Article Text

Download PDFPDF

Original article
Disruption of a long distance regulatory region upstream of SOX9 in isolated disorders of sex development
  1. Sabina Benko1,2,
  2. Christopher T Gordon1,2,
  3. Delphine Mallet3,
  4. Rajini Sreenivasan4,5,
  5. Christel Thauvin-Robinet6,
  6. Atle Brendehaug7,
  7. Sophie Thomas1,2,
  8. Ove Bruland7,
  9. Michel David8,
  10. Marc Nicolino8,
  11. Audrey Labalme9,
  12. Damien Sanlaville9,
  13. Patrick Callier10,
  14. Valerie Malan11,
  15. Frédéric Huet12,
  16. Anders Molven13,14,
  17. Frédérique Dijoud15,
  18. Arnold Munnich1,2,16,
  19. Laurence Faivre6,
  20. Jeanne Amiel1,2,16,
  21. Vincent Harley4,
  22. Gunnar Houge6,17,
  23. Yves Morel3,
  24. Stanislas Lyonnet1,2,15
  1. 1INSERM U-781, Hôpital Necker-Enfants Malades, Paris, France
  2. 2Université Paris Descartes, Faculté de Médecine, Paris, France
  3. 3Endocrinologie Moléculaire et Maladies Rares, Centre de Biologie et Pathologie Est, Hospices Civils de Lyon, Bron, France
  4. 4Molecular Genetics and Development, Prince Henry's Institute of Medical Research, Monash Medical Centre, Clayton, Australia
  5. 5Department of Anatomy and Cell Biology, Faculty of Medicine, Dentistry and Health Sciences, University of Melbourne, Parkvill, Australia
  6. 6Centre de Génétique, Hôpital d'Enfants, Dijon, France
  7. 7Center for Medical Genetics and Molecular Medicine, Haukeland University Hospital, Bergen, Norway
  8. 8Endocrinologie Pédiatrique, HFME, Hospices Civils de Lyon, Bron, France
  9. 9Cytogénétique, Centre de Biologie et Pathologie Est, Hospices Civils de Lyon, Bron, France
  10. 10Cytogénétique, Plateforme de Biologie, CHU Dijon
  11. 11AP-HP, Service d'Histo-embryo-cytogénétique, Hôpital Necker-Enfants Malades, Paris, France
  12. 12Pédiatrie, Hôpital d'Enfants, CHU Dijon
  13. 13The Gade Institute, University of Bergen, Bergen, Norway
  14. 14Department of Pathology, Haukeland University Hospital, Bergen, Norway
  15. 15Service d'anatomopathologie, Centre de Biologie Est, Hospices Civils de Lyon, Bron, France
  16. 16AP-HP, Département de Génétique, Hôpital Necker-Enfants Malades, Paris, France
  17. 17Department of Clinical Medicine, University of Bergen, Bergen, Norway
  1. Correspondence to Professor Stanislas Lyonnet, Département de Génétique et Unité INSERM U-781, Hôpital Necker-Enfants Malades, 149 rue de Sèvres, Paris cedex 15 75743, France; stanislas.lyonnet{at}


Background The early gonad is bipotential and can differentiate into either a testis or an ovary. In XY embryos, the SRY gene triggers testicular differentiation and subsequent male development via its action on a single gene, SOX9. The supporting cell lineage of the bipotential gonad will differentiate as testicular Sertoli cells if SOX9 is expressed and conversely will differentiate as ovarian granulosa cells when SOX9 expression is switched off.

Results Through copy number variation mapping this study identified duplications upstream of the SOX9 gene in three families with an isolated 46,XX disorder of sex development (DSD) and an overlapping deletion in one family with two probands with an isolated 46,XY DSD. The region of overlap between these genomic alterations, and previously reported deletions and duplications at the SOX9 locus associated with syndromic and isolated cases of 46,XX and 46,XY DSD, reveal a minimal non-coding 78 kb sex determining region located in a gene desert 517–595 kb upstream of the SOX9 promoter.

Conclusions These data indicate that a non-coding regulatory region critical for gonadal SOX9 expression and subsequent normal sex development is located far upstream of the SOX9 promoter. Its copy number variations are the genetic basis of isolated 46,XX and 46,XY DSDs of variable severity (ranging from mild to complete sex reversal). It is proposed that this region contains a gonad specific SOX9 transcriptional enhancer(s), the gain or loss of which results in genomic imbalance sufficient to activate or inactivate SOX9 gonadal expression in a tissue specific manner, switch sex determination, and result in isolated DSD.

