Y Genes: Unraveling the Molecular Pathogenesis of Disorders of Sex Development

SeX(X)Y genes: unraveling the molecular pathogenesis of Disorders of Sex Development

Dorien Baetens

Promoter: Prof. dr. Martine Cools

Co-promoter: Prof. dr. Elfride De Baere

2017

Ghent University, Faculty of Medicine and Health Sciences

Thesis submitted to fulfil the requirements for the degree of

Doctor of Health Sciences

Thesis submitted to fulfill the requirements for the degree of Doctor in Health Sciences.

Promoter Prof. dr. Martine Cools Ghent University, Ghent, Belgium

Co-promoter Prof. dr. Elfride De Baere Ghent University, Ghent, Belgium

Members of the examination committee Prof. dr. Jan Gettemans Ghent University, Ghent, Belgium

Prof. dr. Piet Hoebeke Ghent University, Ghent, Belgium

Prof. dr. Paul Coucke Ghent University, Ghent, Belgium

Prof. dr. Björn Heindryckx Ghent University, Ghent, Belgium

Prof. dr. Olaf Hiort University of Lübeck, Lübeck, Germany

Dr. Sarah Vergult Ghent University, Ghent, Belgium

Dr. Remko Hersmus Erasmus University Rotterdam, Rotterdam, The Netherlands

The research described in this thesis was conducted in the Center for Medical Genetics, and at the department of Pediatric Endocrinology Ghent University, Ghent, Belgium. The author and the promoters give the authorization to consult and to copy parts of this work for personal use only. Every other use is subject to copyright laws. Permission to reproduce any material contained in this work should be obtained from the author.

In all things of nature,

there is something of the marvelous

- Aristotle -

TABLE OF CONTENTS

ABBREVIATIONS ...... 1

SUMMARY ...... 5

SAMENVATTING ...... 9

CHAPTER 1: INTRODUCTION...... 13

1 SEX DEVELOPMENT ...... 13 1.1 The undifferentiated stage ...... 15 1.2 Sex determination ...... 16 1.3 Sex differentiation ...... 18 2 DISORDERS OF SEX DEVELOPMENT ...... 21 2.1 Sex chromosome DSD ...... 23 2.2 46,XY Disorders of sex development ...... 25 2.3 46,XX Disorders of sex development ...... 34 2.4 Disorders of ovarian maintenance: primary ovarian insufficiency ...... 37 3 GENETIC PATHWAYS INVOLVED IN SEX DEVELOPMENT AND DSD ...... 41 3.1 Basic genetics ...... 41 3.2 The undifferentiated stage ...... 44 3.3 Male sex developmental pathways ...... 46 3.4 Female sex developmental pathways ...... 60 3.5 Interactions between the male and female pathways and gonadal transdifferentiation ...... 63 4 DIAGNOSIS AND MANAGEMENT OF DSD PATIENTS ...... 66 4.1 Diagnosis ...... 67 4.2 Management ...... 68 5 TOWARDS A FASTER GENETIC DIAGNOSIS FOR DSD PATIENTS ...... 73 5.1 Pre-screening approach ...... 73 5.2 Identification of new DSD genes: whole exome sequencing ...... 77 6 REFERENCES ...... 79

CHAPTER 2: OUTLINE & OBJECTIVES ...... 91

CHAPTER 3: 46,XY DSD ...... 97

PAPER 1: Extensive clinical, hormonal and genetic screening in a large consecutive series of 46,XY neonates and infants with atypical sexual development ...... 99

PAPER 2: Biallelic and monoallelic ESR2 mutations are associated with 46,XY disorders of sex development ...... 131

CHAPTER 4: 46,XX DSD ...... 171

PAPER 3: NR5A1 is a novel disease gene for 46,XX testicular and ovotesticular disorders of sex development ...... 173

PAPER 4: FOG2/ZFPM2 haploinsufficiency causes primary ovarian insufficiency ...... 213

CHAPTER 5: NON-CODING VARIATION IN DSD ...... 233

PAPER 5: Non-coding variation in disorders of sex development ...... 235

CHAPTER 6: DISCUSSION & PERSPECTIVES ...... 261

Discussion ...... 263

Limitations of our research and future perspectives ...... 273

CURRICULUM VITAE ...... 283

DANKWOORD………………………………………………………………………………………………………………………293

ABBREVIATIONS

7-DHC 7-dehydrocholesterol

AIS Androgen insensitivity syndrome

AMH Anti-Müllerian hormone

APOD Apolipoprotein D

AR Androgen receptor

Array-CGH Array-comparative genomic hybridization

BMP15 Bone morphogenetic protein 15

BPES Blepharophimosis – ptosis – epicanthus inversus syndrome

CAH Congenital adrenal hypoplasia

CBX2 Chromobox 2

CMGG Center for Medical Genetics Ghent

CNV Copy number variant

COST Cooperation in Science and Technology

CYP11A1 Cytochrome P450 Family 11 Subfamily A Member 1

CYP17A1 Cytochrome P450 Family 17 Subfamily A Member 1

DAX1 Dosage-Sensitive Sex Reversal

DHCR7 7-DHC reductase

DHH Desert Hedgehog

DHT Dihydrotestosterone dNTP Deoxynucleotide ddNTP Dideoxynucleotide

DNA desoxyribonucleic acid

DPC Days post coitum

DSD Disorder of sex development

ENCODE Encyclopedia of DNA elements

EMS External masculinization scores

EMX2 Empty spiracles 2

ERN European reference network

FGF9 Fibroblast growth factor 9

FGFR2 FGF receptor 2

FOG2 Friend of GATA4 – 2

FOXL2 Forkhead box L2

FMR1 Fragile X mental retardation 1

FSH Follicle stimulating hormone

GATA4 GATA-binding protein 4

GBY Gonadoblastoma gene

GCC Germ cell cancer

GCNIS Germ cell neoplasia in situ

GD Gonadal dysgenesis

GnRH Gonadotropin-releasing hormone

HCG Human chorionic gonadotropin

HMG High mobility group

HSD3B2 Hydroxysteroid 3-β Dehydrogenase 2

HSD17B3 Hydroxysteroid 17-β Dehydrogenase 3

IBD Identity-by-descent

IBS Identity-by-state

2 iPSC Induced pluripotent stem cell

KO Knock-out

KS Klinefelter syndrome

LCH Leydig cell hypoplasia

LH Luteinizing hormone

LHCGR Luteinizing Hormone/Choriogonadotropin Receptor

LHX9 Lim homeobox 9

MAPK Mitogen-activated protein kinase

MDT Multidisciplinary team

MLPA Multiplex Ligation-dependent Probe Amplification

NGS Next-generation sequencing

NR5A1 Nuclear receptor subfamily 5 group A member 1

OT Ovotesticular

PCR Polymerase chain reaction

PIS Polled Intersex

POI Primary ovarian insufficiency

PTGDS Prostaglandine D synthase

RSPO1 R-spondin 1

SBS Sequencing-by-synthesis

SF-1 Steroidogenic factor-1

SLOS Smith-Lemli-Opitz syndrome

SNP Single nucleotide polymorphism

SOX9 SRY-box 9

SRD5A2 Steroid 5 α-Reductase 2

SRY Sex-determining region of Y

StAR Steroidogenic Acute Regulatory Protein

TESCO Testic-specific Enhancer of SOX9 Core region

TRS Testicular regression syndrome

TS Turner syndrome

TSPY Testis-specific protein, Y-encoded

UGT Undifferentiated gonadal tissue

WES Whole exome sequencing

WGS Whole genome sequencing

WNT4 Wingless-Type MMTV Integration Site Family, Member 4

WT1 Wilms tumor 1

SUMMARY

Sex development is without a doubt one of the most intriguing developmental processes in early embryonic life. Given the complexity of the underlying molecular pathways it is not surprising that sex developmental defects are among the most common birth defects. Severe forms of DSD have a prevalence of ~1/2 000 – 4 500 births but milder forms of genital ambiguity, like hypospadias or micropenis, are seen in as much as 1/200 births. Despite significant progress in the identification of new genes and pathways, a molecular diagnosis is only obtained in ~50% of cases. A molecular diagnosis enables refining of a clinical diagnosis and results in improved management.

In addition, unraveling of the molecular pathways behind these rare phenotypes can provide insights in the pathogenesis of more frequent conditions like male infertility and primary ovarian insufficiency (POI).

The main aim of this doctoral project was to unravel pieces of the molecular pathways behind sex development and to identify new genes responsible for DSDs and related conditions like POI.

The first part of this thesis aimed to assess the appropriateness of a standardized genetic work-up in a cohort of 32 neonates and infants with 46,XY undervirilization phenotypes. Our results showed that array-CGH is still a valuable technique in syndromic cases, with a detection rate of 33%, while it is less useful in non-syndromic cases. Sanger sequencing of AR, WT1 and SRY did not reveal any mutations in this cohort, but we did identify three novel NR5A1 mutations. One of these mutations showed preserved male fertility, which is rarely seen in men with NR5A1 mutations.

Given the low diagnostic yield of this sequential Sanger sequencing approach and the high costs, we anticipated that targeted resequencing of gene panels and whole exome sequencing would become the first choice testing methods in the diagnostic work-up of DSD. Furthermore, we concluded that this extensive screening approach had a comparable diagnostic yield in both severe cases (low external masculinization score

(EMS); <7; detection rate 17.6%) and less severe cases (high EMS; ≥ 7; detection rate

26.5%), indicating that the decision for a genetic work-up cannot only be based on phenotypic presentation.

The second part of this dissertation focusses on the identification of new disease genes for DSD and related phenotypes. First of all, we propose ESR2 as a novel disease gene for 46,XY DSD. We identified three novel ESR2 mutations in unrelated patients: one homozygous in-frame deletion in a syndromic 46,XY DSD case, and two heterozygous missense variants in girls with isolated 46,XY DSD. In vitro functional characterization of these variants showed increased transcriptional activation for the in-frame deletion, suggesting a gain-of-function effect. For the heterozygous missense variants, no significant differences were observed. Structural analysis of the mutant proteins showed altered protein conformations, which likely perturb protein functioning.

Furthermore, immunohistochemistry on an 8-week old male embryo revealed a ubiquitous expression of ER-β, which is reflected in the syndromic phenotype of the first syndromic case. The exact working mechanism of these mutations is currently unknown, although we propose a mechanism via early MAPK signaling. Further experimental work is necessary to address these assumptions.

In a third study, we have shown the involvement of a single recurrent NR5A1 mutation, c.274C>T, p.(Arg92Trp), in the pathogenesis of 46,XX testicular and ovotesticular DSD. Genetic and structural data were both in favor of pathogenicity.

The results of luciferase transcriptional activation assays were not in line with our initial gain-of-function hypothesis. We proposed an alternative working mechanism via the so far under-recognized pro-ovarian functions of NR5A1. Almost simultaneously with our paper, two other groups reported on the identification of exactly the same mutation. Experimental work in these papers confirmed the pro- ovarian hypothesis.

The fourth paper of this thesis reports on the identification of FOG2 as a novel disease gene for POI. Previously this gene had been implicated in the pathogenesis of cardiac malformation syndromes in combination or not with 46,XY gonadal dysgenesis. The starting point of this study was the identification of a maternally inherited FOG2 encompassing deletion in a boy with 46,XY DSD. Interestingly, hormonal investigations revealed a lowered AMH value in his mother, suggesting a potential risk for POI development. We decided to screen FOG2 in a French-Belgian POI cohort, resulting in the identification of four additional, likely pathogenic FOG2 mutations.

Luciferase assays to assess the functional implications of these variants showed significantly reduced transcriptional activation for three out of four of these variants.

Finally, we have reviewed the hitherto known non-coding, cis-regulatory defects underlying DSD. Currently, DSD genetic research has mainly been focused on the identification of coding defects. But as for other developmental disorders, development is dependent on a correct spatiotemporal regulation of different developmental factors. With the quickly reducing costs of whole genome sequencing, we anticipate that this technique will quickly be introduced in a clinical setting, allowing the assessment of both coding and non-coding defects with only one test.

SAMENVATTING

De menselijke geslachtsontwikkeling is zonder twijfel één van de meest fascinerende processen tijdens de embryonale ontwikkeling. Gezien de complexiteit van de onderliggende moleculaire signalisatiepaden, is het niet verwonderlijk dat afwijkingen van de geslachtsontwikkeling tot de meest voorkomende aangeboren aandoeningen behoren. Ernstige vormen van DSD worden gezien bij ~1/2000-4500 geboorten, maar mildere vormen van een atypisch genitaal uiterlijk, zoals hypospadie en micropenis, komen voor bij maar liefst 1/200 geboorten. Ondanks de belangrijke vooruitgang van de identificatie van nieuwe DSD ziektegenen en signalisatiepaden, kan men vandaag slechts in ~50% van de gevallen een moleculaire diagnose verkrijgen. Dit kan echter van belang zijn voor de verfijning van een klinisch gestelde diagnose en kan leiden tot een aangepaste behandeling. Verder kan het begrijpen van de moleculaire achtergrond van deze zeldzame aandoeningen bijdragen tot het verkrijgen van meer inzicht in de pathogenese van frequentere aandoeningen zoals mannelijke infertiliteit en primaire ovariële insufficiëntie (POI).

Het voornaamste doel van dit doctoraatsproject was de ontrafeling van de moleculaire signalisatiepaden die betrokken zijn bij geslachtsontwikkeling en de identificatie van nieuwe ziektegenen voor DSD en gerelateerde aandoeningen.

Het eerste deel van deze thesis was erop gericht om de geschiktheid van een gestandaardiseerde genetische uitwerking na te gaan in een cohorte van 32 jonge kinderen met 46,XY ondervirilisatie fenotypes. Deze studie toonde aan dat array-CGH een waardevolle strategie is, met een detectie ratio van 33%, wanneer er sprake is van een syndromaal fenotype. Wanneer er enkel sprake is van ondervirilisatie bleek array-

CGH minder goed te scoren. Sanger sequenering van AR, WT1 en SRY was negatief in deze cohorte, maar er werden drie nieuwe NR5A1 mutaties gevonden. Eén van deze varianten werd eveneens teruggevonden bij een man zonder fertiliteitssproblemen, wat slechts zelden wordt gezien bij NR5A1 mutaties. Gezien de lage detectie ratio van

9 deze sequentiële aanpak en de hoge kost die ermee gepaard gaat, stelden we voorop dat een dergelijke aanpak op korte termijn zou vervangen worden door parallelle hersequenering van genpanels en door exoomsequenering voor de genetische uitwerking van DSD casussen. Onze resultaten toonden eveneens aan dat de beslissing om een uitgebreide genetische uitwerking op te starten niet enkel gebaseerd mag zijn op het fenotype. Er werden gelijkaardige detectieratio’s gezien voor de groep met ernstigere ondervirilisatie (lage EMS; <7; detectie ratio 17.6%) als voor de groep met een milder fenotype (hogere EMS; ≥ 7; detectie ratio 26.5%).

Het tweede deel van dit doctoraatsonderzoek, was gericht op de identificatie van nieuwe ziektegenen voor DSD en gerelateerde fenotypes. Ten eerste, identificeerden we ESR2 als nieuw kandidaat ziektegen voor 46,XY DSD. In totaal werden drie nieuwe, potentieel pathogene ESR2 varianten geïdentificeerd: een homozygote in- frame deletie bij een syndromale 46,XY DSD casus en twee heterozygote missense varianten bij niet verwante meisjes met geïsoleerd 46,XY DSD. In vitro functionele validatie voor deze varianten, toonde een verhoogde transcriptionele activatie aan voor de in-frame deletie, suggestief voor een gain-of-function effect. Voor de heterozygote missense varianten kon geen significant verschil waargenomen worden.

Structurele analyse van de mutante eiwitten toonde wel veranderde eiwitconformaties aan, met een waarschijnlijk effect op eiwitfunctie. Verder toonden we via immuunhistochemie op materiaal van een 8-weken oud mannelijk embryo aan dat ER-

β verspreid tot expressie komt, wat gereflecteerd wordt in het syndromaal fenotype van de syndromale DSD casus. Tot nog toe is het exacte werkingsmechanisme van deze varianten onbekend, onze hypothese is dat dit kan gebeuren via vroege MAPK signalisatie. Verder experimenteel werk zal leiden tot meer inzichten hierin.

In een derde studie toonden we aan dat een specifieke, recurrente NR5A1 mutatie, c.274C>T, p.(Arg92Trp), betrokken is bij de pathogenese van 46,XX testiculair en ovotesticulair DSD. Zowel genetische als structurele argumenten waren suggestief voor een pathogeen effect. De resultaten van luciferase transcriptionele activatie

10 assays konden onze initiële gain-of-function hypothese niet bevestigen. Een alternatieve hypothese waarbij de tot nog toe onderschatte pro-ovariële functies van NR5A1 een rol spelen, lijkt relevanter te zijn. Dit werd bevestigd door twee nagenoeg gelijktijdige publicaties over dezelfde NR5A1 mutatie.

Een vierde paper beschrijft de identificatie van FOG2 als een nieuwe ziektegen voor

POI. Voordien werd dit gen reeds betrokken bij de pathogenese van congenitale hartafwijkingen, al dan niet in combinatie met 46,XY gonadale dysgenesie. Het uitgangspunt van deze studie was de identificatie van een materneel overgeërfde,

FOG2 overspannende deletie bij een jongen met 46,XY DSD. Hormonale uitwerking bij de moeder bracht een verlaagde AMH concentratie aan het licht, suggestief voor een risico op POI. De daaropvolgende screening van een Frans-Belgische POI cohorte resulteerde in de identificatie van vier bijkomende, waarschijnlijk pathogene FOG2 varianten. Luciferase experimenten toonden een gedaalde transcriptionele activatie aan voor drie van deze varianten.

Ten slotte hebben we een overzicht gemaakt van tot nu toe gekende niet-coderende, cis-regulatorische defecten die aan de basis liggen van DSD. Momenteel is genetisch

DSD onderzoek voornamelijk toegespitst op de identificatie van coderende defecten.

Aangezien geslachtsontwikkeling sterk afhankelijk is van een correcte spatiotemporele expressie van verschillende factoren, kan verwacht worden dat een aanzienlijk aantal casussen zal kunnen verklaard worden door cis-regulatorische defecten. Door de dalende kostprijs voor genoomsequenering wordt verwacht dat deze technologie op korte termijn zal geïntroduceerd worden in een klinische setting.

Hierdoor zal men in staat zijn om met één test zowel coderende als niet-coderende veranderingen op te pikken.

CHAPTER 1: CHAPTER 1: INTRODUCTION

Introduction

1 SEX DEVELOPMENT

To understand the pathogenesis of disorders of sex development (DSD) it is essential to have a profound knowledge of the physiological processes driving human sex development, beyond doubt one of the most intriguing processes of embryonic life. In general, sex development can be divided in two important steps: sex determination and sex differentiation. Sex determination comprises the initial developmental decision of the bipotential gonad to follow either the testicular or ovarian path. This step is tightly regulated by the temporospatial expression of different genes, mainly encoding transcription factors. Subsequent sex differentiation involves the development of phenotypic sex, both the internal and external genitalia, which is mediated by hormones that are secreted by the developing gonads, the testes in particular1,2. Gonads possess two distinct functions; first of all, they produce hormones that are responsible for different effects throughout the entire body and second, they produce gametes that are indispensable for reproduction3.

1.1 The undifferentiated stage

Sex development is initiated in the fourth week of embryonic life and starts with the formation of the undifferentiated urogenital ridges, a ventromedial thickening of the coelomic epithelium of the mesodermal mesonephros3,4 (Figure 1). This urogenital ridge is the precursor of a urinary ridge, that is the anlage of the urinary system, and an adreno-genital ridge (week 5), that gives rise to the adrenal glands and the gonads4.

The proliferating cells of the gonadal ridge will penetrate the underlying mesenchyme, resulting in the formation of primitive sex cords3. These genital ridges are subsequently colonized by primordial germ cells that are derived from the ectoderm5,6. At the same time two unipotential duct systems develop from the

(para)mesonephros: the mesonephric ducts (Wolffian duct) and the paramesonephric ducts (Müllerian duct). These ducts are the precursors of the internal genitalia and both of them debouch in the cloaca, the terminal end of the hindgut, which is closed

15 by a cloacal membrane at that time. The cloaca is later divided in a recto-anal sinus and a urogenital sinus, covered by a rectal and a genital membrane respectively.

Underneath this membrane, two mesodermal swellings develop: the urethral folds and the labioscrotal swellings. These structures, together with the genital tubercle that arises as a separate swelling, will give rise to the external genitalia1,4,7.

Figure 1: Development of the undifferentiated gonads (adapted from Gilbert et al.8). (A) The gonadal ridge at 4 weeks. The coelomic epithelium starts proliferating and germ cells start migrating from the yolk sac to the gonadal ridges. (B) The gonadal ridge at 6 weeks. The coelomic epithelium keeps proliferating with the formation of primitive sex cords that invade the underlying mesenchyme.

1.2 Sex determination

Sex determination commences around week 7 and leads to the development of testes in 46,XY embryos and ovaries in 46,XX embryos.

1.2.1 Male sex determination

Testis formation can be divided in three distinct steps: (1) continuous proliferation of the pre-Sertoli cells, (2) compartmentalization into secondary sex cords that are now called testis cords and (3) elongation of these cords to form seminiferous tubules3. At the end these tubules get narrower and form the rete testis that connects the developing testes with the efferent ducts. The seminiferous tubules are separated from the surface epithelium by a layer of extracellular matrix, called the tunica albuginea, and from each other by interstitial tissue in which Leydig cells will develop4,8 (Figure 2). Leydig cell development begins around the eighth week of development and their origin remains

16 a matter of debate as it is currently unclear whether they arise from the coelomic epithelium or from specialized cells in the gonadal-mesonephric border9,10. In general, three cell types are present in the developing and adult testis: (1) germ cells, with an extra-embryonic origin, (2) somatic Sertoli cells and (3) the sex steroid producing

Leydig cells.

1.2.2 Female sex determination

Unlike during male sex determination, the primitive sex cords will regress during female development. The coelomic epithelium will allow the formation of a second generation of sex cords, the so-called cortical cords, which will organize themselves around the oocytes with the formation of the first primordial follicles (Figure 2). The cortical cells surrounding the oocytes are now called granulosa cells, while the mesenchymal cells are called theca cells. These cells form the female counterpart of the Sertoli and Leydig cells respectively3,8.

Figure 2: Sex determination leads to the formation of testes and ovaries (adapted from Gilbert el al.8)

1.3 Sex differentiation

Sex differentiation takes place during weeks 8-14 of development and involves the development of both internal and external genitalia. According to early research by

Jost et al., female development is the default pathway and can even occur independently of the existence of ovaries, while male differentiation depends on the adequate production and action of testicular hormones11.

1.3.1 Male sex differentiation

Once testes are formed, Sertoli cells will start to produce anti-Müllerian hormone

(AMH) resulting in the regression of the paramesonephric ducts in male embryos during weeks 8-9 of development. Apoptosis and epithelial-mesenchymal transformation progresses in a cranio-caudal direction12. Adequate timing of AMH expression is crucial as the Müllerian ducts become insensitive to AMH by the tenth week13. At the same time, cholesterol will be transformed to testosterone in the Leydig cells through coordinated action of different steroidogenic enzymes14. Testosterone action is important for development of the internal genitalia and results in mesonephric duct differentiation into the epididymides, vasa deferentia and the seminal vesicles15 (Figure 3). Subsequently, testosterone is converted to dihydrotestosterone (DHT), which has a higher affinity for the androgen receptor and is responsible for the virilization of the urogenital sinus and external genitalia (week

10-14)1. The bladder, the proximal part of the urethra and the prostate are urogenital sinus’ derivatives. The prostate develops from the dorsal wall of the urethra and has both a meso- and endodermal origin: the stroma and smooth muscle cells are mesodermal derivatives and mesenchymal signals induce proliferating signals on the overlying epithelium, resulting in the formation of the glandular epithelium and the prostatic ducts. Around week 9, virilization of the external genitalia starts with the lengthening of the anogenital distance, later followed by fusion of the labioscrotal folds with formation of the scrotum. Fusion of the urethral folds leads to the formation of the distal part of the urethra. The genital tubercle will form the penis4,16 (Figure 4).

Figure 3: Differentiation of the internal genitalia. On the top, the undifferentiated duct system with the Wolffian duct in dark blue and the Müllerian duct in light blue. In male embryos the developing testis will start to secrete AMH, responsible for Müllerian duct regression, and testosterone, which cause differentiation of the Wolffian duct. In absence of these hormones in female embryos, the Wolffian duct will regress, while the Müllerian duct gives rise to the fallopian tubes, the uterus and the upper part of the vagina.

1.3.2 Female sex differentiation In the female embryo, the paramesonephric ducts are not influenced by AMH and will therefore not regress. The female gonads will start to produce estrogens, which are responsible for differentiation of the paramesonephric duct to the Fallopian tubes

(oviduct), the uterus and the proximal third of the vagina (Figure 3). In the absence of testosterone, the mesonephric ducts on the other hand will regress. The urogenital sinus will differentiate into the bladder, the urethra and the distal part of the vagina, the urethral folds will form the labia minora and the labioscrotal folds will give rise to the labia majora. Finally, the genital tubercle will form the clitoris (Figure 4).

Figure 4: Differentiation of the external genitalia (adapted from www.embryology.ch). In light blue, the labioscrotal folds which differentiate to the scrotum in male embryos and to the labia majora in female embryos. In darker blue, the urethral folds which form the anlage of the penis and the labia minora in males and females respectively.

2 DISORDERS OF SEX DEVELOPMENT

According to the Jost paradigm, sex development consists of four successive steps; first of all, the karyotype is determined at fertilization, followed by the formation of undifferentiated gonads and subsequent sex determination; the last step involves differentiation of both internal and external genitalia (sex differentiation).

Disorders of sex development can occur due to errors in each of these four steps17.

The term DSD was first introduced in the 2006 consensus statement18,19. Before that time, these conditions were referred to as intersex, pseudohermaphroditism or sex reversal. These terms were perceived as stigmatizing by individuals having a DSD and did not give any specifications about the actual condition. Currently, the normative character of the word “disorder” is contested, and is therefore gradually replaced by

“differences (of sex development)”, especially in the clinical setting. The current nomenclature is based on three main classification levels20. A first division is based on the karyotype: sex chromosomal DSD, 46,XY DSD and 46,XX DSD. The sex chromosomal DSD class includes the conditions Klinefelter and Turner syndrome (KS and TS), characterized by 47,XXY and 45,X karyotypes respectively. Secondly, for the

46,XY and 46,XX DSD conditions, a distinction is made between disorders of gonadal development and disorders of steroid hormone synthesis or action, and a rest group.

Thirdly, a genetic diagnosis can be added if available20. Table 1 shows an overview of the DSD nomenclature. The different types of DSD are discussed below, a thorough description of the molecular pathways that are involved in the pathogenesis of these conditions is given in section 3 of this chapter.

Table 1: Overview of the DSD classification system, based on Hiort et al.21

Sex chromosomal DSD 46,XY DSD 46,XX DSD

Disorders of testicular Disorders of ovarian development development

47,XXY (Klinefelter Complete or partial gonadal Testicular DSD syndrome) dysgenesis (monogenic forms)

45,X (Turner syndrome) Ovotesticular DSD Ovotesticular DSD

45,X/46,XY (mixed gonadal Syndromic forms Syndromic forms dysgenesis)

46,XX/46,XY (chimerism)

Disorders of androgen Androgen excess synthesis/action

Syndromic e.g. Smith-Lemli-Opitz Congenital adrenal syndrome hyperplasia (mostly steroid 21-hydroxylase deficiency)

Congenital adrenal hyperplasia Aromatase deficiency

Early androgen biosynthesis defects

Isolated androgen biosynthesis Luteoma defects

Associated with endocrine Iatrogenic disruption

Complete and partial androgen insensitivity syndrome

Hypospadias of unknown genetic origin

Other Other

Syndromic associations of male Mayer–Rokitansky–Küster– genital development Hauser syndrome

Persistent Müllerian duct Complex syndromic

syndrome disorders

Vanishing testis syndrome

Isolated hypospadias

Congenital hypogonadotropic

hypogonadism

Cryptorchidism

2.1 Sex chromosome DSD

All conditions in this group are characterized by numerical changes in the sex chromosomes. The most prevalent of them are KS and TS, other conditions such as

45,X/46,XY and 46,XX/46,XY are less common.

Klinefelter syndrome KS is the most common form of DSD and the most common genetic cause of hypogonadism in males. Prevalence is estimated at 1 in 600 births; however it is believed that over 60% of cases remain undiagnosed due to unpronounced clinical features or low self-reporting because of the social stigma surrounding the condition22,23. KS is mostly characterized by a 47,XXY karyotype although 10-20% of patients display a higher supernumeracy of sex chromosomes and various mosaic forms have been described22,24. In general these variants with increased aneuploidy show more severe clinical features22,25. Many attempts have been undertaken to present a clear phenotype for KS. Klinefelter himself described patients as tall, broad at the hip and narrow at the shoulder, with absent or sparse body hair, gynecomastia, azoospermia, small testes, and elevated follicle stimulating hormone (FSH)26, however multiple parameters like behavioral difficulties and decreased bone mineral density were added to the list27. In general, there is a wide phenotypic variability of the condition.

Turner syndrome TS has a prevalence of 1 in 2 000 newborn girls and is the result of partial or complete

X chromosome monosomy28,29. As in KS, the phenotype associated with TS varies and in general more severe phenotypes are observed in cases with X monosomy compared to other chromosomal rearrangements30. Short stature, webbed neck, cubitus valgus, widely-spaced nipples, infertility, and streak ovaries are among the most common characteristics of TS31. Structural cardiac defects and cardiovascular abnormalities are found in 20-45% of TS cases, however studies with more advanced technologies suggest an even higher number. Besides these cardiac defects, which have an important impact on patients’ life expectancy, neurological and metabolic disturbances, hearing impairment, thyroid dysfunction, diabetes, auto-immune diseases and psychosocial difficulties are frequently observed in TS. Clinical management of TS patients should cover all these different aspects29,31.

Mosaicism and chimerism Less frequent types of sex chromosomal DSD occur when the karyotype is 45,X/46,XY

(mosaicism) or 46,XX/46,XY (chimerism), or variants of these. Both mosaicism and chimerism refer to the presence of two genetically distinct cell lineages in the same individual. However, chimerism is defined by the fusion of two different zygotes to a single embryo while mosaicism is the result of a mitotic error in one zygote32.

45,X/46,XY mosaicism has a prevalence of 1.5-1.7 in 10 000 and is characterized by a broad spectrum of phenotypes ranging from phenotypic women with or without

Turner-like features to men with variable virilization of external genitalia. Besides concerns about gender assignment, genital surgery and gonadal tumor risk, medical care should concentrate on problems related to Turner syndrome, such as short stature and cardiac malformation, due to the presence of a 45,X cell population33,34. The phenotype in 46,XX/46,XY chimerism varies from phenotypically normal males and females to cases with ambiguous genitalia32.

2.2 46,XY Disorders of sex development

2.2.1 Disorder of testicular development

Complete and partial 46,XY gonadal dysgenesis Complete and partial 46,XY gonadal dysgenesis (GD) are the consequence of abnormal gonadal determination in the presence of an XY karyotype. Complete GD has a prevalence of 1 in 20 000 births and was first described by Swyer et al.35,36, partial GD occurs more frequently. Complete GD is characterized by typical female internal and external genitalia, absence of secondary sex characteristics and bilateral non-functional gonads that may appear as streaks or have areas of undifferentiated gonadal tissue37.

Most patients present at adolescence after referral for absent pubertal development, but the diagnosis is increasingly made prenatally with the finding of a prenatal XY karyotype in discordance with the external phenotype. Girls with complete gonadal dysgenesis have a uterus due to absence of intrauterine AMH production. The partial form is characterized at the gonadal level by incomplete testicular differentiation in one or both gonads, most often in combination with areas of undifferentiated gonadal or streak tissue. The dysgenetic testes are somewhat hormonally active, resulting in ambiguous external genitalia with or without Müllerian structures36,38,39. Hormonal investigations in patients with complete and partial GD reveal elevated FSH and luteinizing hormone (LH) levels, while sex steroids, AMH and inhibin-B concentrations are below the male reference range36,38,39. While interpreting these hormonal tests, age-specific reference ranges must be taken into account.

The first gene that was associated with complete GD is Sex-determining region of Y

(SRY) and mutations in this gene account for 10-20% of cases36,40. Several other disease genes for complete and partial GD have been described: ARX, ATRX, CBX2, DHH,

DMRT1, GATA4, MAMLD1, MAP3K1, NR0B1, NR5A1, SOX9, WNT4, WT1 and

WWOX. Some of these genes are associated with syndromic phenotypes, the most important ones are described below.

SRY-box 9 (SOX9) mutations and deletions are associated with campomelic dysplasia, characterized by a skeletal malformation syndrome and 46,XY DSD, the latter in 75% of male cases41–43. Decreased activity of Nuclear receptor subfamily 5 group A member

1 (NR5A1; steroidogenic factor-1, SF-1) can cause a plethora of different phenotypes ranging from undervirilization to complete 46,XY GD with primary adrenal insufficiency44,45. In addition it has been shown that variants in NR5A1 can cause primary ovarian insufficiency (POI) in females and spermatogenic failure in males44.

Wilms Tumor 1 (WT1) mutations can underlie three different syndromes depending on the type and location of the mutation: WAGR syndrome, Denys-Drash and Frasier syndrome. WAGR syndrome can be caused by WT1 encompassing deletions and is characterized by Wilms tumor, aniridia, genitourinary abnormalities and intellectual disability. Denys-Drash syndrome is characterized by 46,XY complete or partial GD, early onset renal failure and the development of a Wilms tumor during the first decade of life. This syndrome is caused by heterozygous missense mutations in the zinc fingers of WT146,47. Frasier syndrome is characterized by partial or complete 46,XY GD, proteinuria and end-stage renal failure in the second decade and a particularly high risk of gonadoblastoma development. This syndrome is caused by specific heterozygous splice site mutations in exon 9, leading to decreased amounts of the

KTS+ transcript and a perturbed KTS-/KTS+ transcript balance48,49.

46,XY ovotesticular DSD Ovotesticular (OT) DSD is a rare condition that is characterized by the coexistence of both testicular and ovarian tissue in one individual. Testicular and ovarian tissue can be found each on one side or in the same gonad. OT DSD is mostly found in patients with 46,XX DSD (60%), the remainder have sex chromosomal DSD (more specifically

46,XX/46,XY DSD. Although a small number of cases with OT DSD and a 46,XY karyotype have been reported, it is currently unclear if this entity exists, given the possible confusion between ovarian tissue and undifferentiated gonadal tissue (UGT), leading to misdiagnosis in a number of cases37,50.

2.2.2 Disorders of androgen synthesis Disorders of androgen synthesis can occur in every step of the testosterone biosynthesis process starting from the transport of cholesterol to the mitochondria in order to start steroidogenesis to the conversion of testosterone to dihydrotestosterone.

Impaired Leydig cell differentiation: Leydig cell hypoplasia

Most cases of Leydig cell hypoplasia (LCH) present as phenotypical females, who have absence of secondary sexual characteristics and undescended, slightly hypotrophic testes. Partial LCH has a less strictly defined phenotype, with micropenis and hypospadias being the most common features51. These patients show no or a subnormal response to prepubertal human chorionic gonadotropin (HCG) stimulation and increased LH levels in puberty52.

LCH is a genetically heterogeneous condition with an autosomal recessive inheritance mode53. Several inactivating mutations in the Luteinizing Hormone/

Choriogonadotropin Receptor (LHGCR) gene, encoding the LH receptor, have been identified in patients with complete or partial LCH51. Recently a new disease mechanism was added to the LCH pathogenesis. Vezzoli et al. described a patient with partial LCH and a compound heterozygous LHGCR mutations. The first variant was a known missense variation, but the second variant was novel and affected the putative signaling peptide of the receptor. Functional and structural validation shows that this signaling peptide is indispensable for correct receptor biosynthesis54. So far, no other genes have been linked to human LCH.

Testosterone biosynthesis defects

DSDs can also be the result of enzymatic defects in each of the steps necessary to convert cholesterol to functional steroid hormones. An overview of the steroid biosynthesis pathway is provided in Figure 5. All of these conditions are characterized by autosomal recessive inheritance51. We will make a distinction between (1) cholesterol synthesis defects, (2) defects affecting adrenal and testicular

27 steroidogenesis and leading to congenital adrenal hyperplasia (CAH), and (3) defects in testicular steroidogenesis.

- DSD associated to cholesterol synthesis defects: Smith-Lemli-Opitz syndrome Cholesterol is the precursor of all steroid hormones, and cholesterol deficiency or decrease of cholesterol precursors are involved in different conditions, among them

46,XY DSD. The final step in cholesterol biosynthesis is the reduction of 7- dehydrocholesterol (7-DHC) by 7-DHC reductase, encoded by the DHCR7 gene.

Defects of this gene cause Smith-Lemli-Opitz syndrome (SLOS) 55. SLOS was first described in three male patients with small stature, failure to thrive, severe feeding disorder, hypospadias, a characteristic facial appearance and microcephaly56.

However, examination of additional cases showed a wider phenotypic variability55.

The condition can be diagnosed on biochemical grounds, based on elevated levels of

7-DHC and reduced cholesterol levels57.

In 1998, three groups independently reported the first DHCR7 mutations58–61. Today, over 150 different mutations have been identified55. More recently, a number of exon deletions were identified as the causal defect in SLOS cases62,63. Although several attempts have been undertaken to find genotype-phenotype correlations, none of them are reliable enough to allow individual counseling and it is highly likely that the SLOS phenotype is influenced by modifier genes55.

- Congenital adrenal hyperplasia Defects at the top of the steroid hormone biosynthesis cascade not only affect sex hormone production, but also lead to impaired gluco- and/or mineralocorticoids. CAH can be caused by mutations in four different genes: Cytochrome P450 Family 11

Subfamily A Member 1 (CYP11A), Steroidogenic Acute Regulatory Protein (StAR),

Hydroxysteroid 3-β Dehydrogenase 2 (HSD3B2) and Cytochrome P450 Family 17 Subfamily

A Member 1 (CYP17A1). These different enzymatic defects are reflected in slightly

Figure 5: Overview of the steroid hormone biosynthesis pathways (based on Mendonca et al. 201051) 29

Table 2: Overview of the different types of congenital adrenal hyperplasia, based on Mendonca et al. and references therein51.

