The Desert Hedgehog Signalling Pathway in Human Gonadal Development and Differences of Sex Development

The highly conserved Hedgehog (HH) family of signalling molecules plays a central role in the development of numerous organs during embryogenesis. Disruption to the ligands or signalling pathway can cause a variety of genetic disorders at different stages [Sasai et al., 2019] as well as cancer [Doheny et al., 2020; Iriana et al., 2021]. The canonical HH signalling pathway has been the subject of numerous excellent reviews [Briscoe and Therond, 2013; Lee et al., 2016] and is only briefly described here with a focus on human. The HH morphogens are growth factors produced and secreted by a subset of cells. Humans, like many higher vertebrates, have 3 HH genes; Sonic (SHH), Indian (IHH), and Desert (DHH) hedgehog. This review focuses on DHH, the only ligand known to be involved in sex differentiation and differences of sex development (DSDs). All HH ligands undergo auto-processing and cleavage with the covalent addition of cholesterol [Porter et al., 1996] and palmitoylation [Pepinsky et al., 1998] by the hedgehog acyltransferase (HHAT) [Chamoun et al., 2001], which increase stability in the extracellular matrix and facilitate long-range spreading [Lewis et al., 2001; Zeng et al., 2001; Callejo et al., 2006]. Secreted HH ligands diffuse in a gradient, binding to the Patched (PTCH) receptor on target cells [Marigo et al., 1996; Motoyama et al., 1998]. PTCH is a constitutive repressor, indirectly destabilising the G-protein coupled receptor, Smoothened (SMO), the main pathway activator [Ayers and Therond, 2010]. Upon HH binding, PTCH is degraded in the endosome [Incardona et al., 2000] and repression of SMO is released, leading to downstream activation of the GLI family of transcription factors, GLI1, GLI2, and GLI3 [Bai et al., 2004].

During embryogenesis the main roles of DHH are in the formation of the peripheral nerve sheath and the development of the testis [Briscoe and Therond, 2013]. Testis development from the bi-potential gonad is triggered by the expression of the Y chromosome linked SRY gene, which induces the differentiation of the Sertoli cells via SOX9. Following this, development of interstitial cell types such as the peritubular myoid (PTM) and fetal Leydig cells (FLCs) is initiated. The germ cells, which have migrated to the gonad early in its development, become encased in the seminiferous tubules along with the supportive Sertoli cells, surrounded by a basement membrane laid down by the PTM cells (Fig. 1a) [for a detailed review see Eggers et al., 2014].

Most of what we know about the role of the DHH signalling pathway during testis development stems from animal models. This is summarised here. In mice, Dhh expression does not overlap with either Shh or Ihh, and Dhh is the main Hh involved in gonad development [Bitgood and McMahon, 1995; Shima et al., 2004]. Dhh expression in the mouse gonad starts from 11.5 dpc (days post coitum) and is maintained into adulthood [Bitgood et al., 1996; Yao et al., 2002; Szczepny et al., 2006]. In the early stages, Dhh expression is restricted to the Sertoli cells, likely inducing PTM and LC differentiation in a paracrine fashion [Bitgood et al., 1996; Clark et al., 2000; Pierucci-Alves et al., 2001; Yao et al., 2002] (Fig. 1b). PTM and LC precursors express the Dhh pathway receptors Ptch1 and Ptch2 from 11.5 dpc [Bitgood et al., 1996; Carpenter et al., 1998; Clark et al., 2000; Yao et al., 2002], and the Gli transcription factors (Gli1-3) are observed in the fetal and adult interstitium from 14.5 dpc [Szczepny et al., 2006; Barsoum and Yao, 2011]. The role of Dhh appears to be conserved in the marsupial animal model, the tamar wallaby, where DHH is the only HH expressed during gonadal development [Chung et al., 2012], found in pre-Sertoli cells at the beginning of cord formation, when its receptors PTCH1 and PTCH2 are also expressed [O'Hara et al., 2011].