  • Disorder of sex development
  • sox9
  • campomelic dysplasia
  • gonad
  • non-coding DNA
  • genetics
  • molecular genetics
  • reproductive medicine
  • clinical genetics
  • copy-number
  • developmental
  • epilepsy and seizures
  • cytogenetics
  • diabetes
  • cancer: dermatological
  • pancreas and biliary tract
  • paediatric oncology
  • genetic screening/counselling
  • parkinson-s disease
  • cytogenetics
  • endocrinology
  • adrenal disorders
  • other endocrinology

Statistics from


Heterozygous loss-of-function mutations in the SOX9 gene are responsible for the human polymalformation syndrome, campomelic dysplasia (CD; MIM 114290), comprising multiple endophenotypes in accordance with the complex pattern of SOX9 expression in tissues that include cartilage, testes, notochord, neural crest, inner ear, and the central nervous system. Precise spatiotemporal SOX9 expression is coordinated by a complex regulatory region extending at least 1.5 Mb upstream and downstream of its coding sequence and harbouring tissue specific transcriptional regulatory elements.1 2

During sex determination, the only known function of the Y chromosome sex determining gene SRY is to stimulate SOX9 expression in the supporting cell lineage of the bipotential gonad, which will consequently differentiate into testicular Sertoli cells.3 Conversely, if SOX9 expression is turned off, the supporting cells of the bipotential gonad will differentiate as ovarian granulosa cells.3 A male-to-female disorder of sex development (46,XY DSD) occurs in ∼75% of XY CD patients,4 while ectopic gonadal expression of SOX9 in XX humans and mice leads to testicular development with subsequent female-to-male sex reversal,5–7 suggesting a critical requirement for appropriate levels of SOX9 expression during gonadal development.

Long distance genomic alteration at the SOX9 locus can be associated with isolated Pierre Robin sequence (PRS),1 a craniofacial anomaly that is a typical feature of CD. It was proposed that the genomic alterations in PRS patients cause a site specific and stage specific loss of transcription, resulting in a restricted reduction of SOX9 expression in human cranial neural crest cells and/or mandibular arch.1 2 We speculated that other endophenotypes of CD could also occur in an isolated form following the same principle of tissue restricted alterations of SOX9 expression due to disruption of tissue specific, long distance regulatory regions. According to this model, tissue specific alteration of SOX9 expression in the developing bipotential gonad may result in an isolated form of DSD. To test this hypothesis we performed copy number variation mapping in SRY negative 46,XX patients with isolated female-to-male DSD and SRY positive 46,XY patients with isolated male-to-female DSD.


Genomic copy number analysis

Genomic copy number analyses for the DSD1 and DSD4 index cases were performed using Agilent Human Genome comparative genomic hybridisation (CGH) 244K oligonucleotide arrays. The DSD3 duplication was mapped using Agilent 180K and 400K oligonucleotide arrays. For the DSD2 case, genomic copy number analyses were initially done on Affymetrix Genome-Wide Human SNP 6.0 arrays, followed by fine-mapping on Affymetrix Cytogenetics Whole-Genome 2.7M arrays. The junction fragment of the DSD2 duplicated region was PCR amplified using primers between markers rs11871027 and rs11871770 at the 5′ end, and rs2429978 and rs2430557 at the 3′ end, followed by sequencing of the PCR product.

Multiplex ligation dependent probe amplification (MLPA) was used to identify the DSD3 duplication and DSD4 deletion, and to confirm the DSD1 duplication. The SALSA MLPA kit P185 intersex was used according to the manufacturer's recommendations (MRC-Holland, Amsterdam, The Netherlands). The probe ligation sites were −1499, −1008, −1007, −483, and −243 kb upstream of the SOX9 gene.


To determine SOX9 haplotypes and allele segregation in the DSD2 family, simple tandem repeat markers (D17S1350, D17S1797, D17S1351, and SOX9M2; two upstream and two downstream of the SOX9 gene) were PCR amplified and their size determined. The primers used for SOX9M2 were 5′-TAA CTC TTG TAG CCA TTG TCA-3′ (forward) and 5′-ACT GAA CAA ATG CCT GGA-3′ (reverse).

Metaphase FISH

To confirm the absence of the SRY gene in the DSD1, DSD2, and DSD3 patients, metaphase chromosome analysis using two colour fluorescent in situ hybridisation (FISH) was performed as previously described,8 using a specific SRY FISH probe (LSI SRY/CEPX Vysis). For the DSD1 patient, metaphase FISH was performed with a probe for the duplicated region (RP11-879D6).

