray et al., 2013).
Zebra finch sexing PCR protocol is a modification of a previously described method (Soderstrom et al., 2007). This method relies upon specific amplification of a fragment of the CDH gene located on the W chromosome. As an internal control, a fragment of the CHD gene located in the Z chromosome was amplified. The reaction was performed in a final volume of 11 µL. W1 (5 -GGGTTTTGACTGACTAACTGATT-3 ), W2 (5 -GTTCAAAGCTACATGAATAAACA-3 ), Z1 (5 -GTGTAGTCCGCTGCTTTTGG-3 ) and Z2 (5 -GTTCGTGGTCTTCCACGTTT-3 ) primers used at a final concentration of 0.1 µM each. 1 µL of digestion buffer was used with 10 µL of the sexing mix. The PCR program consisted of denaturation for 2 min at 94 • C followed by 30 cycles of incubation at 94 • C × 30 s; 56 • C × 45 s; 72 • C × 45 s and final extension at 72 • C for 5 min, followed by 4 • C hold. PCR products and molecular ladder (1 kb plus, Invitrogen) were run on a 2% agarose gel for 30 min at 130 V and visualized with gel red (Biotium).

Immunofluorescence
Whole embryos or urogenital systems were collected, briefly fixed in 4% PFA/PBS for 15 min, cryo-protected in 30% sucrose overnight and blocked in OCT embedding compound for sectioning. Immunofluorescence was carried out as described previously (Estermann et al., 2020). Briefly, 10 µm frozen sections were cut and antigen retrieval was performed for DMRT1 and PAX2 protein immunofluorescence using the Dako PT Link automated system. Sections were then permeabilized in 1% Triton X-100 in PBS for 10 min at room temperature and washed 3 times in PBS. All sections were blocked in 2% BSA in PBS for 1 h at room temperature followed by primary antibody incubation overnight at 4 • C in 1% BSA in PBS. The following primary antibodies were used: rabbit anti-PAX2 (Biolegend 901001, 1;500), rabbit anti-DMRT1 (in house antibody; 1:2000), rabbit anti-AMH (Abexa ABX132175; 1:1000), rabbit anti-Aromatase (in house antibody; 1:4000), rabbit anti-FOXL2 (in house antibody; 1:2000), and rabbit anti-SOX9 (Millipore AB5535, 1:4000). After overnight incubation with primary antibody, sections were then washed 3 times in PBS and incubated for 1 h at room temperature with Alexa Fluor 488 donkey anti-Rabbit (1:1000) and Alexa Fluor 594 donkey anti-Mouse (1:1000) in 1% BSA in PBS. Sections were counterstained in DAPI/PBS and mounted in Fluorsave (Millipore). Images were collected on a Zeiss Axiocam MRC5 microscope using the same exposure time between males and females for expression comparisons.
For double immunofluorescence using two primary antibodies raised in the same species (rabbit anti-PAX2 and rabbit anti-DMRT1), the iterative indirect immunofluorescence imaging (4i) protocol was used on paraffin sections (Gut et al., 2018). Dissected gonads were fixed overnight in 4% paraformaldehyde at 4 • C, paraffin-embedded and sectioned in the transverse plane at 5µm. After deparaffinisation, antigen retrieval was carried out using TE buffer (Moreau et al., 2019). Sections were incubated with anti-DMRT1 antibody (1:2000, in house) and anti-cytokeratin antibody (1:200, Novus Bio NBP2-29429) overnight at 4 • C. Sections were then washed with 1X PBS and incubated with Donkey anti-Rabbit Alexa Fluor R Plus 647 (1:2000, Invitrogen) and Donkey anti-mouse Alexa Fluor R Plus 488 (1:2000, Invitrogen) together with DAPI for 2 h at room temperature, after which the tissue was washed with 1X PBS. Slides were mounted in imaging buffer (Gut et al., 2018) and images were captured using a 3i Marianas spinning disk confocal at low laser power. Sections were washed in 1X PBS and antibodies were eluted following the 4i protocol. After elution, sections were imaged again with the same parameters to ensure that the first round of antibody labels were removed. Slides were then incubated with anti-PAX2 (1:400, Biolegend 901001) and anti-cytokeratin (1:200, Novus Bio NBP2-29429) in 5% BSA in 1X PBS overnight at 4 • C with the same wash, secondary antibody incubation and imaging parameters from the first round of labeling. Brightness and contrast were equally altered across all images to improve data display using ImageJ (Schneider et al., 2012).