In mice, Dhh signalling initiates FLC differentiation and maturation through upregulation of the transcription factor Sf1 (steroidogenic factor 1). Sf1 in turn upregulates the expression of the steroidogenic enzymes Cyp11a1 (cholesterol side-chain cleavage; P450ssc) and Hsd3b1 (hydroxyl-delta-5-steroid dehydrogenase, 3 beta- and steroid delta-isomerase 1) [Yao et al., 2002; Park et al., 2007; Brokken et al., 2009]. Constitutively active Dhh signalling in Sf1-positive LC progenitor cells led to an increased number of FLCs in mouse fetal testis [Barsoum et al., 2013], while the treatment of fetal gonad explant cultures with an Hh signalling inhibitor severely disrupted FLC differentiation [Yao and Capel, 2002]. Dhh has also been implicated in adult LC (ALC) differentiation, increasing their proliferation and transition rate through their stages of development [reviewed in Ye et al., 2017]. Dhh may also maintain ALC steroidogenic function (Fig. 1b), as the loss of testosterone production in conditional Wt1 (Wilms tumor protein) null mice could be partially rescued by treatment with SAG (a SMO agonist) [Chen et al., 2014]. Interestingly, Dhh protein expression has been described in mouse ALCs together with Ptch1 and Smo, leading to the speculation that ALCs are capable of autocrine signalling [Morales et al., 2009], although further evidence is needed (Fig. 1b).

Dhh ablation in animal models confirms its importance during gonadal development. Dhh null XY mice display a range of testicular defects, depending on their genetic background. In some cases, XY nulls have abnormal FLC and ALC development with an interstitium containing primarily undifferentiated cells, leading to testosterone insufficiency [Clark et al., 2000; Yao et al., 2002]. Testis with disorganised and irregularly shaped Sertoli cells, abnormal peritubular tissue, and discontinued basal lamina that result in irregular and anastomotic tubules with germ cells located outside of the cords have also been described in mice lacking Dhh [Clark et al., 2000; Pierucci-Alves et al., 2001]. In some cases, Dhh null testes are significantly smaller, with seminiferous tubules only thinly lined with Sertoli cells [Bitgood et al., 1996]. Dhh mutant rats also show impaired LC development and insufficient androgen production [Kawai et al., 2011]. Wallaby ex vivo testicular cultures with inhibited Hh signalling have disorganised cords with some germ cells outside, likely due to a perturbation in the PTM cells and basal lamina, suggesting conservation of the mechanism [Chung et al., 2012].

Dhh also regulates germ cell maturation/spermatogenesis [reviewed in Walterhouse et al., 2003] (Fig. 1c). Adult XY Dhh null mice are infertile due to a lack of mature sperm, as germ cells do not progress towards the final stages of spermatogenesis [Bitgood et al., 1996]. Dhh signalling likely begins in spermatogonia and primary spermatocytes, which express the Ptch receptors, the Dhh activators Smo and Fu, and the Gli transcription factors [Persengiev et al., 1997; Carpenter et al., 1998; Fan et al., 2007; Morales et al., 2009]. At later germ cell maturation stages, Dhh signalling seems to be switched off in round and elongating spermatids, as indicated by both a shift of Gli1 protein expression from the nucleus to the cytoplasm [Kroft et al., 2001] and the expression of SuFu (Suppressor of Fused), a negative regulator of the Hh signalling pathway, in these cells [Szczepny et al., 2006].

In humans, DHH, composed of 3 exons, is located at 12q12.13 and encodes a protein of 396 amino acids [Tate et al., 2000]. Like the other HH proteins, DHH is likely produced as a 45 kDa precursor with the 25 kDa C-terminal driving autolytic cleavage assisted by the covalent addition of a cholesterol moiety to the 19 kDa N-terminal fragment, thought to be the active molecule [Mann and Beachy, 2004]. However, the extent of DHH processing in the human embryonic gonads is unknown. DHH can be cleaved in an in vitro system [Tajouri et al., 2018], yet DHH expressed in cultured cells such as HEK293T cells is often uncleaved [Pettigrew et al., 2014; Ayers et al., 2019]. Indeed, domain swapping experiments have demonstrated that the DHH C-terminal is less efficient in cleaving the SHH N-terminal than the SHH C-terminal [Pettigrew et al., 2014]. In culture, uncleaved DHH is not effectively secreted and accumulates intracellularly, unable to elicit paracrine pathway activation [Pettigrew et al., 2014], yet these cells are still able to induce HH signalling in a juxtacrine assay, including the activation of LCs derived from immature mouse testes (TM3 cells) [Pettigrew et al., 2014]. Based on this, it has been proposed that DHH is not processed or secreted in the Sertoli cells, instead eliciting a cellular response in the LCs in a juxtacrine fashion. Although based on the position and proposed effects of human DHH genetic variants, others have proposed that DHH auto-processing is necessary in the human gonads [Elzaiat et al., 2020].