Clinical data

DSD1 index patient

Clinical examination of the DSD1 46,XX patient at birth revealed evidence of an isolated male external genitalia malformation, including bifid scrotum with two palpable gonads, incurved short penis (1.3 cm) and hypospadias. There was no dysmorphism, scoliosis or skeletal dysplasia. Skeletal radiographic findings were normal. Pelvic ultrasound detected two gonads in the scrotum, while abdominal and pelvic MRI showed epididymal structures and no pelvic formation resembling uterine structures. At 21 days after birth, hormonal data revealed a male gonadal–pituitary activity termed ‘minipuberty’ with a peak of serum testosterone (7.8 nmol/l) correlated with serum leutinising hormone (LH) (5.5 IU/l). In contrast, low serum anti-Mullerian hormone (AMH) (109 pmol/l; normal values (mean±1SD (value range)): 882±335 (286–2116) pmol/l) and inhibin B (67 pg/ml; normal range: male 190±30 pg/ml; female 20±5 pg/ml) and slightly high serum follicle stimulating hormone (FSH) suggested a degree of testicular dysgenesis. Macroscopic studies performed during surgery for genitoplasty showed bilateral fallopian tubes associated with a gonadal structure, supporting the presence of ovotestes.

DSD2 index patient

The 46,XX index patient was born with a severe DSD involving perineal hypospadias and an asymmetric scrotum that was more developed on the right side. In the right scrotal half a normal testis was present, while the left scrotal half contained an ovarian remnant with fallopian tube structures. The gonadal histology confirmed the macroscopic findings. The patient also had small and rudimentary anlagen for a vagina and uterus that were surgically removed. His psychomotor development, intelligence, pubertal development, and serum testosterone values were normal. He displayed normal growth and bodily proportions, and there was therefore no reason to suspect a skeletal dysplasia.

DSD3 index patient

At birth, the patient had a severe 46,XX DSD, comprised of perineal hypospadias, a ventrally curved phallus (2.5 cm), a hypoplastic and asymmetric scrotum, a right palpable gonad in the scrotum, and no palpable left gonad. At 1 year, hormonal data supported the diagnosis of gonadal dysgenesis: the patient had low AMH (94 pmol/l; normal values (mean±1SD (value range)): 1736±616 (639–4364) pmol/l), normal LH (5.7 IU/l, normal values (mean (value range)): 3.6 IU/l (1.4–6.0 IU/l)), and high FSH (24.7 IU/l, normal values (mean (value range)): 4.3 IU/l (2.3–6.9 IU/l)) after gonadotrophin releasing hormone (GnRH) test, and subnormal serum testosterone after human chorionic gonadotrophin (hCG) test (10.4 nmol/l; normal value >10 nmol/l). A vaginal pouch and uterus were present at genitography. As the decision was made to rear this newborn as a girl, feminising genitoplasty and bilateral gonadectomy were performed at the age of 15 months. The histological examination showed that the right gonad corresponded to a streak gonad partially differentiated towards an ovary (comprising dispersed primordial ovocytes and rare follicles). The left gonad corresponded to a typical ovotestis with a testicular portion having numerous seminiferous tubules lined only with Sertoli cells and a small ovarian portion with abundant primordial follicles. An epididymal structure and a fallopian tube were both present. No bone abnormalities were identified.

DSD3 deceased sibling: A prenatal diagnosis was performed by chorionic villus sampling (CVS) during the second pregnancy. As the karyotype was 46,XX without abnormalities, a follow-up by ultrasonography showed a normal fetus with male external genitalia. A therapeutic abortion was performed. The duplication of the SOX9 region was determined from DNA extracted by CVS.

DSD4 family

Case 1

At birth, the 46,XY index patient had a severe DSD with asymmetric external genitalia, consisting of a urogenital sinus with a phallus (2.0 cm), a right hemiscrotum with palpable gonad, and a left labioscrotal fold with a gonad which was not palpable. At day 45, hormonal data resembled that of a classic minipuberty observed in normal 46,XY males, with normal LH (4.6 IU/l), FSH (3.0 IU/l), and serum testosterone (4.6 nmol/l) concentrations. AMH was low (65 pmol/l), indicating a testicular dysgenesis. As the decision was made to rear this neonate as a girl, feminising genitoplasty and bilateral gonadectomy were conducted at the age of 4 months. On microscopic examination, the right gonad was a small testis with normal architecture and spermatogonia in seminiferous tubules, while the left gonad was a streak gonad associated with a fallopian tube and a hemi-uterus.