StAR CYP11A CYP17A1 HSD3B2

Enzyme Cholesterol Cholesterol 17-OH pregnenolone DHEA function transport within the mitochondria

Pregnenolone DHEA Androstenedione

Phenotype - Phenotypic - Similar to StAR - Female-like, - Ambiguous females deficiency slightly virilized external genetalia external genitalia - External genitalia - No adrenal - Micropenis slightly virilized hyperplasia - Hypoplastic internal genitalia - Hypospadias - Cryptorchidism - Gynaecomastia - Bifid scrotum - Underdeveloped internal genitalia - Cryptorchidism - Gynaecomastia

- Enlarged adrenal - High blood - Possible salt-loss cortex, engorged pressure with cholesterol(esters) - Hypokalaemia

- Salt-wasting

- Hyponatraemia

- Hyperkalaemia

- Hypovolaemia

- Acidosis

- Death in infancy

Biochemistry - High ACTH and - High - High (17-OH) renin corticosterone, pregnenolone, deoxycorticosterone DHEA(S), - Low/undetectable and progesterone gluco- and - High ratio mineralocorticoids - Low aldosterone, 17OHPreg/17OHP 17-OH progesterone, - Low androgens cortisol, androgens - Low cortisol and estrogens - Low androgens

Therapy - Gluco-and - Gluco-and - Glucocorticoids - Glucocorticoids mineralocorticoids mineralocorticoids (for hypertension) - Androgens - Androgens/ - Androgens - Mineralocorticoid Estrogens in case of salt-losing

DHEA(S)= dehydroandrostenedione (-sulphate), ACTH= adrenocorticotropic hormone, LH= luteinizing hormone, FSH= follicle stimulating hormone

different phenotypes and distinctive biochemical patterns. Treatment strategies include gluco- and/or mineralocorticoid supplementation and androgen replacement therapy in affected males. An overview of the phenotypical and biochemical findings in these patients is shown in Table 251.

- Testicular steroidogenesis defects DSDs can also be caused by defects in enzymes that are further down in the steroidogenesis pathway, these conditions are not associated with adrenal insufficiency. Causal genes here are Hydroxysteroid 17-β Dehydrogenase 3 (HSD17B3) and Steroid 5 α-Reductase 2 (SRD5A2)51. The HSD17B3 gene encodes the enzyme that is necessary for the conversion of androstenedione to testosterone. Its deficiency is characterized by abnormal androstenedione/testosterone ratios, both in basal conditions as after HCG stimulation. Phenotypically, patients with HSD17B3 variants present with female-like or ambiguous genitalia and abdominal or inguinal testes.

HSD17B3 deficiency is the most common disorder of androgen biosynthesis. SRD5A2 encodes the 5α-reductase enzyme that is responsible for the metabolization of testosterone to its more potent derivative DHT. Patients with SRD5A2 variants typically present with ambiguous genitalia, micropenis and prostate hypoplasia.

Internal genitalia and testes develop normally. Although spermatogenesis is reduced, in part due to the cryptorchic position of the testis, several cases have offspring, mainly

31 through assisted reproductive techniques. Biochemical diagnosis relies on elevated testosterone/DHT ratios51.

2.2.3 Disorders of androgen action Male gonadal differentiation is not solely depending on correct testosterone and DHT biosynthesis, it is also crucial to have functioning receptors to bind these androgens and thereby facilitate their effects on the developing genitalia. Inactivating mutations in the androgen receptor (AR) gene are known to cause various forms of androgen insensitivity syndromes (AIS)64,65. Very recently, a new mechanism was added to AIS pathogenesis; mutations in the AR 5’ untranslated region resulting in the formation of an upstream open reading frame. These mutations lead to a decreased amount of AR protein without decreasing messenger RNA (mRNA) levels66. Interestingly, for a subset of AIS cases, ~15% of complete AIS (CAIS) cases and ~65% of partial AIS (PAIS) cases, no inactivating AR mutations can be found. Hornig et al. have set-up a study in which the AR transcription activity of this group was tested in genital skin fibroblasts.

This test could distinguish normal male controls from genetically confirmed AIS cases and was also positive for a subset of AR-mutation negative cases that were suggestive for AIS. Further investigation of these cases might result in the discovery of novel AIS genes67. The prevalence of AIS is estimated between 1 in 20 000 and 1 in 99 000 births68.

CAIS is characterized by complete female external genitalia with absence of Müllerian structures. PAIS has a broad phenotypic spectrum but in most cases a short penis and hypospadias are observed. Puberty in AIS patients is characterized by spontaneous breast development often in combination with reduced sexual hair growth69. Lab results for these patients show normal levels of all testosterone precursor molecules;

AMH and inhibin-B are also normal (hence the absence of paramesonephric structures). Gonadotropins and testosterone are elevated, since feedback mechanisms are perturbed69.

2.2.4 Other 46,XY DSD conditions

The rest group contains three important entities, being testicular regression syndrome

(TRS), isolated hypospadias and cryptorchidism. They are included in the DSD classification because they represent other forms of divergent sex development.

However, their etiology is more diverse than the “classic” forms of DSD.

TRS or vanished testes is a condition in which testicular differentiation is initially normal but is followed by degeneration during fetal or early postnatal life70. TRS is estimated to be present in 5% of all cryptorchidism cases71. TRS has been associated with microdeletions of the Y chromosome and has been seen more frequently in boys with persistent Müllerian duct syndrome72. One case has been associated with an

NR5A1 mutation but this observation has never been reported afterwards73. However, it should be recognized that far from all cases of TRS are caused by genetic defects or an underlying endocrinopathy. Other, more common causes of TRS include vascular accidents and prenatal torsion events72.

The same is true for isolated hypospadias and cryptorchidism. While some cases related to monogenic defects in DSD associated genes have been reported, multifactorial genetic changes are more commonly observed74,75. A general observation is that severe hypospadias cases are more likely to be associated with a monogenic defect compared to distal hypospadias74. Furthermore, different associations have been described with environmental factors like exogenous exposure to estrogens or progestins, assisted reproductive technologies, maternal smoking and drug use and exposure to endocrine disrupters, heavy metals or phthalates. Although, for some of these factors the association is inconsistent74,75.

2.3 46,XX Disorders of sex development

2.3.1 Disorders of ovarian development

46,XX ovarian dysgenesis The most common forms of ovarian dysgenesis are TS and other conditions with X- chromosome abnormalities (described above). However, rare cases of ovarian dysgenesis in patients with a 46,XX karyotype have been described. Affected girls present with typical female genitalia, but do not experience puberty. Müllerian structures have fully developed and streak gonads are present; biochemical analyses reveal hypergonadotrophic hypogonadism. Ovarian dysgenesis is a genetically heterogeneous condition with, in most cases, autosomal recessive inheritance; however the molecular cause remains enigmatic in the majority of cases76. Several syndromes like Perrault syndrome and blepharophimosis – ptosis – epicanthus inversus syndrome (BPES), which are respectively attributed to autosomal recessive mutations in Hydroxysteroid 17-β Dehydrogenase 4 (HSD17B4) and autosomal dominant mutations in Forkhead box L2 (FOXL2) respectively, are also characterized by gonadal dysgenesis77–79.

46,XX testicular and ovotesticular DSD 46, XX testicular DSD has a prevalence of 1 in 20 000 births and the majority of patients are infertile males with typical male genitalia (85%); 15% of patients have a variable degree of genital ambiguity, sometimes with gynaecomastia. In men who have typical male genitalia, the diagnosis is usually made due to slow progression of puberty or work-up for infertility. Affected men have hypergonadotropic hypogonadism76.

Gonadal biopsies reveal testicular tissue with small Sertoli-cell only (depending on the timing), SOX9 positive seminiferous tubules, even in SRY-negative cases37,80,81. In general 46,XX testicular DSD can be caused by activating mutations in pro-testis genes and inactivation of ovary-specific genes82. In 90% of men who have a typical male phenotype, an SRY translocation can be identified, however this finding is far less frequent in patients with an ambiguous phenotype. Other causes of 46,XX

34 testicular DSD are SOX9 duplications, genomic rearrangements of the SOX3 locus resulting in ectopic expression, copy number variants (CNVs) of SOX10 and inactivating R-spondin 1 (RSPO1) mutations43,83–86. Management of these affected men involves life-long follow-up and androgen replacement therapy in the presence of low testosterone values. In case of SRY translocation, the presence of the gonadoblastoma region of Y (GBY) region (including testis-specific protein, Y-encoded, TSPY) should be investigated, as it is believed to increase the risk of germ cell tumor development87–89.

46,XX OT DSD also has a prevalence of 1 in 20 000 births and accounts for approximately 3-10% of all DSD cases76. In these individuals the ovotestis do not have germ cells in the testicular part, however follicles are present in the ovarian regions of the gonad and they can undergo ovulatory changes during puberty90. This condition generally results in a varying degree of external genital ambiguity and the presence of both male and female internal genitalia. Diagnosis should be confirmed histologically.

Hormonal profiles may show normal sex steroid production and age-appropriate gonadotropins. Given that both testicular and ovarian parts will start steroidogenesis during puberty, both male and female secondary sex characteristics will emerge76,90.

Ten percent of 46,XX OT DSD cases show a translocation of SRY and RSPO1 mutations have been described in a number of cases with 46,XX testicular DSD, palmoplantar hyperkeratosis and a predisposition for squamous cell carcinoma76,85,86.

Management of these patients includes specialized psychological care and follow-up to assess gender identity before the start of steroid hormone production at puberty.

Once a clear gender identity has developed conservative surgery can be attempted with preservation of the gonadal tissue in agreement with gender identity. As for other types of DSD, it is advisable to delay surgery until the patient him- or herself can participate in this decision. As follicles are present in the ovarian parts of ovotestis and normal menstrual cycles have been observed, there is a chance of fertility. However, as most of these patients develop amenorrhea, they should be counseled for possible infertility and eventually oocyte cryopreservation76,90.

2.3.2 Androgen excess

Apart from defective ovarian development, 46,XX DSD can also be related to defects in steroid hormone production. Female fetuses that are exposed to excess androgen levels will have virilized external genitalia. The origin of these androgens can be fetal, feto-placental or maternal82.

CAH is the most prevalent cause of 46,XX DSD and is actually a disorder of impaired cortisol synthesis (see above). Decreased cortisol levels result in insufficient feedback inhibition and increased secretion of adrenocorticotrope hormone inducing adrenal hyperplasia and increased androgen production. The clinical presentation of CAH is the result of (1) the cortisol deficiency, (2) the accumulation of cortisol precursors and their biological activity and (3) the diversion of these precursors to steroid hormones that are normally present in lower amounts91. CAH is associated with defects in different steroid producing enzymes: CYP21A2, CYP17A1, CYP11B1,

HSD3B2 and POR. The different types of CAH were reviewed in detail by Auchus and Chang91.

Another extremely rare cause of 46,XX DSD due to androgen excess is placental aromatase deficiency, which has to date been described in fewer than 40 cases. During pregnancy CYP19A1, the gene encoding aromatase, is expressed in the placenta.

Aromatase has a critical function in protecting the developing (female) fetus against high concentrations of androgens that are produced by the fetal adrenals and maternal androgens. Aromatase deficiency can be expected in case of maternal virilization during pregnancy in women with low estriol levels. Affected females are born with ambiguous genitalia similar to other forms of 46,XX DSD due to androgen excess91.

In rare cases, fetal virilization is related to a maternal source of androgens. Chronic conditions with androgen excess like polycystic ovaries or maternal CAH, generally do not lead to fetal virilization, underlining the capability of aromatase to protect the fetus from these androgens. However, several other sources that may lead to

36 virilization have been described. Several types of both malignant (granulosa cell tumors,…) and benign tumors (adrenal adenoma, luteoma,…) produce high levels of androgens. Furthermore androgen excess can be related to certain medication like progestins and potassium-sparing diuretics91.

2.4 Disorders of ovarian maintenance: primary ovarian insufficiency POI, previously termed premature ovarian failure is defined by a triad of symptoms before the age of 40: amenorrhea for at least 4-6 months, sex hormone deficiency and elevated levels of FSH on at least two measurements and with an interval of at least one month apart92,93. The prevalence of POI is estimated to be one in 10 000 women by age 20 and reaches approximately one in 100 women by age 4094. POI can be classified based on the pathophysiological mechanism eventually leading to the disorder. The maximum number of follicles a girl will ever have is obtained at 20 weeks of gestation

(6-7 million), followed by physiological follicle atresia reducing their number to approximately 700 000 at birth and 400 000 follicles at the start of puberty. Early menopause can therefore be related to ovarian follicle depletion and it can occur in women who started life with a decreased number of follicles or in women who experience accelerated follicle atresia postnatally. Second, early menopause can be related to ovarian follicle dysfunction, where a correct number of follicles is reached but where they cannot react appropriately on gonadotropic stimuli92,93. The causes of

POI are very diverse ranging from iatrogenic (surgery, chemo- and radiotherapy) to autoimmune disorders or genetic causes. Some of these genetic causes, like NR5A1 mutations, have been identified by studying DSD families. Genetic causes of POI have been recently reviewed by Cordts et al. and Tucker et al.95,96. Taken together, the currently known POI genes can be divided in ten categories: (1) genes involved in gonadogenesis, (2) genes required for cell division/cohesion, (3) homologous recombination genes, (4) genes involved in follicular development, maturation and activation, (5) genes functioning in DNA damage repair processes, (6) genes related to cell death and autophagy, (7) protein translation associated genes, (8) genes

37 supporting hormone production and functioning, (9) genes providing metabolic support and (10) genes to protect against auto-immunity. An overview is given in

Figure 6 and a selection of them is discussed below.

A first group of genetic defects associated with POI is formed by the different numerical X chromosome aberrations ranging from partial and complete X deletions to trisomy X. The phenotypical characteristics of TS (45,X) have been described above

(section 2.1). POI in Turner women is related to accelerated follicle atresia. In general, oocyte depletion occurs already in the first decade of life, allowing only around 10% of TS women to reach menarche. Girls who have a 45,X/46,XX karyotype have a higher chance of menstruating for at least a few years92,95. Several hypotheses concerning the underlying mechanism for ovarian dysgenesis in TS women have been proposed; incorrect pairing of the X chromosome during meiosis could form an obstacle at a meiotic checkpoint leading to apoptosis or alternatively, haploinsufficiency of some X chromosome genes directly causes oocyte depletion92. Bone morphogenic protein 15

(BMP15) is such a candidate gene as heterozygous point mutations have been described in women with POI97,98.

A second X-chromosome-related POI cause are premutations of the Fragile X mental retardation 1 (FMR1) gene. FMR1 is characterized by an expandable CGG repeat in its 5’UTR. Depending on the number of repeats, three different types of alleles can occur: normal and intermediate alleles (from 6 to 54 CGG repeats), premutated alleles

(from 55 to 200 CGG repeats) and full mutations (>200 CGG repeats)99. Full mutations lead to fragile X syndrome, an X-linked recessive disorder causing intellectual disability mainly in boys99. Premutations result in POI in approximately 20% of female carriers100 and it is estimated that they account for 13% of familial POI cases and 1-7.5% of sporadic POI cases101. The exact pathogenic mechanism is not yet elucidated99. Allen et al. proposed that premutated mRNA has a toxic effect during reproductive life resulting in increased follicular atresia. This toxic effect could be attributed to

Figure 6: Overview of genes involved in the pathogenesis of POI (Adapted from Tucker et al. 201696).

‘kidnapping’ of mRNA-binding proteins by the mutated FMR1-mRNA, causing a depletion of these proteins for other cellular processes99.

Several genes involved in gonadogenesis are also involved in POI, for instance the

FOXL2 and NR5A1 genes. FOXL2 is one of the early female defining genes and is not only known to have a function in ovarian development but also in ovarian maintenance102. The gene is expressed in developing eyelids and ovaries and autosomal dominant mutations are associated with the BPES syndrome. This condition is characterized by complex eyelid malformations associated (type I) or not with POI (type II)103. Several attempts have been undertaken to identify FOXL2 mutations in patients with isolated, idiopathic POI; however identification rates are low104–106. The mechanism leading to POI can differ in women with FOXL2 mutations;

39 some of them show a block in follicle maturation, while others show an altered ratio of primordial to primary follicles107. NR5A1 is a second gonadogenesis-related gene involved in the pathogenesis of POI. First associated with 46,XY DSD (and adrenal insufficiency)108, NR5A1 inactivating mutations were later also found in women with

POI109. NR5A1 disruption can disturb ovarian functioning on several levels including reduced germ cell number, impaired stromal integrity, abnormal folliculogenesis and defective steroidogenesis44.

3 GENETIC PATHWAYS INVOLVED IN SEX DEVELOPMENT AND DSD

The following section will summarize some basic genetic concepts, necessary for understanding the genetic signaling pathways underlying DSD pathogenesis, followed by an overview of the different genes and pathways that are involved in gonadal development.

3.1 Basic genetics

3.1.1 The central dogma

(Human) genetics is a rather young research field, as it was only in 1953 that the chemical structure of our genetic material, deoxyribonucleic acid (DNA), was elucidated by James Watson and Francis Crick110. The human genome is build up out of 3 billion repeats of only four building stones or nucleotides: dATP, dGTP, dCTP and dTTP (desoxy-(A,G,C or T)-triphosphate with A = adenine, G= guanine, C= cytosine and T= thymine). These nucleotides are paired up according to the Chargaff’s rule; each basepair contains a purine (A and G) and a pyrimidine (C and T)111. In 2003, the exact sequence of all the nucleotides of our genome was determined by the Human

Genome Project and it was estimated that our genes, functional DNA units that encode proteins, account for only 2% of our DNA112,113. The remaining portion of our DNA was refered to as “junk” DNA, as it did not have an apparent function at that time. To exert its function a gene first has to be transcribed to so-called messenger ribonucleic acid

(mRNA) and translated to the corresponding protein, these processes form the central dogma of molecular biology114.

However, this central dogma cannot explain the complex diversity of cell types (and functions) that are present in our body. If all our cells contain exactly the same genetic material, then how is it possible that a testicular Leydig cell is so different from a vascular endothelial cell. The answer to that question lays within the remaining 98% of non-coding DNA. In 2003, the Encyclopedia Of DNA Elements (ENCODE) project was initiated, aiming to catalogue all functionally important regulatory sequences115.

In general, the expression level of a gene is determined by (1) interaction of the transcription complex with the gene’s minimal promoter, (2) interactions with cis- regulatory regions (enhancers, repressors, insulators) and (3) the surrounding chromatin structure116.

3.1.2 Genetic defects: coding vs. non-coding

Changes in our genetic code can result in altered protein functioning of expression patterns and can therefore cause diseases. Monogenic or Mendelian disorders are caused by a molecular defect in only a single gene and their inheritance follows clearly distinguished patterns (autosomal dominant/recessive, X-linked dominant/recessive and Y-linked). However situtions like de novo mutations, incomplete penetrance and genetic heterogeneity greatly complicate the genetic diagnostic process114.

Coding defects Initially genetic research only focused on the effects of variation in the coding regions of our genome. First of all, a distinction is made between sequence variations, in which only one or a couple of nucleotides have been changed; and structural variations, like copy number variations, in which a larger portion of our genome is disturbed. Each of these classes contains several types of variation and they are classified based on the change on DNA level and their consequence on the RNA or protein level. An overview of the most important types of coding variation (those relevant for this thesis) is given in Figure 7.

Needless to say, that not every genetic variation is pathogenic, as these small changes mostly account for interpersonal phenotypic variations. Interpretation of the identified variants remains challenging despite the publication of classification guidelines by the

American College of Medical Genetics117. Classification criteria include segregation analysis, absence in the general population, the nature of the variant and functional evidence. Based on these criteria a variant can be classified as not disease causing, likely not disease causing, possibly disease causing, likely disease causing, disease causing or as a variant of unknown significance117.

Figure 7: Coding genetic variations. Left: different types of sequence variations and their effect on the encoded protein. Missense substitution lead to the alteration of one amino acid, a nonsense substitution changes a regular codon to a stopcodon, leading to premature ending of the protein and in most cases degradation. In-frame deletion/duplications (multiple of three nucleotides) result in the deletion/insertion of one or several amino acids, without an effect on the reading frame. Frameshif deletions/duplications on the other hand to hamper the reading frame and result in a completely different amino acid sequence and an altered stop codon Right: different types of structural variation. Structural variations involve bigger portions of the genome. The most important ones are duplications, in which additional copies of a genomic region are present; and deletions, in which a genomic region is missing. Inversions of a genomic region can also perturb the correct genomic sequence.

Furthermore, mutations can also be classified based on their functional effects, namely loss-of-function and gain-of-function variation. Loss-of-function mutations result in a diminished protein function and are caused by hypomorphic alleles or null alleles

(in this case all function is lost). Loss-of-function mutations are usualy associated with autosomal recessive disorders as for most genes 50% of the normal gene product is sufficient for normal functioning114. A DSD-related example of loss-of-function variation can be found in the steroid hormone biosynthesis disorders. Both alleles of an enzyme need to be abrogated to have a phenotypic consequence. Gain-of-function mutations result in either increased function of a protein (hypermorphic allele) or in a

43 newly acquired function (neomorphic allele). This type of mutations is generally associated with an autosomal dominant inheritance pattern114.

Non-coding defects Besides an intact coding sequence, appropriate gene expression, and subsequent protein expression, requires correct interactions with the surrounding, non-coding regions as well. These regulatory regions enable expression of the right gene, at the right time. Over the years more and more mechanisms have been elucidated by which these regulatory regions exert their functions and how they can cause phenotypes when perturbed. These mechanisms include disruption of cis-regulatory elements, the interplay with non-coding RNAs and epigenetic modifications. Cis-regulatory DNA elements are the non-coding DNA fragments, located in the proximity of coding genes, that regulate the expression of these neighbouring genes118–123. Most of the regulatory defects in DSD pathogenesis involve defects of these cis-regulatory elements resulting in aberrant gene expression pattern. Specific examples ars reviewed in paper 5 of this thesis and include regulatory structural variation, deep intronic variation resulting in alternative splicing and the introduction of upstream open reading frames124. The importance of non-coding RNAs, genes that are transcribed to RNA but not translated to proteins, and epigenomic regulation in gonadal development and function has recently been reviewed by Wilhelm and Bernard125.

3.2 The undifferentiated stage

Sex development starts with the correct formation of the bipotential gonadal primordium from the intermediate mesoderm. Studies in mice have shown that many genes are implicated in this process; however human mutations leading to gonadal agenesis are very rare. Two groups of genes have been related to these earliest stages of gonadal development: homeobox genes (Lim homeobox 9, Lhx9; Empty spiracles 2,

Emx2; Chromobox 2, Cbx2/M33) and genes encoding transcription factors (Wt1, Sf-1).

Maintenance of the urogenital ridge is achieved by promoting cell proliferation and

44 preventing apoptosis1,2,126. The known and presumed interactions between these early genes are shown in Figure 8.

Figure 8: Overview of the early regulators of sex development. These genes are implicated in the development of the bipotential urogenital ridge. The Wt1-KTS isoform is known to activate Sf-1 expression together with Lhx9. Both Wt1 and Sf-1 have an effect on Dax1 expression. It is currently unclear where M33 and Emx2 fit in this model and what their interaction partners are (based on Wilhelm et al. 2007126).

Lhx9 belongs to the Lim homeobox gene family and is required for proliferation of the genital ridge cells. Lhx9-/- mice are all phenotypically female and are characterized by absence of gonads in both sexes127. These mice also show reduced levels of Sf-1 expression, implicating that Lhx9 is an upstream regulator of this gene. Wilhelm et al. showed that Lhx9 can bind directly on the Sf-1 promoter, both in vitro and in vivo, and that this has an additive effect on Wt1 induced Sf-1 expression128. Emx2 is the mouse orthologue of the Drosophila empty spiracles gene and is expressed in the epithelial parts of the developing urogenital system. Emx2 deficient mice show a complete lack of gonads and genital tract. Up to now, nothing is known about its regulation and interactions with target genes129. In humans rare EMX2 mutations have been associated with schizencephaly and somatic mutations have been found in endometrial cancer, no gonadal defects have been described130. Cbx2 is the mouse orthologue of the

Drosophila gene Polycomb. Deficient male mice show male-to-female sex reversal and ovaries are absent or hypoplastic in female mice. The gene has been implicated in 45 adrenal Sf-1 regulation, but it is unclear whether it has a similar function in the developing gonads131,132. So far, one case with a CBX2 mutation and gonadal dysgenesis has been described. Initially, this girl was reported to have normal ovaries, but reassessment of the gonadal tissue revealed streak gonads, in accordance with the presence of hypergonadotropic hypogonadism observed in this girl (unpublished observations). 133.

Transcription factors Wt1 and Sf-1 have multiple roles during sex development, starting in the earliest stages. Wt1 is expressed in the undifferentiated gonad and acts as a repressor of apoptosis. Wt1-/- mice are characterized by gonadal and adrenal apoptosis, which can be linked to two different pathways134,135. First of all Wt1 is coexpressed with Sf-1 in the gonadal primordium and it was shown that Sf-1 is a direct target gene of Wt1; therefore gonadal apoptosis might be related to the absence of Sf-

1 signaling in Wt1 deficient mice128. A second pathway is Sf-1 independent and acts through Bcl2 signaling, an anti-apoptotic gene and known Wt1 target gene136. Sf-1 deficiency itself also leads to abnormal gonadal development; both female and male mice fail to develop gonads beyond the early differentiated stage. All Sf-1 deficient mice show a female phenotype without adrenals and gonads137,138.

3.3 Male sex developmental pathways

3.3.1 SRY: the master switch and how to turn it on Up to this point of development there is no difference between the expression patterns of the gonadal primordium in 46,XX or 46,XY embryos, nor is there a phenotypical difference. A first molecular event in the testis determination pathway is the upregulation of SRY expression. The single-exon SRY gene encodes a transcription factor with a highly conserved DNA-binding domain, the so-called high mobility group (HMG) box that binds and bends DNA139. In mice, Sry is expressed from 10.5 days post coitum (dpc) until 12.5 dpc starting from the center of the genital ridges and proceeding to the poles; in humans SRY expression is initiated at embryonic day 41 and is maintained at lower levels beyond embryogenesis140–142. Despite intensive

46 research, little is known about SRY’s function and target genes. Some reports suggest a role for SRY in transcriptional activation, others in transcriptional repression143,144. An activating role is supported by transgenic mice in which the trans-activating domain is depleted; these mice fail to undergo normal male development143.

Human SRY mutations

Since the identification of SRY, numerous mutations and structural variants have been described in a plethora of phenotypes145–149. Altogether, translocations of the SRY gene to the X chromosome or the autosomes can explain roughly 90% of 46,XX testicular

DSD and inactivating point mutations, most of which are located in the conserved

HMG-box, account for approximately 15% of 46,XY GD cases2. Besides these coding variants, several SRY regulatory defects have been described150–152, emphasizing the necessity of its spatiotemporal regulation for normal gonadal development. This fits the current belief that SRY expression should reach a certain threshold in a specific time window to direct the precursor supporting cells towards a Sertoli cell fate.

Additional proof for this hypothesis is given by the B6-Ypos mouse, that possesses a

‘weak’ Sry allele, leading to delayed Sry expression and subsequent development of ovotestes. Similar mouse experiments also confirmed the existence of a threshold of

Sry expression levels for the induction of testis development153,154.

The path towards SRY

Despite its functional importance, the pathways regulating SRY expression are far from elucidated. Early insights were gathered via knock-out mouse models and validated by the identification of human mutations in 46,XY GD phenotypes. The most important factors and their working mechanisms are described and depicted below

(Figure 9).

Figure 9: Overview of the most important regulators of SRY expression. (A) SRY expression is regulated by WT1, more specifically the +KTS isoform, the GATA4/FOG2 transcription complex and the MAPK pathway. (B) The WT1+KTS isoform is involved in SRY activation by binding and stabilizing the SRY mRNA155,156. (C) GATA4 together with its co-factor FOG2 upregulates SRY expression. Besides this role in testicular determination, the complex is also believed to be involved in the upregulation of AMH later on. (Based on Zaytouni et al, 2011157.(D) Overview of the mechanism of MAP3K1 action during normal male development. MAP3K1 gain-of-function mutations will lead to increased binding of RHOA and Rock, two female specific factors that can inhibit SOX9 and to increased phosphorylation of p38 and ERK1/2. Hyperphosphorylation will stabilize β-catenin and initiate female developmental pathways. Furthermore these gain-of-funtion mutations will sequestrate AXIN and MAP3K4, thereby abolishing SRY activation. (Based on Ostrer, 2014158).

The previously mentioned WT1 gene is not only important for development of the undifferentiated gonad (specifically the –KTS isoform); a second isoform WT1+KTS is involved in SRY upregulation. Both isoforms are the result of alternative splicing at the end of exon 9 and only differ by the absence (-KTS) or presence (+KTS) of three amino acids; Lysine, Threonine and Serine (KTS). The functional difference between

48 these two isoforms was described by Hammes et al.; transcript specific knock-out mice displayed different phenotypes. Heterozygous mice with reduced Wt1+KTS expression develop a phenotype resembling Frasier syndrome. Homozygous

Wt1+KTS deficient mice die shortly after birth because of renal insufficiency and they show complete male to female sex reversal due to lowered levels of Sry159.

Homozygous Wt1-KTS deficient mice are characterized by hypo- and dysplastic kidneys and streak gonads159. These phenotypical differences are also encountered in human WT1 mutations: mutations perturbing the functioning of all WT1 transcripts result in Denys-Drash syndrome, mutations leading to reduced levels of WT1+KTS and perturbing the –KTS/+KTS balance cause Frasier syndrome (described in section

2.2.1). Additional research revealed that the WT1+KTS isoform is not a direct activator of SRY transcription, but that it preferentially binds to SRY mRNA, functioning as a posttranscriptional stabilizer155,156 (Figure 9B).

A second transcription factor acting on SRY activation is GATA-binding protein 4

(GATA4) and its co-factor Friend of GATA4-2 (FOG2 or ZFPM2) (Figure 9C). Both genes/proteins are highly expressed in the developing gonads and cardiac tissue157.

Their involvement in the regulation of the initial steps of gonadal differentiation was first assessed by evaluation of knock-out models160. Gata4-/- mice die at 7-9.5 dpc due to severe cardiac abnormalities, making it impossible to study the effect on gonadal differentiation161. This restriction was overcome by generation of a Gata4ki/ki knock-in mouse model, that specifically abrogates the interaction between Gata4 and its cofactor Fog2, these mice survive up to 13.5 dpc162. Fog2-/- mice survive until 14.5 dpc, allowing assessment of early gonadal development163. Both Fog2-/- and Gata4ki/ki mice show similar defects in gonadal development: testicular cords fail to form by 13.5 dpc,

Sry expression is reduced, Sry target gene expression is absent as is expression of genes essential for the onset of testosterone biosynthesis157,160. Besides this early role in sex determination, the GATA4/FOG2 complex is also believed to play a role in regulation of SOX9 and AMH expression157. In humans, GATA4 and FOG2

49 mutations/translocations were first identified in patients with heart septation defects and tetralogy of Fallot respectively164–166. Later on, identification of point mutations in patients with 46,XY DSD confirmed their importance in human sex development167,168.

More recently, the Mitogen-activated protein kinase (MAPK) pathway has been implicated in these early stages of testicular development. Using a forward genetic screen in mice, Bogani et al. identified a homozygous boygirl (byg) mutation, a missense variant in the mouse Map3k4 gene leading to male-to-female sex reversal. On the cellular level, this mutation leads to decreased somatic cell proliferation in the coelomic epithelium and to decreased mesonephric cell migration. On the molecular level, a dramatic reduction in the expression of Sox9 and Sry was noticed on both mRNA and protein level169. One year later, Pearlman et al. described MAP3K1 mutations in two families and 11 sporadic cases with 46,XY DSD. Follow-up studies aimed to unravel the molecular mechanism behind these mutations which are different in mice and humans (reviewed in Ostrer et al. 2014 158). In mice, Map3k4 together with its binding partner Gadd45γ induce Sry expression via p38 phosphorylation and subsequent Gata4 activation. Homozygous loss-of-function mutations perturb this pathway and cause male-to-female sex reversal170,171. So far, no homozygous MAP3K4 mutations have been identified in humans, most likely because they are embryonically lethal. The identified human MAP3K1 variants on the other hand have a gain-of-function effect. During normal male development, MAP3K1 interacts with AXIN1 and MAP3K4 which in its turn interacts with GADD45γ to augment SRY expression. At the same time MAP3K1 phosphorylates p38 and ERK1/2 that have a suppressing effect on female development via β-catenin inhibition. In case of MAP3K1 gain-of-function mutations, the female developmental situation is mimicked. These mutations lead to an increased phosphorylation of downstream targets like p38 and ERK1/2 and increased binding of female specific co-factors like

RHOA and Rock. Hyperphosporylation of p38 and ERK1/2 leads to inhibition of SOX9 and stabilization of β-catenin and upregulation of its downstream targets like FOXL2.

A possible sequestration of male specific proteins AXIN and MAP3K4 abolishes

MAP3K4-induced SRY expression158. Figure 9D summarizes the MAP3K1 related events during normal male development.

3.3.2 Downstream events in testicular development Once SRY is activated, a male specific gene expression cascade is initiated resulting in

Sertoli and Leydig cell differentiation. The following paragraphs will focus on the most important genes in male development: NR5A1 and the SOX genes. Figure 10 schematically shows the different interactions of these genes during development. An overview of other involved genes is given in Table 3 at the end of this section, genes implicated in the undifferentiated stage are also mentioned.

Figure 10: Overview of the most important genes for male sex development and their interactions (based on Ono et al. 2014172)

NR5A1: a gene with many faces

NR5A1 or steroidogenic factor-1 (SF-1) was first mentioned as one of the genes responsible for the development of the undifferentiated gonad, but later it turned out to be crucial in several subsequent steps of gonadal development and functioning such as Sertoli cell differentiation and steroidogenesis. Furthermore, NR5A1 plays a role in the regulation of the hypothalamic-pituitary axis.

The NR5A1 protein belongs to the nuclear receptor family and has a broad expression range, consistent with a role in different processes. In mice, Sf-1 expression is observed in the adrenogonadal primordium as early as 9 dpc108. In humans, expression starts in 51 the bipotential gonad at embryonic day 32-33, and it is consistently maintained in the somatic cells of the early testis, where it first serves a critical role in Sertoli cell differentiation108,173. A study by Sekido and Lovell-Badge has demonstrated that Nr5a1, together with Sry, can bind on multiple elements of a Sox9 gonad-specific enhancer in mice. Their work is also in favor of a feedforward, self-reinforcing signaling loop in which Sox9 together with Nr5a1 bind the same enhancer elements to maintain Sox9 expression174. In humans, adequate SOX9 expression is critical for Sertoli cell differentiation (discussed below). Once Sertoli cells are functional, a second role for

NR5A1 emerges. Starting at approximately 7 weeks of gestation, NR5A1 activates the expression of AMH resulting in the regression of Müllerian ducts in male fetuses175.

NR5A1 can also be seen as the initial trigger for virilization of the external genitalia, as it is a key activator of the CYP and the HSD enzymes that are crucial for steroid hormone production176. Besides expression in the developing gonads and adrenals,

NR5A1 expression also occurs in the ventromedial hypothalamus and the pituitary where it contributes to the reproductive axis by transcriptional activation of the genes encoding gonadotropin-releasing hormone (GnRH), FSH and LH177,178. Most recently, a role for NR5A1 in the development of the spleen has been postulated179,180. Despite extensive knowledge on NR5A1 target genes, far less is known about regulation of

NR5A1 itself108.

In the mid-80’s three different Nr5a1 knock-out (KO) mice were independently generated. Despite using different approaches, all three groups reported a similar phenotype. Homozygous NR5A1 KO mice exhibit complete gonadal and adrenal agenesis resulting in severe adrenal failure and a female phenotype with persistent

Müllerian ducts in male mice. Moreover, these animals have an abnormal spleen and ventromedial hypothalamus, and reduced gonadotropin secretion138,181,182.

Heterozygous KO mice show a less severe phenotype including hypoplastic adrenals and gonads during development. Remarkably, the adrenal hypoplasticity is partially compensated at later developmental stages and in adult life by increased proliferation

52 and these adrenals show an increased steroidogenic capacity. In the hypoplastic gonads no compensatory mechanisms have been observed. This disparity might by related to upregulation of a related transcription factor Ngfib in the adrenals which could compensate for Nr5a1 haploinsufficiency and which is not expressed in gonadal tissue183. This mechanism might also explain why human heterozygous mutations have a more distinct gonadal phenotype and apparently normal adrenal functioning.

Zhao et al. created a conditional KO mice, exclusively targeting NR5A1 expression in the central nervous system. These pituitary-specific KO mice are sterile and fail to develop normal male secondary characteristics. Ovaries and testes are significantly hypoplastic and FSH and LH levels are diminished. These studies established an essential role for NR5A1 in pituitary functioning178.

This broad functional spectrum of NR5A1 is mirrored in the plethora of phenotypes seen in patients with NR5A1 mutations. The first human NR5A1 mutation was a homozygous mutation identified in a patient with 46,XY DSD and primary adrenal failure45 and was quickly followed by the identification of heterozygous mutations in numerous patients with isolated 46,XY DSD. Screening of larger patient populations allows a prevalence estimation of NR5A1 mutations in 46,XY DSD cohorts; several studies report identification rates of ~15%184,185. For unknown reasons, the negative impact of NR5A1 mutations on steroidogenesis seems to resolve at least partially in puberty, leading to marked virilization of female and undermasculinised male patients186. Moreover, NR5A1 mutations have also been associated with distinct reproductive phenotypes such as hypospadias, bilateral anorchia and male infertility

(reviewed by Ferraz-de-Souza)44. In women, NR5A1 mutations are associated with

POI, and very recently a specific mutation, c.274C>T p.(Arg92Trp), has been implicated in 46,XX (ovo)testicular DSD by our group and two other teams (Chapter

4)109,187–189

SOX genes

To date, the only genuine candidate target gene for SRY is SOX9, also a transcription factor belonging to the SOX family characterized by the presence of a HMG-box. The first indication of its role in male sex development is its expression pattern; in the undifferentiated gonad Sox9 is expressed at low levels in both sexes, but it is extensively upregulated immediately after Sry expression in Sertoli cells. In addition,

Sox9 expression is observed in developing chondrocytes190. As mentioned above Sox9 is activated by a synergistical interaction of Sry and Nr5a1 on the Sox9 promoter174. The importance of Sox9 in testis development was confirmed by the phenotype of a Sox9 knock-out mouse; gonads of this XY KO mice are characterized by ovarian differentiation191. Furthermore, SOX9 overexpression in XX mice revealed that Sox9 is not only indispensable but also sufficient for testis development192.

Once SOX9 is expressed, it initiates multiple feedforward loops to maintain its own expression. A first loop starts with Sox9-induced upregulation of Fibroblast growth factor 9 (Fgf9). The Fgf9 protein activates the FGF receptor 2 (FGFR2), which has a positive effect on Sox9 expression193. Loss-of-function mouse models of both Fgf9 and

Fgfr2 are characterized by male-to-female sex reversal194,195. A second loop acts via the prostaglandin D synthase (Ptgds) gene, resulting in increased prostaglandin D2 secretion which acts in paracrine and autocrine ways and promotes the nuclear translocation of Sox9196,197. Remarkably, both feedforward loops can also induce Sox9 expression in neighbouring cells without endogenous Sry, showing that not Sry expression but indeed Sox9 expression is a prerequisite for Sertoli cell differentiation84.