Recombinant human DHH shows an affinity to its PTCH receptor complex comparable to SHH and IHH [Pathi et al., 2001] and is therefore thought to activate the same pathway during development, and DHH likely plays a conserved role in the testis although some interesting differences may exist in humans. Single cell RNA seq data from human fetal gonads [Li et al., 2017] (https://rgv.genouest.org/) [Darde et al., 2015, 2019] show DHH expression in the Sertoli cells with a peak in expression at 12 gestational weeks. Pathway receptor/targets PTCH1 and PTCH2 become strongly expressed at 10 weeks of gestation in Leydig precursors, mature LCs, and primordial germ cells (PGCs). Similarly, the GLI genes can be observed in the same cell clusters with the strongest expression between 20 and 25 weeks of gestation. Of note, the PTCH receptor complex shows weak expression in Sertoli cells [Li et al., 2017], and the pathway activator SMO is expressed throughout the human gonad but absent from the Sertoli cells. This might suggest DHH produced by Sertoli cells can bind these cells via PTCH yet cannot activate signalling due to the absence of SMO. Just one study has looked at DHH protein in human fetal testis using immunohistochemistry and describes cytoplasmic expression of DHH predominantly in interstitial LCs but also in Sertoli cells at 9 weeks post conception [Buonocore et al., 2019] (Fig. 2). It is not clear whether this Leydig DHH originated in the secreting Sertoli cells or whether LCs are also a source of DHH, and DHH staining is still needed in earlier development phases (i.e., 5–7 weeks gestation) to truly understand its role in human testis differentiation. At later time points (19 gestational weeks), DHH protein expression is reportedly predominantly localised to Sertoli cells [Fowler et al., 2008], and maternal smoking during pregnancy appears to lead to a decrease in both DHH and PTCH1 expression in male fetuses. Since maternal smoking is also associated with reduced testis size with decreased sperm count and quality in male offspring, a phenotype intriguingly similar to that of the Dhh knockout mice, it has been speculated that diminished DHH expression is responsible for this phenotype [Fowler et al., 2008].

It is well established that pathogenic variants in DHH are a rare cause of 46,XY DSD in humans, in particular causing 46,XY partial or complete gonadal dysgenesis with or without peripheral neuropathies (OMIM 607080, OMIM 233420). The DHH gene is highly conserved and shows 20 different DHH pathogenic variants that have been published in 24 cases (Fig. 3, Table 1) with variants falling in both the N- and C-terminal domains (Fig. 3) [Umehara et al., 1999, 2000; Sugie et al., 2002; Canto et al., 2004, 2005; Das et al., 2011; Werner et al., 2015; Eggers et al., 2016; Paris et al., 2017; Sato et al., 2017; Baldinotti et al., 2018; Rothacker et al., 2018; Tajouri et al., 2018; Ayers et al., 2019; Buonocore et al., 2019; Neocleous et al., 2019]. Given that DHH is a potent growth factor that signals in a concentration-dependent manner, it is not surprising that the vast majority of pathogenic cases reported are bi-allelic (homozygous or compound heterozygous), and the contribution of heterozygous variants in DHH to DSD is unclear. Two patients reported to carry a heterozygous variant (p.Leu363Cysfs*4) in DHH also had sex chromosome mosaicism (45,X/46,XY) [Canto et al., 2005] (Table 1), and our group has described a number of DSD patients with heterozygous DHH variants although our functional studies have suggested that these are unlikely to be the major cause [Ayers et al., 2019]. Most patients with bi-allelic DHH variants are 46,XY female, with just 2 reportedly raised as males (Table 1). Complete gonadal dysgenesis has been described in 13 patients with streak gonads and/or Müllerian structures indicating no functional gonadal tissue (Table 1). Ten cases of confirmed or suspected partial gonadal dysgenesis have also been reported. One patient had histological findings suggestive of Leydig cell hypoplasia (Table 1). Five patients with DHH variants have reported gonadal cancer, and 7 had DHH-associated neuropathies (Table 1).