Chicken
Previous chicken single-cell RNA-seq identified the gonadal supporting cell precursors as a mesenchymal population expressing the transcription factors PAX2, DMRT1, OSR1 and the signaling molecule, WNT4 (Estermann et al., 2020). PAX2 and DMRT1 immunofluorescence was performed to evaluate PAX2 expression pattern before, during and after gonadal differentiation (Figure 1 and Supplementary Figure 1). At E4.5 (HH24) and E5.0 (HH26), PAX2 positive cells were detected in the gonadal medulla in both sexes ( Figure 1A). In males, from E5.5 (HH28) to E6.0 (HH29), PAX2 expression continued to be present in the basal region of the gonad but was absent in the most apical ( Figure 1A). In females, a similar pattern occurred at E6.0 ( Figure 1A). This data suggests that PAX2 is down-regulated at the onset of gonadal sex differentiation.
We previously found that PAX2 mRNA was co-expressed with DMRT1 protein in gonadal sections (Estermann et al., 2020), but the co-localization of PAX2 and DMRT1 proteins in the same cells/nucleus was not assessed, because both primary antibodies were raised in rabbit. Iterative indirect immunofluorescence imaging (4i) (Gut et al., 2018) was used here to detect DMRT1 and PAX2 proteins in the same undifferentiated chicken gonad just prior to sexual differentiation at E5.0 (HH26) (Figure 1B and Supplementary Figure 1A). PAX2 positive cells in the gonadal mesenchyme were also DMRT1 positive. Interestingly, a gradient of PAX2 expression was noted; high in some cells (basal), lower in others (apical). This presumably reflects the gene being downregulated among cells. Additionally, some DMRT1 positive cells were negative for PAX2, as expected, due to DMRT1 being expressed in the left coelomic epithelium and in the germ cells.
To evaluate the expression pattern of PAX2 in differentiating embryonic chicken gonads, immunofluorescence was performed on E6.5 (HH30) and E8.5 (HH34) male and female gonads (Supplementary Figures 1B-D). PAX2 was not expressed in E6.5 testis (Supplementary Figure 1B), consistent with previous reports of PAX2 down-regulation upon sexual differentiation (Estermann et al., 2020). DMRT1, AMH, and SOX9 immunofluorescence confirmed that these gonads were presumptive testes (Supplementary Figure 1B). In females, PAX2 expression was absent from the apical region of the gonad, whereas it was still expressed in the basal region of the gonad at E6.5 (Supplementary Figure 1C). This expression pattern was complementary to FOXL2 expression pattern, more strongly expressed in the gonadal apical than the basal region (Supplementary Figure 1C). Only few aromatase positive cells were detected in the gonad at this stage, suggesting that the gonadal differentiation had just commenced. As for the male, the female gonad also showed PAX2 being down-regulated after sex differentiation, on a cell-to-cell basis. By E8.5, PAX2 expression was excluded from the gonad in both sexes (Supplementary Figure 1D). To quantify these changes in PAX2 expression, qRT-PCR was performed in male and female gonads at E4.5 (HH24), E6.5 (HH30) and E8.5 (HH34). These time points correspond to the period before, at the onset and after the onset of morphological gonadal sex differentiation. Consistent with the immunofluorescence data, PAX2 expression was significantly reduced after gonadal sex differentiation in both sexes (E6.5 and E8.5) ( Figure 1C). In females, expression reduction was delayed, occurring by E8.5 ( Figure 1C). This suggests that male gonad sex differentiation commences prior to female gonadal sex differentiation. In conclusion, PAX2 is expressed in chicken undifferentiated gonadal supporting cells, co-localizing with DMRT1 in the medulla, and its expression is down-regulated during sexual differentiation.