Besides activation of these maintenance target genes, different SOX9 target genes have been discovered, as described below126. SOX9 expression also seems to be responsible for maintenance of NR5A1 expression in Sertoli cells198. Together with NR5A1, SOX9 has a critical role in upregulation of AMH. Based on experiments by Arango et al, it is

SOX9 that effectuates the actual transcriptional activation, while NR5A1 has a modulatory function199. Likewise, the SOX9-NR5A1 tandem is required for

54 upregulation of Vanin-1, a gene implicated in protection of primordial germ cells against oxidative stress200.

In humans, loss-of-function mutations and regulatory defects are a known cause of campomelic dysplasia, a semi-lethal skeletal malformation syndrome with 46,XY DSD in 75% of male cases42. Duplications, both coding and regulatory, cause 46,XX DSD or

Cooks syndrome (brachydactyly and anonychia), depending on the duplication size and position43,201. Disruption of specific long-range cis-regulatory elements are a known cause for Pierre-Robin-sequence, a condition characterized by craniofacial malformations202,203. This phenotypic spectrum emphasizes the need of an intact SOX9 regulatory region, known to be over 1 Mb long. The regulatory mechanisms leading to these different phenotypes were reviewed recently by our group (see this thesis,

Chapter 5)124.

SOX9 appears not to be the only SOX gene involved in sex development. Gain-of- function variants of SOX3 are known to cause 46,XX DSD and ectopic expression of

Sox3 in a transgenic mouse resulted in a comparable phenotype84,204. In mice, ectopic expression of Sox10 was also reported to induce Sry-independent female-to-male sex reversal. Sox10 KO mice on the other hand did not have a phenotype, suggesting that

Sox10 is not necessary for testis development but that it can function as a testis- determining gene205. A few SOX10 encompassing duplications have been described in patients with 46,XX DSD, however, it remains debatable whether the sex seversal is attributed to SOX10 as these rearrangements also contained other genes2. Based on its expression pattern and its ability to induce in vitro Amh expression, a sex developmental function was also suggested for Sox8. Despite the absence of a severe gonadal phenotype in Sox8-/- mice, a double Sox8/Sox9 KO model confirmed a role for

Sox8 as a reinforcer of Sox9. Both genes are functionally redundant206.

Table 3: Overview of the genes involved in the development of the undifferentiated gonad and genes involved in testicular development (based on reviews by Ono et al. and Eggers et al.2,172)

Gene Locus Function Inheritance Mouse Phenotype Human Phenotype

DSD related Other

Genes involved in the development of the bipotential gonad

Emx2 10q26.11 TF ? Emx2 −/− lacking kidneys, ureters, * * gonads and genital tracts

Lhx9 1q31.3 TF ? Lhx9 −/− fail to develop bipotential * * gonad

M33/Cbx2 17q25 TF AR Cbx2 −/− 46,XY DSD Normal

XY male-to-female sex reversal

XX impaired ovary development

Nr5a1 9q33 NR/TF AD/AD Nr5a1 −/− : fail to develop 46,XY gonadal dysgenesis, Primary adrenal failure bipotential gonad and adrenal hypospadias, micropenis

Nr5a1+/–: impaired adrenal stress 46,XX (ovo)testicular DSD, XX POI response, small gonads

Wt1 11p13 TF AD Wt1 −/− mice fail to develop bi- 46,XY DSD Wilms tumor, renal potential gonad abnormalities, gonadal tumours

Genes involved in testis development

Arx Xp22.13 TF X ARX– : Arrest of Leydig cell Dysgenetic testis, no Müllerian Lissencephaly, epilepsy, differentiation, abnormal neuronal structures, ambiguous external temperature instability

migration genitalia

Atrx Xq13.3 Chromatin X ATRX–: embryonically lethal Dysgenetic testis, no Müllerian α-Thalassaemia, mental remodeling structures, female/ambiguous/male retardation, dysmorphic face Sertoli-cell-specific ATRX–: small protein external genitalia testes and discontinuous tubules

(fetal period), delayed onset of

spermatogenesis (postnatal)

Dhh 12q13.1 Signaling AD/AR Dhh −/− disruption of testis cord 46,XY partial or complete gonadal Minifascicular neuropathy molecule formation due to abnormal dysgenesis

peritubular tissue

Dmrt1 9p24.3 TF Monosomic Dmrt1 −/− impaired testis 46,XY gonadal dysgenesis Facial abnormality, intellectual deletion development and loss of Sertoli and disability, microcephaly

germ cells

Fgf9 13q12.11 Signaling ? Fgf9 −/− XY male-to-female sex * molecule reversal and impaired development

of Sertoli cells

Fog2 8q22.3 Co-factor AD Fog2 −/− Reduced Sry levels, XY 46,XY DSD Congenital heart malformations male-to-female sex reversal

Gata4 8p23.1– TF AD Gata4 −/− embryo lethality (E7–E9.5) 46,XY DSD Congenital heart malformations p22

Gata4ki severe anomalies of testes

Mamld1 Xq28 Co- X Normal Hypospadias Normal activator

Map3k1 5q11.2 Kinase AD ° 46,XY partial or complete gonadal Normal dysgenesis

Map3k4 Kinase XY male-to-female sex reversal *

Nr0b1/Dax1 Xp21.3 NR/TF X XY impaired testis cord formation LOF: congenital adrenal hypoplasia Cleft palate, intellectual and spermatogenesis (genetic disability Duplications: 46,XY gonadal background dependent dysgenesis with hypogonadotropic

hypogonadism

Sox3 Xq27.1 TF X ° Duplications (including SOX3) and regulatory deletion: 46,XX testicular

DSD

Sox8 16p13.3 TF ? Sox8 −/− reduced fertility *

Sox8 −/−, Sox9 −/− XY variable

degree of male-to-female sex

reversal

Sox9 17q24– TF AD Sox9 −/− XY male-to-female sex 46,XY gonadal dysgenesis Campomelic dysplasia q25 reversal

Sox10 22q13.1 TF ? ° Duplication encompassing SOX10 (includes also other genes): 46,XX

(ovo)testicular DSD

Sry Yp11.3 TF Y Sry- : XY male-to-female sex LOF: 46,XY DSD GOF/Translocation: reversal 46,XX testicular DSD

Sry translocation: XX female-to-

male sex reversal

AD= autosomal dominant, AR= autosomal recessive, GOF = gain of function, LOF = loss of function, NR=nuclear receptor, TF= transcription factor, X= X- linked, *= no mutations described so far, °= no gonadal phenotype observed

3.4 Female sex developmental pathways For a long time, ovarian development has been seen as the default developmental pathway in the absence of SRY, for which no active genetic steps are required.

However, several lines of evidence contradict these views and are more in favor of the existence of ovary-specific genes and pathways. First, like any other organ, ovaries are build up out of different interacting cell types and as for other tissues it would be more likely that these complex differentiation processes require specific gene expression patterns. Second, human 46,XY DSD phenotypes, in the absence of Sry (e.g.

SRY deletions), result in complete gonadal dysgenesis rather than ovarian development and/or function, suggesting that active processes are necessary for ovarian development. Thirdly, observations of sex reversed mice showed that Sry needs to act during a specific time window to prevent ovarian differentiation, implying that competitive ovarian factors are present and need to be inhibited in a timely fashion126. Two different mechanisms underlying ovarian development were proposed: the existence of an SRY counterpart, an ovary-determining gene, and the

Z-factor hypothesis207. The first mechanism hypothesized that, given the importance of SRY in testis determination, the existence of a parallel ovary-determining gene is very likely208. Given the ovary-specific expression and its location on the X chromosome, Dosage-Sensitive Sex Reversal (DAX1) was initially considered a good candidate. Transgenic and KO mouse studies counter this assumption; Dax1 depletion has no or little effect on ovarian development and causes severe testicular dysgenesis207. The Z-factor model, on the other hand, could form an explanation for those cases in which 46,XX individuals in the absence of SRY, develop testes. The Z- factor is supposed to be a gene expressed in the XX gonad to repress testis development, thereby allowing ovarian differentiation. In XY gonads, the Z-factor is repressed by SRY expression. Inactivation of the Z-factor in XX individuals, would abrogate inhibition of the male pathways, thereby causing female-to-male sex reversal144. Due to their ability to antagonize testis development, genes like Foxl2 and

Wingless-Type MMTV Integration Site Family, Member 4 (Wnt4) have been suggested as

60 candidate Z-factor genes, however KO mice do not show female-to-male sex reversal, which is considered to be a critical feature of the Z-factor. The absence of one clear candidate gene suggests that a more complex pathway probably exists in which several genes play a role, but none of them are sufficient on their own207. Below, the two most studied pathways in ovarian differentiation will be elaborated: the

Wnt/RSPO1/β-catenin pathway and FOXL2. An overview of their interactions is given in Figure 11.

Figure 11: Overview of the most important genes for female sex development and their interactions (based on Ono et al. 2014172).

3.4.1 The WNT/RSPO1/β-catenin pathway The “canonical” WNT signaling pathway is widely used during developmental processes, including ovarian development. WNT protein ligands can bind with the membrane-associated Frizzled/LRP5/6 receptor complexes leading to receptor dimerization. Subsequently, β-catenin is detached from a β-catenin destruction complex, allowing nuclear translocation and interaction with target gene promoters.

In the absence of Wnt ligand, β-catenin is retained in the destruction complex, and can therefore not act as a transcriptional activator209.

In mice, Wnt4 expression starts in the undifferentiated gonads of both female and male mice. At 11.5 dpc its expression is upregulated in differentiating ovaries and the mesenchyme surrounding the Müllerian ducts and downregulated in males210. Wnt4

KO mice display partial female-to-male sex reversal, suggesting that Wnt4 has a positive role in ovarian development. Wnt4-/- XX gonads show both masculine and feminine characteristics; they display testis-like vascularization and androgen production, but on the other hand oocytes are still present although drastically

61 reduced in number and gonads lack testis cords210,211. Transgenic XY mice overexpressing Wnt4 do not show male-to-female sex reversal; suggesting that although Wnt4 plays an important role in at least some aspects of ovarian development, it is not sufficient to trigger these processes212. Inactivating WNT4 mutations in humans mimic to a certain extent the observed mouse phenotype, indicating a conserved role for WNT4 in mice and humans213–215.

Work by Chassot et al. has shown that other factors like RSPO1 also play a role in β- catenin upregulation. They developed a Rspo1 KO mouse characterized by masculinized females. These Rspo1-/- mice fail to activate Wnt4 and display testicular vascularization and steroidogenesis as a consequence. Furthermore gonadal Sox9 expression is observed around birth, resulting in sex cord formation216. Expression analysis experiments in Wnt4-/- and Rspo1-/- mice proved that Rspo1 is necessary for robust Wnt4 activation in the developing gonad and that it functions as an upstream regulator of β-catenin stabilization. The first human RSPO1 mutations were found in families with an autosomal recessive syndrome that includes 46,XX testicular DSD and skin abnormalities85

Conditional β-catenin KO mice confirm its importance in ovarian development, as they display ovarian defects comparable to those observed in Wnt4 and Rspo1 KO mice217.

Overexpression of β-catenin in the somatic cells of developing XY gonads, results in disruption of testis development and male-to-female sex reversal. These data underscore the importance of β-catenin as a key ovarian-promoting gene and an anti- testis gene. No human β-catenin mutations have been identified so far218.

3.4.2 FOXL2 A fourth important gene in ovarian development is FOXL2, first implicated in sex development with the identification of a deletion in the proximity of Foxl2 in the polled intersex (PIS) goat. These goats are characterized by polledness in both sexes and female-to-male sex reversal in XX goats219. FOXL2 mutations in humans are associated with BPES syndrome, as previously mentioned, characterized by eyelid malformations 62 and POI in a number of cases. To date, no FOXL2 mutations have been found in 46,XX testicular or OT DSD cases. The BPES phenotype is mimicked in Foxl2-/- mice and no sex reversal has been reported78,220,221. Curiously, Foxl2-/-, Wnt4-/- double knock-out mice do display testis differentiation, underlining the relevance of a synergistic action of the two genes for ovarian development and inhibition of testicular development222.

Moreover, conditional Foxl2 KO in adult XX mice results in transdifferentiation of ovarian tissue to testicular tissue with upregulation of male specific genes like Sox9.

More information on transdifferentiation and interaction between male and female pathways is presented below102.

3.5 Interactions between the male and female pathways and gonadal transdifferentiation In contrast to environmentally regulated sex determination, that has been studied in species like fish, amphibians, reptiles and birds, genetically determined sex development has been believed to be irreversible once the decision has been taken. In recent years however, it has become increasingly clear that active gene expression throughout life is required for gonadal maintenance. These findings have led to the elucidation of several inhibitory interactions between male and female specific genes

(Figure 12) 6,207,223.

Ottolenghi et al. were the first to attribute an ovarian maintenance role to FOXL2224.

Gene expression analysis in Foxl2-/- adult mice revealed upregulation of male specific genes like Sox9. Furthermore, these mice had ovaries containing a cord-like structure, resembling testicular cords. These results are in favor of ongoing requirement of Foxl2 to maintain ovarian cell fate224. Creation of a conditional Foxl2 KO mouse model confirmed this hypothesis. Adult mice showed upregulation of Sox9 and other genes and transdifferentiation of granulosa and theca cells to Sertoli and Leydig cells respectively. They showed that Foxl2, together with estrogen receptors, inhibits Sox9 expression by interaction with Testic-specific Enhancer of SOX9 Core region (TESCO) in the ovary102. Similarly, testicular cell fate is also unstable; loss of Dmrt1 in mouse

Sertoli cells reprograms them to a granulosa fate through activation of Foxl2225. These examples suggest that antagonism between male and female specific genes is required at least for gonadal maintenance in mice. Comparable antagonistic pathways have been described during the fetal period; Sox9 and Fgf9 expression in male mice represses expression of Wnt4/β-catenin and vice versa, however the molecular mechanism regulating this repression remains enigmatic193.

Figure 12: Summary of the interactions and antagonisms between male and female genes during gonadal development and after birth. Black arrows: established pathways leading to male and female development. Orange lines and boxes: female specific genes and their repressing roles on male specific genes. Blue lines and boxes: male specific genes and their repressing actions on female specific genes. (Adapted from Wilhelm et al. 20136).

The existence of these mutual antagonistic interactions, led to the hypothesis that abrogation of both pathways should result in neither testicular nor ovarian development. This hypothesis was tested by generation of two double KO mice: a Fgf9-

/-, Wnt4-/- mouse and a Sox9-/-, Rspo1-/- mouse226,227. Remarkably, neither of the XY mice showed sex reversal. Male Sox9-/-, Rspo1-/- mice are characterized by delayed testicular development resulting in hypoplastic testes with well-defined seminiferous cords, the 64 early stages of spermatogenesis are supported as well however these germ cells fail to initiate meiosis226. For male Fgf9-/-, Wnt4-/- mice similar observations were made227. In both double KO models, the XX phenotype resembled the one observed in Wnt4 and

Rspo1 single KO mice226,227. The results in both XX and XY double KO mice suggest that

Rspo1 and Wnt4 are truly necessary for ovarian development and that Sox9 functions can be taken over by redundant Sox genes. Furthermore the Fgf9, Wnt4 double KO model shows that Fgf9 signaling has a role in suppressing female differentiation rather than in supporting male development6.

4 DIAGNOSIS AND MANAGEMENT OF DSD PATIENTS

Today it is generally accepted that care for DSD patients requires the professional expertise of a multidisciplinary team (MDT) in which each member has specific know- how about the different aspects of these conditions21. These multidisciplinary teams are ideally coordinated by a “patient navigator” who forms a bridge between the team and the patient. Other members include an endocrinologist, urologist, gynaecologist, psychologist, biochemist, and clinical and molecular geneticist (Figure 13)228.

Furthermore DSD care should be centered in so-called centers of expertise that are capable of providing care for the entire spectrum of conditions and all age groups to ensure effective transition from childhood to adolescence and adulthood21. The sections below provide a concise overview of the current views regarding different aspects of diagnosis and management for DSD patients.

Figure 13: Overview of the multidisciplinary DSD team. (Adapted from Brain et al. and Hiort et al.21,228)

4.1 Diagnosis

Diagnosis of DSD conditions is based on the interplay of three important types of information: physical examination, biochemical analysis and genetic tests (section

5).

Structured clinical assessment of the appearance of the genitals is usually based on

Prader and/or external masculinization scores (EMS) (Figure 14). The study described in chapter 3 questions if a diagnostic strategy can be stratified according to the degree of virilization as assessed by the EMS. The EMS provides a tool for standardized and objective description of external genitalia, based on the meatus urethrae, phallic size, fusion of the labioscrotal folds and location of the gonads229,230. Previous studies have linked the EMS at diagnosis with important clinical outcome parameters such as sex assignment and tumor risk37,231. Next to a thorough general physical examination, it is important to screen for the presence of dysmorphic features, that may direct the diagnostic process towards a specific condition (e.g. ear abnormalities in the Mowat-

Wilson syndrome, see chapter 3, and structural organ anomalies such as cardiac defects, suggestive of GATA4 or FOG2 mutations).

Figure 14: The external masculinization score ( adapted from Ahmed et al. 2016232)

Laboratory investigations guide further diagnostic (genetic) tests as they permit an evaluation of gonadal and adrenal function21. Decreased levels of Sertoli cell products

AMH and Inhibin-B direct towards conditions associated with gonadal dysgenesis, whereas these markers are within the age-specific reference range in disorders of androgen synthesis or action233,234. When suspecting a defect of androgen synthesis, liquid chromatography–tandem mass spectrometry is superior to antibody-based methods to assess at what level steroid synthesis is perturbed, due to reduced cross- reactivity of steroids235,236. In case of normal steroid precursor levels in an individual with female or ambiguous genitalia, androgen insensitivity is expected. However, in some cases, expected to result from AIS, especially in case of partial AIS, no AR mutations can be found. In those cases the apolipoprotein D (APOD) test, which assesses AR-mediated APOD transcription, can be beneficial. These cases might be explained by defects in currently unknown AR cofactors. The main disadvantage of this test is the requirement of genital skin fibroblasts, which requires invasive sampling methods67,237.

As important during the diagnostic phase as the quest for the underlying genetic etiology and its and endocrine consequences is to document the internal and external anatomy. A first step is imaging of internal genital structures by ultrasound or eventually magnetic resonance imaging. Both techniques have limitations and in complex cases, more invasive procedures like laparoscopic evaluation of the gonads and gonadal biopsies are necessary 238. Though these procedures are more invasive, they can give an irrefutable answer about the internal genital situation e.g. presence of a uterus and the degree of gonadal differentiation (dysgenetic testis, ovotestis,

UGT,…), important for predicting hormonal function and tumor risk238,239.

4.2 Management The management of DSD patients and their families poses many challenges, and the initial step in this process is sex assignment. In the past sex assignment was mainly based on the appearance of the external genitalia and was accompagnied by early

68 surgical adjustment of the child’s genitalia to fit into a standard male or female phenotypic box. At that point it was believed that gender identity was primarily influenced by the gender of rearing and that early surgical interventions, in combination with a non-disclosing policy, would be in the best interst of the patient.

However, testimonials from adults who had had these early interventions and reports on gender change following these interventions have challenged this concept of gender neutrality at birth. Current believes allocate a role for early hormonal imprinting to gender development240,241. Today the gender assignment decision is mainly based on the expected gender outcome and several aspects like the karyotype and genetic diagnosis, the appearance of internal and external genitalia, gonadal hormone production and sensitivity and fertility aspect are taken into account. Given the low prevalence of each separate DSD condition, it is currently not possible to make straight-forward recommendations concerning gender development for all of them, although some general trends have been reported21. For instance, all individuals with

CAIS are raised as girls and have no gender issues. This situation is in contract with

PAIS, were gender assignment and acceptance is particularly difficult242,243. Regardless of the situation, the gender assignment decision is always taken by the MDT, together with the parents.

A second point of concern in DSD management is the risk for the development of gonadal germ cell tumors, which is significantly increased in DSD patients who have

Y chromosomal material, especially the GBY (gonadoblastoma region on Y). Germ cell cancer (GCC) development is initiated with the retention of pluripotent primordial germ cells, probably related to malfunctioning of Sertoli cells that are needed for the intrinsic germ cell maturation program. These pluripotent germ cells may expand in later life, resulting in the development of GCC244. Two types of precursor lesions have been observed depending on the cellular context; in case of dysgenetic gonads the GCC precursor is called a gonadoblastoma, in a testicular contex the precursor stage is referred to as germ cell neoplasia in situ (GCNIS), formerly known as carcinoma in

69 situ245,246. The main goal in DSD-related GCC research is to obtain a straight-forward clinical decision model based on reliable parameters and risk factors. These risk factors are: the presence of Y chromosomal material in the (gonadal) karyotype (specifically the GBY region), the presence of immature germ cells, a low degree of gonadal differentiation (“testicularization” in the context of an XY karyotype), and aberrant expression of KITLG (stem cell factor). The initial location of the gonad is a modifying factor, with abdominally or inguinally positioned testes increasing the risk. The ultimate decision to perform gonadectomy is based on germ cell cancer risk assessment, (expected) gonadal function, modalities for non-conservative follow-up and patient/parent preferences. Alternatives for gonadectomy are self-examination and regular imaging of gonads in situ. Serum tumor markers are currently of limited use in a prophylactic setting244.

A third major aspect of DSD management lies in the preservation of potential fertility.

For most of the individuals with a DSD condition, sub- or infertility is inherent to their condition and their reproductive potential can be affected by several medical and surgical therapies. In general, DSD patients can benefit from commonly available reproductive techniques like ovarian or testicular tissue cryopreservation, ovarian stimulation and oocyte/embryo cryopreservation and testicular sperm extraction.

However, today infertility treatment in DSD patients is limited as their gonadal abnormalities limit the success rate of these technologies. Most reports of succesfull reproduction involve men with KS, PAIS or 5-alfa reductase deficiency247.

Furthermore, women with CAH appear to have pregnancy rates comparable with the general population. When fertility preservation is impossible or unsuccessfull, options like egg of sperm donation, surrogacy or adoption can fulfill the desire to have children248. Although these reproductive options contribute greatly to the management plan for DSD patients, some ethical concerns should be considered as well. For instance, some families postpone tumorrisk-related gonadectomy for possible fertility purposes. As counseling for the risk of malignancies and fertility are both uncertain, it

70 is currently unclear which arguments are decisive and where the trade-off should be.

Furthermore, as most DSD conditions are the result of a genetic mutation, which could be passed on to the offspring, genetic counseling and pre-implantation genetic diagnosis are crucial for DSD individuals247.

One of the most debatable aspects of DSD management is surgery. Today, surgical interventions on genitalia and gonads in neonates and infants who have a DSD are addressed with extreme caution and postponing genital surgery until the child can give consent has become a valuable alternative in recent years249,250. Long term data on satisfaction with surgery indeed show that informed consent and realistic expectations are of utmost importance21. However, given the heterogeneity of DSD conditions and the different arguments for both sides of the dilemma, thusfar no consensus on the indications, timing, procedure or outcome evaluation has been reached251. The general aims of surgery in DSD are extensive and include both functional (restoration of genital function anatomy, facilitation of future reproduction, reduction of urological hazards like urogenital infection of incontinence, avoidance of late virilization, reduction of gonadal tumor risk) and social (avoidance of social stigmatization and the address the parents desire to raise their child in the best possible way) considerations252. In general, DSD specialists do agree that each decision concerning surgery should be taken in agreement of a multidisciplinary expert DSD team and that it can only been undertaken at centers of expertise, given the complexity and irreversibility of these procedures. Furthermore, decisions concerning gonadectomy should be conservative unless a significant risk for gonadal tumor development necessitates an early intervention251,252. Even in case of augmented tumor risk, a decision can not only be based on estimated group risks but should also consider pathological examination of a gonadal biopsy, the age of the patient, the anatomical position and possible residual function of the gonad and possible side effects of the surgery253. As a final remark in this debate, it should also be noted that despite several reports on dissatisfaction with early childhood surgical interventions, only very

71 limited data is available on the impact on non-treated DSD during childhood252. This ongoing debate and the lack of clear scientific evidence on the preferential approach clearly emphasizes the importance of case-specific decision making addressing the needs and quality of life of all stake-holders.

5 TOWARDS A FASTER GENETIC DIAGNOSIS FOR DSD PATIENTS

Due to the development of new sequencing techniques, substantial progress has been made in the molecular diagnosis of DSD conditions, both at the level of mutation detection and of gene identification. Previously, with Sanger sequencing being the only available method, a sequential (gene-by-gene) approach was used, making the diagnostic process both laborious and expensive. With the advent of novel, high- throughput sequencing techniques, such as next generation sequencing (NGS), it has become possible to screen gene panels in parallel in a flexible manner. Today, techniques like whole exome and whole genome sequencing (WES/WGS) are greatly facilitating the discovery of new disease genes and step-by-step they are being introduced in a clinical setting as a broad screening test for genetically heterogeneous disorders, and thus also in the DSD diagnostic workflow254,255. In the following sections a brief overview is given of the different strategies that are used in our department

(Center for Medical Genetics Ghent, CMGG) to unravel the underlying genetic cause in individuals who have a DSD.

5.1 Pre-screening approach

5.1.1 Targeted resequencing of known DSD genes

Sanger sequencing In the seventies of the previous century, Fred Sanger developed a groundbreaking method for DNA sequencing256. Due to Sanger sequencing it became possible to screen the sequence of one or a small set of genes for variants. This polymerase chain reaction (PCR)-based technique relies on cyclic repeats of template denaturation, primer annealing and elongation. By using a combination of standard deoxynucleotides (dNTPs) and modified fluorescently-labeled dideoxynucleotides

(ddNTPs), fragments are generated that differ in length by one nucleotide. These fragments are subsequently separated by capillary electrophoresis and their fluorescent signal is detected by laser excitation257 (Figure 15). The per-base accuracy

73 of Sanger sequencing reaches 99.999%258, making it still a valuable technique for validation of identified variants. However due to the high-cost and its labor intensity, new techniques are quickly replacing Sanger sequencing for mutation screening in genetically heterogeneous conditions like DSD.

Figure 15: Sanger sequencing is based on a modified PCR reaction where fluorescently- labeled ddNTPs are added to the reaction mix. PCR fragments are generated that differ in length by one nucleotide. Separation of these fragments by capillary electrophoresis followed by laser excitation of the fluorescent label allows base calling.

Targeted resequencing using next-generation sequencing technologies Given the disadvantages of Sanger sequencing, the quest for better sequencing technologies has generated several so-called NGS technologies; the most successful of them being the sequencing-by-synthesis (SBS) method, for instance developed by

Illumina259. This method is build up by three sequential steps: (1) library preparation,

(2) bridge PCR amplification and (3) SBS (Figure 16). A comprehensive video on how this technology works is available at http://www.illumina.com/technology/next- generation-sequencing/sequencing-technology.html. This sequencing process is followed by read mapping against the human reference genome and variant calling using different software tools258,259. The biggest advantage of this technology is its ability to generate millions of DNA reads in parallel, thereby increasing cost- efficiency and decreasing labor intensity260.

In the CMGG this NGS technology is used for several applications, amongst them targeted resequencing of gene panels. De Leeneer et al. developed an elegant system based on standardized singleplex PCR reactions for target enrichment261. This method

74 is highly suitable for sequencing small gene panels and allows flexibility as new genes can easily be added. An in-house DSD gene panel was developed containing relevant

DSD genes like SRY, NR5A1, SOX9, WT1, DMRT1, GATA4 and AR.

Figure 16: The sequencing-by-synthesis NGS method. (Adapted from Illumina.com). The sequencing by synthesis method involves three major steps. First of all, DNA is fragmentized during the library preparation, followed by adaptor ligation. Subsequently, these DNA fragments are loaded on a flowcell, where the adaptors hybridize with complementary probes on the surface. Bridge amplification results in clustering of identical DNA fragments. As a final step, modified nucleotides are added and incorporated with the formation of double stranded molecules. Laser excitation of the chromofores enables base calling.

5.1.2 Copy number detection: MLPA

Several reports on the identification and importance of CNVs in DSD, underscore the importance of CNV screening in the diagnostic package for DSD262,263. Multiplex

Ligation-dependent Probe Amplification (MLPA), is a very accurate technique for targeted CNV detection. MLPA distinguishes itself from other techniques by amplifying MLPA probes instead of target DNA and by the fact that only a single pair of PCR primers is used264. The principle behind MLPA is shown in Figure 17.

Advantages of MLPA are its cost-effectiveness, technical simplicity and high resolution, however this technique is not suited for genome-wide CNV detection264.

In our diagnostic setting two different commercially available MLPA kits are used. The first is the Intersex MLPA kit with probes for DAX1, NR5A1, SOX9, SRY and WNT4

(P185-Intersex, MRC Holland), the second kit contains probes for WT1 (P118-WT1,

MRC Holland). This combination allows us to detect CNVs in a subset of important

DSD genes.

Figure 17: The MPLA-principle. MLPA reactions are based on a standard PCR reaction. For each target specific probes are designed, consisting of three parts. (1) the hybridization sequence, complementary with the target sequence, (2) a stuffer sequence that differs in length between the different targets and allows a separation of the different fragments and (3) the universal primers X and Y, that enable multiplex fragment amplification. After DNA denaturation, probes are hybridized on targed DNA, followed by a ligation step. In case of deletions of duplications, this ligation is perturbed. Probe ligation is followed by a PCR amplification using primers X and Y. Subsequent capillary electrophoresis allows fragment separation and detection of the different fragments.

5.2 Identification of new DSD genes: whole exome sequencing A second NGS application is WES, by which the coding regions of the human DNA are enriched and sequenced. This technique has proven to be a valuable tool to identify new disease genes underlying Mendelian disorders. However, the use of WES in a clinical setting should be applied with caution as complex filtering procedures are needed and false-positive and false-negative rates are considerably higher as compared to targeted approaches. From an ethical point-of-view it is important to notice that WES might also uncover secondary findings beyond the scope of the diagnostic question, emphasizing the necessity of detailed informed consent265.

During this thesis, two different WES approaches were used to identify novel disease genes in our DSD patient cohort. In cases with a consanguineous background, where we expect a homozygous causal mutation, data filtering was facilitated by use of identity-by-descent (IBD) regions. To understand this technology, knowing the difference between identity-by-state (IBS) and identity-by-descent is crucial, shown in Figure 18A and B respectively. In both situation A and B, two brothers share the same allele A1, however in situation A the allele originates from a different ancestor in both brothers while in situation B both brothers have inherited the A1 allele from their father. Alleles originating from the same ancestor are called IBD-alleles, alleles originating from a different ancestor are called IBS-alleles. Children originating from a consanguineous marriage have a higher chance of being homozygous for the same allele, because both parents share a larger proportion of their genome as compared to non-related parents. This is shown in Figure 18C: both parents share allele A1, which they both inherited from their common ancestor. The chance that their offspring are homozygous for A1 is 25%. Off course, this concept does not apply only to the A1 allele but also to the surrounding alleles, resulting in homozygous regions by descent. In general the size of such regions is directly proportional to the degree of consanguinity266. Homozygosity-mapping enables the identification of those regions

77 and when combined with WES, it can reduce the number of variants that need to be filtered out267.

In our center homozygosity-mapping is performed by using the genome-wide SNP- chip analysis with the Illumina HumanCytoSNP-12 DNA Analysis Beadchip, allowing genotyping of over 200 000 single nucleotide polymorphisms (SNPs). The obtained data are subsequently analyzed with the PLINK software package which is incorporated in ViVar, an online platform for data storage and analysis268,269. The homozygous regions are then applied for filtering of WES data.

Figure 18: Identity-by-state (IBS) and identity-by-descent (IBD)

When no consanguinity is suspected, a trio-sequencing strategy is applied. DSD phenotypes are usually associated with infertility, implying that most cases will be due to a de novo mutation. However, when looking for a causal variant for DSD phenotypes it is to be kept in mind that a variant can be present in one of the parents without causing a phenotype. For example, if a disease-causing GATA4 variant is suspected in an individual who has 46,XY DSD, one would not expect the same variant to be present in the patient’s healthy father. However, the same variant can be allowed in the patients mother given that “feminizing” mutations would not harm her. An important pitfall is that for a lot of DSD genes, incomplete penetrance in same sex individuals has been described270–272.

6 REFERENCES

1. Biason-lauber, A. Control of sex development. Best Pract. Res. Clin. Endocrinol. Metab. 24, 163–186 (2010). 2. Eggers, S. & Sinclair, A. Mammalian sex determination—insights from humans and mice. Chromosome Res. 20, 215–38 (2012). 3. Ungewitter, E. K. & Yao, H. H. C. How to make a gonad: Cellular mechanisms governing formation of the testes and ovaries. Sex. Dev. 7, 7–20 (2012). 4. Rey, R. a & Grinspon, R. P. Normal male sexual differentiation and aetiology of disorders of sex development. Best Pract. Res. Clin. Endocrinol. Metab. 25, 221–238 (2011). 5. Nikolic, A., Volarevic, V., Armstrong, L., Lako, M. & Stojkovic, M. Primordial germ cells: Current knowledge and perspectives. Stem Cells Int. 2016, (2016). 6. Wilhelm, D., Yang, J. X. & Thomas, P. Mammalian Sex Determination and Gonad Development. Current Topics in Developmental Biology 106, (Elsevier Inc., 2013). 7. Öçal, G. Current concepts in disorders of sexual development. JCRPE J. Clin. Res. Pediatr. Endocrinol. 3, 105–114 (2011). 8. Gilbert, S. in Developmental biology (2000). 9. DeFalco, T., Takahashi, S. & Capel, B. Two distinct origins for Leydig cell progenitors in the fetal testis. Dev. Biol. 352, 14–26 (2011). 10. O’Shaughnessy, P. J., Baker, P. J. & Johnston, H. The foetal Leydig cell-- differentiation, function and regulation. Int. J. Androl. 29, 90-5-8 (2006). 11. Jost, A. Problems of fetal endocrinology: the gonadal and hypophyseal hormones. Recent Prog. Horm. Res. 8, 379–418 (1953). 12. Rey, R., Lukas-croisier, C., Lasala, C. & Bedecarrás, P. AMH / MIS : what we know already about the gene , the protein and its regulation. Mol. Cell. Endocrinol. 211, 21–31 (2003). 13. Taguchi, O., Cunha, G. R., Lawrence, W. D. & Robboy, S. J. Timing and irreversibility of M??llerian duct inhibition in the embryonic reproductive tract of the human male. Dev. Biol. 106, 394–398 (1984). 14. Parker, K. et al. Steroidogenic factor 1: an essential mediator of endocrine development. Recent Prog. Horm. Res. 57, 19–36 (2002). 15. Hannema, S. E. & Hughes, I. A. Regulation of Wolffian duct development. Horm. Res. 67, 142– 151 (2007). 16. Zalel, Y., Pinhas-Hamiel, O., Lipitz, S., Mashiach, S. & Achiron, R. The development of the fetal penis--an in utero sonographic evaluation. Ultrasound Obstet. Gynecol. 17, 129–131 (2001). 17. Jost, A. Becoming a male. Adv. Biosci. 10, 3–13 (1973). 18. Hughes, I. A., Houk, C., Ahmed, S. F. & Lee, P. A. Consensus statement on management of intersex disorders. Endocrine 554–562 (2006). doi:10.1136 19. Hughes, I. A. Consequences of the Chicago DSD consensus: A personal perspective. Horm. Metab. Res. 47, 394–400 (2015). 20. Hughes, I. A. Disorders of sex development : a new definition and classification. Best Pract. Res. Clin. Endocrinol. Metab. 22, 119–134 (2008). 21. Hiort, O. et al. Management of disorders of sex development. Nat. Rev. Endocrinol. 10, 520–9 (2014). 22. Lanfranco, F., Kamischke, A., Zitzmann, M. & Nieschlag, E. Klinefelter’s syndrome. Lancet 364, 273–83 (2004). 23. Stochholm, K., Bojesen, A., Jensen, A. S., Juul, S. & Gravholt, C. H. Criminality in men with Klinefelter’s syndrome and XYY syndrome: a cohort study. BMJ Open 2, e000650 (2012). 24. Linden, M., Bender, B. & Robinson, A. Sex Chromosome Tetrasomy and Pentasomy. Pediatrics 96, 672–682 (1995). 25. Tartaglia, N., Ayari, N., Howell, S., D'Epagnier, C. & Zeitler, P. 48,XXYY, 48,XXXY and 49,XXXXY syndromes: Not just variants of Klinefelter syndrome. Acta Paediatr. Int. J. Paediatr. 79