Several groups have used functional genomics to understand the underlying pathology of variants (Table 2). Overexpression constructs of tagged DHH proteins have been used to assay cellular localisation and protein cleavage with varying results which appear to depend on the cells used. Our group found that DHH expressed in HEK293T cells was mostly full length, and expression levels/cleavage did not significantly differ for any variant tested with the exception of the frameshift variant p.Leu363CysFs*4 where no full length protein was observed (Table 2) [Ayers et al., 2019]. In these cells, transfected DHH was expressed in the cytoplasm and cell membrane. Cellular localisation of the 11 DHH variants tested did not differ to the wild type, although the p.Arg164Pro variant had a slightly reduced staining [Ayers et al., 2019]. In contrast, another study found that DHH is cleaved when expressed in HeLa cells, and that a number of patient variants had reduced or absent self-cleavage [Elzaiat et al., 2020] (Table 2). In addition, HeLa cells transfected with DHH-GFP variants had 2 patterns of intracellular localisation, termed by the authors “granular” (cytoplasmic expression) and “ring-type” (homogenous expression surrounding the nucleus). A significantly lower proportion of the granular-type staining was described in some of the variants with impaired protein processing [Elzaiat et al., 2020] (Table 2). A cell-free in vitro cleavage assay has also shown that a fusion human DHH protein can undergo cleavage (with the addition of DTT or β-mercaptoethanol) [Tajouri et al., 2018] and 2 C-terminus patient variants led to reduced [p.(Glu212Lys)] or absent [p.(Asn337Lysfs*24)] cleavage efficiency in this system [Tajouri et al., 2018]. Interestingly, using HEK293T cells expressing wild type or variant DHH co-cultured with TM3 mouse LCs (known to be responsive to DHH [Pathi et al., 2001]), our group found reduced or absent pathway activity for 9 DHH variants (Table 2). This suggests that despite the lack of cleavage, variants in DHH affect its activity on LCs in this system.

HHAT is a palmitoyl acyltransferase for HH proteins [Buglino and Resh, 2008], and its loss in mice affects Hh secretion and long-range diffusion, resulting in holoprosencephaly, acrania, and agnathia, likely due to reduced Shh signalling [Dennis et al., 2012], as well as disrupted testis cord formation and reduced LC differentiation [Callier et al., 2014]. In humans, HHAT is expressed in embryonic ovaries and testes at 9 weeks gestation, where it is higher in Sertoli cells compared to LCs [Callier et al., 2014]. Recessive variants in HHAT cause Nivelon-Nivelon-Mabille syndrome (NNMS; OMIM 600092), characterised by progressive microcephaly, vermis hypoplasia, and skeletal dysplasia. Just one 46,XY patient with NNMS has been described [Nivelon et al., 1992; Thauvin-Robinet et al., 2005; Callier et al., 2014], and this individual was born with normal external female genitalia. Histology confirmed testicular dysgenesis with dysplastic/immature seminiferous tubules and a dense interstitial compartment with reduced numbers of LCs. Persistent Müllerian and Wolffian duct structures were noted. The patient carried a homozygous HHAT p.Gly287Val missense variant inherited from unaffected carrier parents, which fell in the O-acyltransferase domain and abolished palmitoylation of HH proteins [Callier et al., 2014]. This suggests that loss of HHAT disrupts DHH signalling in the human testis.

An interesting feature to vertebrate HH signalling is the importance of the primary cilia [Huangfu et al., 2003; Bangs and Anderson, 2017]. In the absence of HH, PTCH is present in the cilia and when HH binding inactivates PTCH, SMO accumulates in the cilia [Corbit et al., 2005; Rohatgi et al., 2007] where it controls the activation of the GLIs which are enriched at the cilia tip. Dhh signalling is thought to be regulated via the primary cilium in mouse LCs, as shown in TM3 culture cells [Nygaard et al., 2015], and cilia are also likely to play a role in human testis DHH signalling. PMCs are positive for primary cilia, and while mature LCs do not commonly possess primary cilia, less differentiated LCs often do [Nygaard et al., 2015]. This would indicate that DHH signalling proceeds via the primary cilia in these cells. This idea is supported by findings in humans where defects in ciliary assembly or trafficking lead to numerous developmental disorders collectively referred to as ciliopathies. These include Bardet-Biedl (BBS), Joubert, and Meckel-Gruber syndromes [Reiter and Leroux, 2017]. Reproductive disorders are not uncommon in ciliopathies. For example, BBS patients often present with phenotypes such as LC or general testis hypoplasia [Green et al., 1989]. It has not yet been established whether these phenotypes are due to primary failure in testis differentiation or later from a disrupted adrenal-pituitary-gonadal axis [Soliman et al., 1996], and therefore, further work is required to understand the exact role of primary cilia in DHH signalling in the testis.

In addition, enzymes involved in lipid biosynthesis in the ciliary microenvironment that promote Hedgehog signalling may also be important contributors to DSD. DHCR7 (7-dehydrocholesterol reductase) is an enzyme that synthesises cholesterol, and recent studies have shown that DHCR7 localises near the ciliary base to prime the cilia for HH pathway activation [Findakly et al., 2021]. Children with variants in DHCR7 develop Smith-Lemli-Opitz syndrome (SLOS; OMIM 270400), with elevated 7-DHC and low cholesterol levels and a broad phenotypic spectrum including underdeveloped external genitalia and poorly differentiated testicular tissue. These studies were not carried out in gonadal tissue/cell lines; however, given that HH molecules like DHH are cholesterol-modified and that cilia-based activation may require local cholesterol synthesis, it would be interesting to investigate whether variants in DHCR7 lead to DSD via a decrease in gonadal DHH signalling.