Quail
To evaluate if the PAX2 + mesenchymal origin of gonadal supporting cells is conserved among birds or is specific to chicken, gonads were analyzed from all three main bird clades. The Japanese quail (Coturnix japonica) belongs to the Galloanserae clade, the same group as chicken. This means that gonadal differentiation is likely to be very conserved between the two species. All previous studies on quail gonadal sex differentiation have relied upon histology to define the onset of sexual differentiation. Consequently, the timing of gonadal sex differentiation in this species has been variably reported, from between E6 (Kannankeril and Domm, 1968), E7 (Mohamed et al., 2017), and E8 (Intarapat and Satayalai, 2014). Quail gonadal sex differentiation was analyzed by immunofluorescence from E3.5 (stage 21) to E6.0 (stage 30), in half day incubation intervals (Figure 2). PAX2 was used as a (presumed) undifferentiated supporting cell precursor marker. Meanwhile, DMRT1, SOX9 and AMH were used as Sertoli cell markers in the testis and aromatase as pre-granulosa cell marker in developing ovary. Consistent with the chicken data, PAX2 positive cells started colonizing the region underlying the coelomic epithelium in both sexes of quail gonads at E3.5 (stage 21) and E4.0 (stage 24) (Figure 2). Unlike in chicken, DMRT1 was not detected in the undifferentiated quail gonads. Instead, by E4.0, DMRT1 expression was first detected and already sexually dimorphic, showing higher intensity levels in males than females (Figures 2A,B). By E4.5, PAX2 expression was turned off in the gonadal cells and its expression was excluded from the gonad during subsequent time points, when DMRT1 and other markers of sexual differentiation were activated (Figure 2). Some PAX2 positive cells were still visible in the basal region of the gonad, adjacent to the mesonephros (Figure 2). In addition, higher DMRT1 expression in males also suggests that the supporting cells commenced differentiation into Sertoli cells (Figure 2A). Some AMH positive cells were observed in the gonadal mesenchyme at E4.5 and E5.0 (stage 27) in both male and female gonads, but no SOX9 (Sertoli cell marker) (Figure 2A) or aromatase (pre-granulosa cell marker) (Figure 2B). The latter two were detected from E5.5 (stage 28). This quail data indicates that the PAX2 + mesenchymal origin of supporting cells observed in chicken is conserved among Galliformes. In addition, PAX2 down-regulation indicates that gonads commence gonadal sex differentiating at E4.5 in quail, much earlier than reported. This indicates that gene expression is a better predictor than morphological markers in defining the precise onset of gonadal sex determination.

Zebra Finch
The other major clade of the Neoganthae is the Neoaves (perching birds). This group contains almost 95% of all living modern birds and is the result of early and rapid diversification around the Cretaceous mass extinction event (Claramunt and Cracraft, 2015;Prum et al., 2015). One of the most widely studied models in this clade is the zebra finch (Taeniopygia guttata), primarily in the field of neurobiology. Due to its popularity, the zebra finch genome was the second avian genome to be sequenced (Warren et al., 2010;Mak et al., 2015;Patterson and Fee, 2015). In addition, embryonic gonadal sex differentiation and primordial germ cell colonization have been studied in zebra finch, showing some differences between previous chicken reports (Hirst et al., 2017a;Jung et al., 2019). Zebra finch gonads have been reported to be sexually differentiated at E6.5, evidenced by SOX9 and FOXL2 mRNA expression in males and females, respectively (Hirst et al., 2017a). At E4.5 these markers are not expressed, suggesting an undifferentiated state (Hirst et al., 2017a). To evaluate if PAX2 + mesenchymal origin of supporting cells is conserved in Neoaves, PAX2, DMRT1, FOXL2, and AMH immunofluorescence was performed in male and female zebra finch gonads at E4.5 (stage 24), E5.5 (stage 28) and E6.5 (stage 31). PAX2 positive mesenchymal cells were detected in both male and female gonads at E4.5 (Figure 3). DMRT1, AMH and FOXL2 were not detected in the gonads at this stage. Altogether, this suggests that the zebra finch gonads are undifferentiated at E4.5. This is consistent with previous reports (Hirst et al., 2017a).
By E5.5, PAX2 expression was extinguished from both male ( Figure 3A) and female (Figure 3B) gonads. DMRT1 and AMH positive Sertoli cells were identified in the male testicular medulla at E5.5 ( Figure 3A). The downregulation of PAX2 and up-regulation of supporting cell markers indicates that gonadal sex differentiation in zebra finches commences at E5.5. By E6.5, FOXL2 expression was detected in the ovarian medulla ( Figure 3B). In males, DMRT1 and AMH positive testicular cords were evident in the gonadal medulla ( Figure 3A). Altogether, this data confirms the conservation of PAX2 + mesenchymal cell origin of supporting cells in Neoaves, and in particular in zebra finch. In addition, using PAX2 as a predictor of sex differentiation we were able to Frontiers in Cell and Developmental Biology | www.frontiersin.org 9 August 2021 | Volume 9 | Article 735203 determine that zebra finch gonadal sex differentiation begins at E5.5 (stage 28).