100, 851–860 (2011). 26. Klinefelter, H., Reifenstein, E. & Albright, F. CLINICAL Syndrome Characterized by Gy-. J. Clin. Endocrinol. Metab. 2, 615–627 (1942). 27. Bird, R. J. & Hurren, B. J. Anatomical and clinical aspects of Klinefelter’s syndrome. Clin. Anat. 619, 606–619 (2016). 28. Nielsen, J. & Wohlert, M. Sex chromosome abnormalities found among 34,910 newborn children: results from a 13-year incidence study in Arhus, Denmark. Birth Defects Orig. Artic. Ser. 26, 209– 223 (1990). 29. Donaldson, M. D. C., Gault, E. J., Tan, K. W. & Dunger, D. B. Optimising management in Turner syndrome: from infancy to adult transfer. Arch. Dis. Child. 91, 513–520 (2006). 30. Al Alwan, I. et al. Turner Syndrome Genotype and phenotype and their effect on presenting features and timing of Diagnosis. Int. J. Health Sci. (Qassim). 8, 195–202 (2014). 31. Granger, A., Zurada, A., Zurada-Zielińska, A., Gielecki, J. & Loukas, M. Anatomy of turner syndrome. Clin. Anat. 642, 638–642 (2016). 32. Malan, V., Vekemans, M. & Turleau, C. Chimera and other fertilization errors. Clin. Genet. 70, 363–373 (2006). 33. El Moussaif, N., Haddad, N. El, Iraqi, N. & Gaouzi, A. 45,X/46,XY mosaicisme: Report of five cases and clinical review. Ann. Endocrinol. (Paris). 72, 239–243 (2011). 34. Lindhardt Johansen, M. et al. 45,X/46,XY mosaicism: phenotypic characteristics, growth, and reproductive function--a retrospective longitudinal study. J. Clin. Endocrinol. Metab. 97, E1540-9 (2012). 35. Swyer, G. I. M. Male Pseudohermaphroditism: A Hitherto Undescribed Form. Br. Med. J. 2, 709– 712 (1955). 36. Michala, L., Goswami, D., Creighton, S. M. & Conway, G. S. Swyer syndrome: Presentation and outcomes. BJOG An Int. J. Obstet. Gynaecol. 115, 737–741 (2008). 37. Cools, M., Wolffenbuttel, K. P., Drop, S. L. S., Oosterhuis, J. W. & Looijenga, L. H. J. Gonadal development and tumor formation at the crossroads of male and female sex determination. Sex Dev. 5, 167–80 (2011). 38. Mendonca, B. B., Domenice, S., Arnhold, I. J. P. & Costa, E. M. F. 46,XY disorders of sex development (DSD). Clin. Endocrinol. (Oxf). 70, 173–87 (2009). 39. King, T. F. J. & Conway, G. S. Swyer syndrome. Curr. Opin. Endocrinol. Diabetes. Obes. 21, 504–10 (2014). 40. Jäger, R. J., Anvret, M., Hall, K. & Scherer, G. A human XY female with a frame shift mutation in the candidate testis-determining gene SRY. Nature 348, 452–4 (1990). 41. Wagner, T. et al. Autosomal sex reversal and campomelic dysplasia are caused by mutations in and around the SRY-related gene SOX9. Cell 79, 1111–20 (1994). 42. Wunderle, V. M., Critcher, R., Hastie, N., Goodfellow, P. N. & Schedl, A. Deletion of long-range regulatory elements upstream of SOX9 causes campomelic dysplasia. Proc. Natl. Acad. Sci. U. S. A. 95, 10649–10654 (1998). 43. Benko, S. et al. Disruption of a long distance regulatory region upstream of SOX9 in isolated disorders of sex development. J. Med. Genet. 48, 825–30 (2011). 44. Ferraz-de-Souza, B., Lin, L. & Achermann, J. C. Steroidogenic factor-1 (SF-1, NR5A1) and human disease. Mol. Cell. Endocrinol. 336, 198–205 (2011). 45. Achermann, J., Masafumi, I., Ito, M., Hindmarsh, P. & Jameson, J. A mutation in the gene encoding steriodogenic factor -1 causes XY sex reversal and adrenal failure in humans. Nat. ge 22, 1–2 (1999). 46. Drash, A., Sherman, F., Hartmann, W. & Blizzard, R. A syndrome of pseudohermaphroditism, Wilm’s tumor, hypertension and degenerative renal disease. J.Pediat. 76, 1970 (1970). 47. Pelletier, J. et al. Germline mutations in the Wilms’ tumor suppressor gene are associated with abnormal urogenital development in denys-drash syndrome. Cell 67, 437–447 (1991). 48. Frasier, S. D., Bashore, R. A. & Mosier, H. D. Gonadoblastoma Associated With Pure Gonadal

Dysgenesis in Monozygous Twins. J. Pediatr. 64, 740–5 (1964). 49. Barbaux, S. et al. Donor splice-site mutations in WT1 are responsible for Frasier syndrome. Nat. Genet. 17, 467–470 (1997). 50. Krob, G., Braun, A. & Kuhnle, U. True hermaphroditism: Geographical distribution, clinical findings, chromosomes and gonadal histology. Eur. J. Pediatr. 153, 2–10 (1994). 51. Mendonca, B. B. et al. 46 , XY DSD due to impaired androgen production. Best Pract. Res. Clin. Endocrinol. Metab. 24, 243–262 (2010). 52. Brown, D. M., Markland, C. & Dehner, L. P. Ley dig Cell Hypoplasia: A Cause of Male Pseudohermaphroditism. J Clin Endocrinol Metab 46, 1–7 (1978). 53. Zenteno, J. C., Canto, P., Kofman-Alfaro, S. & Mendez, J. P. Evidence for genetic heterogeneity in male pseudohermaphroditism due to Leydig cell hypoplasia. J. Clin. Endocrinol. Metab. 84, 3803–3806 (1999). 54. Vezzoli, V. et al. A newvariant in signal peptide of the human luteinizing hormone receptor (LHCGR) affects receptor biogenesis causing leydig cell hypoplasia. Hum. Mol. Genet. 24, 6003– 6012 (2015). 55. Bianconi, S. E., Cross, J. L., Wassif, C. A. & Porter, F. D. Pathogenesis, epidemiology, diagnosis and clinical aspects of Smith–Lemli–Opitz syndrome. Expert Opin. Orphan Drugs 3, 267–280 (2015). 56. DW, S., Lemli, L. & Opitz, J. A newly recognized syndrome of multiple congenital anomalies. J Pediatr 1964;64:210-17 . J. Pediatr. 64, 210–17 (1964). 57. Irons, M., Roy Elias, E., Salen, G., Tint, G. S. & Batta, A. Defective cholesterol biosynthesis in Smith-Lemli-Opitz syndrome. Lancet 341, 1414 (1993). 58. Fitzky, B. U. et al. Mutations in the Delta7-sterol reductase gene in patients with the Smith-Lemli- Opitz syndrome. Proc. Natl. Acad. Sci. U. S. A. 95, 8181–6 (1998). 59. Waterham, H. R. et al. Smith-Lemli-Opitz syndrome is caused by mutations in the 7- dehydrocholesterol reductase gene. Am. J. Hum. Genet. 63, 329–38 (1998). 60. Wassif, C. A. et al. Mutations in the Human Sterol Δ7-Reductase Gene at 11q12-13 Cause Smith- Lemli-Opitz Syndrome. Am. J. Hum. Genet. 63, 55–62 (1998). 61. Moebius, F. F., Fitzky, B. U., Lee, J. N., Paik, Y. K. & Glossmann, H. Molecular cloning and expression of the human delta7-sterol reductase. Proc. Natl. Acad. Sci. U. S. A. 95, 1899–1902 (1998). 62. Aradhya, S. et al. Exon-level array CGH in a large clinical cohort demonstrates increased sensitivity of diagnostic testing for Mendelian disorders. Genet. Med. 14, 594–603 (2012). 63. Lanthaler, B. et al. Characterization of large deletions in the DHCR7 gene. Clin. Genet. 88, 149– 154 (2015). 64. Quigley, C. et al. Androgen receptor defects: Historical, clinical, and molecular perspectives. Endocr Rev. 16, 271–321 (1995). 65. Gottlieb, B., Beitel, L. K., Nadarajah, A., Paliouras, M. & Trifiro, M. The androgen receptor gene mutations database: 2012 update. Hum. Mutat. 33, 887–894 (2012). 66. Hornig, N. C. et al. A Recurrent Germline Mutation in the 5’UTR of the Androgen Receptor Causes Complete Androgen Insensitivity by Activating Aberrant uORF Translation. PLoS One 11, e0154158 (2016). 67. Hornig, N. C. et al. Identification of an AR mutation-negative class of androgen insensitivity by determining endogenous AR activity. J. Clin. Endocrinol. Metab. 101, 4468–4477 (2016). 68. Boehmer, A. et al. Phenotype in Families with Androgen Insensitivity Syndrome ¨. J. Clin. Endocrinol. Metab. 86, 4151–4160 (2009). 69. Hiort, O., Werner, R. & Gro, H. 46 , XY disorders of sex development – the undermasculinised male with disorders of androgen action. Best Pract. Res. Clin. Endocrinol. Metab. 24, 263–277 (2010). 70. Hegarty, P. K., Mushtaq, I. & Sebire, N. J. Natural history of testicular regression syndrome and consequences for clinical management. J. Pediatr. Urol. 3, 206–208 (2007). 71. Spires, S. E., Woolums, C. S., Pulito, A. R. & Spires, S. M. Testicular regression syndrome: a

clinical and pathologic study of 11 cases. Arch. Pathol. Lab. Med. 124, 694–698 (2000). 72. Pirgon, Ö. & Dündar, B. N. Vanishing testes: a literature review. J. Clin. Res. Pediatr. Endocrinol. 4, 116–20 (2012). 73. Philibert, P. et al. Mutational analysis of steroidogenic factor 1 (NR5a1) in 24 boys with bilateral anorchia: a French collaborative study. Hum Reprod 22, 3255–3261 (2007). 74. George, M., Schneuer, F. J., Jamieson, S. E. & Holland, A. J. A. Genetic and environmental factors in the aetiology of hypospadias. Pediatr. Surg. Int. 31, 519–527 (2015). 75. van der Zanden, L. F. M. et al. Aetiology of hypospadias: a systematic review of genes and environment. Hum. Reprod. Update 18, 260–83 (2012). 76. Ahmed, S. F., Warne, G., Quigley, C. & Cools, M. in Biology of Reproduction 681–715 (2016). doi:10.1016/B0-7216-9777-1/50018-4 77. Pierce, S. B. et al. Mutations in the DBP-deficiency protein HSD17B4 cause ovarian dysgenesis, hearing loss, and ataxia of perrault syndrome. Am. J. Hum. Genet. 87, 282–288 (2010). 78. Crisponi, L. et al. The putative forkhead transcription factor FOXL2 is mutated in blepharophimosis/ptosis/epicanthus inversus syndrome. Nat. Genet. 27, 159–66 (2001). 79. Baere, E. De et al. Spectrum of FOXL2 gene mutations in blepharophimosis-ptosis-epicanthus inversus ( BPES ) families demonstrates a genotype – phenotype correlation. Hum. Mol. Genet. 10, 1591–1600 (2001). 80. Hersmus, R. et al. FOXL2 and SOX9 as parameters of female and male gonadal differentiation in patients with various forms of disorders of sex development (DSD). J. Pathol. 215, 114–125 (2008). 81. Kojima, Y. et al. Up-regulation of SOX9 in human sex-determining region on the Y chromosome (SRY)-negative XX males. Clin. Endocrinol. (Oxf). 68, 791–799 (2008). 82. Grinspon, R. P. & Rey, R. A. Disorders of Sex Development with Testicular Differentiation in SRY-Negative 46,XX Individuals: Clinical and Genetic Aspects. Sex. Dev. 1–11 (2016). doi:10.1159/000445088 83. Huang, B., Wang, S., Ning, Y., Lamb, a N. & Bartley, J. Autosomal XX sex reversal caused by duplication of SOX9. Am. J. Med. Genet. 87, 349–53 (1999). 84. Sutton, E. et al. Identification of SOX3 as an XX male sex reversal gene in mice and humans. Jounal Clin. Investig. 121, 328–341 (2011). 85. Parma, P. et al. R-spondin1 is essential in sex determination, skin differentiation and malignancy. Nat. Genet. 38, 1304–9 (2006). 86. Tomaselli, S. et al. Syndromic True Hermaphroditism Due to an R-spondin1 ( RSPO1 ) Homozygous Mutation. Hum. Mutat. 29, 220–226 (2008). 87. Looijenga, L. H. J. et al. Gonadal tumours and DSD. Best Pract. Res. Clin. Endocrinol. Metab. 24, 291–310 (2010). 88. Cools, M., Drop, S. L. S., Wolffenbuttel, K. P., Oosterhuis, J. W. & Looijenga, L. H. J. Germ cell tumors in the intersex gonad: old paths, new directions, moving frontiers. Endocr. Rev. 27, 468– 84 (2006). 89. Lau, Y., Chou, P., Iezzoni, J., Alonzo, J. & Kömüves, L. Expression of a candidate gene for the gonadoblastoma locus in gonadoblastoma and testicular seminoma. Cytogenet. Cell Genet. 91, 160–4 (2000). 90. Verkauskas, G. et al. The Long-Term Followup of 33 Cases of True Hermaphroditism: A 40-Year Experience With Conservative Gonadal Surgery. J. Urol. 177, 726–731 (2007). 91. Auchus, R. J., Ph, D. & Chang, A. Y. 46 , XX DSD : the masculinised female. Best Pract. Res. Clin. Endocrinol. Metab. 24, 219–242 (2010). 92. Welt, C. K. Primary ovarian insufficiency: a more accurate term for premature ovarian failure. Clin. Endocrinol. (Oxf). 68, 499–509 (2008). 93. De Vos, M., Devroey, P. & Fauser, B. C. J. M. Primary ovarian insufficiency. Lancet 376, 911–21 (2010). 94. Beck-Peccoz, P. & Persani, L. Premature ovarian failure. Orphanet J. Rare Dis. 1, 9 (2006). 95. Cordts, E. B., Christofolini, D. M., Dos Santos, A. A., Bianco, B. & Barbosa, C. P. Genetic aspects

of premature ovarian failure: a literature review. Arch. Gynecol. Obstet. 283, 635–43 (2011). 96. Tucker, E. J., Grover, S. R., Bachelot, A., Touraine, P. & Sinclair, A. H. Premature Ovarian Insufficiency: New Perspectives on Genetic Cause and Phenotypic Spectrum. Endocr Rev 37, 609– 635 (2016). 97. Di Pasquale, E., Beck-Peccoz, P. & Persani, L. Hypergonadotropic ovarian failure associated with an inherited mutation of human bone morphogenetic protein-15 (BMP15) gene. Am. J. Hum. Genet. 75, 106–11 (2004). 98. Di Pasquale, E. et al. Identification of new variants of human BMP15 gene in a large cohort of women with premature ovarian failure. J. Clin. Endocrinol. Metab. 91, 1976–1979 (2006). 99. Allen, E. G. et al. Examination of reproductive aging milestones among women who carry the FMR1 premutation. Hum. Reprod. 22, 2142–2152 (2007). 100. Schwartz, C. E. et al. Obstetrical and gynecological complications in fragile X carriers: A multicenter study. Am. J. Med. Genet. 51, 400–402 (1994). 101. Wittenberger, M. D. et al. The FMR1 premutation and reproduction. Fertil. Steril. 87, 456–465 (2007). 102. Uhlenhaut, N. H. et al. Somatic sex reprogramming of adult ovaries to testes by FOXL2 ablation. Cell 139, 1130–42 (2009). 103. Verdin, H. & De Baere, E. FOXL2 impairment in human disease. Horm. Res. pædiatrics 77, 2–11 (2012). 104. Bodega, B., Porta, C., Crosignani, P. G., Ginelli, E. & Marozzi, a. Mutations in the coding region of the FOXL2 gene are not a major cause of idiopathic premature ovarian failure. Mol. Hum. Reprod. 10, 555–7 (2004). 105. Ni, F. et al. Mutation analysis of FOXL2 gene in Chinese patients with premature ovarian failure. Gynecol. Endocrinol. 26, 246–9 (2010). 106. De Baere, E. et al. FOXL2 mutation screening in a large panel of POF patients and XX males. J. Med. Genet. 7–10 (2002). 107. Méduri, G. et al. FOXL2 mutations lead to different ovarian phenotypes in BPES patients: Case Report. Hum. Reprod. 25, 235–243 (2010). 108. Lin, L. & Achermann, J. C. Steroidogenic Factor-1 (SF-1, Ad4BP, NR5A1) and Disorders of Testis Development. Sex 2, 200–209 (2009). 109. Lourenco, D. et al. Mutations in NR5A1 associated with ovarian insufficiency. N. Engl. J. Med. 360, 1200–1210 (2009). 110. Watson, J. D. & Crick, F. H. C. Molecular structure of nucleic acids. Nature 171, 737–738 (1953). 111. Chargaff, E. Chemical specificity of nucleic acids and mechanism of their enzymatic degradation. Experientia 6, 201–209 (1950). 112. The international Human Genome Sequencing Consortium. Initial sequencing and analysis of the human genome. 409, (2001). 113. Consortsium, I. H. G. S. Finishing the euchromatic sequence of the human genome. Nature 50, 162–168 (2005). 114. Strachan, T. & Read, A. Human Molecular Genetics. (Garland Science, 2011). 115. Consortium", T. E. project. An integrated encyclopedia of DNA elements in the human genome. Nature 489, 57–74 (2012). 116. Maston, G. a, Evans, S. K. & Green, M. R. Transcriptional regulatory elements in the human genome. Annu. Rev. Genomics Hum. Genet. 7, 29–59 (2006). 117. Richards, S. et al. Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology. Genet. Med. 17, 405–423 (2015). 118. Kleinjan, D.-J. & Coutinho, P. Cis-ruption mechanisms: disruption of cis-regulatory control as a cause of human genetic disease. Brief. Funct. Genomic. Proteomic. 8, 317–32 (2009). 119. Zhang, F. & Lupski, J. R. Non-coding genetic variants in human disease. Hum. Mol. Genet. 24, R102–R110 (2015).

120. Spielmann, M. & Mundlos, S. Looking beyond the genes: the role of non-coding variants in human disease. Hum. Mol. Genet. 0, ddw205 (2016). 121. Taft, R., Pang, K., Mercer, T., Dinger, M. & Mattick, J. Non-coding RNAs: regulators of disease. J. Pathol. 220, 114–125 (2010). 122. Berdasco, M. & Esteller, M. Genetic syndromes caused by mutations in epigenetic genes. Hum. Genet. 132, 359–383 (2013). 123. Schenkel, L. C., Rodenhiser, D., Siu, V., Ainsworth, P. & Sadikovic, B. Constitutional Epi / Genetic Conditions : Genetic , Epigenetic , and Environmental Factors. J. Pediatr. Genet. 6, 30–41 (2017). 124. Baetens, D., Mendonça, B. B., Verdin, H., Cools, M. & Baere, E. De. Non-coding variation in Disorders of Sex Development. Clin. Genet. (2016). doi:10.1111/cge.12911 125. Wilhelm, D. Non-coding RNA and the Reproductive System. (2016). 126. Wilhelm, D., Palmer, S. & Koopman, P. Sex determination and gonadal development in mammals. Physiol. Rev. 87, 1–28 (2007). 127. Birk, O. S. et al. The LIM homeobox gene Lhx9 is essential for mouse gonad formation. Nature 403, 909–913 (2000). 128. Wilhelm, D. & Englert, C. The Wilms tumor suppressor Wt1 regulates early gonad development by activation of Sf1. Genes Dev. 16, 1839–1851 (2002). 129. Miyamoto, N., Yoshida, M., Kuratani, S., Matsuo, I. & Aizawa, S. Defects of urogenital development in mice lacking Emx2. Development 124, 1653–64 (1997). 130. Noonan, F. C., Mutch, D. G., Mallon, M. A. & Goodfellow, P. J. Characterization of the homeodomain gene EMX2: sequence conservation, expression analysis, and a search for mutations in endometrial cancers. Genomics 76, 37–44 (2001). 131. Katoh-fukui, Y., Tsuchiya, R., Shiroishi, T., Nakahara, Y. & Hashimoto, N. Male-to-female sex reversal in M33 mutant mice. Nature 393, 688–692 (1998). 132. Katoh-Fukui, Y. et al. Mouse Polycomb M33 is required for splenic vascular and adrenal gland formation through regulating Ad4BP/SF1 expression. Blood 106, 1612–1620 (2005). 133. Biason-Lauber, A., Konrad, D., Meyer, M., deBeaufort, C. & Schoenle, E. J. Ovaries and Female Phenotype in a Girl with 46,XY Karyotype and Mutations in the CBX2 Gene. Am. J. Hum. Genet. 84, 658–663 (2009). 134. Hatano, O., Takakusu, A., Nomura, M. & Morohashi, K. Identical origin of adrenal cortex and gonad revealed by expression profiles of Ad4BP / SF-1. Genes to Cells 1, 663–671 (1996). 135. Moore, A. W., McInnes, L., Kreidberg, J., Hastie, N. D. & Schedl, A. YAC complementation shows a requirement for Wt1 in the development of epicardium, adrenal gland and throughout nephrogenesis. Development 126, 1845–1857 (1999). 136. Mayo, M., Wang, C. & Drouin, S. WT1 influences apoptosis through transcriptional regulation of Bcl-2 family members. Cell Cycle 5, 1249–1253 (2006). 137. Ingraham, H. A. et al. The nuclear receptor steroidogenic factor 1 acts at multiple levels of the reproductive axis. Genes Dev. 8, 2302–2312 (1994). 138. Sadovsky, Y. et al. Mice deficient in the orphan receptor steroidogenic factor 1 lack adrenal glands and gonads but express P450 side-chain-cleavage enzyme in the placenta and have normal embryonic serum levels of corticosteroids. PNAS 92, 10939–10943 (1995). 139. Sinclair, A. et al. A gene from the human sex-determining region encodes a protein with homology to a conserved DNA-binding motif. Nature 374, 685–689 (1990). 140. Hacker, A., Capel, B., Goodfellow, P. & Lovell-Badge, R. Expression of [i]Sry[/i], the mouse sex determining gene. Development 121, 1603–1614 (1995). 141. Bullejos, M. & Koopman, P. Spatially Dynamic Expression of. Int. J. Dev. Biol. 205, 201–205 (2001). 142. Hanley, N. . et al. SRY, SOX9, and DAX1 expression patterns during human sex determination and gonadal development. Mech. Dev. 91, 403–407 (2000). 143. Cohen, D. R., Sinclair, A. & McGovern, J. SRY Protein Enhances Transcription of Fos-Related Antigen 1 Promoter Constructs. Proc. Natl. Acad. Sci. 91, 4372–4376 (1994). 144. McElreavey, K., Vilain, E., Abbas, N., Herskowitz, I. & Fellous, M. A regulatory cascade

hypothesis for mammalian sex determination: SRY represses a negative regulator of male development. Proc. Natl. Acad. Sci. U. S. A. 90, 3368–72 (1993). 145. Jäger, R. J., Harley, V. R., Pfeiffer, R. A., Goodfellow, P. N. & Scherer, G. A familial mutation in the testis-determining gene SRY shared by both sexes. Hum. Genet. 90, 350–355 (1992). 146. Cameron, F. J. & Sinclair, A. H. Mutations in SRY and SOX9: testis determining genes. Hum. Mutat. 9, 388–395 (1997). 147. Maier, E. M., Leitner, C., Löhrs, U. & Kuhnle, U. True hermaphroditism in an XY individual due to a familial point mutation of the SRY gene. J. Pediatr. Endocrinol. Metab. 16, 575–80 (2003). 148. Margarit, E. et al. SRY gene transferred to the long arm of the X- chromosome in a Y- positive XX true hermaphrodite. Am. J. Med. Genet. 90, 25–28 (2000). 149. Braun, A. et al. True hermaphroditism in a 46,XY individual, caused by a postzygotic somatic point mutation in the male gonadal sex-determining locus (SRY): molecular genetics and histological findings in a sporadic case. Am. J. Hum. Genet. 52, 578–85 (1993). 150. McElreavey, K. et al. Loss of sequences 3’ to the testis-determining gene, SRY, including the Y pseudoautosomal boundary associated with partial testicular determination. Proc. Natl. Acad. Sci. U. S. A. 93, 8590–8594 (1996). 151. Kwok, C. et al. Mutation analysis of the 2 kb 5’ to SRY in XY females and XY intersex subjects. J. Med. Genet. 33, 465–468 (1996). 152. Poulat, F. et al. Mutation in the 5’ noncoding region of the SRY gene in an XY sex-reversed patient. Hum. Mutat. Suppl. 1, S192-4 (1998). 153. Bullejos, M. & Koopman, P. Delayed Sry and Sox9 expression in developing mouse gonads underlies B6-YDOM sex reversal. Dev. Biol. 278, 473–481 (2005). 154. Nagamine, C. M., Morohashi, K., Carlisle, C. & Chang, D. K. Sex Reversal Caused by Mus musculus domesticus Y Chromosomes Linked to Variant Expression of the Testis-Determining Gene Sry. Dev. Biol. 216, 182–94 (1999). 155. Kennedy, D., Ramsdale, T., Mattick, J. & Little, M. An RNA recognition motif in Wilms’ tumour protein (WT1) revealed by structural modelling. Nat Genet. 14, 353–6 (1996). 156. Caricasole, A. et al. RNA binding by the Wilms tumor suppressor zinc finger proteins. Proc. Natl. Acad. Sci. U. S. A. 93, 7562–6 (1996). 157. Zaytouni, T., Efimenko, E. E. & Tevosian, S. G. GATA transcription factors in the developing reproductive system. Advances in genetics 76, (Elsevier Inc., 2011). 158. Ostrer, H. Disorders of Sex Development (DSDs): An Update. J. Clin. Endocrinol. Metab. 99, 1503– 9 (2014). 159. Hammes, A. et al. Two splice variants of the wilms’ tumor 1 gene have distinct functions during sex determination and nephron formation. Cell 106, 319–329 (2001). 160. Tevosian, S. G. et al. Gonadal differentiation, sex determination and normal Sry expression in mice require direct interaction between transcription partners GATA4 and FOG2. Development 129, 4627–34 (2002). 161. Molkentin, J. D., Lin, Q., Duncan, S. A. & Olson, E. N. Requirement of the transcription factor GATA4 for heart tube formation and ventral morphogenesis. Genes Dev. 11, 1061–1072 (1997). 162. Crispino, J. D. et al. Proper coronary vascular development and heart morphogenesis depend on interaction of GATA-4 with FOG cofactors. Genes Dev. 15, 839–44 (2001). 163. Tevosian, S. G. et al. FOG-2, a cofactor for GATA transcription factors, is essential for heart morphogenesis and development of coronary vessels from epicardium. Cell 101, 729–39 (2000). 164. Garg, V. et al. GATA4 mutations cause human congenital heart defects and reveal an interaction with TBX5. 21, 443–447 (2003). 165. Pizzuti, A. et al. Mutations of ZFPM2/FOG2 gene in sporadic cases of tetralogy of Fallot. Hum. Mutat. 22, 372–7 (2003). 166. Finelli, P. et al. Disruption of friend of GATA 2 gene (FOG-2) by a de novo t(8;10) chromosomal translocation is associated with heart defects and gonadal dysgenesis. Clin. Genet. 71, 195–204 (2007).

167. Lourenço, D. et al. Loss-of-function mutation in GATA4 causes anomalies of human testicular development. Proc. Natl. Acad. Sci. U. S. A. 108, 1597–602 (2011). 168. Bashamboo, A. et al. Mutations in the FOG2/ZFPM2 gene are associated with anomalies of human testis determination. Hum. Mol. Genet. 1–9 (2014). doi:10.1093/hmg/ddu074 169. Bogani, D. et al. Loss of mitogen-activated protein kinase kinase kinase 4 (MAP3K4) reveals a requirement for MAPK signalling in mouse sex determination. PLoS Biol. 7, e1000196 (2009). 170. Warr, N. et al. Gadd45?? and Map3k4 Interactions Regulate Mouse Testis Determination via p38 MAPK-Mediated Control of Sry Expression. Dev. Cell 23, 1020–1031 (2012). 171. Gierl, M. S., Gruhn, W. H., von Seggern, A., Maltry, N. & Niehrs, C. GADD45G Functions in Male Sex Determination by Promoting p38 Signaling and Sry Expression. Dev. Cell 23, 1032–1042 (2012). 172. Ono, M. & Harley, V. R. Disorders of sex development: new genes, new concepts. Nat. Rev. Endocrinol. 9, 79–91 (2013). 173. Hanley, N. A. et al. Expression of steroidogenic factor 1 and Wilms’ tumour 1 during early human gonadal development and sex determination. Mech. Dev. 87, 175–180 (1999). 174. Sekido, R. & Lovell-Badge, R. Sex determination involves synergistic action of SRY and SF1 on a specific Sox9 enhancer. Nature 453, 930–4 (2008). 175. De Santa Barbara, P. et al. Direct interaction of SRY-related protein SOX9 and steroidogenic factor 1 regulates transcription of the human anti-Müllerian hormone gene. Mol. Cell. Biol. 18, 6653–65 (1998). 176. Parker, K. L. & Schimmer, B. P. Steroidogenic Factor 1 : A Key Determinant of Endocrine Development and Function. Endocr. Rev. 18, 361–377 (1997). 177. Hoivik, E. a, Lewis, A. E., Aumo, L. & Bakke, M. Molecular aspects of steroidogenic factor 1 (SF- 1). Mol. Cell. Endocrinol. 315, 27–39 (2010). 178. Zhao, L. et al. Steroidogenic factor 1 ( SF1 ) is essential for pituitary gonadotrope function. Development 154, 147–154 (2006). 179. Zangen, D. et al. Testicular differentiation factor SF-1 is required for human spleen development. J. Clin. Invest. 124, 2071–2075 (2014). 180. Colson, C. et al. Short report: SF1 and spleen development: new heterozygous mutation, literature review and consequences for NR5A1 - mutated-patient’s management. Clin. Genet. (2016). doi:10.1111/cge.12957 181. Luo, X., Ikeda, Y. & Parker, K. L. A cell-specific nuclear receptor is essential for adrenal and gonadal development and sexual differentiation. Cell 77, 481–490 (1994). 182. Shinoda, K. et al. Developmental Defects of the Ventromedial Hypothalamic Nucleus and Pituitary Gonadotroph in the Ftz-Fl Disrupted Mice. Dev. Dyn. 9, 22–29 (1995). 183. Bland, M. L., Fowkes, R. C. & Ingraham, H. A. Differential Requirement for Steroidogenic Factor- 1 Gene Dosage in Adrenal Development Versus Endocrine Function. Mol. Endocrinol. 18, 941–952 (2004). 184. Lin, L. et al. UKPMC Funders Group Author Manuscript Ad4BP , NR5A1 ) are associated with 46 , XY disorders of sex development with normal adrenal function. 92, 991–999 (2007). 185. Köhler, B. et al. Five Novel Mutations in Steroidogenic With Severe Underandrogenization But Without Adrenal Insufficiency. Hum. Mutat. 29, 59–64 (2008). 186. Cools, M. et al. Pubertal androgenization and gonadal histology in two 46,XY adolescents with NR5A1 mutations and predominantly female phenotype at birth. Eur. J. Endocrinol. 166, 341–9 (2012). 187. Baetens, D. et al. NR5A1 is a novel disease gene for 46,XX testicular and ovotesticular disorders of sex development. Genet. Med. (2016). doi:10.1038/gim.2016.118 188. Bashamboo, A. et al. A recurrent p.Arg92Trp variant in steroidogenic factor-1 (NR5A1) can act as a molecular switch in human sex development. Hum. Mol. Genet. 25, 3446–3453 (2016). 189. Igarashi, M. et al. Identical NR5A1 Missense Mutations in Two Unrelated 46,XX Individuals with Testicular Tissues. Hum. Mutat. 1–4 (2016). doi:10.1002/humu.23116

190. Morais Da Silva, S. et al. Sox9 expression during gonadal development implies a conserved role for the gene in testis differentiation in mammals and birds. Nat Genet. 14, 353–6 (1996). 191. Barrionuevo, F. J., Burgos, M., Scherer, G. & Jiménez, R. Genes promoting and disturbing testis development. Histol. Histopathol. 27, 1361–1383 (2012). 192. Bishop, C. E. et al. A transgenic insertion upstream of sox9 is associated with dominant XX sex reversal in the mouse. Nat. Genet. 26, 490–494 (2000). 193. Kim, Y. et al. Fgf9 and Wnt4 act as antagonistic signals to regulate mammalian sex determination. PLoS Biol. 4, e187 (2006). 194. Bagheri-Fam, S. et al. Loss of Fgfr2 leads to partial XY sex reversal. Dev. Biol. 314, 71–83 (2008). 195. Colvin, J. S., Green, R. P., Schmahl, J., Capel, B. & Ornitz, D. M. Male-to-female sex reversal in mice lacking fibroblast growth factor 9. Cell 104, 875–889 (2001). 196. Wilhelm, D. et al. Sertoli cell differentiation is induced both cell-autonomously and through prostaglandin signaling during mammalian sex determination. Dev. Biol. 287, 111–124 (2005). 197. Wilhelm, D. et al. SOX9 regulates prostaglandin D synthase gene transcription in vivo to ensure testis development. J. Biol. Chem. 282, 10553–10560 (2007). 198. Shen, J. H.-C. & Ingraham, H. a. Regulation of the orphan nuclear receptor steroidogenic factor 1 by Sox proteins. Mol. Endocrinol. 16, 529–540 (2002). 199. Arango, N. A., Lovell-Badge, R. & Behringer, R. R. Targeted Mutagenesis of the Endogenous Mouse Mis Gene Promoter: In Vivo Definition of Genetic Pathways of Vertebrate Sexual Development. Cell 99, 409–419 (1999). 200. Wilson, M. J., Jeyasuria, P., Parker, K. L. & Koopman, P. The transcription factors steroidogenic factor-1 and SOX9 regulate expression of Vanin-1 during mouse testis development. J. Biol. Chem. 280, 5917–5923 (2005). 201. Kurth, I. et al. Duplications of noncoding elements 5’ of SOX9 are associated with brachydactyly- anonychia. Nat. Genet. 41, 862–863 (2009). 202. Benko, S. et al. Highly conserved non-coding elements on either side of SOX9 associated with Pierre Robin sequence. Nat. Genet. 41, 359–364 (2009). 203. Jakobsen, L. P. et al. Pierre Robin sequence may be caused by dysregulation of SOX9 and KCNJ2. J. Med. Genet. 44, 381–6 (2007). 204. Haines, B. et al. Interchromosomal insertional translocation at Xq26.3 alters SOX3 expression in an individual with XX male sex reversal. J. Clin. Endocrinol. Metab. 100, E815-20 (2015). 205. Polanco, J. C., Wilhelm, D., Davidson, T.-L., Knight, D. & Koopman, P. Sox10 gain-of-function causes XX sex reversal in mice: implications for human 22q-linked disorders of sex development. Hum. Mol. Genet. 19, 506–16 (2010). 206. Chaboissier, M.-C. et al. Functional analysis of Sox8 and Sox9 during sex determination in the mouse. Development 131, 1891–1901 (2004). 207. Yao, H. H.-C. The pathway to femaleness: current knowledge on embryonic development of the ovary. Mol. Cell. Endocrinol. 230, 87–93 (2005). 208. Eicher, E. M. & Washburn, L. L. Genetic control of primary sex determination in mice. Annu Rev Genet 20, 327–360 (1986). 209. Bernard, P. & Harley, V. R. Wnt4 action in gonadal development and sex determination. Int. J. Biochem. Cell Biol. 39, 31–43 (2007). 210. Vainio, S., Heikkilä, M., Kispert, a, Chin, N. & McMahon, a P. Female development in mammals is regulated by Wnt-4 signalling. Nature 397, 405–9 (1999). 211. Heikkilä, M. et al. The partial female-to-male sex reversal in Wnt-4 deficient females involves induced expression of testosterone biosynthetic genes and testosterone production and depends on androgen action. Endocrinology 146, 4016–4023 (2005). 212. Correa, S. M. et al. Sex reversal in C57BL/6J XY mice caused by increased expression of ovarian genes and insufficient activation of the testis determining pathway. PLoS Genet. 8, (2012). 213. Philibert, P. et al. Identification and functional analysis of a new WNT4 gene mutation among 28 adolescent girls with primary amenorrhea and Müllerian duct abnormalities: A French

collaborative study. J. Clin. Endocrinol. Metab. 93, 895–900 (2008). 214. Biason-Lauber, A. et al. WNT4 deficiency-a clinical phenotype distinct from the classic Mayer- Rokitansky-Kuster-Hauser syndrome: A case report. Hum. Reprod. 22, 224–229 (2007). 215. Philibert, P. et al. Molecular analysis of WNT4 gene in four adolescent girls with mullerian duct abnormality and hyperandrogenism (atypical Mayer-Rokitansky-K??ster- Hauser syndrome). Fertil. Steril. 95, 2683–2686 (2011). 216. Chassot, A. A. et al. Activation of ??-catenin signaling by Rspo1 controls differentiation of the mammalian ovary. Hum. Mol. Genet. 17, 1264–1277 (2008). 217. Manuylov, N. L., Smagulova, F. O., Leach, L. & Tevosian, S. G. Ovarian development in mice requires the GATA4-FOG2 transcription complex. Development 135, 3731–43 (2008). 218. Maatouk, D. M. et al. Stabilization of beta-catenin in XY gonads causes male-to-female sex- reversal. Hum. Mol. Genet. 17, 2949–55 (2008). 219. Pailhoux, E. et al. A 11.7-kb deletion triggers intersexuality and polledness in goats. Nat. Genet. 29, 453–8 (2001). 220. Schmidt, D. et al. The murine winged-helix transcription factor Foxl2 is required for granulosa cell differentiation and ovary maintenance. Development 131, 933–42 (2004). 221. Uda, M. et al. Foxl2 disruption causes mouse ovarian failure by pervasive blockage of follicle development. Hum. Mol. Genet. 13, 1171–81 (2004). 222. Ottolenghi, C. et al. Loss of Wnt4 and Foxl2 leads to female-to-male sex reversal extending to germ cells. Hum. Mol. Genet. 16, 2795–804 (2007). 223. Veitia, R. a. FOXL2 versus SOX9: a lifelong ‘battle of the sexes’. Bioessays 32, 375–80 (2010). 224. Ottolenghi, C. et al. Foxl2 is required for commitment to ovary differentiation. Hum. Mol. Genet. 14, 2053–62 (2005). 225. Matson, C. K. et al. DMRT1 prevents female reprogramming in the postnatal mammalian testis Supp. . Nature 9, 1–5 (2011). 226. Lavery, R. et al. Testicular Differentiation Occurs in Absence of R-spondin1 and Sox9 in Mouse Sex Reversals. PLoS Genet. 8, (2012). 227. Jameson, S. A., Lin, Y. T. & Capel, B. Testis development requires the repression of Wnt4 by Fgf signaling. Dev. Biol. 370, 24–32 (2012). 228. Brain, C. E. et al. Holistic management of DSD. Best Pract. Res. Clin. Endocrinol. Metab. 24, 335– 354 (2010). 229. Prader, A. Genital findings in the female pseudo-hermaphroditism of the congenital adrenogenital syndrome; morphology, frequency, development and heredity of the different genital forms. Helv. Paediatr. Acta 9, 231–238 (1954). 230. Ahmed, S. F., Khwaja, O. & Hughes, I. a. The role of a clinical score in the assessment of ambiguous genitalia. BJU Int. 85, 120–4 (2000). 231. Kolesinska, Z. et al. Changes Over Time in Sex Assignment for Disorders of Sex Development. Pediatrics 134, E710-15 (2014). 232. Ahmed, S. F. et al. Society for Endocrinology UK guidance on the initial evaluation of an infant or an adolescent with a suspected disorder of sex development (Revised 2015). Clin. Endocrinol. (Oxf). 84, 771–788 (2016). 233. Josso, N., Rey, R. A. & Picard, J. Y. Anti-Müllerian hormone: A valuable addition to the toolbox of the pediatric endocrinologist. Int. J. Endocrinol. 2013, (2013). 234. Kubini, K. et al. Basal inhibin B and the testosterone response to human chorionic gonadotropin correlate in prepubertal boys. J. Clin. Endocrinol. Metab. 85, 134–138 (2000). 235. Kulle, A. E., Welzel, M., Holterhus, P. M. & Riepe, F. G. Principles and clinical applications of liquid chromatography - Tandem mass spectrometry for the determination of adrenal and gonadal steroid hormones. J. Endocrinol. Invest. 34, 702–708 (2011). 236. Kulle, A. E., Welzel, M., Holterhus, P. M. & Riepe, F. G. Implementation of a liquid chromatography tandem mass spectrometry assay for eight adrenal C-21 steroids and pediatric reference data. Horm. Res. Paediatr. 79, 22–31 (2013).