Despite the overwhelming evidence that DHH variants cause 46,XY gonadal dysgenesis, the exact role of DHH and its signalling pathway in the human gonad are still not well understood. Outstanding questions surround the expression, cleavage, and secretion of the DHH ligand in the human gonads. Further research into the relative importance of juxtacrine, paracrine, and even autocrine signalling is required and may reveal differential mechanisms for each gonadal cell type. Thus far, this research has been severely impeded by the difficulty in obtaining primary human embryonic gonadal tissue at a relevant time point for fetal gonadal development (i.e., weeks 5–9 of gestation). To add to this, an adequate human embryonic gonadal cell line is currently lacking. We and others have begun to address this issue using differentiation methods for pluripotent stem cells into several cell types of the embryonic gonad, that is, bipotential cells, early Sertoli cells, and immature LCs [Bucay et al., 2009; Buganim et al., 2012; Yang et al., 2015, 2017; Sepponen et al., 2017; Rodriguez Gutierrez et al., 2018; Chen et al., 2019; Knarston et al., 2020]. We have observed the induction of DHH expression in 3D organoid cultures, thus providing a valuable future tool to investigate the role of key testicular genes such as DHH during human fetal testis development [Knarston et al., 2020].

Interestingly, animal models suggest that the major role of Sertoli-secreted Dhh is to drive differentiation of the interstitial LC and PTM cells, and mice nulls often have testes with reduced populations of these cells and disrupted testis cords. However, in humans, the majority of patients with bi-allelic DHH variants demonstrate complete gonadal dysgenesis, a more severe phenotype. This suggests that DHH may play an additional role in human testis, for instance in Sertoli cell differentiation or maintenance. Alternatively, human gonads may show a greater sensitivity to disruption of the testis cord basement membrane, observed in several animal models, perhaps resulting in complete degeneration/loss of the gonad. Interestingly, DHH has been suggested to play a role in vascular endothelial cells, in particular in maintaining the integrity of the blood-brain and the blood-nerve barrier by regulating endothelial tight junction expression [Chapouly et al., 2016; Caradu et al., 2018; Hollier et al., 2020]. DHH has also been shown to regulate the expression of proangiogenic factors in Schwann cells [Renault et al., 2013]. A similar role for DHH in the formation of the blood-testis barrier or testis vasculature is therefore feasible.

A DHH genetic diagnosis can guide clinical management and patient monitoring, in particular, in relation to gonadal malignancies and neuropathies. Indeed, 7 of the 24 patients previously reported with DHH-associated DSD also had neuropathy, and 5 patients had reported gonadal cancers (including seminoma in situ, dysgerminoma, and gonadoblastoma) [Canto et al., 2004, 2005; Werner et al., 2015]. However, functional studies are yet to reveal which DHH variants carry greater risk of these additional features or whether this is due to additional environmental or genetic factors. These genotype-phenotype correlations require more investigation and may be revealed only with additional reported cases and a greater understanding of the DHH pathway role and regulation in the human embryo.

It is also interesting that very few additional DHH pathway genes have been implicated in DSD thus far. This is likely explained by the fact that DHH, SHH, and IHH all converge on a central pathway, mutations in which lead to severe phenotypes or even embryonic lethality. In these cases, a fertility or sex-reversal phenotype may be overlooked. Currently, less than 50% of patients born with a DSD will receive a genetic indication [Eggers et al., 2016]. Future advances in sequencing technologies may identify variants in additional DHH signalling pathway components with gonadal-specific mosaicism as genetic risk factors or as oligogenetic contributors to DSD.

The HH signalling pathway is a well-studied regulator of embryonic development and when disrupted causes severe genetic defects. The HH morphogen DHH specifically regulates fetal gonadal development, and DHH variants cause 46,XY gonadal dysgenesis in humans, sometimes combined with peripheral neuropathy. A detailed understanding of how DHH signalling works in human is still lacking, and more research is required to understand the pathological mechanisms underlying DHH-related DSDs. This is required for better diagnosis, prognosis, and clinical management of these patients.

The authors have no conflicts of interest to declare.

K.L.A and S.P. are funded by an Ideas Grant (APP1156942) from the National Health and Medical Research Council of Australia (NHMRC).

All authors wrote the review.