Emu
The Palaeognathae superorder includes the flightless ratites and the volant neotropical tinamou. Among the ratites, gonadal sex differentiation has only been described in the emu (Dromaius novaehollandiae) (Hirst et al., 2017a). Previous histological data suggested that emu gonadal differentiation commences at E16, evidenced by the presence of seminiferous cords in male gonads, containing DMRT1 + Sertoli cells (Hirst et al., 2017a). As noted previously, histological analysis is not the best methodology for defining the onset of sex differentiation. To gain insight into the specific timeframe of gonadal sex differentiation in emu and assess if PAX2 mesenchymal origin of supporting cells is conserved in ratites, gonadal immunofluorescence was performed at E9.5 (HH24), E11.5 (HH27) and E13.5 (HH29). PAX2 positive cells were detected at E9.5 in the medullary mesenchyme of both male and female emu gonads (Figure 4). This expression pattern is similar to the previous data shown for chicken, quail and zebra finch. FOXL2, SOX9 and DMRT1 were not expressed at E9.5 (HH24) (Figure 4), indicating that the gonads were undifferentiated and bipotential. By E11.5 (HH27), PAX2 expression was extinguished from the gonad in both sexes. In males DMRT1 and, to a lesser extent, SOX9, were expressed in the E11.5 (HH27) testis, indicating activation of the testicular differentiation pathway (Figure 4A). Similarly, in females, FOXL2 expression was up-regulated, indicating that the ovarian differentiation program had commenced ( Figure 4B). By E13.5, in males, DMRT1 + /SOX9 + testicular cords were identified in the gonadal medulla ( Figure 4A). In females, FOXL2 was expressed in pre-granulosa cells in the ovarian medulla ( Figure 4B). This data indicates that emu gonadal sex differentiation commences at E11.5 (HH27), earlier than previous reports based on histology. In addition, a PAX2 + mesenchymal origin of supporting cells is also conserved in the Palaeognathae clade.

DISCUSSION
The data presented here support a conserved origin of gonadal supporting cells in birds, distinct from that reported in the mouse. In the mouse model, the supporting cell lineage derives from the coelomic epithelium (Stevant et al., 2018. The gonadal supporting cells in birds do not derive from the coelomic epithelium but rather from a DMRT1 and PAX2 positive mesenchymal population. In this study, we show that PAX2 is expressed in the bipotential supporting cells of the gonadal mesenchyme in members of the Galloanserae (chicken and quail), Neoaves (zebra finch) and Paleognathae (emu), suggesting a conserved mechanism among all birds. In addition, this is the first systematic evaluation of gonadal sex differentiation in quail, emu and zebra finch using expression of gonadal marker proteins.
Previous reports based in histological and morphological analysis of the quail gonad have not consistently determined an embryonic stage of gonadal sex differentiation. Previous estimates of gonadal sex differentiation onset ranged from E5.5 to E8.0 (Kannankeril and Domm, 1968;Intarapat and Satayalai, 2014;Mohamed et al., 2017). The results reported here indicate that quail gonad sex differentiation commences at E4.5, earlier than previously suggested. This is shown by down-regulation of the undifferentiated supporting cell marker PAX2 and the up-regulation of DMRT1 in male gonads (Figure 2). These results show that gene expression analysis is more accurate than morphological and histological analysis in determining the onset of gonadal sex differentiation. Similarly, previous histological analysis of emu gonads suggested that sex differentiation commences at E16 (Hirst et al., 2017a). The data presented here indicates that emu gonadal sex determination commences at E11.5 (Figure 4), earlier than previous histological data suggests. In zebra finch, previous reports suggested the onset of sex differentiation occurs between E4.5 (undifferentiated) and E6.5 (differentiated). Our results agree with this data, showing that zebra finch gonadal sex differentiation commences at E5.5 (Figure 3). Downregulation of PAX2 expression precisely predicted the onset of sexual differentiation in the three avian clades, more accurately than previous histological analysis. This research identifies PAX2 as a new marker for evaluating gonadal differentiation in birds.