Case 2

The DSD4 family ‘case 2’, with a normal external female phenotype, was independently diagnosed with a 46,XY DSD. Dispersed pubic hair was noted at the age of 8 years. A subsequent examination of 24 h urinary gonadotrophins showed increased concentrations of FSH. These findings prompted karyotyping, the result of which was 46,XY. Successive ultrasonographies showed the presence of a uterus with ‘ovaries’. At the age of 9 years, serum hormonal values, especially low AMH and high FSH, suggested gonadal dysgenesis: serum testosterone 0.18 nmol/l (normal values 2±0.9 nmol/l), LH 0.5 IU/l (normal values <0.2–2.1 IU/l), FSH 23.8 IU/l (normal values <0.3–3.0 IU/l), AMH 6 pmol/l (normal values (mean±1SD (value range)) 1736±616 (639–4364) pmol/l). During the coeloscopy, a bilateral gonadectomy was performed. A left gonad compatible with an ovary was noted. The right gonad corresponded to a streak gonad with gonadoblastoma. The gonadoblastoma was well limited and contained calcifications and two types of cells: germ cells expressing placental-like alkaline phosphatase and CD117, and sex cord cells expressing inhibin and WT1.

A careful investigation of the DSD4 index cases and their parents excluded any bone (scoliosis or skeletal dysplasia) or craniofacial abnormalities reminiscent of campomelic dysplasia.


We report here on three unrelated SRY negative cases (DSD1–DSD3) with a 46,XX karyotype and isolated female-to-male ovotesticular DSD, and a fourth (DSD4) familial case comprising two 46,XY patients with isolated male-to-female DSD of variable expression in which we identified copy number variations at the SOX9 locus.

In the sporadic DSD1 patient, we identified by CGH and confirmed by MLPA a de novo duplication of ∼605–695 kb of a region at least ∼353 kb upstream of the SOX9 transcription start site (figures 1 and 2A; maximum duplication size Hg19 coordinates: chr17:69069079–69764059, minimum duplication size Hg19 coordinates: chr17: 69107506–69712726). Metaphase FISH with a probe for the duplicated region demonstrated the absence of this region elsewhere in the genome, and an asymmetry in signal intensity for the chromosome 17 pairs, suggesting the duplication was likely to be in situ. In the second case (DSD2, figure 2B) a ∼148 kb tandem duplication of the region −595 to −447 kb upstream of SOX9 was identified (figure 1) by CGH and further confirmed by PCR amplification and sequencing of the junction fragment (duplicated region, Hg19 coordinates: chr17:69521863–69670036). The duplication was paternally inherited. The father inherited the duplication from his unaffected mother. The duplication was shared with two of the proband's 46,XY brothers and was not transmitted to the healthy 46,XX sister (figure 2B). In the index case of the third family (DSD3, figure 2C), a duplication upstream of SOX9 was identified by MLPA and confirmed by 180K CGH. Fine-mapping by 400K CGH indicated maximum duplication size coordinates of chr17:68829028–69609453 (Hg19) and minimum duplication size coordinates of chr17:68838024–69599915 (Hg19), corresponding to a duplication of ∼762–780 kb, at least ∼508 kb upstream of SOX9. Two affected 46,XX probands carried the duplication, inherited from a healthy 46,XY father and grandfather. All 46,XY individuals carrying this duplication were normal.

Figure 1

Rearrangements in the SOX9 centromeric gene desert are associated with isolated 46,XX and 46,XY disorders of sex development (DSD). The region surrounding SOX9 is represented as a grey bar with genes depicted in black, and proximal and distal campomelic dysplasia (CD) translocation breakpoint clusters (BPCs) as shadows within the locus. The copy number gains associated with ovotesticular 46,XX DSD (DSD1,DSD2 and DSD3) and with complete 46,XX DSD reported by Cox et al. (duplication; Dup4) and Vetro et al. (triplication; Trip1) are depicted as bars above the proximal SOX9 gene desert. Brachydactyly-anonychia duplications are indicated as BA1, BA2, BA3 and BA4.13 The XX DSD translocation breakpoint is indicated by an arrow.9 The deletions associated with syndromic 46,XY DSD (CD1, CD2, and CD3, 10–12) and isolated 46,XY DSD (DSD4) are represented as bars below the proximal SOX9 gene desert. The TESCO element and the mandibular enhancer HCNE-F2 are represented as dots. The minimal 78 kb RevSex region, shared among DSD1, DSD2, DSD3, Dup4 and Trip1 copy number gains and DSD4, CD1, CD2 and CD3 deletions is indicated as a shadow within the locus. Coordinates are according to the GRCh37/hg19 genome assembly.