237. Appari, M. et al. Apolipoprotein D (APOD) is a putative biomarker of androgen receptor function in androgen insensitivity syndrome. J. Mol. Med. 87, 623–632 (2009). 238. Chertin, B. et al. The use of laparoscopy in intersex patients. Pediatr. Surg. Int. 22, 405–408 (2006). 239. Wünsch, L., Holterhus, P. M., Wessel, L. & Hiort, O. Patients with disorders of sex development (DSD) at risk of gonadal tumour development: Management based on laparoscopic biopsy and molecular diagnosis. BJU Int. 110, (2012). 240. Fisher, A. D. et al. Gender identity, gender assignment and reassignment in individuals with disorders of sex development: a major of dilemma. J. Endocrinol. Invest. 39, 1207–1224 (2016). 241. Sudai, M. Changing ethical and legal norms in the management of differences of sex development. Lancet Diabetes Endocrinol. 8587, 8–10 (2017). 242. Lucas-Herald, A. et al. The long-term outcome of boys with partial androgen insensitivity syndrome and a mutation in the androgen receptor gene. J. Clin. Endocrinol. Metab. 101, 3959– 3967 (2016). 243. Hughes, I. A. et al. Androgen insensitivity syndrome. Lancet 380, 1419–1428 (2012). 244. Cools, M. Germ cell cancer risk in DSD patients. Ann. Endocrinol. (Paris). 75, 67–71 (2014). 245. van der Zwan, Y. G., Biermann, K., Wolffenbuttel, K. P., Cools, M. & Looijenga, L. H. J. Gonadal maldevelopment as risk factor for germ cell cancer: towards a clinical decision model. Eur. Urol. 67, 692–701 (2015). 246. Berney, D. M. et al. Germ cell neoplasia in situ (GCNIS): evolution of the current nomenclature for testicular pre-invasive germ cell malignancy. Histopathology 69, 7–10 (2016). 247. Johnson, E. K. & Finlayson, C. Health Preservation of Fertility Potential for Gender and Sex Diverse Individuals. Transgender Heal. 1, 41–44 (2016). 248. Beale, J. M. & Creighton, S. M. Long-term health issues related to disorders or differences in sex development/intersex. Maturitas 94, 143–148 (2016). 249. Liao, L.-M., Wood, D. & Creighton, S. M. Parental choice on normalising cosmetic genital surgery. Bmj 5124, 2 (2015). 250. Cools, M. et al. Response to the Council of Europe Human Rights Commissioner’s Issue Paper on Human Rights and Intersex People. Eur. Urol. 70, 407–409 (2016). 251. Lee, P. A. et al. Global disorders of sex development update since 2006: Perceptions, approach and care. Horm. Res. Paediatr. 85, 158–180 (2016). 252. Mouriquand, P. D. E. et al. Surgery in disorders of sex development (DSD) with a gender issue: If (why), when, and how? J. Pediatr. Urol. 12, 139–149 (2016). 253. Wolffenbuttel, K. P. et al. Gonadal dysgenesis in disorders of sex development: Diagnosis and surgical management. J. Pediatr. Urol. 6–11 (2016). doi:10.1016/j.jpurol.2016.08.015 254. Baxter, R. M. et al. Exome sequencing for the diagnosis of 46, XY disorders of sex development. J. Clin. Endocrinol. Metab. 100, E333–E344 (2015). 255. Eggers, S. et al. Disorders of sex development: insights from targeted gene sequencing of a large international patient cohort. Genome Biol. 17, 243 (2016). 256. Sanger, F. & Coulson, A. R. A Rapid Method for Determining Sequences in DNA by Primed Synthesis with DNA Polymerase. J. Mol. Biol. 94, 441–448 (1975). 257. Tucker, T., Marra, M. & Friedman, J. M. Massively parallel sequencing: the next big thing in genetic medicine. Am. J. Hum. Genet. 85, 142–54 (2009). 258. Shendure, J. & Ji, H. Next-generation DNA sequencing. Nat. Biotechnol. 26, 1135–45 (2008). 259. Metzker, M. L. Sequencing technologies - the next generation. Nat. Rev. Genet. 11, 31–46 (2010). 260. Mardis, E. R. Next-generation DNA sequencing methods. Annu. Rev. Genomics Hum. Genet. 9, 387–402 (2008). 261. De Leeneer, K. et al. Flexible, scalable, and efficient targeted resequencing on a benchtop sequencer for variant detection in clinical practice. Hum. Mutat. 36, 379–87 (2015). 262. White, S. et al. Copy number variation in patients with disorders of sex development due to 46,XY gonadal dysgenesis. PLoS One 6, e17793 (2011). 263. Tannour-Louet, M. et al. Identification of de novo copy number variants associated with human

disorders of sexual development. PLoS One 5, e15392 (2010). 264. MLPA® - an introduction. Available at: https://www.mlpa.com/WebForms/WebFormMain.aspx?Tag=_wl2zCji- rCGANQgZPuTixsEyIW1MscfzuKj2NDFYc-g. 265. Bamshad, M. J. et al. Exome sequencing as a tool for Mendelian disease gene discovery. Nat. Rev. Genet. 12, 745–55 (2011). 266. Lander, E. S. & Botstein, D. Homozygosity mapping: a way to map human recessive traits with the DNA of inbred children. Science (80-. ). 236, 1567–1570 (1987). 267. Alkuraya, F. S. The application of next-generation sequencing in the autozygosity mapping of human recessive diseases. Hum. Genet. (2013). doi:10.1007/s00439-013-1344-x 268. Purcell, S. et al. PLINK: a tool set for whole-genome association and population-based linkage analyses. Am. J. Hum. Genet. 81, 559–75 (2007). 269. Sante, T. et al. ViVar: A comprehensive platform for the analysis and visualization of structural genomic variation. PLoS One 9, 1–12 (2014). 270. Phillips, N. B. et al. Mammalian testis-determining factor SRY and the enigma of inherited human sex reversal: Frustrated induced fit in a bent protein-DNA complex. J. Biol. Chem. 286, 36787–36807 (2011). 271. Slaney, S. F., Chalmers, I. J., Affara, N. A. & Chitty, L. S. An autosomal or X linked mutation results in true hermaphrodites and 46 , XX males in the same family. J Med Genet 35, 17–22 (1998). 272. Quinonez, S. C. et al. 9P Partial Monosomy and Disorders of Sex Development: Review and Postulation of a Pathogenetic Mechanism. Am. J. Med. Genet. A 161A, 1882–96 (2013).

CHAPTER 2: OUTLINE & OBJECTIVES

CHAPTER 2:

Outline &

91 Objectives

Given the complexity of the many underlying molecular pathways it is not surprising that disorders/differences of sex development (DSD) are among the most common birth defects. Severe forms of DSD have a prevalence of ~1/2 000 – 4 500 births but milder genital variations, like hypospadias or micropenis, are seen in as much as 1/200 births. Despite significant progress in the identification of new genes and pathways, a

(likely) molecular diagnosis is only obtained in ~50% of 46,XY cases. Obtaining a molecular diagnosis enables refining of a clinical diagnosis and can in some cases facilitate clinical decision making with respect to sex assignment, gonadal tumor risk and lifelong care. For affected individuals and families, understanding the etiology of their condition can contribute to a better acceptance and it opens perspectives for genetic counseling and eventually reproductive options. In addition, unraveling of the molecular pathways behind these rare phenotypes can greatly enhance translational research (e.g. with respect to associated anomalies, germ cell cancer risk or outcome) and provide insights in human-typical gonadal development and the pathogenesis of more frequent conditions like male infertility and primary ovarian insufficiency (POI).

The aim of this doctoral project was to unravel pieces of the molecular pathways behind the sex developmental processes and to identify new genes implicated in DSD and/or related conditions such as POI.

Objective 1: Assessing the appropriateness and diagnostic yield of a standardized work-up in a large consecutive cohort of 46,XY neonates and infants with atypical gonadal development

From a clinical point of view and given that DSD (in the sense of Differences of Sex

Development) represent a spectrum of genital and gonadal atypicalities, it is not always clear which individuals with a variant sex development need to be referred for extensive molecular genetic diagnostic work-up and if persons with less severe phenotypes like isolated hypospadias or micropenis may benefit from such a referral.

Milder genital variations have been attributed to environmental factors, low birth weight and multiple gene polymorphisms rather than to single gene mutations.

However, there have been reports associating these milder phenotypes to mutations in genes known to cause genital ambiguity in other cases. In study 1, we aimed to assess the appropriateness of a standardized work-up in a large series of 46,XY cases with atypical gonadal development. Our screening approach involved structured clinical examination, biochemical assessment and genetic work-up consisting of consecutive Sanger sequencing of selected DSD genes and copy number analysis using array-comparative genomic hybridization (array-CGH) and MLPA. Results were interpreted in the light of clinical and hormonal findings.

Objective 2: Identification of novel disease genes underlying 46,XY and 46,XX disorders of sex development

The main aim of this PhD project was to identify novel disease genes and mechanisms for both 46,XX and 46,XY DSD by use of whole exome sequencing (WES). In case of known consanguinity WES was combined with homozygosity mapping, if this was not the case a trio-sequencing strategy was applied. This approach led to the identification of ESR2 as a novel candidate disease gene for 46,XY DSD and the identification of a recurrent NR5A1 mutation in cases with 46,XX DSD.

In study 2, the identification of three ESR2 mutations is reported in three unrelated cases with syndromal and isolated 46,XY DSD. Functional characterization of these variants was done using transcriptional activation assays. Structural analysis provided additional evidence for functional consequences on the protein level. Furthermore, the expression of ER-β was assessed in an 8-week old male fetus.

In study 3, we implicated a recurrent NR5A1 mutation in the pathogenicity of 46,XX testicular and ovotesticular DSD. Immunohistochemistry on patient-derived gonadal tissue showed SRY-independent SOX9 expression in the testicular regions of the gonads and FOXL2 expression in the ovarian parts. Transcriptional activity and subcellular localization experiments showed no consistent changes however.

Therefore, we proposed a theoretical working hypothesis for this mutation, which we aim to explore in future research.

Objective 3: To explore the concept that DSD genes may underlie a spectrum of reproductive disorders in males and females: identification of a novel disease gene for POI starting from a case of 46,XY DSD

Insights in the pathogenesis of rare conditions such as DSD can extend our knowledge on the mechanisms underlying more common reproductive disorders like POI. For instance, NR5A1 mutations were first identified in 46,XY DSD cases, but later implicated in POI and male infertility. In study 4, we report on the identification of

FOG2/ZFPM2 mutations, previously implicated in the pathogenesis of 46,XY DSD, in a large cohort of POI women. First, a maternally inherited FOG2 deletion was identified in a boy with isolated 46,XY DSD. AMH assessment in the mother suggested a decreased ovarian follicle pool, leading to the hypothesis that FOG2 is a disease gene for POI, which is supported by literature data. Targeted resequencing of FOG2 in a large POI cohort revealed four heterozygous variants with likely pathogenic effect

(4/388, 1%), for which we assessed transcriptional activity using luciferase experiments.

Objective 4: To explore the role of non-coding genetic variation as a mechanism for DSD

Despite the identification of novel DSD disease genes, the majority of cases remains unsolved, suggesting the involvement of other mechanisms in the pathogenesis of

DSD. Over the years several reports on non-coding variation in DSD have been published. Study 5 provides a review of these findings and the different action mechanisms involved.

CHAPTER 3: 46,XY DSD CHAPTER 3:

46,XY DSD

EXTENSIVE CLINICAL, HORMONAL AND GENETIC SCREENING IN A EXTENSIVE CLINICAL, HORMONAL AND GENETIC SCREENING IN A LARGE LARGE CONSECUTIVE SERIES OF 46,XY NEONATES AND INFANTS CONSECUTIVE SERIES OF 46,XY NEONATES AND INFANTS WITH ATYPICAL WITH ATYPICAL SEXUAL DEVELOPMENT SEXUAL DEVELOPMENT

Orphanet Journal of Rare Disease (2014), 9; 209

Dorien Baetens1, Wilhelm Mladenov2,3, Barbara Delle Chiaie1, Björn Menten1, An

Desloovere2, Violeta Iotova3, Bert Callewaert1, Erik Van Laecke4, Piet Hoebeke4, Elfride

De Baere1 and Martine Cools2

1 Center for Medical Genetics Ghent, Ghent University and Ghent University Hospital, Ghent, Belgium

2Department of Pediatric Endocrinology, Ghent University and Ghent University Hospital, Ghent,

Belgium

3Department of Pediatrics and Medical Genetics, Medical University of Varna, University Hospital

“Sveta Marina”, Varna, Bulgaria

4Department of Urology, Ghent University and Ghent University Hospital, Ghent, Belgium

100

ABSTRACT

Background: One in 4 500 children is born with ambiguous genitalia, milder phenotypes occur in one in 300 newborns. Conventional time-consuming hormonal and genetic work-up provides a genetic diagnosis in around 20-40% of 46,XY cases with ambiguous genitalia. All others remain without a definitive diagnosis. The investigation of milder cases, as suggested by recent reports remains controversial.

Methods: Integrated clinical, hormonal and genetic screening was performed in a sequential series of 46, XY children, sex-assigned male, who were referred to our pediatric endocrine service for atypical genitalia (2007–2013).

Results: A consecutive cohort of undervirilized 46,XY children with external masculinization score (EMS) 2–12, was extensively investigated. In four patients, a clinical diagnosis of Kallmann syndrome or Mowat-Wilson syndrome was made and genetically supported in 2/3 and 1/1 cases respectively. Hormonal data were suggestive of a (dihydro)testosterone biosynthesis disorder in four cases, however no

HSD17B3 or SRD5A2 mutations were found. Array-CGH revealed a causal structural variation in 2/6 syndromic patients. In addition, three novel NR5A1 mutations were found in non-syndromic patients. Interestingly, one mutation was present in a fertile male, underlining the inter- and intrafamilial phenotypic variability of NR5A1- associated phenotypes. No AR, SRY or WT1 mutations were identified.

Conclusion: Overall, a genetic diagnosis could be established in 19% of non- syndromic and 33% of syndromic cases. There is no difference in diagnostic yield between patients with more or less pronounced phenotypes, as expressed by the external masculinisation score (EMS). The clinical utility of array-CGH is high in syndromic cases. Finally, a sequential gene-by-gene approach is time-consuming, expensive and inefficient. Given the low yield and high expense of Sanger sequencing, we anticipate that massively parallel sequencing of gene panels and whole exome

101 sequencing hold promise for genetic diagnosis of 46,XY DSD boys with an undervirilized phenotype.

Keywords: 46,XY DSD, Undervirilization, Integrated screening, Diagnosis, Array-

CGH, MLPA, NR5A1 mutations

102

BACKGROUND Receptor (AR), Nuclear Receptor Subfamily

The birth of a child with ambiguous 5 Group A Member 1 (NR5A1) and Wilms genitalia is a rare event with a Tumor 1 (WT1) genes - classically prevalence of one in 4 500 live births and associated with genital ambiguity or poses a challenge to the parents and the more severe forms of undervirilization - medical team1. Specialized have recently been identified in cases multidisciplinary medical care, aiming with isolated proximal or even distal at addressing concerns and hypospadias, combined cryptorchidism uncertainties regarding gender and (distal) hypospadias or anorchia, assignment, underlying etiology and and sequencing of these genes has been 9-14 management, as well as providing advocated in such cases . On the other adequate psychological support is hand, copy number variations in genes essential2. Extensive and time- involved in the process of sexual consuming hormonal and genetic work- development have effectively been up provides a genetic diagnosis in 20- detected by whole genome (array

40% of cases3,4. Less pronounced comparative genomic hybridization, atypical development of male external array-CGH) or targeted (multiplex genitalia is more prevalent and is ligation-dependent probe amplification, 15-17 noticed in the newborn period in around MLPA) copy number analysis , and one in 300 males; 75% of cases are both techniques have become widely associated with hypospadias5. These available in recent years. However, milder forms of undervirilization, such whether a systematic extensive genetic as isolated or combined cryptorchidism work-up is indicated in the 46,XY and hypospadias have been related to newborn with a milder degree of environmental factors, low birth weight undervirilization, as indicated by a and multiple gene polymorphisms higher Prader or External rather than single gene mutations6-8. Masculinization Score (EMS) remains a 18 However, mutations in the Androgen matter of debate . Current screening methods are time consuming and have a 103 low efficiency. The introduction of dysgenesis, and sequencing of genome-wide technologies such as Hydroxysteroid 17-β Dehydrogenase whole exome sequencing (WES) holds (HSD17B3) or Steroid-5-α-Reductase, promise for future clinical decision (SRD5A2) was performed in cases with making in a routine diagnostic setting suspicion of a (dihydro)testosterone for these rare, genetically heterogeneous biosynthesis defect. Results were conditions. In order to gain insight in the interpreted in the light of clinical and appropriateness and diagnostic yield of hormonal findings. a systematic genetic work-up in 46,XY RESULTS infants with atypical external genitalia, Clinical investigation we performed a standardized genetic screening panel in all 46,XY neonates Consanguinity was present or suspected and infants who were referred to our in 4/32 cases (12.5%). Another four cases pediatric endocrine service for atypical had a family history of subfertility or male or ambiguous genitalia in the atypical genitalia. Nine children (28.1%) period 2007–2013 and who received were born small for gestational age male sex assignment. This screening (SGA), defined as a BW< −2 Standard consisted of consecutive Sanger Deviation (SD) for gestational age, with sequencing of the AR, NR5A1 and WT1 a mean BW of −2.8 SD; mean BW of genes, high-resolution (180 K) array- children born appropriate for GA was

CGH and a commercially available −0.36 SD. EMS scores ranged from 2/12

MLPA kit with probes for Sex to 12/12. In 6/32 children (18.7 %)

Determining Region Y (SRY), SRY-box 9 dysmorphic features were noticed.

(SOX9), Nuclear Receptor Subfamily 0 Patient details are represented in Table

Group B Member 1 (NR0B1), Wingless type 1). Three out of 32 patients (P26, P28,

4 (WNT4) and NR5A1. Additionally, P29) were diagnosed with KS based on sequencing of SRY was performed in clinical and hormonal data (day 14–90). cases with hormonal results consistent Patient 26 (EMS 12) was referred for an with the presence of (partial) gonadal atypically looking short penis (with 104 bilateral descended testes). At physical on the X chromosome including the examination, stretched penis length Kallmann syndrome 1 (KAL1) gene, as

(SPL) measured 30 mm, but his penis discussed below. Patient 29 (EMS 9) was was extremely thin and weak, diagnosed with KS based on the reminiscent of agenesis of the corpora presence of micropenis (SPL 21 mm) and cavernosa, which was excluded by a positive family history for KS: the

Magnetic Resonance Imaging (MRI) of father had been diagnosed with KS and the penile structures. Hormonal data was able to conceive following concordant with hypogonadotropic gonadotrophin therapy. Hormonal data hypogonadism (HoH) (Table 2) and confirmed HoH in the index patient. The

MRI revealing a hypoplastic bulbus diagnosis was supported genetically by olfactorius were both consistent with a the identification of a heterozygous diagnosis of Kallmann syndrome. An FGFR1 mutation, c.1042G>A (p.G348R), etiological diagnosis was sought by in both the patient and his father. This targeted resequencing of several known mutation has been described

KS genes (KAL1, FGFR1, FGF8, CHD7, previously273. Patient 30 was diagnosed

PROK2, PROKR2, HS6ST1, WDR11, with Mowat-Wilson syndrome (MWS),

SEMA3A, GNRH1, GNRHR, KISS1, he presented with typical external ear

KISS1R, TAC3 and TACR3); no causal abnormalities (Figure 1), hypotonia, mutations were identified. The second persistent ductus arteriosus, ventricular patient with KS (P28, EMS 8) presented septum defect, facial dysmorphism, with mild craniofacial dysmorphism Hirschsprung disease, penoscrotal

(ptosis, plagiocephaly), general inversion and hypospadias. MWS is hypotonia, developmental delay, caused by heterozygous de novo micropenis (SPL 15 mm) and bilateral mutations in ZEB2. Sequencing of this inguinal testes. Low gonadotrophins in gene revealed a heterozygous one association with a low AMH was basepair frameshift deletion, c.2856delG suggestive of HoH. Array-CGH (p.Arg953Glufs*24). revealed a causal hemizygous deletion 105

Table 1: Medical history and phenotypic details of patients

Code GA BW BW EMS Pregnancy Dysmorphic Consanguinity Family history

(weeks) (g) (SD) features

1 32 900 -3.15 2/12 CS IUGR No Unremarkable

2 41 3.260 -1.12 3/12 Normal Large ears No Unremarkable

Broad nose

Mild frontal

bossing

3 40 4.150 1.12 2,5/12 Normal / No Maternal aunt:

difficulties to

get pregnant,

one child with

congenital

abnormalities

4 41 3.060 -1.6 3/12 Normal / No Unremarkable

Minoxidil

treatment

5 34 1.320 -2.93 7,5/12 CS IUGR No Unremarkable

6 40 3.380 -0.59 9/12 Normal / No Grandfather

with

hypospadias,

maternal aunt

with POF

7 30 510 -3.8 6/12 Induced IUGR No Paternal

delivery grandmother:

Twins cleft lip

8 39 3.690 0.45 10/12 IVF / No Unremarkable

9 38 3.310 0.01 8/12 Normal / No Unremarkable

10 39 2.850 -1.53 9/12 Placental Microcepha No Unremarkable

infarction ly

106

Mild facial

dysmorphis

11 40 3.000 -1.51 3/12 Normal Macrocepha No Unremarkable

Facial

dysmorphis

Short neck

Developme

ntal delay

(speech)

12 40 3.850 0.47 6/12 Normal / Yes Mother: fertility

problems,

irregular

menses

13 35 2.530 -0.26 6/12 Preeclampsia / No Unremarkable

and

hypertension

Obesity

Twins

Sectio

14 32 945 -2.98 3/12 Bleeding / No Unremarkable

15 40 NA NA 3/12 Normal / No Unremarkable

16 28 860 -1.26 7/12 Eclampsy Atrial No Unremarkable

Prematurity septum

defect

17 34 1.450 -2.52 6/12 CS Ventricular Yes Cousin

septum (deceased)with

defect Jeune

syndrome

18 34 1.400 -2.68 2/12 CS / No Unremarkable

107

Antidepressa

nt Paroxetine

19 41 2.805 -2.24 3/12 Normal / No Unremarkable

20 39 2.880 -1.45 6/12 Normal / Possible Unremarkable

21 36 1.585 -3.39 7/12 Preeclampsia / No Unremarkable

22 39 2.640 -2.07 6/12 Normal / No Unremarkable

23 40 3.250 -0.9 8/12 Normal / No Unremarkable

24 37 3.290 -0.43 6/12 Normal / Yes Unremarkable

25 38 2.800 -1.20 6/12 Preeclampsia / No Mother: brother

deceased from

SIDS

Maternal

grandfather:

depression

Maternal

grandmother:

recurrent

miscarriage

Father: late

puberty

Paternal uncle:

retractile testes,

normal fertility

26 39 3.770 -0.63 12/12 Preterm Hypoplastic / Unremarkable

contractions bulbus

olfactorius

27 26 700 -1.26 10/12 IVF Persisting / Unremarkable

Twins ductus

Placental arteriosus

rupture Periventricu

lar

leukcomala

108

28 36 2.570 -0.72 8/12 Previous X-linked / First child was

abortions ichthyosis stillborn

Preterm Hypotonia Three

contractions Developme miscarriages

Bleeding ntal delay between month

Abnormal 1 and 2

liver

function

tests

29 39 3.230 -0,6 9/12 Pregnancy / / Father with

after Kallmann

gonadotroph syndrome,

in treatment pregnancy after

father gonadotrophins

30 38 3.782 1.04 8/12 Normal Mowat- / Unremarkable

Wilson

syndrome

32 39 2780 -1.70 7/12 Normal No No Unremarkable

GA: gestational age, BW: birth weight, EMS: external masculinization score, IUGR: intra uterine growth retardation, POF: premature ovarian failure, SIDS: sudden infant death syndrome; CS: Caesarian section,

IVF: in vitro fertilization

109

Table 2: Hormonal and genetic data of patients

Code FSH AMH T AR NR5A1 WT1 SRY Array-CGH MLPA Other

(U/l)# (µg/l) (ng/dl)*/*

(ref) (ref)$ *

1 3.2 59.8 136* Nl Nl Nl FISH Nl Nl HSD1

(1-12) (46.8- Nl 7B3

173)

2 14.0 62.7 63.8*/579 / Nl Nl Nl Nl Nl HSD1

(1-12) (46.8- ** 7B3,

173) ATRX

3 6.6 10.8 4.7* / c.253_254del / Nl / /

(1-12) (105-

270)

4 1.6 118 195* Nl Nl Nl / Nl Nl

(1-12) (46.8-

173)

5 NA 152 526** Nl c.437G>C / FISH / /

(67.4- (tolerated) Nl

197)

6 3.3 64.5 184* Nl c.630_637del / FISH / Nl

(1-12) (62-130) Nl

7 1.5 57.6 222* Nl Nl Nl / 7q36.3q36.3 Nl

(1-12) (105- (158189154-

270) 158343770)x1

, maternal

8 NA 10.6 275** Nl Nl Nl Nl Nl Nl

(38-180)

9 NA 231 NA Nl Nl Nl / Xp22.33p22.3 Nl

(46.8- 3 (839417-

173) 1179089)x3,

maternal

10 NA 244 NA Nl Nl Nl / Nl Nl ATRX

(46.8- Nl

173)

110

11 5.5 NA 158*/511* / / / / 3p25(RP11- /

(1-12) * 385A18→RP

11-

334L22)x3,

9p24.3

(RP11-

48M17→RP1

1-320E16)x1

12 2.5 147 109* / Nl Nl / / Nl

(1-12) (46·8-

173)

13 NA 72.7 280* / Nl Nl FISH Nl Nl

(46.8- Nl

173)

14 1.2 8.6 248* Nl / / / / /

(1-12) (105-

270)

15 NA 132.9 NA Nl c.1109 G>A / FISH 16p12.3p12.3 /

(42-185) (p Nl (18894303-

Cys370Tyr) 19162153)x3,

maternal

(normal

variant)

16 1.3 298 337* Nl Nl Nl FISH Xq13.3q13.3 Nl

(1-12) (67.4- Nl (74285912-

197) 75325119)x2,

maternal

17 NA 72 NA / Nl Nl FISH / Nl HSD1

(38-180) Nl 7B3

18 2.6 82.2 NA Nl Nl Nl / 2p16.3p16.3 Nl

(1-12) (23.8- (5073244-

124) 50894316)x1;

111

16p13.11p13.

11(15830681-

16270149)x3

19 1.8 49.1 35.3*/390 Nl Nl Nl FISH Xq13.3q13.3( Nl

(23.8- ** Nl 74380482-

124) 74567915)x2,

maternal

20 2.8 194 104* Nl Nl Nl FISH Nl Nl

(1-12) (105- Nl

270)

21 1.9 28.8 NA Nl Nl Nl Nl / Nl

(1-12) (55.3-

187)

22 2.1 94.1 151* Nl Nl Nl Nl 5p14.3p14.3( Nl

(1-12) (105- 21438696-

270) 21490654)x1,

maternal,

14q21.2q21.3

(42908541-

43293564)x3,

maternal

23 NA 11.27 404** Nl Nl Nl Nl Nl Nl

(55.3-

187)

24 NA 43 152*/474* / Nl Nl Nl Nl Nl SRD5

(105- * A2 N

270)

25 1.0 156 207.3* Nl Nl Nl / Nl Nl

(1-12) (105-

270)

26 0.17 65.9 3.3*/90.8 / / / / Nl /

(1-12) **

112

(105-

270)

27 NA 245 951** Nl Nl Nl / / Nl

(55.3-

187)

28 NA 14.03 NA / / / / Xp22.32p22.3 /

(55.3- 1(5405569-

187) 9222059)x0,

maternal

29 0.57 NA 3.2* / / / / / / FGFR

(1-12) 1

c.1042

G>A

30 1.2 159 502* / / / / Nl / ZEB2

(1-12) (105- c.2856

270) delG

32 11 25 191.9* Nl Nl Nl Nl Nl Nl

(1-12) (105-

270)

Symbols and abbreviations: NA: not available, FSH: follicle stimulating hormone; AMH: anti-Müllerian hormone; ref: age-specific reference value; T: testosterone; Nl: normal

#: Determined between day 14-90; $: age-specific AMH reference values may differ according to the commercial kits that have been used during the course of the study; *: Basal testosterone value between day 14-90 / ** : Testosterone value after

HCG stimulation (1500 U, blood sampling after 72h).

Genomic coordinates based on build hg18 (2006), except for patient 32, where build hg19 (2009) is used.

113

diagnosed with a NR5A1 mutation and

P26, with KS). Two of three patients

with NR5A1 mutations had an AMH

value within the reference for age.

Ratios of T/A and T/DHT were

determined to identify possible cases of

(dihydro)testosterone biosynthesis

disorders. The T/A ratio, measured

during mini-puberty was suggestive of

17β-HSD deficiency in two patients

(case 1: T/A ratio 0,19; case 2: T/A ratio

0,52) and after HCG stimulation in one case (case 17: T/A ratio 0,08)20,21. Figure 1: Mowat-Wilson syndrome, facial characteristics. The typical large HSD17B3 sequencing was performed in and uplifted earlobes in Patient 30, who all three cases but revealed no causal was diagnosed with Mowat-Wilson syndrome based on clinical data. mutations. In patient 2 a heterozygous

Hormonal work-up missense variant was identified, c. 866G>A (p.Gly289Asp), although With the exception of cases with KS, mutation prediction programs indicated where FSH was low, serum FSH was this variant to be tolerated. In patient 24, within the reference range in all cases. a T/DHT ratio of 10,8 was found at basal AMH, representing Sertoli cell function, sampling during mini-puberty but was low in 11/32 cases (34,3%), SRD5A2 sequencing revealed no including 2/3 cases with KS (in the third mutations. KS case, AMH could not be determined) and 4/9 cases (44%) born SGA. Low Genetic work-up

AMH was associated with low T values Array-CGH was done in 23/32 patients

(a marker for Leydig cell function) in to screen for larger genomic only two cases (P3, subsequently rearrangements. In 10 of them, CNVs

114 were identified as shown in Table 2. Identification of three novel NR5A1

Seven of these rearrangements were mutations maternally inherited, making their NR5A1 sequencing revealed three novel clinical significance questionable. In mutations (Figure 2A). In patient 3 a patient 11, we identified a partial heterozygous frameshift deletion was chromosome 9 deletion (9p24.3), identified: c.253_254del, resulting in a encompassing the Doublesex and Mab3 premature stopcodon (p.Ala85*). No related transcription factor 1 (DMRT1) other family members were available for gene. In patient 28, a deletion was found segregation analysis. A second on the X chromosome (Xp22.31- heterozygous frameshift deletion of 8 bp

Xp22.32). This region includes the STS was identified in patient 6, c.630_637del, region and the genes KAL1 and (p.Tyr211Profs*12). RT-qPCR in the

Neuroligin 4, X-linked (NLGN4X). This patient’s lymphoblasts indeed showed a deletion was also present in the patient’s lower expression of NR5A1 mRNA mother. In addition, we performed (Figure 2B). Segregation analysis

MLPA for 23/32 patients to screen for indicated that this mutation was present deletions and/or duplications on the in (1) the asymptomatic patient’s exon level, however no additional CNVs mother, (2) maternal aunt, who had were identified. AR (20/32) and WT1 been diagnosed with POF at the age of

(22/32) sequencing did not reveal any 35, and (3) grandfather, who had been mutations. NR5A1 sequencing was done operated for proximal hypospadias, but in 26/32 patients, leading to the spontaneously fathered two children identification of three novel mutations, (pedigrees in Figure 2C). The third which will be discussed below. In cases mutation was found in patient 15, c.1109 with serum AMH below the reference G > A, (p.Cys370Trp). This mutation was value for age (8/32), suggestive of predicted to have a deleterious effect on gonadal dysgenesis, SRY was sequenced, however no mutations were found. 115

Figure 2: Three novel NR5A1 mutations. (A) Schematic overview of the positions of the mutations and electropherograms. (B) RT-qPCR showed a lower NR5A1 expression in the maternal grandfather of the index patient (I:1), and in the mother of the index patient (II:2). We did not include the index case in this experiment as no fresh blood could be collected. Two negative control samples (NC) without the mutation were included for comparison. To exclude technical variations, expression of the reference genes GADPH, HMBS and TBP were also measured, showing stable expression in all patients. (C) Pedigrees for the patients with a NR5A1 mutation. The genotype of the analysed individuals is shown under their symbol. Full black squares indicate affected males with hypospadias, partially black circles indicate females with POF and circles with a black dot correspond with asymptomatic carrier females. 116 protein function according to several Clinical investigation and hormonal data prediction programs (SIFT, Polyphen were sufficient to diagnose Kallmann syndrome and Mowat-Wilson syndrome in and MutationTaster). The affected amino respectively three and one patients acid is located in the ligand-binding domain and is highly conserved (up to Familial, hormonal and/or phenotypic zebrafish). Segregation analysis revealed data were sufficient to suspect KS in that the mutation was present in the three patients (P26, 28 and 29) and MWS patient’s mother, who had no symptoms in patient 30. As suggested in a paper by of POF at the age of 24. Grumbach et al. our study confirms that in boys, the period of physiological DISCUSSION gonadotrophin surge (the so-called

To gain insight into the appropriateness “mini-puberty”) represents a unique and diagnostic yield of a systematic opportunity to diagnose KS early in cases integrated work-up in 46,XY with a suggestive phenotype (micropenis undervirilized cases who are sex- +/− cryptorchidism in the absence of assigned male, we used a standardized hypospadias)23. In these patients a screening panel in a series of 32 cases targeted approach to identify the referred to our DSD clinic. An overview underlying molecular cause was used. of the approach is shown in Figure 3A. Here, we ended up with a higher

Difficulties in blood collection in diagnostic success rate, the molecular newborns and infants made it impossible cause was identified in 75% (3/4) of to perform the complete screening in patients. every case, resulting in missing data. Despite suggestive hormonal results, we Low EMS scores (EMS < 7, n = 17) did not could not identify any HSD17B3, SRD5A2 or yield a higher diagnostic success as SRY mutations compared to higher EMS scores (EMS ≥ 7, Accumulation of A or T due to 17β-HSD n = 15). As reported earlier, no causal deficiency or 5α-reductase deficiency genetic variations were identified in respectively may lead to markedly low children born SGA (n = 9) in our series22.

117

T/A (in case of 17β-HSD deficiency) or SRY was performed but revealed no elevated T/DHT (in case of 5α-reductase mutations, confirming that SRY deficiency) ratios. In contrast to previous mutations are a relatively rare cause of reports, sequencing of the HSD17B3 and 46,XY partial gonadal dysgenesis in

SRD5A2 genes in cases with aberrant T/A contrast to 46,XY complete gonadal and T/DHT ratios respectively revealed dysgenesis, where SRY mutations are no mutations20,21,24,25. However, for thought to account for up to 15% of practical reasons, stimulated A and T cases3,30. values, which are generally considered Genetic screening: targeted resequencing more accurate than basal values during and copy number analysis mini-puberty, had been obtained in only Following a number of recent reports in one of three patients with T/A < 1. which NR5A1, AR and WT1 mutations Another possible explanation might be and CNVs have been identified as the the different detection methods used for cause of isolated hypospadias and/or the various androgens cryptorchidism10,12,31-37, a standardized (Radioimmunoassay for A versus genetic screening protocol was applied to LC/MSMS for T). Simultaneous detection identify the underlying genetic cause of of A, T and DHT by LC/MSMS, as the observed atypical genital described recently, is expected to be more development in all cases where clinical reliable but is not routinely available and hormonal data did not suggest a yet26. Low serum AMH has been specific diagnosis, irrespective of the reported previously in infants with KS3,27, EMS scores. The screening consisted of and has been attributed to a lack of FSH- array-CGH, MLPA and SRY-specific driven stimulus28,29. In all cases in which FISH to screen for genomic serum AMH was below the age-specific rearrangements, and sequencing of the reference values (n = 10), sequencing of AR, WT1 and NR5A1 genes.

118

Array-CGH is a valuable diagnostic tool in kidney stones and developmental delay.