In female chicken embryos, FOXL2 and aromatase proteins were expressed in the apical supporting cells at E6.5 (HH 30) (Supplementary Figure 1C), suggesting that those are the first supporting cells to differentiate. This apical-basal wave of differentiation is consistent with the concentration of PAX2 positive cells in the basal region of the gonad (Figures 1A-C). A similar pattern was observed in male quail gonads. At E5.5 (stage 28), SOX9 and AMH proteins were expressed in the apical testicular cords of the quail, suggesting that those are the first supporting cells to differentiate, and by E6.0 all Sertoli cells were SOX9, AMH and DMRT1 positive (Figure 2A). Recently, two transcriptionally distinctive Sertoli cell populations in E10.5 chicken testis were identified. One expressed lower levels of SOX9 and DMRT1 and higher levels of CBR4 and GSTA2 and was located in the peripheral testicular cords. The second population was located in the basal region, expressing higher levels of SOX9 and DMRT1 but no GSTA2 (Estermann et al., 2020). This suggests that there may be two distinctive stages of Sertoli cell maturation, inner immature and outer mature populations. Mouse Sry is expressed as a wave across the male gonad (Larney et al., 2014), starting from the central region of the genital ridges, and then extending to cranial and caudal poles (Polanco and Koopman, 2007). 3D imaging approaches would be crucial to be understand how supporting cell differentiation occurs in birds and to understand if gonadal development also follows a longitudinal wave of differentiation, as in mouse.
Despite conservation of cell types, recent single-cell RNAseq data from embryonic chicken gonads has shown that cell lineage specification in the gonad may also vary substantially between birds and mammals (Estermann et al., 2020). This research has shown that there are two main sources of gonadal cells; the coelomic epithelium and the mesonephric mesenchyme (Sekido and Lovell-Badge, 2007;Estermann et al., 2020). The current study confirms that this is the case for members of all three major bird clades. It is unclear why birds exhibit this different developmental origin of the key supporting cell lineage. A mesenchymal origin of supporting cells could be an ancestral feature, lost in mammals, or a feature acquired by the avian lineage. Further research is required to evaluate if the mesenchymal origin of supporting cells also occurs among reptiles, amphibians or fish. A key difference between placental mammals and birds is the genetic sexdetermination system (XY vs. ZW). It would be interesting to evaluate if the supporting cell origin is correlated with the genetic sex-determination system (XY vs. ZW), sex determining genes (Sry vs. DMRT1 vs. others) and if it is also present in environmental sex determining species. Reptiles have diverse sex determining systems, ranging from pure GSD with either XX/XY or ZZ/ZW sex chromosome systems, to GSD modifiable by egg temperature, through to complete temperature dependent sex determination (TSD) (Warner, 2011;Ge et al., 2017;Whiteley et al., 2021). In at least one turtle species with TSD, the supporting cell lineage has been shown to derive form the coelomic epithelium (Yao et al., 2004). At present, there is not observable correlation between the type of sex determining system and the sources of gonadal supporting cells. Gonadal epithelium lineage tracing by GFP electroporation is feasible in oviparous reptiles, which would shed light on the origin of gonadal cell lineages in these groups (Hirst et al., 2017b). Among reptiles, crocodilians are the closest living clade to birds (Green et al., 2014), making them an ideal model to test the possible synapomorphy of gonadal PAX2 in birds and to study the evolution of the mesenchymal epithelial supporting cell origin. A previous ultrastructural study suggested that the supporting cell lineage in the American alligator (Alligator mississippiensis) may be of coelomic epithelial origin (Smith and Joss, 1993).
In summary, this study demonstrates a conserved gonadal PAX2 positive mesenchymal expression pattern in representatives of all three bird clades. Analysis should be expanded to other avian species to evaluate the degree of conservation among birds more broadly. Among the Galloanserae, only Galliformes have been studied in any detail (chicken and quail). Anseriformes (ducks, geese, and swans) could be examined. In addition, among the Neovaes, only the zebra finch has been studied in any detail. During the rapid diversification that characterize the Neoaves, birds could exhibit other mechanisms of gonadal formation, diverging from the PAX2 + mesenchymal origin. It would be interesting to expand this study to more members of this diverse clade. Given the monophyly of birds and their conserved ZZ/ZW sex determining system, we postulate that the gonadal PAX2 mesenchymal expression pattern is prevalent among avians.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author.

ETHICS STATEMENT
The animal study was reviewed and approved by Animal Ethics Office, Monash University (AEC approval not required for avian embryos less than mid embryogenesis as per state legislation). The zebra finch colony is a captive population derived from wild caught birds under Deakin University Animal Ethics # G23-2018. The birds used in this study were several generation-captive birds derived from this initial population. As for emu, quail, and chicken, animal ethics was not required for the finch embryos, as they were harvested less than mid-gestation (permitted by Australian law).