Figure 2

Pedigrees of 46,XX and 46,XY disorder of sex development (DSD) families. (A) A ∼605–695 kb duplication occurred de novo in the DSD1 sporadic case of 46,XX DSD. (B) A ∼148 kb duplication in the DSD2 case with a 46,XX DSD is inherited from the father and from the unaffected paternal grandmother. (C) In the DSD3 case, a ∼762–780 kb duplication in the two affected 46,XX probands was inherited from their healthy 46,XY father and grandfather. (D) The ∼240 kb DSD4 deletion is shared between two 46,XY DSD patients of the same family (one with a female phenotype, the other with severe ambiguous and asymmetric external genitalia). The deletion is transmitted by unaffected 46,XX mothers. dup: duplication within chromosome 17q24.3; del: deletion within chromosome 17q24.3. Open squares: normal males; open circles: normal females; black filled symbols, isolated 46,XY or 46,XX DSD.

Conversely, in the two patients of the fourth family (DSD4, figure 2D) with isolated 46,XY DSD, one with a female phenotype and the other with severe ambiguous asymmetric external genitalia, a ∼240 kb deletion located between ∼405 and ∼645 kb upstream of the SOX9 transcription start site was detected by MLPA and delimited by CGH analysis (figure 1). No dysmorphism or skeletal abnormalities reminiscent of CD were detected. The mothers of the two subjects were sisters and carried the same deletion (figure 2D).

DSD4 was the only case from a 46,XY DSD patient cohort (29 cases with complete female phenotype, 118 with undermasculinised external genitalia) that were screened by MLPA and quantitative PCR for copy number variation in the SOX9 proximal gene desert and in which a disruption upstream of SOX9 was identified. DSD1 and DSD3 were the only cases, from a cohort of 14 cases of 46,XX DSD patients (ovotesticular or SRY negative testicular DSD), with testicular tissue and high AMH, in which a copy number variation upstream of SOX9 was identified.

None of the genomic alterations reported here were recorded as copy number polymorphisms in the Database of Genomic Variants. Moreover, the sequencing of several genes (SRY, NR5A1, WT1, DHH, FGF9, MAP3K1) whose mutations are associated with 46,XY DSD did not reveal mutations in DSD4. No abnormalities of the WNT4 and RSPO1 genes were detected in DSD1, DSD2, and DSD3 46,XX patients.


We report here on copy number variations located a long distance from the SOX9 gene in four unrelated families with isolated DSD (DSD1-DSD4). Three cases (DSD1-DSD3) are SRY negative patients with a 46,XX karyotype and isolated female-to-male ovotesticular DSD, associated with duplications in the centromeric SOX9 gene desert. Those duplications are overlapping with two recently reported copy number gains at the SOX9 locus identified in families with isolated complete 46,XX DSD (Dup4 and Trip1 in figure 1).14 15 Conversely, the fourth case, DSD4, is a familial case comprising two 46,XY patients with isolated male-to-female DSD in which we identified a small deletion overlapping the duplications identified in DSD1-DSD3. We observed incomplete penetrance in the DSD2 family and phenotypic variability in the DSD4 family, which is concordant with the variability of the phenotype previously reported in familial cases of 46,XX DSD,16 17 as well as with the variable penetrance of the female-to-male sex reversal phenotype in Ods/+ transgenic mice on different genetic backgrounds,18 and the fact that 46,XY DSD is variable and is penetrant in only ∼75% of CD patients.

Collectively, our data and the previously reported genomic alterations at the SOX9 locus associated with syndromic 46,XY DSD (CD1, CD2, and CD3 deletions in figure 1)10–12 or isolated 46,XX DSD (Dup4 and Trip1 in figure 1)14 15 point towards the existence of a 78 kb sex determining region (RevSex; figure 1) located ∼517–595 kb 5′ to the SOX9 gene. These cases thus suggest that one or more regulatory elements essential for gonadal development map to this region. The gain-of-function of such elements resulting from genomic gains in the DSD1, DSD2, DSD3, Dup4, and Trip1 patients might activate gonadal SOX9 expression in 46,XX embryos and subsequently induce testis differentiation. Conversely, their loss-of-function resulting from a deletion in the DSD4 family patients could be sufficient to cause tissue specific loss of gonadal SOX9 expression in 46,XY embryos, inhibiting the formation of a normal testis and favouring the female developmental pathway, while SOX9 expression would remain unaffected in other tissues. These observations suggest that the putative regulatory element(s) within the RevSex region and disrupted in DSD1–DSD4 would be testicular transcriptional enhancers critical for testicular development.