46,XY undervirilization newborns with Liver function tests showed abnormal dysmorphic features and allowed the results, of hitherto unknown etiology. In identification of two causal CNVs in our this patient a part of the X-chromosome, cohort including the genes KAL1 and NLGN4X,

Array-CGH was used to screen for was deleted. KAL1 deletions or larger genomic rearrangements and led mutations are an established cause of X- to the identification of two deletions linked KS and can explain the genital with clinical significance, both found in phenotype seen in this patient40. syndromic patients. Patient 11 (EMS = 3) NLGN4X, has been associated with X- presented with penoscrotal linked mental retardation and X-linked hypospadias and transposition. Besides autism spectrum disorders41, and might these genital characteristics, this patient explain the observed developmental also showed macrocephaly, facial delay. Previously, a link between KS, dysmorphism and developmental ichthyosis and Xp deletions has been delay. Hormonal results revealed described by Bick et al.42. No evident normal T levels, AMH was not available; association could be found between the array-CGH revealed a partial identified deletion and the elevated liver chromosome 9 deletion, encompassing enzymes and recurrent kidney stones. the DMRT gene cluster. These genes This deletion was inherited from the encode transcriptional regulators mother, who had mild mental delay but involved in sex development, and no symptoms of KS. This deletion is monosomy of the distal part of therefore characterized by incomplete chromosome 9p, mostly DMRT1, has penetrance. In total, array-CGH been associated with 46,XY DSD in revealed 10 CNVs in 22 patients, seven 38,39 several cases . Patient 28 (EMS = 8) of them were inherited from the mother; showed symptoms of KS. Other making their clinical relevance phenotypic characteristics included: X- questionable. Array-CGH resulted in a linked ichthyosis, hypotonia, recurrent definite genetic diagnosis in 2/22 119 patients, (9%). When only considering We identified three novel NR5A1 the syndromic cases, array-CGH mutations, one of them was present in an affected male with preserved fertility renders a diagnostic yield of 2/6 patients

(33%). Although our series is small we Recently Kohler et al. reported a WT1 can conclude that array-CGH is a mutation rate of 7.5% in children with valuable diagnostic tool in 46,XY DSD severe hypospadias and Wang et al. with associated dysmorphic features identified AR mutations in 6.6% of their however larger patient groups should patient cohort with isolated be investigated to make more definite hypospadias and micropenis, indicating conclusions. Because of the limited a role for both WT1 and AR in minor resolution of array-CGH, we performed forms of undervirilization32,44. Sanger

MLPA to screen for deletions or sequencing of AR and WT1 was done in duplications on the exon level for SOX9, respectively 20 and 22 patients of our

NR5A1, WNT4 and NR0B1. In total 23 cohort. In contrast to these series, no patients were screened, however no significant sequence changes in these additional CNVs were identified. genes were identified. The relatively

Likewise, FISH analysis of SRY could high frequency in previous cohorts not reveal any deletions. Although the might be attributed to a selection bias. mutation uptake of targeted CNV Therefore, we conclude that the detection (MLPA) was limited in our incidence of mutations in AR and WT1 cohort, it still remains an important mutations is probably overestimated in addition to a genetic work-up of 46,XY patients with milder forms of undervirilized or 46,XY DSD patients. undervirilization. On the other hand,

Different reports showed NR5A1 NR5A1 was sequenced in 26 patients microdeletions as a cause of both 46,XY and revealed mutations in three of them

DSD and POF43,16. (11.5%). This is in line with other series, where mutations were identified in

approximately 15% of patients. In our

cohort, two frameshift mutations and

120 one missense mutation were identified. have regular menses. Interestingly, the

The missense mutation, c.1109G > A, grandfather had been treated for found in patient 15 (EMS = 3), targets an hypospadias as a child. Preserved amino acid in the functionally important fertility in males with NR5A1 mutations ligand binding domain (p.Cys370Trp) has only exceptionally been reported so and is predicted to alter protein function far46,47. These findings support the

(SIFT, Polyphen, MutationTaster). This extreme intrafamilial variability seen mutation was also found in the patient’s with NR5A1 mutations. At the moment mother. In addition to causing 46,XY the mechanism behind this phenotypic

DSD, NR5A1 mutations are a known variability and incomplete penetrance cause of premature ovarian failure resulting from NR5A1 mutations

(POF)45. The patient’s mother had remains elusive; they likely result from regular menses at the age of 30, however the effects of multiple genetic variations she is at risk for developing POF. The (modifiers) and/or their interactions first frameshift mutation (patient 3), with environmental factors. Variable c.253_254del induces a premature stop expressivity, reduced penetrance and codon at position 85 (p.Ala85*). There even more complex inheritance patterns were no additional family members such as digenic models have been available for segregation analysis. The reported in other developmental second frameshift mutation (patient 6), conditions such as Kallmann syndrome c. 630_637del, also leads to a premature and may be explained in part by the stop codon (p. Tyr211Profs12*). This overall ‘mutational load’ in different mutation was also present in the mother genes playing arole in common of the patient, a maternal aunt and the signaling pathways48-50. maternal grandfather. The aunt had The integrated story: clinical, hormonal recently been diagnosed with POF at the and genetic data age of 35 years and underwent several in Taken together, in spite of extensive vitro fertilization (IVF) cycles, the clinical, hormonal and genetic patient’s mother (age 39) declared to 121 screening, the molecular cause of 46,XY diagnostic yield of this sequential atypical male genital development sequencing approach, cost and time could only be identified in seven out of efficiency should be considered. Sanger

32 patients (21.8%). When comparing sequencing has an average cost of $2400 the diagnostic success rate between per million bases, whereas the emerging patients with low (<7, n = 17) or high (≥7, next generation sequencing n = 15) EMS scores, we identified the technologies (NGS) are much cheaper. underlying molecular defect in With the Illumina platform, there is only respectively three and four patients, a $0.07 sequencing cost per million bases leading to a diagnostic success rate of (number based on Hiseq2000)51. The respectively 17.6% and 26.5% for next step in the diagnostic work-up of patients with low versus higher EMS 46,XY boys wit atypical genitalia should scores, suggesting that the decision to be the implementation of targeted NGS perform a detailed diagnostic work-up panels covering clinically relevant genes in 46,XY patients with atypical genitalia with a known role in sex development should not be based on the severity of and steroid biosynthesis pathways. A the phenotype alone. Array-CGH flexible and automated NGS workflow revealed the causal CNV in two out of used for targeted resequencing of six syndromic patients, leading to a disease gene panels has been reported diagnostic yield of 33% in patients with by us and allows parallel and cost- additional phenotypic characteristics. effective analysis of a sizeable number of

When we included non-syndromic genes in a clinical setting52. While this cases, the success rate drops to 9%, approach seems to be very useful in indicating that array-CGH is still an some heterogeneous disorders, their appropriate diagnostic tool in clinical utility in 46, XY DSD is syndromic forms of 46,XY DSD, but is debatable, since the known disease less efficient in non-syndromic cases. genes in these phenotypes only account

Sequencing of AR, WT1 and SRY did not for 20–40% of patients. Therefore we reveal any mutations. Besides the low anticipate that whole exome sequencing 122

(WES), which is increasingly put where associated phenotypic forward as a clinical diagnostic test in characteristics or cases where clinical genetically heterogeneous disorders53,54, and hormonal data suggest a specific will gain importance in the diagnostic gene defect, it remains advisable to work-up of 46,XY DSD, both in a clinical perform targeted resequencing of the and research context. However, in cases specific disease gene(s).

CONCLUSION

In this study we examined a large consecutive cohort of undervirilized 46,XY neonates and infants. Following this protocol, we were able to genetically diagnose 19% of non- syndromic patients and one third of the syndromic cases. There was no significant difference between the diagnostic success rate in patients with low EMS compared to higher EMS. In syndromic cases, array-CGH had a high diagnostic yield. Serial gene screening resulted in several novel NR5A1 mutations, although the overall diagnostic yield was rather low. Interestingly, we identified a novel NR5A1 mutation that was also present in a related male with preserved fertility, which has only exceptionally been reported. Given the low diagnostic yield of the sequential approach, parallel screening technologies such as targeted resequencing of clinically relevant disease genes and WES will be a preferred choice in future screening protocols. However, in cases where associated phenotypes are present, a more targeted approach remains the preferential strategy.

123

124

Figure 3: Overview of the integrated investigation approach. (A) Results in the 46,XY undervirilization cohort. Clinical and hormonal investigation was sufficient to suspect a diagnosis in 4/32 cases. For two Kallmann syndrome patients the diagnosis was genetically confirmed, as shown in the CNV analysis and targeted resequencing boxes. A ZEB2 mutation was identified in the Mowat-Wilson syndrome patient. Subsequently a genetic work-up was performed for the remaining patients, guided by hormonal results. Sequencing of HSD17B3 and SRD5A2 in patients with a possible testosterone biosynthesis disorder did not reveal mutations. Genetic screening consisting of array-CGH, DSD MLPA and sequential gene-by-gene sequencing led to the identification of two causal CNVs (of which one KS, see above) and three novel NR5A1 mutations, respectively. (B) Suggested clinical algorithm for the investigation of 46,XY male neonates or infants referred for atypical genitalia. Upper section (orange): clinical investigation, including pregnancy history, medical history and physical examination, enables categorization in cases with and without syndromic features. Mid-section (blue): In all cases, clinical investigation should be followed by a hormonal work-up, which in turn can be suggestive of gonadal dysgenesis (GD), disorders of the steroid hormone biosynthesis pathway and/or rare forms of CAH (*: Only forms characterized by defective androgen production are implicated here), partial androgen receptor defects or KS. Insights in hormone levels can guide selection of target candidate genes. Lower section (green): After thorough evaluation of clinical and hormonal data, a decision can be made to sequence specific gene panels or to proceed to clinical whole exome sequencing to identify the underlying molecular cause and thereby support the clinical diagnosis. The boxes between brackets (with squared filling) represent single gene tests which can be replaced be the aforementioned gene panels In cases with syndromic features, array-CGH is still a recommended method to identify CNVs.

PATIENTS & METHODS

Patients

All 46,XY children younger than two years who were referred to our pediatric endocrinology service for the evaluation of atypical genitalia (e.g. hypospadias, micropenis) and who were sex assigned male, between 2007-2013 were included (n=32)

(Table 1). Medical history included pregnancy details, birth weight (BW), consanguinity and a familial history of disorders of sex development (DSD), sub- or infertility, premature ovarian failure (POF) or atypical genitalia. Phenotypic description consisted of a physical examination with special attention to

125 dysmorphism; EMS scores were calculated based on the aspect of the external genitalia55. None of the patients had proteinuria or renal insufficiency.

Biochemical analyses

Hormonal levels were obtained between day 14–90 after birth or after HCG stimulation (Pregnyl®, 1500U, with blood sampling at baseline and after 72 hours).

The following hormone levels were measured: anti-Müllerian Hormone (AMH) by enzyme linked immunosorbent assay (Beckman Coulter Company), Androstenedione

(A) by Radioimmunoassay (DiaSource Company), Testosterone (T) and

Dihydrotestosterone (DHT) by liquid chromatography/tandem mass spectrometry

(UPLC Waters Quattro premier). LH and FSH by electrochemoluminescence assay

(Roche Diagnostics E170 Modular).

Genetic analyses

Array-CGH using the Agilent 180 K array was used as a genomewide screen for copy number variations (CNVs) with an overall mean probe spacing of 14 kb, or 11 kb when only taking into account the Refseq genes. Hybridization was done according to the manufacturer’s protocol, followed by visualization of the results in arrayCGHbase56.

Fluorescent in situ hybridization (FISH) was performed for SRY to search for SRY rearranging translocations and mosaicism. To screen for CNVs on the exon level,

MLPA was done using the SALSA MLPA P185 Intersex probemix (MRC-Holland) containing probes for NR0B1, NR5A1, SOX9, SRY and WNT4. Sanger sequencing of the coding exons and untranslated regions (UTRs) was used to identify mutations in

AR, NR5A1 and WT1. SRY sequencing was included for patients suspected to have gonadal dysgenesis, based on an AMH level below the reference range. HSD17B3 and

SRD5A2 were sequenced in cases with suspicion of a testosterone biosynthesis disorder based on a T/A ratio <1 for 17β-HSD deficiency and a T/DHT ratio > 8.5 for

5α Reductase Deficiency (Table 2)274,275. Primers for AR, WT1 and SRY were designed using PrimerXL (http://www.primerxl.org/, available on request). Primer sequences for NR5A1, HSD17B3 and SRD5A2 can be found in supplemental data (Additional file 126

1: Table S1). Zinc Finger E-Box Binding Homeobox 2 (ZEB2) sequencing and sequencing of the Kallmann syndrome (KS) gene panel, consisting of six genes (KAL1, CHD7,

FGFR1, PROK2, PROKR2, FGF8) was done at the Henri Mondor Hospital (Paris,

France). Fibroblast Growth Factor Receptor 1 (FGFR1) sequencing was performed at the

CHU Hospital Cochin (Paris, France).

Cell culture, RNA extraction and cDNA synthesis

Lymphocytes were isolated by Lymphoprep™ (STEMCELL Technologies) and cultured in RPMI medium with 10% FCS; interleukin-2 and phytohemagglutin were added. Cells were incubated at 37°C and 5% CO2. RNA was extracted using the

RNeasy Plus Mini kit (Qiagen), followed by cDNA synthesis with the iScript™ cDNA synthesis kit (Biorad).

Expression analysis

Expression levels of NR5A1 were measured through real-time quantitative PCR (rt- qPCR), using following primers: NR5A1-F 5′ caggagtttgtctgcctcaa 3′ and NR5A1-R 5′ agtggcacagggtgtagtca 3′. After in silico validation primers were tested using a dilution series. The experiment was done with the SsoAdvanced SYBR supermix (Bio-rad).

Analysis of rt-qPCR results was done with qbase + software (Biogazelle).

The study was approved by the local medical ethical committee (Registration number

B670201110608) and all parents signed a written informed consent.

REFERENCES

1. Thyen, U., Lanz, K., Holterhus, P.-M. & Hiort, O. Epidemiology and initial management of ambiguous genitalia at birth in Germany. Horm. Res. 66, 195–203 (2006). 2. Brain, C. E. et al. Holistic management of DSD. Best Pract. Res. Clin. Endocrinol. Metab. 24, 335– 354 (2010). 3. Ono, M. & Harley, V. R. Disorders of sex development: new genes, new concepts. Nat. Rev. Endocrinol. 9, 79–91 (2013). 4. Köhler, B. et al. Analysis of the Wilms’ tumor suppressor gene (WT1) in patients 46,XY disorders of sex development. J. Clin. Endocrinol. Metab. 96, E1131-6 (2011). 5. Ahmed, S. F. Prevalence of hypospadias and other genital anomalies among singleton births, 1988-1997, in Scotland. Arch. Dis. Child. - Fetal Neonatal Ed. 89, 149F–151 (2004).

127

6. Toppari, J., Virtanen, H. E., Main, K. M. & Skakkebaek, N. E. Cryptorchidism and hypospadias as a sign of testicular dysgenesis syndrome (TDS): environmental connection. Birth Defects Res. A. Clin. Mol. Teratol. 88, 910–9 (2010). 7. van der Zanden, L. F. M. et al. Common variants in DGKK are strongly associated with risk of hypospadias. Nat. Genet. 43, 48–50 (2011). 8. Sharpe, R. M. Sperm counts and fertility in men: a rocky road ahead. Science & Society Series on Sex and Science. EMBO Rep. 13, 398–403 (2012). 9. Hughes, I. a, Martin, H. & Jääskeläinen, J. Genetic mechanisms of fetal male undermasculinization: a background to the role of endocrine disruptors. Environ. Res. 100, 44– 9 (2006). 10. Köhler, B. et al. The spectrum of phenotypes associated with mutations in steroidogenic factor 1 (SF-1, NR5A1, Ad4BP) includes severe penoscrotal hypospadias in 46,XY males without adrenal insufficiency. Eur. J. Endocrinol. 161, 237–42 (2009). 11. Ferraz-de-Souza, B., Lin, L. & Achermann, J. C. Steroidogenic factor-1 (SF-1, NR5A1) and human disease. Mol. Cell. Endocrinol. 336, 198–205 (2011). 12. Allali, S. et al. Mutation analysis of NR5A1 encoding steroidogenic factor 1 in 77 patients with 46, XY disorders of sex development (DSD) including hypospadias. PLoS One 6, e24117 (2011). 13. Brauner, R. et al. Clinical, biological and genetic analysis of anorchia in 26 boys. PLoS One 6, e23292 (2011). 14. Köhler, B., Schumacher, V., l’Allemand, D., Royer-Pokora, B. & Grüters, a. Germline Wilms tumor suppressor gene (WT1) mutation leading to isolated genital malformation without Wilms tumor or nephropathy. J. Pediatr. 138, 421–4 (2001). 15. Ledig, S. et al. Array-CGH analysis in patients with syndromic and non-syndromic XY gonadal dysgenesis: evaluation of array CGH as diagnostic tool and search for new candidate loci. Hum. Reprod. 25, 2637–46 (2010). 16. Barbaro, M., Cools, M., Looijenga, L. H. J., Drop, S. L. S. & Wedell, a. Partial deletion of the NR5A1 (SF1) gene detected by synthetic probe MLPA in a patient with XY gonadal disorder of sex development. Sex Dev. 5, 181–7 (2011). 17. Barbaro, M. et al. Gene dosage imbalances in patients with 46,XY gonadal DSD detected by an in-house-designed synthetic probe set for multiplex ligation-dependent probe amplification analysis. Clin. Genet. 73, 453–64 (2008). 18. Ahmed, S. F. & Rodie, M. Investigation and initial management of ambiguous genitalia. Best Pract. Res. Clin. Endocrinol. Metab. 24, 197–218 (2010). 19. Miraoui, H. et al. Mutations in FGF17, IL17RD, DUSP6, SPRY4, and FLRT3 are identified in individuals with congenital hypogonadotropic hypogonadism. Am. J. Hum. Genet. 92, 725–43 (2013). 20. Twesten W, Holterhus P, Sippell WG, Morlot M, Schumacher H, Schenk B, H. O. Clinical, endocrine, and molecular genetic findings in patients with 17beta-hydroxysteroid dehydrogenase deficiency. Horm. Res. 53, 26–31 (2000). 21. Walter KN, Kienzle FB, Frankenschmidt A, Hiort O, Wudy SA, van der Werf-Grohmann N, Superti-Furga A, S. K. Difficulties in diagnosis and treatment of 5alpha-reductase type 2 deficiency in a newborn with 46,XY DSD. Horm. Res. Paediatr. 74, 67–71 (2010). 22. Cox, K. et al. Novel associations in disorders of sex development: findings from the I-DSD Registry. J. Clin. Endocrinol. Metab. 99, E348-55 (2014).

128

23. Grumbach, M. M. A window of opportunity: the diagnosis of gonadotropin deficiency in the male infant. J. Clin. Endocrinol. Metab. 90, 3122–7 (2005). 24. Boehmer, A. et al. 17beta-hydroxysteroid Dehydrogenase-3 Deficiency: Diagnosis, Phenotypic Variability, Population Genetics, and Worldwide Distribution of Ancient and de novo Mutation*. J. Clin. Endocrinol. Metab. 84, 4713–4721 (1999). 25. Baldinotti, F. et al. Molecular characterization of 6 unrelated Italian patients with 5alpha- reductase type 2 deficiency. J. Androl. 29, 20–8 (2008). 26. Kulle, a E., Riepe, F. G., Melchior, D., Hiort, O. & Holterhus, P. M. A novel ultrapressure liquid chromatography tandem mass spectrometry method for the simultaneous determination of androstenedione, testosterone, and dihydrotestosterone in pediatric blood samples: age- and sex-specific reference data. J. Clin. Endocrinol. Metab. 95, 2399–409 (2010). 27. Vizeneux, A. et al. Congenital hypogonadotropic hypogonadism during childhood: presentation and genetic analyses in 46 boys. PLoS One 8, e77827 (2013). 28. Lindhardt Johansen, M. et al. Anti-Müllerian Hormone and Its Clinical Use in Pediatrics with Special Emphasis on Disorders of Sex Development. Int. J. Endocrinol. 198698 (2013). doi:10.1155/2013/198698 29. Young, J. et al. Testicular anti-mullerian hormone secretion is stimulated by recombinant human FSH in patients with congenital hypogonadotropic hypogonadism. J. Clin. Endocrinol. Metab. 90, 724–8 (2005). 30. Bashamboo, A. & McElreavey, K. Gene mutations associated with anomalies of human gonad formation. Sex. Dev. 7, 126–146 (2013). 31. Wang, Y. et al. Mutation analysis of five candidate genes in Chinese patients with hypospadias. Eur. J. Hum. Genet. 12, 706–12 (2004). 32. Alléra, A. et al. Mutations of the androgen receptor coding sequence are infrequent in patients with isolated hypospadias. J Clin Endocrinol Metab 80, 2697–9 (1995). 33. Hiort, O. et al. Molecular characterization of the androgen receptor gene in boys with hypospadias. Eur. J. Pediatr. 153, 317–21 (1994). 34. Sutherland, R. W. et al. Androgen receptor gene mutations are rarely associated with isolated penile hypospadias. J. Urol. 156, 828–31 (1996). 35. Muroya, K. et al. Hypospadias and the androgen receptor gene: mutation screening and CAG repeat length analysis. Mol. Hum. Reprod. 7, 409–13 (2001). 36. Radpour, R., Rezaee, M., Tavasoly, A., Solati, S. & Saleki, A. Association of long polyglycine tracts (GGN repeats) in exon 1 of the androgen receptor gene with cryptorchidism and penile hypospadias in Iranian patients. J. Androl. 28, 164–9 (2007). 37. Nordenskjöld, a et al. Screening for mutations in candidate genes for hypospadias. Urol. Res. 27, 49–55 (1999). 38. Onesimo, R. et al. Chromosome 9p deletion syndrome and sex reversal: novel findings and redefinition of the critically deleted regions. Am. J. Med. Genet. A 158A, 2266–71 (2012). 39. Quinonez, S. C. et al. 9P Partial Monosomy and Disorders of Sex Development: Review and Postulation of a Pathogenetic Mechanism. Am. J. Med. Genet. A 161A, 1882–96 (2013). 40. Beate, K., Joseph, N., Nicolas, D. R. & Wolfram, K. Genetics of isolated hypogonadotropic hypogonadism: role of GnRH receptor and other genes. Int. J. Endocrinol. 2012, 147893 (2012). 41. Laumonnier, F. et al. X-linked mental retardation and autism are associated with a mutation in the NLGN4 gene, a member of the neuroligin family. Am. J. Hum. Genet. 74, 552–7 (2004).

129

42. Bick, D. et al. Male infant with ichthyosis, Kallmann syndrome, chondrodysplasia punctata, and an Xp chromosome deletion. Am. J. Med. Genet. 33, 100–7 (1989). 43. Harrison, S. et al. Screening and Familial Characterization of Copy-Number Variations in NR5A1 in 46,XY Disorders of Sex Development and Premature Ovarian Failure. Am J Med Genet 161, 2487–2494 (2013). 44. Köhler, B. et al. Analysis of the Wilms’ tumor suppressor gene (WT1) in patients 46,XY disorders of sex development. J. Clin. Endocrinol. Metab. 96, E1131-6 (2011). 45. Lourenco, D. et al. Mutations in NR5A1 Associated with Ovarian Insufficiency. N. Engl. J. Med. 360, 1200–1210 (2009). 46. Philibert, P. et al. Predominant Sertoli cell deficiency in a 46,XY disorders of sex development patient with a new NR5A1/SF-1 mutation transmitted by his unaffected father. Fertil. Steril. 95, 1788.e5-9 (2011). 47. Ciaccio, M. et al. Preserved Fertility in a Patient with a 46 , XY Disorder of Sex Development due to a New Heterozygous Mutation in the NR5A1 / SF-1 Gene : Evidence of 46 , XY and 46 , XX Gonadal Dysgenesis Phenotype Variability in Multiple Members of an Affected Kindred. Horm. Res. Paediatr. 119–126 (2012). doi:10.1159/000338346 48. Dodé, C. & Hardelin, J.-P. Kallmann syndrome. Eur. J. Hum. Genet. 17, 139–146 (2009). 49. Quaynor, S. D. et al. The prevalence of digenic mutations in patients with normosmic hypogonadotropic hypogonadism and Kallmann syndrome. Fertil. Steril. 96, 1424–1430 (2013). 50. Costa-Barbosa, F. A. et al. Prioritizing genetic testing in patients with Kallmann syndrome using clinical phenotypes. J. Clin. Endocrinol. Metab. 98, E943-53 (2013). 51. Liu, L. et al. Comparison of next-generation sequencing systems. J. Biomed. Biotechnol. 2012, 251364 (2012). 52. De Leeneer, K. et al. Flexible, scalable, and efficient targeted resequencing on a benchtop sequencer for variant detection in clinical practice. Hum. Mutat. 36, 379–87 (2015). 53. Biesecker, L. G. & Green, R. C. Diagnostic clinical genome and exome sequencing. N. Engl. J. Med. 370, 2418–25 (2014). 54. Yang, Y. et al. Clinical whole-exome sequencing for the diagnosis of mendelian disorders. N. Engl. J. Med. 369, 1502–11 (2013). 55. Ahmed, S. F., Khwaja, O. & Hughes, I. a. The role of a clinical score in the assessment of ambiguous genitalia. BJU Int. 85, 120–4 (2000). 56. Menten, B. et al. arrayCGHbase: an analysis platform for comparative genomic hybridization microarrays. BMC Bioinformatics 6, 124 (2005).

130

BIALLELICBIALLELIC AND AND MONOALLELIC MONOALLELIC ESR2 MUTATIONSESR2 MUTATIONS ARE ASSOCIATED ARE ASSOCIATED WITH 46, XY WITH 46, XY DISORDERS OF SEX DEVELOPMENT DISORDERS OF SEX DEVELOPMENT

Under review in Genetics in Medicine

Dorien Baetens1, Tülay Güran2, Berenice Mendonca3, Lode De Cauwer4, Frank

Peelman4, Marnik Vuylsteke5, Malaïka Van der Linden6, ESR2 study group, Hans

Stoop7, Leendert Looijenga7, Karolien De Bosscher4, Martine Cools8*, Elfride De Baere1*

1Center for Medical Genetics Ghent, Ghent University and Ghent University Hospital, Ghent, Belgium

2Zeynep Kamil Maternity and Childrens’ Diseases Training and Research Hospital, Division of Pediatric

Endocrinology and Diabetes, Istanbul, Turkey

3Disciplina de Endocrinologia, Laboratorio de Hormonios e Genetica Molecular LIM/42, Unidade de

Adrenal, Disc. de Endocrinologia e Metabologia, Hospital das Clínicas, Faculdade de Medicina da

Universidade de São Paulo, São Paulo, Brazil

4Receptor Research Laboratories, Nuclear Receptor Lab, Department of Medical Protein Research, VIB,

Ghent, Belgium; Department of Biochemistry, Ghent University, Ghent, Belgium

5GNOMIXX ltd, Statistics for Genomics, Melle, Belgium

6Department of Medical and Forensic Pathology, Ghent University and Ghent University Hospital,

Ghent, Belgium

7Department of Pathology, Erasmus MC, University Medical Center Rotterdam, Rotterdam 3015 AA,

The Netherlands

8Departmant of Pediatric Endocrinology, Ghent University and Ghent University Hospital, Ghent,

Belgium 131

132

ABSTRACT

Background: Disorders/differences of sex development (DSDs) are rare congenital conditions characterized by atypical sex development. The molecular cause remains unknown in 50% of cases. Whole exome sequencing (WES) can uncover novel genes implicated in sexual differentiation.

Methods: Homozygosity mapping and WES revealed an ESR2 variant in an individual with syndromic 46,XY DSD. Additional cases with 46,XY DSD underwent WES and targeted next-generation sequencing of ESR2. Functional characterization of the identified variants included luciferase assays and protein structure analysis.

Immunostaining of ER-β was performed in an eight-weeks old human male embryo.

Results: We identified a homozygous ESR2 variant, c.541_543del p.(Asn181del), located in the highly conserved DNA-binding domain of ER-β, in an individual with syndromic 46,XY DSD. Two additional heterozygous missense variants, c.251G>T p.(Gly84Val) and c.1277T>G p.(Leu426Arg), located in the N-terminus and the ligand- binding domain of ER-β, were found in unrelated, non-syndromic 46,XY DSD cases.

A significantly increased transcriptional activation and an impact on protein conformation was shown for the p.(Asn181del) and p.(Leu426Arg) variants. ER-β expression was demonstrated in developing intestine and eyes.

Conclusion: Our study supports a role for ESR2 as a novel disease gene for 46,XY DSD.

133

134

INTRODUCTION

In humans, gonadal fate is determined granulosa and Sertoli cells respectively by the presence or absence of sex- result in trans-differentiation to an specific gene expression. In the opposite cell fate6,7. bipotential stage of development, Subtle imbalances of this strictly transcription factors such as Wilm’s regulated sex-specific gene network can Tumor 1 (WT1), Nuclear Receptor lead to disorders/differences of sex Subfamily 5 Group A member 1 development (DSD)1,5. Currently, (NR5A1) and GATA binding protein 4 mutations in genes known to be (GATA4) lead to upregulation of Sex involved in sex development explain the determining Region Y (SRY) in the male molecular pathogenesis in only a embryo, thereby initiating the male minority of 46,XY DSD1,8. Next- developmental signaling cascade, with generation sequencing-based the expression of other male-specific approaches such as whole exome genes like SRY-box 9 (SOX9)1. In sequencing (WES) improve the addition, several reports indicate an identification of causal genetic defects, important role for the Mitogen and the discovery of new genes and Activated Protein Kinase (MAPK) signaling pathways implicated in sex pathway in the early phases of (male) development9,10. Elucidating the sex determination2-4. In female embryos, underlying causes of DSD can facilitate lacking the Y chromosome, female- sex assignment, and lead to improved specific genes like R-spondin 1 (RSPO1) management, a better prognosis and a and Forkhead box L2 (FOXL2) are more accurate evaluation of gonadal expressed amongst others1,5. Animal function and gonadal germ cell cancer studies have revealed multiple and risk11,12. long-lasting interactions between these male and female pathways; for instance So far, mutations in genes encoding loss of Foxl2 or Dmrt1 in murine adult nuclear receptors such as NR5A1 (also

135 known as Steroidogenic Factor 1 or SF1)13 genes, none of them are known DSD and Androgen Receptor (AR, NR3C4)14, genes. The gene encoding 5α-reductase, have been implicated in DSD SRD5A2, was present in another small pathogenesis as well as the duplication homozygous region (3.6 Mb), however of Nuclear Receptor Subfamily 0 Group B no potentially pathogenic variants were

Member 1 (NR0B1, DAX1)15. Here, we found in its coding region. Variant identified biallelic and monoallelic prioritization ranked ESR2 as the top

Estrogen receptor 2 (ESR2) mutations in candidate gene (Supplemental data, one syndromic and two non-syndromic Table S7). Sequencing of ESR2 revealed

46,XY DSD cases respectively, putting a homozygous 3-bp deletion in exon 9, forward ESR2 as a novel disease gene c.541_543del, leading to an in-frame for 46,XY DSD and expanding the deletion, p.(Asn181del) (Figure 1A-B). spectrum of nuclear receptors This deletion affects a highly conserved implicated in 46,XY DSD etiology. amino acid, located in the DNA-binding

domain (DBD) (Figure 1A-B). The RESULTS deletion was found to be absent in 92 Identification of a homozygous ESR2 tested ethnically matched control variant in Case 1 with syndromic 46,XY DSD chromosomes and in 320 other tested Given the possible parental control chromosomes. This variant is consanguinity, we performed reported at very low frequencies in homozygosity mapping to identify runs publicly available databases (ESP, ExAC of homozygosity containing the 0.3.1, gnomAD), however no putative underlying genetic cause of the homozygotes have previously been syndromic 46,XY DSD phenotype in identified (variant details in Case 1. Twenty-one homozygous Supplemental data, Table S8). regions larger than 1 Mb were identified Segregation analysis showed that both (Supplemental data, Table S6). The parents (Figure 1B, I:1 and I:2) and one largest one, 20 Mb in length, is located healthy 46,XX sibling (Figure 1B, II:2) on chromosome 14, and contains 465 are heterozygous carriers of the 136 deletion, another 46,XX sibling (Figure of 109). Several prediction programs

1B, II:3) is homozygous for the wild type (SIFT, PolyPhen-2 and MutationTaster) allele. WES made the presence of other suggest a deleterious effect on protein potentially pathogenic coding variants function. The variant is absent in 412 in the index case and her parents very tested control chromosomes and in unlikely. An overview of the variant genomic databases such as dbSNP, ESP, filtering, performed as previously and ExAC. In gnomAD it occurs in described16, is given in Supplemental 0,000396% (no homozygotes) (Table 3).

Tables S3-4. WES did not reveal any other potentially

pathogenic variants in Case 2, an Identification of two heterozygous ESR2 missense variants in Case 2 and 3 with non- overview of the different filtering steps syndromic 46,XY DSD and variants is given in Supplemental Tables S3 and S5. In Case 3 a Targeted resequencing of the ESR2 heterozygous missense variant was coding region in a Belgian 46,XY DSD found, c.1277T>G in exon 8 of ESR2, cohort (n=73) and in a Brazilian 46,XY leading to the missense variant DSD cohort (n=40) revealed two p.(Leu426Arg) in the ligand-binding heterozygous missense variants in two domain (LBD) and affecting a highly unrelated cases with non-syndromic conserved amino acid (up to Tetraodon). 46,XY DSD. In Case 2 a heterozygous The Grantham distance between Leu ESR2 variant c.251G>T was found in and Arg is moderate (i.e. 102) and exon 2, leading to a missense variant several prediction programs show a p.(Gly84Val) affecting a highly potential deleterious effect. This variant conserved amino acid (up to fruitfly) is not present in the consulted genomic located in the N-terminal domain, which databases (dbSNP, ESP, ExAC, contains the first activation function of gnomAD) nor in 214 Brazilian controls the receptor (Figure 1A). The (Table S8). Figure 1 shows a schematic physicochemical difference between Gly overview of the location and and Val is moderate (Grantham distance

137

Figure 1: Identification of three ESR2 variants. (A) Schematic overview of the ER-β protein with detailed representation of the DNA-binding domain (DBD). Light grey circles: P-box amino acids. Dark grey circles: dimerization interface. Red circle: deleted amino acid in Case1. (B) Protein alignment (left) indicating the deleted amino acid Asn181 (N181) found in Case 1, conserved up to Tetraodon and in human ER-α. Electropherogram (middle) and pedigree (right). Affected individual: black symbol. M: mutation.(C) Protein alignment (left) indicating the affected amino acid in Case 2, Gly84 (G84) conserved up to fruitfly. Electropherogram (middle) and pedigree (right). NA: not available.(D) Protein alignment (left) indicating the affected amino acid in Case 3, Leu426 (L426) conserved up to Tetraodon. Electropherogram (middle) and pedigree (right). NA: not available.

138 characteristics of the identified ESR2 calculations of ΔΔG with foldX and tried variants. to model the best rotamer position for

each variant with UCSF Chimera17. Structural analysis of the ER-β variants There was no template available for The ER-β DBD binds to DNA as a remodeling of the p.(Gly84Val) variant, homodimer. We built homology models and no decent homology models could for the wild type and p.(Asn181del) be obtained. The p.(Leu426Arg) variant mutant homodimer. Residue Asn181 is is located in the LBD, which was part of a longer extended loop that previously characterized by X-ray connects the main DNA binding helix to crystallography18. Leu426 is a buried a zinc coordination site, which is part of residue of helix 9 (Figure 2A). FoldX the homodimer interface. The Asn181 energy calculations using the deletion shortens this connecting loop BuildModel command indicate that the and causes a 15° tilt of the main DNA- p.(Leu426Arg) mutation affects the binding helix (Figure 2, panel A-B). folding energy of the protein and has a Molecular dynamics simulation, destabilizing effect (in silico ΔΔG= 2.3 followed by energy minimization kcal/mol)18. Leu426 fits in a hydrophobic suggests that this slightly affects the pocket and engages in interactions with orientation between the two DBDs in the Leu322, Leu330, Leu406 and Met410 homodimer, with an increased number (Figure 2B). Modeling the best fitting of contacts between the two domains Arg rotamer at the position of Leu426 in (Figure 2, panel A-B). It is therefore very UCSF Chimera shows that this mutation likely that the Asn181 deletion affects abolishes most of these hydrophobic homodimerization and/or DNA binding interactions, while even the best fitting of the ER-β. rotamer of the Arg sidechain introduces

To assess the possible influence of the strong steric clashes in the absence of p.(Leu426Arg) missense variant on ER-β drastic energy minimization (green protein structure, we made in silico lines) (Figure 2C). The p.(Leu426Arg) 139

140

Figure 2: Structural analysis of mutated and WT ER-β. (A) Effect of the Asn181 deletion on a dimer of the ER-β DNA binding domain. A model for a dimer of both the wild type (blue) and Asn181del mutant (red) DNA-binding domain (DBD) are shown. Monomer 1 of both dimers is superposed. The Asn181 (N181) deletion shortens a long loop between the DNA-binding helix and the zinc-binding site at the dimer interface. As a result, the DNA-binding helix tilts towards the dimer interface. The red arrow indicates an upward shift of the second monomer versus the first monomer in the mutant dimer. The molecules are rotated 90° compared with panel B and the DNA-binding helices are directed towards the viewer. (B) Model of the wild type ER-β DBD bound to DNA. The position of Asn181 is indicated. (C) Position of the Leu270 (L270) and Leu426 (L426) residues in the ER-β LBD structure. Atoms of L270 and L426 are shown as grey spheres. (D) L426 is buried in a hydrophobic environment in the LBD structure. (E) When L426 is mutated to the best fitting Arg rotamer in UCSF chimera, the Arg sidechain clashes with residues D326 and M410 (red lines)17.

141 variant is likely to affect the stability or Transcriptional activation assays and folding of the LBD domain. Folding of localization studies the p.(Leu426Arg) variant estrogen- The impact of the identified variants on bound LBD domain into a wild type-like transcriptional activation was assessed structure would at least require drastic using luciferase assays with TK-3xERE- structural changes in the environment of promoter constructs. The p.(Asn181del) residue 426, which could hamper LBD variant resulted in a significantly dimerization, as helix 9 and the loop increased transcriptional activation after between helix 8 and 9 (which interacts stimulation with the ER-β specific with Leu426 via Met410) are both part of ligand, DPN (p<0.001). For the two the symmetrical LBD dimer interface. missense variants, p.(Gly84Val) and

p.(Leu426Arg), no significant alterations ER-β expression in a human male embryo of ligand-dependent transcriptional In a human male embryo, eight weeks activation were observed (Figure 3). gestational age, ER-β expression was Interestingly, the p.(Leu426Arg) variant demonstrated in the developing eye, shows a higher transcriptional especially the corneal epithelium and activation compared to WT ER-β when endothelium, primitive lens neurons, the cells are not stimulated with DPN primitive retinal cells and the retinal (p=0.0014), suggesting a higher baseline pigment epithelium, the mucosa of the activity. Localization studies for wild developing intestine and in the neuronal type, p.(Asn181del) and p.(Gly84Val) system, specifically the primitive constructs in HEK293T cells showed no neurons of the germinal matrix. No ER- differences in subcellular localization β expression could be demonstrated in between the wild type and mutagenized the developing testis at this stage of constructs. Both ER-β variants showed development (Supplemental data, nuclear translocation upon stimulation Figure S1).

142

Figure 3: Transcriptional activation assays. Results are shown as means  standard error for relative luciferase intensity units (R.L.U.). The analysis involved the fitting of a hierarchical generalized linear model (HGLM) to the R.L.U. data, as implemented in Genstat v1819, with transfection and stimulation as fixed terms (Poisson distribution; log link) and replicates (n =

4 or higher) nested within experimenter as additional random terms (Gamma distribution; log link). **: p < 0.01; ***: p < 0.001. with the ER-βspecific ligand DPN (Supplemental data,

Figure S2).

DISCUSSION sexual development such as

20-24 The identification of three distinct hypospadias and micropenis and a potentially pathogenic variants in ESR2 heterozygous ESR2 variant c.1295C>A in unrelated individuals with 46,XY leading to a missense change gonadal dysgenesis suggests an p.(Ala432Asp) in the LBD of ER-β was unanticipated role for ER-β in early recently reported in a girl with primary

25 gonadal development. Previously, amenorrhea . polymorphic ESR2 variants have been Here, we identified a biallelic in-frame associated with milder variations in deletion in a girl with syndromic 46,XY 143

DSD and two monoallelic missense extragonadal organs affected in Case 1. variants in two unrelated cases with However, no testicular expression was non-syndromic 46,XY DSD. The data of observed at this time of development,

Case 1 is suggestive of absent gonadal suggesting an earlier role of ER-β, at the development, while data of Case 2 and 3 stage of gonadal determination, in line are compatible with partial and with the absence of gonads observed in complete gonadal dysgenesis. Both Case 1. Non-classical estrogen signaling population frequencies and in silico has been shown previously to interact predictions support a pathogenic role of with the MAPK pathway27; a signaling the three variants found (Table S8). pathway involved in the earliest stages

Structural analyses support a of gonadal development, although the deleterious effect of the p.(Asn181del) operating mechanisms are different in and the p.(Leu426Arg) variant on humans and mice. In humans, MAPK protein conformation, but are not signaling is essential for tilting the gene available for the p.(Gly84Val) variant expression balance towards SOX9 due to the absence of a suitable model expression; gain-of-function mutations for this region. A significant increase in of MAP3K1 lead to a higher transcriptional activation was shown for phosphorylation of downstream targets, p.(Asn181del) and the p.(Leu426Arg) causing a decreased expression of SOX9 variant resulted in increased activity in and an increased expression of female- the absence of an ER-β specific ligand; specific genes like β-catenin2,28,29. In mice, both effects might hint towards a gain- the Mapk pathway acts through direct of-function effect. upregulation of Sry; homozygous loss-

of-function variants of Map3k4 are Immunohistochemistry in an eight- reflected in a male-to-female sex weeks old male human embryo reversal phenotype3,29. confirmed ER-β expression in the developing intestine, eye and neuronal Nuclear receptor family members, to systems26 which reflects the which the ERs belong, are known to

144 exert their functions through different activation and failure of testicular action mechanisms. The classical, development (Figure 4). For the genomic pathway involves ligand p.(Leu426Arg) variant, we observed a binding and receptor dimerization, significantly higher activation of the followed by nuclear translocation and unstimulated ER-β receptor compared activation or repression of target gene to WT unstimulated ER-β, suggestive of expression27,30. Estrogen binds on both a hyperactive ER-β receptor that can be

ER-α and ER-β receptors, however, activated by estrogen levels that are binding results in partially overlapping otherwise insufficient. Maternal but also differential transcriptional estrogen that typically has no obvious effects31,32. Alternative action effect on male sex development, might mechanisms involve interactions with have activated the hyperactive receptor other transcription factors, ligand- in this case and induce a gain-of- independent receptor activation and function effect as well. Experimental activation of different signaling validation of these hypotheses is pathways. For instance, it has been challenging given the cell-type shown that ER-β can activate specific specificity of these alternative pathways

MAPK pathways in human colon cancer and the fact that no relevant cell line cell lines. The scarce experimental data resembling the undifferentiated gonad on these alternative pathways indicate is available. Furthermore, it remains that these processes are highly cell complex to acquire an experimental set- specific27. We hypothesize that ESR2 up that resembles the earliest phases of gain-of-function variants, as assumed sex development as all available for p.(Asn181del), can lead to increased gonadal cell lines have already activation of the MAPK pathway via a differentiated. non-genomic working mechanism, The differential action modes of MAPK thereby shifting the expression balance signaling in mice and humans can also from SOX9 to β-catenin signaling, and explain the absence of a sex reversal resulting in resulting in decreased SOX9 145

Figure 4: Suggested action mechanism for gain-of-function mutations in ESR2. Green boxes: testis promoting factors, orange boxes: ovary promoting factors, arrows represent interactions that promote action of downstream targets, bars represent interactions that inhibit action of downstream targets. (A) Suggested role of the ESR2 gene in normal male development. Besides the classical genomic action mechanism, ESR2 can also exert its effect by cell-type specific, non-genomic responses via other pathways like the MAPK pathway. The MAPK pathway has previously been implicated in gonadal development by the identification of gain- of function mutations in MAP3K1. This pathway shifts the gene expression balance towards SOX9 upregulation via interactions with downstream targets (e.g. MAP3K4) and their cofactors (GADD45γ). Once SOX9 is activated, it maintains its own expression via a FGF9 regulated positive feedback loop. In physiological circumstances, MAP3K1 activity also leads to phosphorylation of p38 and ERK1/2, leading to inhibition of the female specific β-catenin pathway. (B) We hypothesize that ESR2 gain-of function variants can act on gonadal development in a similar way as MAP3K1 variants. Previous research has shown that gain-of- function MAP3K1 mutations may mimic the ovary determining pathway and lead to increased binding of co-factors RHOA and Rock, inhibiting SOX9 expression. In addition, it induces hyperphosphorylation of p38 and ERK1/2, neutralizing and even reversing β-catenin inactivation. Beta-catenin signaling can then activate FOXL2, thereby further repressing SOX9 activation. This way MAP3K1 mutations, and hence ESR2 mutations according to our hypothesis, can attenuate the testis-determining signal induced by SRY (grey box). Hypothesis and figure are based on data reported by Loke et al. (2014) and Ostrer et al. (2014) 28,29.

146 phenotype in ER-β knock-out mouse NR0B1 (DAX1), where duplications lead models. Over the years, several ER-β KO to male-to-female sex reversal in mice have been generated. A first mouse humans, but to delayed testis knockout model was characterized by development in mice34. female sub- or infertility and normal Differences in zygosity have been reproductive capacity in males33. This shown for the missense ESR2 variants mouse was generated through insertion found here, and may relate to the milder of a neomycin resistance gene into exon and different phenotype in the two

3 of the Esr2 gene by homologous monoallelic cases compared to the recombination in embryonic stem cells. biallelic case (non-syndromic versus

Several transcripts were still being syndromic; complete and partial processed however, leading to an testicular dysgenesis versus gonadal incomplete knockout. Moreover, ER-β agenesis). Moreover, a different knock-out mice originating from the working mechanism cannot be excluded same initial ER-β mutant but raised and for the monoallelic missense variants examined in different laboratories found, although they may all act via the presented with a range of different same dysregulated MAPK pathway. phenotypes. A new ER-β knock-out Differences in zygosity and mutation- mouse used Cre/LoxP mediated dependent action mechanisms have excision of the third exon of Esr2, been described for other nuclear resulting in complete absence of any receptors, such as NR2E3, implicated in

Esr2 transcript. These mice show both autosomal dominant and recessive infertility in both sexes, but no sex retinal dystrophies with clinically reversal. A similar situation has been distinct phenotypes35. reported for another nuclear receptor

147

GENERAL CONCLUSION

We identified biallellic and monoallelic variants in ESR2 in patients with syndromic and non-syndromic 46,XY DSD respectively, putting forward ESR2 as a novel candidate gene for 46,XY DSD etiology. It is interesting to speculate about an unanticipated role for ER-β in early gonadal development, i.e. the development and stabilization of the bipotential gonad and/or its differentiation into a functional testis, for instance through the MAPK signaling pathway, as has been demonstrated for

MAP3K1 mutations.

PATIENTS AND METHODS

Study subjects

Laboratory and clinical investigations of cases 1-3 are summarized in Table 1. Cases and/or parents provided written informed consent for the study. Detailed case reports can be found in Supplemental data, Case reports S1-S3.

Genetic studies

Homozygosity mapping was performed by genome-wide SNP genotyping in Case 1 and in two unaffected siblings, using the HumanCytoSNP-12 BeadChip platform

(Illumina). Plink software, integrated in ViVar, identified homozygous regions > 1 Mb which were ranked according to their length and the number of SNPs, as described16.

Only homozygous regions unique to the index case were selected. The presence of relevant known DSD genes was excluded using Biomart filtering. Next, different gene prioritization tools (Endeavour, GeneDistiller, GeneWanderer, PosMed, SUSPECTS and ToppGene) were used to identify candidate genes in the selected regions36. For this purpose, known DSD genes were used as training genes (Supplemental data,

Table S1).

Sanger sequencing and targeted next-generation sequencing. Primers were developed for PCR amplification and sequencing of all coding exons, intron-exon boundaries and

148 untranslated regions of ESR2 (Supplemental data, Table S2). Sanger sequencing (Case

1 and 2) was done on the ABI 3730XL DNA Analyzer (Applied Biosystems) using the

BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems) followed by data- analysis with SeqScape v2.5 (Life Technologies). Targeted NGS (Case 3) with a MiSeq instrument was performed using a workflow as described37. CLC Genomics

Workbench (CLCBio, Aarhus, Denmark) was used for the mapping of the reads and data-analysis of the variants. Segregation analysis for the ESR2 variant found in Case

1 was performed in the available family members. Apart from the mother, no family members were available for further analysis in case 2. For case 3, no parental DNA was available for segregation analysis.

Whole exome sequencing was performed to exclude the presence of other possibly pathogenic variants in Cases 1 and 2. Exome enrichment was performed with the

Nextera Rapid Capture Exome kit (Illumina), followed by paired-end sequencing on

HiSeq 2000 (2x100 cycles) (Illumina). Reads were mapped against the human reference genome sequence (NCBI, GRCh37/hg19) with the CLC Genomics Workbench v6.4

(CLC bio) followed by post-mapping duplicate read removal, coverage analysis, and quality-based variant calling. More thorough variant annotation was executed with the Alamut-HT v1.1.5 software (Interactive Biosoftware) and complementary literature search after extensive screening (Supplementary Table S3-5).

Structural analysis of the ER-β variants

Models for the wild type ER-β DNA binding domain were built using the standard homology modeling procedure in YASARA structure (http://yasara.org/products). We used the structure of the ER-β bound to a DNA response element (pdb code 1HCQ) as a template38. We built models for dimers of the wild type DNA binding domain with and without DNA. A model for the p.(Asn181del) (N181del) mutation (found in Case

1) was built in YASARA structure based on the wild type model, by deleting residue

N181, introducing a new peptide bond between residues 180 and 182, followed by energy minimization (em_runclean protocol in YASARA structure). The homodimer

149 models for the wild type and N181del mutant DNA binding domains were further subjected to 500 ps of molecular dynamics with the YASARA2 forcefield, followed by energy minimization using the md_refine procedure39.

Using FoldX and the estrogen-bound ER-β ligand binding domain (LBD) structure with pdb code 3OLL, we calculated an in silico estimate of the Gibbs energy of protein folding for the mutant and wild type LBD using the RepairPDB and BuildModel commands18,40. For each mutant, we report the average ΔΔG for folding of mutant versus wild type of 20 BuildModel calculations. The 3OLL protein structure was visualized in UCSF Chimera17. The effect of the p.(Leu426Arg) (L426R) mutation

(found in Case 3) was visualized by replacing L426 with the best fitting arginine rotamer in the Dunbrack backbone dependent rotamer library.

Immunohistochemistry

Tissue slides (4 µm) were prepared from a formalin-fixed, paraffin embedded eight- weeks old male fetus obtained after spontaneous abortion. Prior to immunohistochemistry a heat induced antigen retrieval (HIAR) step (Tris-EGTA pH9.00) was included. We used a standard ABC-protocol for detection of ER-β. After blocking of endogenous peroxidase and biotin (Avidin/Biotin Blocking Kit Catalog

Number: SP-2001), and overnight incubation with mouse anti-ER-β (dilution 1/50,

Dako M7292), detection was performed with a biotinylated rabbit anti-mouse antibody

(dilution 1/150, 30 minutes, Dako E413) followed by 30 minutes of incubation with

Avidin-Biotin-HRP labeled complex (VECTASTAIN Elite ABC HRP Kit Catalog

Number: PK-6100). Visualization was done with DAB/H2O2 and hematoxylin counterstaining.

Luciferase assays and localization studies

Plasmid construction. Plasmids containing the ESR2 variants were constructed with site-directed mutagenesis starting from a pcDNA-flag-ER-β plasmid (35562,

Addgene). Mutagenesis primers were designed using the NEBaseChanger software

150

(http://nebasechanger.neb.com/). After amplification, the plasmids were digested with

DpnI to remove non-mutated DNA. After verifying that mutations had correctly been inserted via sequence analysis, mutated plasmids were transformed in chemically competent E. coli (DH10b cells) and subsequently isolated with the NucleoBond Xtra

Midi/Maxi kit (Macherey-Nagel).

Cell culture and luciferase assays. HEK293T cells were cultured in Dulbecco’s minimal essential medium (DMEM) without phenol red, supplemented with 10% charcoal stripped fetal bovine serum (DCC-FBS) (Life Technologies) and 10 µg/ml Gentamycin at 37°C and 8% CO2. A total of 300 000 cells per well were plated in 6-well plates. Cells were transfected with a total of 2 µg of DNA using a standard calcium phosphate transfection method. After 24h, cells were washed with phosphate buffered saline

(PBS) and subsequently stimulated with 2,3-Bis(4-hydroxyphenyl)propionitrile (DPN) to a final concentration of 10 nM. Following lysis with 1x CCLR lysis buffer (25mM

Tris-phosphate pH7.8, 2mM DTT, 2mM CDTA, 10% Glycerol, 1% Triton X-100), luciferase activity measurements were done 24 hours after stimulation on a TopCount luminometer (Canberra-Packard). To normalize for differences in transfection efficiency, a plasmid constitutively expressing β-galactosidase was co-transfected in all experiments (Tropix Galacto-star system, Applied Biosystems). Each transfection was replicated in at least three independent experiments.

Subcellular localization experiments. Coverslips were coated with poly-L-Lysine and

150,000 cells were seeded per well in DMEM (without phenol red), supplemented with

10% DCC-FBS. Flag-tagged ER-β constructs were transfected as described for the luciferase assays and cells were stimulated with DPN after 4 hours of starvation. After

24 hours, cells were fixed using 4% paraformaldehyde, permeabilized with methanol and blocked with blocking buffer (20mM phosphate buffer, 100mM NaCl, 0.23% Triton

X-100, 10% donkey serum). This was followed by overnight incubation with a primary mouse anti-FLAG antibody (monoclonal anti-flag M2, Sigma #F3165, 20µg/ml). The next day, subsequently to washing steps with PBS with 0.1% TX-100, an anti-mouse secondary antibody (Alexa Fluor 488 donkey anti-mouse IgG H+L, 2µg/ml) (Life 151

Technologies) was added and nuclei were counterstained with DAPI. Visualization was done with a motorized inverted IX81 FluoView FV1000 laser scanning confocal microscope (Olympus) and three fields of on average 15 cells were scored.

ACKNOWLEDGEMENTS

We are most grateful to the family members who participated in this study. This work was supported by grants from the Research Foundation Flanders (FWO) (Bilateral

Scientific Cooperation Brazil NXT-DSDG0D6713N to E.D.B. and M.C.), by Ghent

University Special Research Fund (BOF15/GOA/011 to E.D.B.), by Belspo IAP project

P7/43 (Belgian Medical Genomics Initiative: BeMGI) to E.D.B, by Ghent University

Special Research Fund (BOF Starting Grant to M.C.). E.D.B. and M.C. are Senior

Clinical Investigators of the FWO.

REFERENCES

1. Ono M, Harley VR. Disorders of sex development: new genes, new concepts. Nat Rev Endocrinol. 2013;9(2):79-91. doi:10.1038/nrendo.2012.235. 2. Pearlman A, Loke J, Le Caignec C, et al. Mutations in MAP3K1 cause 46,XY disorders of sex development and implicate a common signal transduction pathway in human testis determination. Am J Hum Genet. 2010;87(6):898-904. doi:10.1016/j.ajhg.2010.11.003. 3. Bogani D, Siggers P, Brixey R, et al. Loss of mitogen-activated protein kinase kinase kinase 4 (MAP3K4) reveals a requirement for MAPK signalling in mouse sex determination. PLoS Biol. 2009;7(9):e1000196. doi:10.1371/journal.pbio.1000196. 4. Warr N, Bogani D, Siggers P, et al. Minor abnormalities of testis development in mice lacking the gene encoding the MAPK signalling component, MAP3K1. PLoS One. 2011;6(5):e19572. doi:10.1371/journal.pone.0019572. 5. Biason-lauber A. Control of sex development. Best Pract Res Clin Endocrinol Metab. 2010;24(2):163-186. doi:10.1016/j.beem.2009.12.002. 6. Uhlenhaut NH, Jakob S, Anlag K, et al. Somatic sex reprogramming of adult ovaries to testes by FOXL2 ablation. Cell. 2009;139(6):1130-1142. doi:10.1016/j.cell.2009.11.021. 7. Matson CK, Murphy MW, Sarver AL, Griswold MD, Bardwell VJ, Zarkower D. DMRT1 prevents female reprogramming in the postnatal mammalian testis Supp. . Nature. 2011;9(7358):1-5. doi:10.1038/nature10239. 8. Hughes IA, Houk C, Ahmed SF, Lee PA. Consensus statement on management of intersex disorders. Endocrine. 2006;(91):554-562. doi:10.1136. 9. Bashamboo A, Ledig S, Wieacker P, Achermann J, McElreavey K. New Technologies for the Identification of Novel Genetic Markers of Disorders of Sex Development ( DSD ). Sex Dev. 2010;15(4):213-224. doi:10.1159/000314917.

152

10. Baxter RM, Arboleda VA, Lee H, et al. Exome sequencing for the diagnosis of 46, XY disorders of sex development. J Clin Endocrinol Metab. 2015;100(2):E333-E344. doi:10.1210/jc.2014-2605. 11. Cools M, Wolffenbuttel KP, Drop SLS, Oosterhuis JW, Looijenga LHJ. Gonadal development and tumor formation at the crossroads of male and female sex determination. Sex Dev. 2011;5(4):167-180. doi:10.1159/000329477. 12. van der Zwan YG, Biermann K, Wolffenbuttel KP, Cools M, Looijenga LHJ. Gonadal maldevelopment as risk factor for germ cell cancer: towards a clinical decision model. Eur Urol. 2015;67(4):692-701. doi:10.1016/j.eururo.2014.07.011. 13. Ferraz-de-Souza B, Lin L, Achermann JC. Steroidogenic factor-1 (SF-1, NR5A1) and human disease. Mol Cell Endocrinol. 2011;336(1-2):198-205. doi:10.1016/j.mce.2010.11.006. 14. Mongan N, Tadokoro-Cuccaro R, Bunch T, Hughes I. Androgen insensitivity syndrome. Best Pract Res Clin Endocrinol Metab. 2015;29:569-580. doi:10.3109/09513590.2012.705378. 15. Orekhova a. S, Rubtsov PM. DAX1, an unusual member of the nuclear receptor superfamily with diverse functions. Mol Biol. 2015;49(1):65-76. doi:10.1134/S0026893315010124. 16. Coppieters F, Van Schil K, Bauwens M, et al. Identity-by-descent-guided mutation analysis and exome sequencing in consanguineous families reveals unusual clinical and molecular findings in retinal dystrophy. Genet Med. 2014;16(9):671-680. doi:10.1038/gim.2014.24. 17. Pettersen EF, Goddard TD, Huang CC, et al. UCSF Chimera - A visualization system for exploratory research and analysis. J Comput Chem. 2004;25(13):1605-1612. doi:10.1002/jcc.20084. 18. Guerois R, Nielsen JE, Serrano L. Predicting changes in the stability of proteins and protein complexes: A study of more than 1000 mutations. J Mol Biol. 2002;320(2):369-387. doi:10.1016/S0022-2836(02)00442-4. 19. VSN International. No Title. Genstat for Windows 18th Edition. Genstat.co.uk. 20. Ban S, Sata F, Kurahashi N, et al. Genetic polymorphisms of ESR1 and ESR2 that may influence estrogen activity and the risk of hypospadias. Hum Reprod. 2008;23(6):1466-1471. doi:10.1093/humrep/den098. 21. Beleza-Meireles A, Omrani D, Kockum I, Frisén L, Lagerstedt K, Nordenskjöld A. Polymorphisms of estrogen receptor β gene are associated with hypospadias. J Endocrinol Invest. 2006;29:5-10. 22. Beleza-Meireles A, Kockum I, Lundberg F, Söderhäll C, Nordenskjöld A. Risk factors for hypospadias in the estrogen receptor 2 gene. J Clin Endocrinol Metab. 2007;92(9):3712-3718. doi:10.1210/jc.2007-0543. 23. van der Zanden LFM, van Rooij I a LM, Feitz WFJ, Franke B, Knoers NV a M, Roeleveld N. Aetiology of hypospadias: a systematic review of genes and environment. Hum Reprod Update. 2012;18(3):260-283. doi:10.1093/humupd/dms002. 24. van der Zanden LFM, van Rooij I a LM, Feitz WFJ, et al. Genetics of hypospadias: are single- nucleotide polymorphisms in SRD5A2, ESR1, ESR2, and ATF3 really associated with the malformation? J Clin Endocrinol Metab. 2010;95(5):2384-2390. doi:10.1210/jc.2009-2101. 25. Asadi M, Ghafouri-Fard S, Zare-Abdollahi D, Ebrahim-Habibi a, Matin N. Estrogen receptor mutation in a girl with primary amenorrhea. Clin Genet. 2013;83(5):497-498. doi:10.1111/cge.12083. 26. Drummond AE, Fuller PJ. The importance of ERbeta signalling in the ovary. J Endocrinol. 2010;205(1):15-23. doi:10.1677/JOE-09-0379.

153

27. Marino M, Galluzzo P, Ascenzi P. Estrogen signaling multiple pathways to impact gene transcription. Curr Genomics. 2006;7(8):497-508. http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=2269003&tool=pmcentrez&render type=abstract. 28. Loke J, Pearlman A, Radi O, et al. Mutations in MAP3K1 tilt the balance from SOX9/FGF9 to WNT/β-catenin signaling. Hum Mol Genet. 2014;23(4):1073-1083. doi:10.1093/hmg/ddt502. 29. Ostrer H. Disorders of Sex Development (DSDs): An Update. J Clin Endocrinol Metab. 2014;99(5):1503-1509. doi:10.1210/jc.2013-3690. 30. Pettersson K, Gustafsson JA. Role of Estrogn Receptor Beta in Estrogen Action. Annu Rev Physiol. 2001;64:65-92. 31. Segars JH, Driggers PH. Estrogen action and cytoplasmic signaling cascades. Part I: membrane- associated signaling complexes. Trends Endocrinol Metab. 2002;13(8):349-354. http://www.ncbi.nlm.nih.gov/pubmed/12217492. 32. Driggers PH, Segars JH. Estrogen action and cytoplasmic signaling pathways. Part II: the role of growth factors and phosphorylation in estrogen signaling. Trends Endocrinol Metab. 2002;13(10):422-427. http://www.ncbi.nlm.nih.gov/pubmed/12431838. 33. Krege JH, Hodgin JB, Couse JF, et al. Generation and reproductive phenotypes of mice lacking estrogen receptor beta. Proc Natl Acad Sci U S A. 1998;95(26):15677-15682. http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=28103&tool=pmcentrez&renderty pe=abstract. 34. Swain A, Narvaez V, Burgoyne P, Camerino G, Lovell-badge R. Dax1 antagonizes Sry action in mammalian sex determination. Nature. 1998;391:761-767. 35. Coppieters F, Leroy BP, Beysen D, et al. Recurrent mutation in the first zinc finger of the orphan nuclear receptor NR2E3 causes autosomal dominant retinitis pigmentosa. Am J Hum Genet. 2007;81(1):147-157. doi:10.1086/518426. 36. Tranchevent L-C, Capdevila FB, Nitsch D, De Moor B, De Causmaecker P, Moreau Y. A guide to web tools to prioritize candidate genes. Brief Bioinform. 2011;12(1):22-32. doi:10.1093/bib/bbq007. 37. De Leeneer K, Hellemans J, Steyaert W, et al. Flexible, scalable, and efficient targeted resequencing on a benchtop sequencer for variant detection in clinical practice. Hum Mutat. 2015;36(3):379-387. doi:10.1002/humu.22739. 38. Schwabe JWR, Chapman L, Finch JT, Rhodes D. The crystal structure of the estrogen receptor DNA-binding domain bound to DNA: How receptors discriminate between their response elements. Cell. 1993;75(3):567-578. doi:10.1016/0092-8674(93)90390-C. 39. Krieger E, Joo K, Lee J, et al. Improving physical realism, stereochemistry, and side-chain accuracy in homology modeling: Four approaches that performed well in CASP8. Proteins Struct Funct Bioinforma. 2009;77(SUPPL. 9):114-122. doi:10.1002/prot.22570. 40. Möcklinghoff S, Rose R, Carraz M, Visser A, Ottmann C, Brunsveld L. Synthesis and crystal structure of a phosphorylated estrogen receptor ligand binding domain. Chembiochem. 2010;11(16):2251-2254. doi:10.1002/cbic.201000532.

154

SUPPLEMENTAL DATA

SUPPLEMENTAL NOTES: Case reports

Case report S1. Clinical case report of Case 1.

Case 1 was referred to the pediatric endocrinology service at the age of 15 because of absence of pubertal development. She was born after an uneventful pregnancy, birth weight was 2850 g. She has two healthy sisters, family history was unremarkable but both parents originated from the same village, suggesting a potential consanguineous background (Figure 1B). Anal atresia, a rectovestibular fistula and a vaginal dimple were observed at birth, external genitalia appeared female. Anal reconstructive surgery was performed eight days after birth and at the same time sigmoidovaginoplasty was performed. Retrospective data on the basis of ultrasound

(US) imaging regarding the presence of Müllerian structures are inconclusive. The girl was additionally diagnosed with blepharophimosis and ptosis for which she underwent oculoplastic surgery at the age of three. She developed nystagmus, severe degenerative myopia and left esotropia, and was operated for retinal detachment at 18 years. At the time of evaluation (15 years), dysmorphism was observed (Table 1).

Breast and pubic hair were concordant with Tanner stages B1 and P3 respectively and skeletal maturation was markedly delayed. Breast development under estradiol supplementation reached a small sized Tanner 5 after 2 years and at follow-up consultations. Laboratory work-up showed normal electrolytes and renal function and elevated liver enzymes (about 2-3 times) of unknown etiology at repeated occasions.

Hepatitis, CMV, EBV serology and autoimmune and tumor markers were negative.

Abdominal US revealed minimal hepatomegaly. A pelvic US was repeated at the age of 10 years, no gonads could be visualized at that occasion but instead, a big cystic lesion was seen in the right adnexal region, which was confirmed by pelvic magnetic resonance imaging. Excision biopsy of the lesion revealed a serous cyst containing

Wolffian duct remnants, liver biopsy at that time was unremarkable and no gonads were found at laparoscopy. Hormonal investigation suggested absence of functional

155 gonads, results are summarized in Table 1. Karyotyping and FISH revealed an SRY positive 46,XY karyotype. In view of the ocular manifestations, sequencing and copy number analysis of the FOXL2 gene was performed but revealed no mutation. Gender identity developed as female and puberty was induced at 15 years with 17-β-estradiol in increasing doses.

Case report S2. Clinical case report of Case 2.

Case 2 is a 20-year old woman from Bosnian origin, born with atypical genitalia

(clitoromegaly, urogenital sinus, bilateral palpable gonads in the groins). Birth weight was 3550 g, birth length was 52 cm. A female sex was assigned at birth. She underwent gonadectomy, vulvovaginoplasty and clitoris reduction at the age of seven years, histology of the gonads revealed dysgenetic testicular tissue. Unfortunately, no medical records or gonadal specimen from this period are available; the family moved to Belgium when she was ten years old. Upon arrival, US showed absence of a uterus.

The girl had no dysmorphic features and had a normal psychomotor development.

Gender identity was female and from 12.5 years onwards, puberty was induced, first with ethinylestradiol, then with 17-β-estradiol in increasing doses. Bone age at the start of puberty induction was 11.5 years and had increased with one year after three years of hormone supplementation. Gonadotropins were mostly within the high-normal follicular range. In spite of good compliance with medication, breast development was poor and the patient sought breast augmentation surgery at the age of 18 years.

Vaginoscopy performed at the age of 16 years revealed a 5 cm long blind ending vagina. Vaginal dilation therapy was proposed but the patient preferred intercourse as a first-line approach. She started university studies and withdrew from further medical follow-up at the age of 19 years. Karyotyping and FISH revealed an SRY positive 46,XY karyotype. No mutations were found in the AR and NR5A1 genes.

156

Case report S3. Clinical case report of Case 3.

Case 3 is a 24 years old Brazilian woman who was referred at the age of 24 for primary amenorrhea. Her parents were first cousins. There were no remarkable medical events during infancy or at birth. The woman had no dysmorphic features and had a normal psychomotor development. At first evaluation, she had breast development Tanner 1, female external genitalia and pubic hair Tanner 4. Her karyotype was 46,XY and her hormonal profile was characterized by high serum gonadotropins and undetectable testosterone and estrogen levels, compatible with a diagnosis of complete gonadal dysgenesis. Pelvic US showed an infantile uterus and no visible gonads. Gonadectomy was performed and bilateral dysgenetic gonads were confirmed by histology analysis.

The left gonad had a 1 cm gonadoblastoma in combination with germ cell neoplasia in situ.

157

SUPPLEMENTAL TABLES

Table S1: List of DSD genes used as training genes for gene prioritization in IBD-regions

AKR1C1 CHD7 DHCR7 FGFR1 GPR54 LEP MAP3K1 PROP1 SRD5A2

AKR1C2 CTNNB1 DHH FKBP4 HESX1 LEPR NELF PSMC3IP SRD5A2

AKR1C4 CYB5 DMRT1 FOG2 HHAT LGR8 NOBOX RSPO1 SRY

AMH CYP11A1 EAP1 FOG2 HOXA13 LHCGR NR5A1 SEMA3A STAR

AMHR2 CYP11B1 EMX1 FOXL2 HSD17B1 LHX1 NR5A2 SOX2 TSPYL1

AR CYP17A1 EMX2 FSHR HSD17B3 LHX3 NR0B1 SOX3 WDR11

ARX CYP19A1 FAM58A FSHβ HSD3B2 LHX4 PBX1 SOX8 WNT4

ATRX CYP1B2 FBLN2 GATA4 INSL3 LHX9 POR SOX8 WT1

BMP15 CYP21A2 FGF8 GATA4 KAL1 LHβ PROK2 SOX9 WWOX

CBX2 DAX1 FGF9 GnRHR KISS1 MAMLD1 PROKR2 SRD5A1

158

Table S2: Primer sequences for PCR amplification and sequencing of ESR2.

Exon Forward 5'->3' Reverse 5'-> 3' UCSC Length

ESR2_1 ccgggagggaggacagttt cgggtgacaaaatccagact chr14:64804890-64805346 457 bp

ESR2_2 cgctggaaggagtgtcagagc accgtcagaaacggcttgta chr14:64804548-64804996 449 bp

ESR2_3 cgctcccacttagaggtcacg ctcaacacgactggtttctca chr14:64760670-64761230 561 bp

ESR2_4 cggatgtatttgtaatctcatacaaacg tttgggtgctgtaagtaaactatg >chr14:64749208-64749908 701 bp

ESR2_5 cggaaccctggaggaaaaaga gtgccaaacaggccagtagt >chr14:64746596-64746989 394 bp

ESR2_6 ctcagctggaaagtgttatcagtg ctcttttcccggctacctgt >chr14:64735424-64735724 301 bp

ESR2_7 ccccacaggctccagaaaata tctaggcacagctcatggac >chr14:64727063-64727557 495 bp

ESR2_8 cttgcgcagcttaacttcaaa cagcacccaggactttgttc >chr14:64723863-64724263 401 bp

ESR2_9 ctcagatgaacatgttacaagatgaaa cgaagtccaaaaggaaacca >chr14:64716134-64716512 379 bp

ESR2_10 ccaaagcaacccagatcacct ttgcagacacttttcccaaa >chr14:64701598-64701974 377 bp

ESR2_11 cgtggggtagactggctctga tcaaatgtgccctctgctaa >chr14:64699642-64700142 501 bp

ESR2_12 tgcagctcacaaaagtcctg catggattacaatgatcccaga >chr14:64694167-64694742 576 bp

ESR2_13 ctcggtggcctaaagaaaaa ctcaactgatctgcccacct >chr14:64693640-64694238 599 bp

159

Table S3: Whole Exome Sequencing (WES) data analysis in Case 1 and Case 2. After quality control, variants with a coverage (cov) below 5 and a minor allele frequency above 5% were discarded. Variants were filtered and listed based on their coding effect. Only splicing variants in the 20 bp proximity of regular splice sites (SS) were included. Three different strategies were used: (1) variants in known Disorders of Sex Development (DSD) genes, (2) variants in homozygous regions (only for patient 1) en (3) variants in the entire exome.

DSD genes (heterozygous and homozygous) Case 1 Case 2

Nonsense variants (cov>5, MAF<0.05) 0 0

Frameshift variants (cov>5, MAF<0.05) 2 1

In-frame variants (cov>5, MAF<0.05) 3 4

Missense variants (cov>5, MAF< 0.05) 18 10

Splicing variants (cov>5, MAF<0.05, nearest SS: +- 20) 4 6

Homozygous regions (homozygous)

Nonsense variants (cov>5, MAF<0.05) 0 /

Frameshift variants (cov>5, MAF<0.05) 1 /

In-frame variants (cov>5, MAF<0.05) 4 /

Missense variants (cov>5, MAF< 0.05) 22 /

Splicing variants (cov>5, MAF<0.05, nearest SS: +- 20) 4 /

Whole exome (heterozygous and homozygous)

Nonsense variants (cov>5, MAF<0.05, VAF>0.2) 44 9

Frameshift variants (cov>5, MAF<0.05, VAF>0.2) 116 76

In-frame variants (cov>5, MAF<0.05, VAF>0.2) 115 128

Missense variants (cov>5, MAF<0.05) 2527 1016

Splicing variants (cov>5, MAF<0.05, VAF>0.2, nearest SS: +- 20) 204 157

160

Table S4: Sanger validation of variants found by Whole Exome Sequencing (WES) in Case 1. Variants that passed all filtering steps were confirmed with Sanger sequencing. When a variant was present in the patient’s father, no confirmation was done. *: variant present in other exomes (variant excluded without confirmation sequencing) +: variant with benign prediction in SIFT, Polyphen and Mutation Taster (variant excluded without confirmation sequencing) °: variant with a high MAF (3%

Gene Coverage VAF Reference sequence: cDNA notation Protein notation Comments Confirmed

SRD5A2 151 95 NM_000348.3:c.93dup p.Gly32Argfs*104 / Not confirmed

MAP3K1 56 32 NM_005921.1:c.2845_2847del p.Thr949del HeZ father° No Sanger

CYP21A2 74 34 NM_000500.6:c.26_28del p.Leu9del rs61338903, *° No Sanger

TSPYL1 42 60 NM_003309.3:c.526_528dup p.Val176dup rs78371471, ExAC 0.7628° No Sanger

ESR2 9 100 NM_001040275.1:c.541_543del p.Asn181del / Yes

HOXA13 5 20 NM_000522.4:c.869A>C p.Tyr290Ser / Not confirmed

POR 5 20 NM_000941.2:c.568T>G p.Tyr190Asp / Not confirmed

NOBOX 5 20 NM_001080413.3:c.1292G>T p.Ser431Ile / Not confirmed

CYP11B1 36 30 ENST00000377675.3:c.1370C>T p.Ala457Val rs4541+° No Sanger

CYP11B1 27 63 ENST00000377675.3:c.593C>A p.Thr198Lys rs9657020+ No Sanger

NR5A1 5 20 NM_004959.4:c.707A>C p.Asp236Ala / Not confirmed

LHX3 5 20 NM_014564.3:c.643G>C p.Ala215Pro / Not confirmed

LHX1 5 60 NM_005568.4:c.616A>C p.Thr206Pro rs201087752, HeZ father No Sanger

AR 5 20 NM_000044.3:c.343C>A p.Gln115Lys /* No Sanger

161

AR 5 20 NM_000044.3:c.412G>T p.Ala138Ser /* No Sanger

ATRX 5 20 NM_000489.3:c.18G>T p.Met6Ile ExAC: 0.00003096 No Sanger

SOX3 5 20 NM_005634.2:c.1174G>T p.Ala392Ser / Not confirmed

SOX3 5 20 NM_005634.2:c.760G>T p.Val254Leu / Not confirmed

MAMLD1 5 20 NM_001177465.2:c.1109G>A p.Ser370Asn / Not confirmed

GPHB5 12 100 NM_145171.2:c.156dup p.Phe53Leufs*79 HoZ F, HoZ M No Sanger

TMIE 28 75 NM_147196.2:c.391_393del p.Lys131del HoZ M, benign No Sanger

ZNF816 37 84 NM_001031665.1:c.552_553insAGT p.Ile184_Gly185insSer / No Sanger

ZBTB8A 15 100 NM_001040441.1:c.1253G>C p.Gly418Ala rs704886, HoZ F No Sanger

NBPF14 497 83 NM_015383.1:c.1711C>T p.Arg571Cys rs11554534+ No Sanger

NBPF16 11 100 NM_001102663.1:c.1844A>C p.Gln615Pro rs6695216, HoZ F No Sanger

NBPF16 12 83 NM_001102663.1:c.1847C>T p.Ser616Leu rs75121840, HoZ F No Sanger

ANO2 32 96 NM_020373.2:c.2077A>G p.Ile693Val + No Sanger

FAM90A1 52 100 NM_018088.3:c.1043C>T p.Thr348Ile rs9668474, HoZ F No Sanger

KIAA0586 20 100 NM_014749.3:c.2300T>C p.Leu767Pro rs1748986, HoZ F No Sanger

KIAA0586 6 100 NM_014749.3:c.2938C>G p.Pro980Ala rs1617510, HoZ F No Sanger

TRMT5 10 100 NM_020810.2:c.649T>C p.Ser217Pro rs7142228+ No Sanger

AKAP5 23 100 NM_004857.3:c.608C>T p.Thr203Ile rs1256149, HoZ F No Sanger

GPX2 37 97 NM_002083.2:c.377C>T p.Pro126Leu rs17881652, HeZ F, HoZ No Sanger reported SIPA1L1 23 100 NM_015556.1:c.119G>A p.Arg40Gln rs78621209, HeZ F, HoZ No Sanger reported

162

ABCD4 14 100 NM_005050.3:c.1042C>T p.Arg348Trp rs147795328, HoZ F+ No Sanger

YLPM1 5 80 NM_019589.2:c.2431A>G p.Met811Val rs200333188, HoZ F No Sanger

MLH3 159 100 NM_001040108.1:c.2476A>G p.Asn826Asp rs175081, HoZ F No Sanger

ZSCAN29 11 100 NM_152455.3:c.595G>A p.Gly199Ser rs3917221, HoZ reported + No Sanger

PATL2 15 100 NM_001145112.1:c.262A>G p.Met88Val rs8026845, HoZ F No Sanger

ZNF611 28 96 NM_001161499.1:c.754_755delinsAT p.Pro252Ile / No Sanger

163

Table S5: Validation of variants found by Whole Exome Sequencing (WES) in known DSD genes in Case 2. Variants that passed all filtering steps were confirmed with Sanger sequencing. *: variant present in other exomes (variant excluded without confirmation sequencing) +: variant with benign prediction in SIFT, Polyphen and Mutation Taster (variant excluded without confirmation sequencing) °: variant with a high MAF (3%

Gene Coverage VAF Reference sequence: cDNA Protein notation Comments Confirmed notation SRD5A2 151 95 NM_000348.3:c.93dup p.Gly32Argfs*104 / Not confirmed

FAM58A 4 100 NM_152274.3:c.55_56insG p.Gln19Argfs*39 / Not confirmed

FAM58A 3 100 NM_152274.3:c.20dup p.Gly8Argfs*50 / Not confirmed

LHCGR 6 66 NM_000233.3:c.50_55dup p.Leu17_Gln18dup rs376653903° No Sanger

MAP3K1 612 78 NM_005921.1:c.2845_2847del p.Thr949del rs5868032 * Yes

CYP21A2 33 45 NM_000500.6:c.26_28del p.Leu9del rs61338903° No Sanger

TSPYL1 286 83 NM_003309.3:c.526_528dup p.Val176dup rs78371471 * Yes

LHCGR 73 46 NM_000233.3:c.872A>G p.Asn291Ser rs12470652 +° No Sanger

MAP3K1 670 5+ NM_005921.1:c.2816C>G p.Ser939Cys rs45556841 +° No Sanger

CYP21A2 51 35 NM_000500.6:c.305G>A p.Arg102Lys rs6474 +° No Sanger

CYP21A2 60 23 NM_000500.6:c.800C>T p.Pro267Leu rs142028935 +° No Sanger

NOBOX 106 37 NM_001080413.3:c.1991A>G p.Lys664Arg rs77802098 +° No Sanger

LHX1 145 47 NM_005568.4:c.1162G>T p.Ala388Ser rs145694637 +° No Sanger

164

Table S6: Overview of the homozygous regions larger than 1 Mb resulting from homozygosity mapping in Case 1.