Of note, the distal translocation breakpoint cluster (distal BPC; figure 1) located centromeric to RevSex includes breakpoints that resulted in 46,XY DSD19 and 46,XX DSD cases,9 most probably through their opposite positional effects on a region telomeric to the distal BPC. Also, three large duplications (BA1, BA2, and BA3; figure 1) including the SOX9 centromeric gene desert and overlapping with the duplications identified in the DSD1–DSD3 cases have been reported in brachydactyly-anonychia syndrome patients without any DSD phenotype.13 We thus hypothesise that RevSex contains testicular enhancer(s) whose activity is regulated by the chromatin environment. The chromatin environment in which RevSex is found would thus be either permissive for enhancer activity in males or repressive in females. A similar situation, involving regulation of enhancer activity through global chromatin effects, was described at the Hoxd locus, at which the activity of tissue specific enhancers depends on the balance between, on the one hand, a repressive chromatin state emanating from the centromeric side and, on the other hand, an activating chromatin state that progresses from the telomeric side of the Hoxd cluster.20 We therefore suggest that in the case of DSD1–DSD3 duplications, it is not simply the third copy of RevSex that would cause the increased SOX9 gonadal expression in the 46,XX gonads, but rather that the putative testicular enhancer(s) localised in the RevSex region become active due to their location in a permissive ‘male’ chromatin environment achieved by a local chromatin alteration caused by the duplications. Conversely, in the non-DSD BA1–BA3 patients with large duplications of the SOX9 proximal gene desert, the resulting chromatin environment of RevSex in XX gonadal cells would remain in its repressive ‘female’ form and thus those alterations would not result in DSDs.

To date, TESCO (Testis-specific enhancer of SOX9 core), located in humans at -13 kb upstream of SOX9, is the only element identified within the SOX9 regulatory domain that is able to drive tissue specific reporter gene expression in the early developing testes of transgenic mice,21 while its activity is repressed in the developing and adult ovary.22 While in vitro studies demonstrate that human TESCO can be activated by human SRY and SOX9 in cooperation with SF1,23 TESCO function in humans is not known and mutations in TESCO are likely to be rare in XY DSD, if they occur at all.24 Yet in a number of patients, CD is associated with 46,XY DSD caused by translocations with breakpoints far upstream of both TESCO and SOX9 (distal and proximal BPCs; figure 1).2 19 In addition, in the DSD4 case and in three reported cases of CD with male-to-female DSD (CD1, CD2, and CD3; figure 1), the DSD is associated with deletions centromeric to both SOX9 and TESCO.10–12 Moreover, the translocations of the proximal BPC are always associated with male-to-female DSD, while the translocations of the distal BPC are associated with male-to-female as well as female-to-male DSD.2 9 These data argue against a position effect on a region telomeric to the proximal BPC comprising TESCO and SOX9, and favour the existence of one or more testis specific enhancers of SOX9 expression essential for male sex determination and positioned between the proximal and distal BPCs, ∼500 kb upstream of SOX9 and TESCO in the RevSex region.

In conclusion, our study, in conjunction with previously reported data, supports a model of long distance genomic disruption25 at the SOX9 locus as the genetic basis of several isolated 46,XX and 46,XY DSDs. More broadly, the SOX9 locus continues to serve as a paradigm in which analysis of genomic lesions in non-coding DNA provides insight into fundamental mechanisms of developmental gene regulation.


We are grateful to the families for participating in the study. We thank the Centre de Référence Maladies Rares Anomalies du Development, the clinicians for the sample collection, Stefan Bagheri-Fam for his contribution and critical discussions, and E Sapin, JB Cotton and P Mouriquand for their contributions to surgical data, samples and discussions.



  • YM and SL are senior co-authors.

  • Funding This work was supported by INSERM, ANR (MRare 2007 and EvoDevoMut 2010), the Hospices Civils de Lyon, and the NHMRC (Australia). S Benko was supported by the Fondation pour la Recherche Médicale (FRM).

  • Competing interests None to declare.

  • Patient consent Obtained.

  • Ethics approval CCP Ile-de-France II.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.