Chromosome SNP start SNP stop Start (hg19) Stop (hg19) Length(bp)

1 rs35595182 rs655552 32,183,736 33,194,672 1,010,936

1 rs12117013 rs3855959 45,107,235 46,633,874 1,526,639

1 rs507630 rs12041774 77,149,207 81,482,945 4,333,738

1 rs1532399 rs7517566 147,381,444 15,0850,035 3,468,591

2 rs6729421 rs745007 30,575,898 34,272,097 3,696,199 SRD5A2

2 rs4449134 rs2848274 88,846,961 90,109,161 1,262,200

3 rs13325398 rs9874474 45,161,070 49,326,178 4,165,108

4 rs1504496 rs10517174 45,951,996 46,982,826 1,030,830

6 rs4502981 rs2474878 61,891,118 63,457,733 1,566,615

8 rs7839348 rs1562634 91,421,822 92,458,202 1,036,380

9 rs11795294 rs10907973 40,421,411 42,345,595 1,924,184

11 rs11038867 rs11570034 46,357,159 47,375,889 1,018,730

12 rs2041377 rs2719020 4,984,601 8,410,202 3,425,601

12 rs10770704 rs35939864 20,855,761 23,167,969 2,312,208

12 rs2204849 rs12831116 38,094,338 39,401,456 1,307,118

13 rs7320567 rs4150834 113,123,999 114,296,219 1,172,220

14 rs1189277 rs11623512 56,951,496 77,199,730 20,248,234 ESR2

15 rs12595426 rs12907793 26,029,193 27,982,953 1,953,760

15 rs1863601 rs12101736 43,495,212 46,289,459 2,794,247

165

15 rs985868 rs10083634 71,915,278 73,165,722 1,250,444

19 rs2290745 rs7260411 52,909,303 54,444,972 1,535,669

Table S7: Output of gene prioritization in the largest homozygous region on chromosome 14 resulting from homozygosity mapping in Case 1. The different greens correlate with the amount of times a gene was highly ranked by the prioritization algorithms. ESR2 popped out 6 times, FOS 5 times, TGFB3 4 times, PSEN1 and ESRRB both popped out 2 times.

Rank PosMed Endeavour ToppGene GeneWanderer SUSPECTS GeneDistiller

1 ESR2 ESR2 ESR2 FOS ESR2 KCNH5

2 SGPP1 ESRRB FOS PSEN1 FOS PARP1P2

3 PSEN1 PSEN1 ESRRB MED6 ESRRB RHOJ

4 LTBP2 TGFB3 TGFB3 ESR2 TGFB3 GPHB5

5 FOS CHX10 MAX EIF2S1 LTBP2 PPP2R5E

6 VSX2 SFRS5 RBM25 SLC10A1 ADAM20 RPL31P5

7 GPHN ZBTB25 VSX2 TGFB3 CHX10 GCATP1

8 RGS6 MAX MED6 FUT8 PGF WDR89

9 MTHFD1 FOS DPF3 DLST RDH12 HSPE1P2

10 NPC2 MTHFD1 EIF2B2 EIF2B2 ADAM21 SGPP1

11 ABCD4 EIF2S1 SRSF5 RDH11 RDH11 EIF2S2P1

12 SLC10A1 ACTN1 ACTN1 RDH12 TM21_Human RPS28P1

13 GPX2 ZFP36L1 C14orf169 ATP6V1D SLC39A9 SYNE2

14 PGF ENSG00000140044 CHURC1 MLH3 C14orf114 MIR548H1

15 HSPA2 C14orf1 C14orf43 MTHFD1 HSPA1 ESR2

16 EIF2S1 BATF SIPA1L1 AKAP5 MTHFD1 TEX21P

17 RHOJ NEK9 YBX1P1 BATF ENTPD5 MTHFD1

18 ARG2 DLST|DLSTP RAP1AP SYNJ2BP SLC8A3 LOC100506245

19 GPHB5 RBM25 ADAM21P1 NUMB CHURC1 AKAP5

20 MAP3K9 COX7A2P1 HSPA2 PCNX LOC100506266

166

Table S8: Detailed overview of the identified ESR2 mutations

Variant cDNA Protein Conservation Amino acid CADD-score SIFT PolyPhen-2 Mutation Grantham Public databases

conservation Taster distance

1 c.541_543del p.(Asn181del) Nucleotide: / Highly / / / / / dbSNP: rs750091675

conserved: ESP: Eur. Am.: C=0.04% Zebrafish

ExAC: ALL:0.041%, no

homozygous cases

reported

GnomAD: ALL:

0.0368%, no

homozygous cases

reported

2 c.251G>T p.(Gly84Val) Moderately Highly 26.9 Tolerated Probably Disease 109 Absent in dbSNP, ESP

conserved: conserved: damaging causing and ExAC

phyloP: 4.22 Fruitfly GnomAD: ALL:

0.000396% (singleton)

3 c.1277T>G p.(Leu426Arg) Highly Highly 23.4 Deleterious Probably Disease 102 Absent in dbSNP, ESP

conserved: conserved: damaging causing and ExAC, GnomAD

phyloP: 8.70 Tetraodon

167

SUPPLEMENTAL FIGURES

Figure S1: Immunohistochemical analysis of ER-β in an 8-week old male embryo. Brown staining indicates ER-β expression. (A- B) No expression is seen in the developing testis. (A: 10x, B: 20x) (C-D) Positive staining is present in mucosa of the developing gut. (C: 1.25x, D: 10x) (E-F) Developing eye, especially the corneal epithelium and endothelium, primitive lens neurons, primitive retinal cells and the retinal pigment epithelium. (E: 5x, F: 20x) (G) Developing brain, specifically the primitive neurons of the germinal matrix (20x) (H) Placental tissue positive control for ER-β-expression. (20x) (I) Overview of the entire embryo (0.5x).

168

Figure S2: Subcellular localization of wild-type and mutant (p.(Asn181del) and p.(Gly84Val)) ER-β proteins using a fluorescent anti-ER-β antibody. The left panel shows the localization of wild type (WT) ER-β, the middle and right panel show the unaltered subcellular localization of both mutant ER-β proteins. 169

170

CHAPTER 4: 46,XX DSD

CHAPTER 4:

46,XX171 DSD

172

NR5A1NR5A1 IS A IS NOVEL A NOVEL DISEASE DISEASE GENE FOR GENE 46,XX FORTESTICULAR 46,XX AND TESTICULAR OVOTESTICULAR AND OVOTESTICULAR DISORDERS OF SEX DEVELOPMENT DISORDERS OF SEX DEVELOPMENT

Genetics in Medicine (2016)

Dorien Baetens1, Hans Stoop2, Frank Peelman3, Anne-Laure Todeschini4, Toon

Rosseel1, Frauke Coppieters1, Reiner A. Veitia1, Leendert H.J. Looijenga2, Elfride De

Baere1, Martine Cools5

1Center for Medical Genetics Ghent, Ghent University and Ghent University Hospital, Ghent, Belgium

2Department of Pathology, Erasmus MC, University Medical Center Rotterdam, Rotterdam 3015 AA,

The Netherlands

3 Flanders Institute for Biotechnology (VIB), Department of Medical Protein Research, Faculty of

Medicine and Health Sciences, Ghent University, Ghent, Belgium

4Molecular Oncology and Pathology, Institut Jacques Monod, France; Université Paris Diderot, Paris

VII, France

5Department of Pediatric Endocrinology, Ghent University and Ghent University Hospital, Ghent,

Belgium

173

174

ABSTRACT

Purpose: We aimed to identify the genetic cause in a cohort of 11 unrelated cases and two sisters with 46,XX SRY-negative (ovo)testicular disorders of sex development

(DSD).

Methods: Whole-exome sequencing (n = 9), targeted resequencing (n = 4), and haplotyping were performed. Immunohistochemistry of sex-specific markers was performed on patients’ gonads. The consequences of mutation were investigated using luciferase assays, localization studies, and RNA-seq.

Results: We identified a novel heterozygous NR5A1 mutation, c.274C>T p.(Arg92Trp), in three unrelated patients. The Arg92 residue is highly conserved and located in the

Ftz-F1 region, probably involved in DNA-binding specificity and stability. There were no consistent changes in transcriptional activation or subcellular localization.

Transcriptomics in patient-derived lymphocytes showed upregulation of MAMLD1, a direct NR5A1 target previously associated with 46,XY DSD. In gonads of affected individuals, ovarian FOXL2 and testicular SRY-independent SOX9 expression observed.

Conclusions: We propose NR5A1, previously associated with 46,XY DSD and 46,XX primary ovarian insufficiency, as a novel gene for 46,XX (ovo)testicular DSD. We hypothesize that p.(Arg92Trp) results in decreased inhibition of the male developmental pathway through downregulation of female antitestis genes, thereby tipping the balance toward testicular differentiation in 46,XX individuals. In conclusion, our study supports a role for NR5A1 in testis differentiation in the XX gonad.

Keywords: gonadal development; NR5A1; ovotesticular DSD; testicular DSD; 46,XX

DSD

175

176

INTRODUCTION mutations of female sex-determining

Although several genetic mechanisms genes can give rise to 46,XX DSD. This is leading to 46,XY disorders of sex illustrated by biallelic loss-of-function development (DSD) have been mutations in R-Spondin 1 (RSPO1), elucidated in recent years, little is which have been reported in a rare known about the developmental syndrome characterized by 46,XX pathways that may induce testicular testicular DSD, palmoplantar development in 46,XX individuals when hyperkeratosis, and susceptibility to dysregulated. Approximately 80% of all development of squamous cell

10,11 46,XX testicular DSD cases are explained carcinoma . Interestingly, by a translocation of the Sex Determining intrafamilial phenotypical variability

Region Y (SRY) gene1,2. Duplications of and incomplete penetrance have been

SRY-box 9 (SOX9) or its regulatory described, ranging from unaffected region are a rare cause of 46,XX mutation carriers to cases with 46,XX testicular DSD3,4. Copy number testicular DSD or ovotesticular DSD, variations disrupting the regulatory indicating that these conditions may region of SRY-box 3 (SOX3), a putative represent a phenotypic spectrum of the

12-14 ancestor of SRY, have also been found in same underlying defect . Here, we patients with 46,XX DSD5,6. In line with aimed to identify the underlying this, 46,XX DSD phenotypes have been molecular genetic cause in 11 unrelated associated with partial duplications of individuals and 2 sisters with 46,XX human chromosome 22q13 containing testicular and ovotesticular DSD; the

SRY-box 10 (SOX10)7,8. Studies in mice latter was defined as the combined overexpressing Sox10 strongly implicate presence of testis and ovarian tissue in gain-of-function of SOX10 in 46,XX the same individual. In three of them,

DSD9. In addition to these gain-of- we identified a novel and recurrent function changes of male sex- heterozygous variant, c.274C>T determining genes, loss-of-function p.(Arg92Trp), in the Nuclear Receptor family 5 subfamily A member 1 gene 177

(NR5A1, AD4BP). The impact of this variant in exon 4 of NR5A1 c.274C>T variant was further studied by p.(Arg92Trp) in cases 1–3. Prediction structural remodeling of the affected programs SIFT, Polyphen-2, and amino acid, luciferase and localization MutationTaster suggest a possible effect studies in cellular systems, RNA-seq on on protein function. The patient-derived lymphocytes, and physicochemical distance between Arg immunohistochemistry of patients’ and Trp is moderate (Grantham score of gonads. 101) and the variant is absent in genomic

databases such as dbSNP, ESP, ExAC, RESULTS GoNL, and 1000 Genomes and in an in- Novel and recurrent missense mutation in house exome database. The affected NR5A1 in three patients with 46,XX amino acid and surrounding region are (Ovo)Testicular DSD highly conserved up to zebrafish and is Eleven unrelated patients and two 15 part of the Ftz-F1 region (Figure 1B) . sisters with 46,XX ovotesticular or Segregation analysis of the three testicular DSD were included in this families showed incomplete penetrance study. Nine underwent WES and four in four 46,XX family members (Figure underwent resequencing of the coding 1A). In case 1, the mutation was region of NR5A1 because they were paternally inherited (family 1: II:2); in included at a later time. When exome the other two families, the mutation was data were available, all variants were transmitted by the mother (family 2: I:2; filtered against a list of known 46,XX family 3: I:2). The gynecological history and 46,XY DSD genes (Supplementary for these unaffected women as well as Data, Supplementary Table S2), for the older sister of case 1, who also annotated, and manually curated. All carries the mutation (family 1: III.1), was variants in DSD genes, identified in unremarkable (Supplementary Data, cases 1–9, are listed in Supplementary Supplementary Table S12). Data and Supplementary Tables S3–S11. Microsatellite analysis showed a We identified a heterozygous missense haplotype of approximately 1.5 Mb 178

Figure 1: Pedigrees and haplotypes of mutation-positive families and schematic representation of the NR5A1 protein, mutation location, and conservation. (A) Pedigrees for families 1–3. Affected individuals heterozygous for NR5A1 mutation c.274C>T p.(Arg92Trp) are indicated by filled symbols. Asymptomatic individuals with NR5A1 mutation p.(Arg92Trp) are represented by a black spot in the symbol. In total, 29 microsatellite markers were used to assess the haplotypes; only the ones building up the common haplotype and flanking markers are shown here. The common haplotype (in red box) shared by the three mutation carriers has a maximum size of 1.5 Mb. (B) The different domains found in NR5A1, a DNA-binding domain at the C-terminus, followed by the Ftz-F1 region, which is important for DNA-binding stability, a hinge region, and a ligand-binding domain. An electropherogram showing the c.274C>T substitution is depicted below. The protein alignment of p.(Arg92Trp) and its surrounding amino acids is shown on top. The affected amino acid and the surrounding region are conserved up to zebrafish.

179 shared among cases 1–3 (Figure 1A), a deoxycytosine base, no hydrogen pointing to a founder effect. Given the bonds are formed by the mutant side incomplete penetrance, we assessed the chains. The (enthalpic) in silico binding

WES data for other variants that might energy for interaction of the DNA- contribute to the phenotype. Variant binding domain of NR5A1 with the lists with de novo variants for case 1 and DNA segment is 9% lower for both rare variants (minor allele frequency < mutants. This suggests that both

0.01) for cases 2 and 3 were added to the mutations can affect the interaction of

Supplementary Data and NR5A1 with its response elements,

Supplementary Tables S13–S16. although effects may be variable and

specific for different response elements Protein structure modeling of NR5A1 p.(Arg92Trp) and p.(Arg92Gln) (Figure 2).

We performed protein structural Transcriptional activity and subcellular modeling, starting from a model of localization of NR5A1 mutants p.(Arg92Trp) and p.(Arg92Gln) NR5A1 bound to the α-inhibin promoter, for the following mutations: We assessed the impact of following p.(Arg92Trp) found here in the three mutations on transcriptional cases with 46,XX DSD and p.(Arg92Gln) activity and subcellular localization in

16 previously found in 46,XY DSD . FoldX different cellular systems: c.274C>T calculations predict that the mutations p.(Arg92Trp) and the previously do not affect the stability of the isolated reported adjacent mutation c.275G>A

17 NR5A1 DNA-binding domain . In the p.(Arg92Gln). For the latter mutation, models of mutants bound to DNA, the partial loss of DNA-binding and tryptophan and glutamine side chains transcriptional activity have previously

16 do not penetrate as deeply into the been shown . We cotransfected WT or minor groove as does the wild-type mutagenized NR5A1 plasmids with arginine side chain. Although the Arg92 different NR5A1 responsive promoters side chain makes a hydrogen bond with (TESCO, SOX9, AMH, CYP11B1) in

180

Figure 2: Structural analysis of NR5A1 (PDB code 2FF0). Position of Arg92 (R92) in the wild- type NR5A1 DNA-binding domain (DBD). (A) R92 is part of the C-terminal A-box and its side chain fits in the minor groove, making a hydrogen bond with a deoxycytosine base. Models of the p.(Arg92Trp) (Trp=W) (B) and p.(Arg92Gln) (Gln=Q) (C) mutants show how this mutated side chain binds in the DNA minor groove without hydrogen bond formation to the DNA. The affected amino acid R92 is located in a region C-terminal to the classical DBD with its two zinc (Zn) fingers. This additional domain is found in family members of NR5A1, including NGFI-B, and was called the A-box or Ftz-F1 box18. This A-box domain has been implicated in interactions with the minor groove of the target DNA. Prior to structure determination, binding studies with NR5A1 and NGFI-B mutants showed that this interaction is important for their binding specificity toward different DNA targets. For NR5A1, this specificity is determined by amino acids 91–93, which are predicted to interact with three specific basepairs 5ʹ of the classical half-site bound by the Zn fingers. The results of this study are in line with those for the later NMR structure of the NR5A1 DBD bound to a DNA fragment of the inhibin-alpha subunit promoter. In this structure, the side chain of R92 inserts deeply into the minor groove, and its guanidine head group binds to two of the three basepairs at the 5ʹ of classical half-sites. Mutation R92A abolished DNA-binding and signaling via the inhibin- alpha subunit promoter in this study18 by removing the interaction with the basepairs of the minor groove. The p.(Arg92Trp) mutation is not allowed in the structure without gross structural rearrangements; the W side chain cannot insert into the minor groove and cannot be accommodated in the current DNA-bound structure without inducing steric clashes 181 different cell lines (HeLa, KGN). These expression analysis with default settings experiments were repeated with FOXL2 showed a total of 1,420 genes with cotransfection because it was shown significant differential expression previously that FOXL2 inhibits NR5A1- (corrected value < 0.05) when mediated SOX9 expression during comparing the patient and control

19 female development . The results of groups (Supplementary Data, these assays were inconclusive, and only Supplementary Table S17). This gene set the results for the TESCO-luc constructs was filtered against a list of DSD- are shown (Supplementary Data, associated genes (Supplementary Data,

Supplementary Figure S1). These Supplementary Table S2), the coverage experiments reproduced the previously of which was low in general. This described lower transcriptional activity filtering resulted in three (LEPR, of the c.275G>A p.(Arg92Gln) variant. MAMLD1, and IRF2BPL) genes with

The c.274C>T p.(Arg92Trp) variant, differential expression (Supplementary however, led to decreased activity in Data, Supplementary Table S18).

HeLa cells but not in KGN cells, MAMLD1, a direct target gene of showing that the transcriptional activity NR5A120 that was previously reported of a variant might be influenced by the to be involved in 46,XY DSD21, was cellular environment. Subcellular upregulated (log 2-fold change = 1.637, localization assays in HeLa cells showed = 0.03) in patients compared with nuclear expression without aggregate controls. BioGPS, an online available formation for both WT and mutant expression database, shows expression

NR5A1 (Supplementary Data, in both testis and ovaries in mice. LEPR

Supplementary Figure S2). and IRF2BPL (EAP1) are genes involved

in hypogonadotropic hypogonadism, Transcriptome analysis in patients’ lymphocytes which we considered not directly relevant in the context of the phenotypes RNA-seq generated an average of 39 observed here. million reads per sample. Differential

182

Figure 3: Immunohistochemistry on gonadal biopsies. (A) Case 1. Testis. Supportive cells have differentiated as Sertoli cells as indicated by SOX9 positivity. SOX9 staining, 200×. (B) Case 1. Scarce germ cells (stained in brown) populate the testis tubules. DDX4 staining, 200×. (C) Case 2. Ovotestis. Sertoli cells in the testicular component (right) express SOX9; no SOX9 expression is seen in the ovarian part (left). SOX9 staining, 200×. (D) Same gonad. Ovarian granulosa cells but not testicular Sertoli cells display FOXL2 positivity. FOXL2 staining, 200×. (E,F) Germ cells (ova, black arrows) are only present in the ovarian and not in the testicular part. DDX4 staining, 200×. (G) Case 3. In this patient, bilateral infantile testes with a Sertoli cell–only pattern were observed. All Sertoli cells are SOX9-positive. SOX9 staining, 200×. (H) Same gonad. FOXL2 staining is negative. FOXL2, 200×. (I) Case 3. DDX4 staining indicates absence of germ cells, thus confirming the Sertoli cell–only pattern. DDX4 staining, 200×

Immunostaining of SOX9, FOXL2, and DDX4 was found. Biopsy specimens for case 3 on patients’ gonads revealed bilateral immature testes with

Gonadal biopsy specimens from cases 1 Sertoli cell–only tubules (i.e., no germ and 3 and a gonadectomy specimen cells present). Case 2 had bilateral from case 2 were available. In case 1, a ovotestes, including the presence of dysgenetic testis with scarce germ cells, primordial and primary follicles in the as identified by DDX4 immunostaining, ovarian part, located at the pole of the

183 gonad, and next to immature shown for p.(Arg92Gln), suggesting a

16 seminiferous tubules without germ loss-of-function effect . This mutation cells, located in the central testicular part was also recently identified in a 46,XX of both gonads. Immunohistochemistry female with adrenal insufficiency but no

22 with SOX9 and FOXL2 antibodies, gonadal fenotype . Subsequently, identifying Sertoli and granulosa cell numerous NR5A1 mutations and copy differentiation, respectively, revealed number variations have been associated exclusive SOX9 expression in testicular with isolated 46,XY gonadal dysgenesis, parts in all three cases and exclusive 46,XY undervirilization, and male

23-26 FOXL2 expression in ovarian parts of infertility . In 46,XX individuals, these case 2 (Figure 3). Hematoxylin and eosin loss-of-function mutations may cause

27 staining of ovarian tissue, removed after primary ovarian insufficiency (POI) . ovarian torsion at age 41 years in the This broad range of phenotypes mother of case 2, was unremarkable and emphasizes that correct NR5A1 showed an aging ovary with some functioning is essential for both male residual primordial follicles and stromal and female gonadal development and tissue (Supplementary Data, maintenance.

Supplementary Figure S3). Few data exist regarding the NR5A1

DISCUSSION expression pattern in human female and

A few NR5A1 mutations have been male fetuses. In humans, NR5A1 is identified in females with 46,XY upregulated in the undifferentiated complete gonadal dysgenesis and stage and gonadal expression is

16 maintained during the entire period of adrenal failure . Interestingly, one of these was an adjacent homozygous fetal development in both male and mutation, c.275G>A p.(Arg92Gln), in the female fetuses, whereas in rodents it has same codon as the c.274C>T been shown that Nr5a1 expression p.(Arg92Trp) mutation identified here. decreases in the ovary as sex

28,29 Lower transcriptional activity was differentiation progresses . Indeed,

184

NR5A1 plays an important role in been shown to result from differences in

37 multiple stages of gonadal development spatiotemporal expression . Of note, the because it is a key regulator of sex geographical distribution of testicular determination via SOX9 upregulation and ovarian regions in the ovotestes of during male development and of sex case 2 supports the involvement of differentiation in both sexes via analogous (temporospatial) upregulation of AMH and the different mechanisms and corresponds to

15,30,31 37 steroidogenic enzymes . Both male observations in the B6 XYpos mouse . and female Nr5a1 knockout (KO) mice During normal ovarian development, lack gonadal and adrenal FOXL2 actively represses SOX9, thereby 32,33 development . Recently, a inhibiting the male developmental CRISPR/Cas9 amino acid substituted 38 cascade . During male development, it mouse model was generated for is well established that NR5A1, together p.(Arg92Trp); however, no phenotypic with other transcription factors, binds to data were provided34. the regulatory domain of SOX9 thereby

Here, the NR5A1 mutation p.(Arg92Trp) inducing a positive feedback loop in was found in three unrelated patients which SOX9 maintains and enhances its with 46,XX (ovo)testicular DSD. A own expression, resulting in Sertoli cell

30 potential founder effect was suggested recruitment and differentiation . by haplotype analysis. The mutation NR5A1 loss-of-function mutations lead was found in XX unaffected individuals to a decrease in activation of SOX9 and in each of the three families, suggesting other transcriptional targets, resulting in incomplete penetrance, as has been undervirilization and 46,XY DSD. Our observed in other NR5A1-associated initial hypothesis that p.(Arg92Trp)

35,36 phenotypes and in other large leads to more stable binding of NR5A1 families with 46,XX (ovo)testicular to the SOX9 promoter, thereby directly

12,13 DSD . In a 46,XY DSD mouse model inhibiting FOXL2-mediated repression

(B6 XYpos), incomplete penetrance has of , as described by Uhlenhaut et

185

19 al. , is not supported by the cellular levels, leading to transient activation of

39 transactivation assays performed. This the male pathway in XX gonads . It is is possibly because the adult cell lines conceivable that in our patients, that were used are not representative for p.Arg92Trp interferes with this early early stages of gonadal development. In ovary-promoting function of NR5A1, addition, transcriptome analysis of ultimately resulting in insufficient patient-derived lymphocytes did not and/or delayed upregulation of FOXL2 reveal an upregulation of early male- and, thus, (partial) escape from its testis- specific genes. However, it did show suppressing activity. In our patients, upregulation of another NR5A1 target, NR5A1 is directly affected, which might

MAMLD1. Whether this finding explain the permanent and more severe represents a direct effect resulting from gonadal phenotype as compared to the p.(Arg92Trp) or whether it is secondary mouse model. Given the fact that the to the male environment from which mutation affects the Ftz-F1 domain, cells were taken is currently unclear. which is important for DNA-binding

specificity and stabilization, an altered A more attractive hypothesis is that expression of specific targets might be NR5A1, apart from its well-known role expected. In addition, this hypothesis in inducing testicular development, also fits with the incomplete penetrance and triggers the activation of early ovarian phenotypic heterogeneity, as seen in our genes such as WNT4, a pathway that families, depending on timing and ultimately leads to FOXL2 activation spatial thresholds. and definitive suppression of the

NR5A1-SOX9 tandem. Repression of the Two animal models support this

Nr5a1 cofactor Cited2 in XX mice entails hypothesis. It was shown very recently reduced Nr5a1 activity, which has been that CRISPR/Cas9-induced Nr5a1 shown to result in decreased Wnt4, mutations cause female-to-male sex

Rspo1, and Foxl2 levels and increased reversal in the XX Nile Tilapia. Nr5a1 mesonephric cell migration and Fgf9 depletion leads to reduced Cyp19A1

186

41 expression and low estradiol levels. development similarly to the fish

Because ovarian differentiation in fish is model. However, if testis formation in mediated by estradiol, the lack of sex this species is, similarly to the Nile steroid production here was proposed Tilapia, mediated by reduced Nr5a1 as the possible trigger for changing the activity, then its influence on the female

40 initial gonadal fate . Temporal gonadal pathway remains to be expression studies of Nr5a1 and Foxl2 in elucidated. These two alternative this model underscore the critical hypotheses may explain the different importance of threshold levels of Foxl2 phenotypes observed for the two to downregulate Nr5a1 in the XX gonad. adjacent NR5A1 mutations. Mutation

A second animal model is the polled c.275G>A (p.Arg92Gln) has a direct intersex syndrome (PIS) goat, which is negative influence on SOX9 expression characterized by absence of horns in and leads to aberrant male gonadal

16 heterozygous and homozygous animals development and adrenal failure . and female-to-male sex reversal in Variant p.Arg92Trp has no direct effect homozygous XX animals. This on SOX9 expression but is hypothesized phenotype is caused by the 11.7-kb to interfere with hitherto unrecognized regulatory deletion that is located 300 kb female-promoting activity of NR5A1, upstream of the goat FOXL2 gene. ultimately resulting in loss of stable

Knockout experiments have shown that SOX9 repression by FOXL2. Both

FOXL2 depletion leads to reduced mutations are located in the same codon

CYP19A1 expression and that it is that is part of the Ftz-F1 box, which is responsible for inhibiting ovarian

187

Figure 4: Schematic overview of the sex development gene regulation network. Blue: testis- promoting activity of NR5A1. Orange: counteractive connections to suppress the opposite pathway. Green: ovary-promoting activity of NR5A1 as hypothesized here and supported by other works39,40. (A) General scheme: NR5A1 is known to initiate the male developmental pathway through upregulation of SOX9 (synergistically with SRY). SOX9 then maintains its own expression via Fgf9 signaling. In the female embryo, in the absence of SRY, NR5A1 induces WNT4 and RSPO1 expression as shown by Combes, leading to upregulation of FOXL2 and other ovary-specific genes. High FOXL2 expression results in stable repression of SOX9 and, possibly, NR5A1, and hence the male pathway. (B) Possible mechanism by which c.274C

188 important for DNA-binding specificity. quantitative changes disturbing the

The newly introduced amino acid might balance between the male and female determine which NR5A1 targets are pathways in the XX gonad may shift the affected by the mutation and which ovarian developmental program to the pathways are perturbed. Structural testicular pathway, probably through models show that both mutations have autocrine and paracrine reinforcing

37 problems entering the DNA minor signals (Figure 4). The exact groove, although they cannot predict mechanisms behind these processes which targets will be affected. Our need to be substantiated by additional results suggest that temporal, spatial, or experimental work.

CONCLUSION

Taking these findings together, we propose NR5A1 as a novel disease gene for 46,XX

(ovo)testicular DSD. Our study suggests SRY-independent SOX9 expression at the gonadal (testicular) level, most likely not through the canonical NR5A1-SOX9 interaction, but rather through downregulation of the pro-ovarian Wnt4/β-catenin pathways, thus tipping the balance toward male development. In addition, we demonstrated that testicular and ovotesticular DSD may represent different phenotypes resulting from a common cause. We conclude that NR5A1, a well- established gene for male gonadal development, is also involved in correct female gonadal development, the exact molecular mechanisms of which are still elusive.

189

MATERIALS AND METHODS

Subjects

Case 1, the second child of two healthy, non-consanguineous parents, was born after an uneventful pregnancy. Mild hypertrophy of the clitoris was noticed at birth.

Vaginoscopy and laparoscopy detected the presence of a cervix and possibly a right hemi-uterus, a left epididymis, and male vasculature. Both inguinal gonads were biopsied. The left gonad revealed testicular differentiation with germ cells present; the right gonad contained only fibrotic tissue. A human chorionic gonadotropin (hCG) stimulation test (1,500 IU administrated intramuscularly with blood sampling after 72 h) was performed on the seventh day after birth, resulting in a testosterone level of

2.52 nmol/l (baseline: 0.72 nmol/l); anti-Müllerian hormone (AMH) was between the male and female references (99.9 pmol/l). Based on these results and the child’s phenotype, a female sex was assigned. Microarray-based comparative genomic hybridization (array CGH) showed a normal female pattern, and fluorescence in situ hybridization (FISH) with SRY-specific probes excluded SRY translocation. The results of quantitative polymerase chain reaction for the regulatory RevSex locus, upstream of SOX9, and MLPA for different sex development–related genes (P185 Intersex, MRC

Holland, Amsterdam, The Netherlands) were normal.

Case 2 is a 22-year-old woman who had been diagnosed with 46,XX ovotesticular DSD.

She had ambiguous genitalia at birth and a short, blind-ending vagina. At age 4, she underwent clitoroplasty and bilateral gonadectomy of abdominal gonads, which both turned out to be ovotestes, including primordial and primary follicles in the ovarian part and Sertoli cell–only tubules in the testicular part. SRY-specific FISH was negative, and array CGH revealed a small duplication reported to be a benign variant

(hg19: chr 8: 132161619 – 132812388). RevSex locus-specific quantitative polymerase chain reaction was normal. She has one younger sister, who displayed normal puberty; family history was unremarkable.

190

Case 3 is a 23-year-old man with 46,XX testicular DSD. At birth, a micropenis, penile hypospadias, and bilateral scrotal but atrophic testes were noticed. Gonadal biopsies at the age of 2 years revealed bilateral testicular differentiation with immature seminiferous tubules without germ cells and limited intertubular fibrosis. When he was 12 years old, a prolonged hCG test (2 × 1,500 IU per week for 3 consecutive weeks) led to a peak increase of serum testosterone to 6.0 nmol/l. Hypergonadotrophic hypogonadism developed rapidly thereafter, and testosterone supplementation was started at the age of 13. FISH with SRY and TSPY probes was negative, and array CGH and quantitative polymerase chain reaction for the RevSex locus were normal. He is the oldest of three brothers; familial history is unremarkable. All three patients are

Caucasian. Cases 1 and 3 are from East-Flanders in Belgium, and case 2 originates from the south of the Netherlands, close to the Belgian-Dutch border. All families live near

(within 70 km) each other. Pedigrees of the three families are shown in Figure 1a and phenotypic data for cases 1–3 are summarized in Table 1; phenotypic data for cases 4–

9 are presented in the Supplementary Data (Supplementary Table S1).

Table 1: Phenotypic data cases 1-3.

Case 1 Case 2 Case 3

Clinical Mild clitoris hypertrophy Ambiguous genitalia, Micropenis, hypospadias

phenotype blind-ending vagina Scrotal, atrophic testes at birth

Gonadal L: infantile testis L and R: ovotestes L and R: infantile testes

phenotype R: fibrous streak

Germ cells Scarce Yes, in ovarian region No

191

Genetic FISH SRY: negative FISH SRY: negative FISH SRY: negative

testing Array CGH: normal Array CGH: normal Array CGH: normal

female pattern female pattern, no female pattern

pathogenic CNVs found

Family Father, Unaffected mother and Unaffected mother and

members sister two brothers paternal uncle, paternal carrying the grandfather mutation

Unaffected sister

Genetic study

Whole-exome sequencing (WES) was performed in nine patients with 46,XX

(ovo)testicular DSD; in four additional patients, included later, the coding region of

NR5A1 was sequenced. Enrichment for WES was performed with the SureSelectXT

Human All Exon V5 kit (Agilent), followed by paired-end sequencing on HiSeq 2000

(2 × 100 cycles) (Illumina). Reads were mapped against the human reference genome sequence (NCBI, GRCh37/hg19) with the CLC Genomics Workbench v6.4 (Qiagen), followed by postmapping duplicate-read removal, coverage analysis, and quality- based variant calling. More thorough variant annotation was executed with Alamut-

HT v1.1.5 and Alamut Visual v2.7 software (Interactive Biosoftware). For case 1, because we had access to parental DNA, a trio-sequencing strategy was used. Variant filtering was performed with the Ingenuity Variant Analysis platform (Qiagen).

Sanger sequencing was used to confirm potentially pathogenic variants identified by

WES and to perform segregation analysis. Primers were designed with primer3plus.

Sequencing was performed on the ABI 3730XL DNA Analyzer (Applied Biosystems) using the BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems), followed by data analysis with SeqScape v2.5 (Life Technologies). Microsatellite

192 analysis with 29 markers (deCODE, Généthon, and Marshfield) was performed to assess disease haplotypes in the three mutation-positive cases. Data analysis was performed with the GeneMapper v3.7 software (Applied Biosystems).

In silico modeling of the p.(Arg92Trp) and p.(Arg92Gln) mutations

All modeling and calculations were performed using model 1 of the NMR structure of

18 the NR5A1 DNA binding domain (PDB code: 2ff0) . The effect of the mutations on the stability of the protein was calculated using the FoldX RepairPDB and BuildModel

17 commands, with 20 replicate calculations . Models for p.(Arg92Trp) and, for comparison, p.(Arg92Gln) mutants of the NR5A1 DNA-binding domain in complex with DNA were built using the YASARA structure with the swap and optimize commands to replace R92 by a tryptophan or glutamine, followed by energy

42 minimization with the YASARA forcefield in explicit solvent . The in silico binding energy between the NR5A1 LBD and the DNA fragments was calculated using the

42 YASARA BindEnergy command . Models and structures were visualized via UCSF chimera43.

Plasmid construction

We constructed a GFP-tagged NR5A1 wild-type (WT) construct starting from the

Gateway pcDNA-DEST47 vector (Invitrogen, Life Technologies) and a WT NR5A1 open reading frame (ORF)-containing entry clone using a Gateway LR-recombination reaction (Invitrogen, Life Technologies). The mutations undergoing study were inserted using the Q5 site-directed mutagenesis kit (New England Biolabs), and mutation-specific primers were designed with the NEBaseChanger software (New

England Biolabs). Mutated plasmids were transformed in One Shot TOP10 competent cells (Invitrogen, Life Technologies) and subsequently isolated with the NucleoBond

Xtra Midi/Maxi kit (Macherey-Nagel). The entire NR5A1 insert and the surrounding backbone were sequenced and selected plasmids were grown to obtain larger quantities.

193

Luciferase assays

Luciferase assays were performed in two cell lines: HeLa and KGN. Cells were grown in DMEM-F12 medium supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin. One day before transfection, 5,000 cells/well were seeded in a

96-well plate. Each experiment was performed in six independent replicates. Cells were transfected with 187 ng of a DNA mixture using calcium phosphate transfection, after which they were washed with phosphate-buffered saline and ethylene glycol tetra-acetic acid to remove remaining precipitates. After 48 h, cells were washed with phosphate-buffered saline again before lysis and luciferase activities were assessed with the Dual-Glo Luciferase Assay System (Promega). Results are reported as relative light units (R.L.U., i.e., the ratio of the Renilla luciferase to firefly luciferase activities).

Each value is the mean of six independent experiments, and the standard error bar represents the standard error of the mean.

Subcellular localization

Cells were seeded in 24-well plates containing sterile coverslips at a concentration of

30,000 cells per well 1 day before transfection. A total of 1,250 ng DNA per well was transfected using the calcium phosphate method. At 48 h after transfection, cells were washed with phosphate-buffered saline and fixed with 4% paraformaldehyde at room temperature for 15 min. Nuclei were stained with Hoechst (1/500) and coverslips were mounted on slides with Vectashield mounting medium (Vector Laboratories).

Subcellular localization was assessed using fluorescence microscopy. All transfections were performed in duplicate, and at least 100 cells were counted per transfection.

RNA-seq

RNA-seq was performed on cultured lymphocytes from two patients (cases 1 and 3) and four female control samples. Lymphocytes were isolated from EDTA blood using

Lymphoprep (Stemcell Technologies), cultured in Roswell Park Memorial Institute medium with 10% fetal bovine serum, and substituted with interleukin-2 and

194 phytohemagglutinin. RNA was isolated with the QiaAmp RNeasy Mini-Kit (Qiagen), followed by sample preparation with the TruSeq Stranded mRNA sample preparation kit (Illumina), subsequent cluster generation, and single-end sequencing with the

NextSeq 500 High Output Kit V2 (75 cycles) (Illumina). To obtain equally large patient and control groups, patient samples were run in duplicate. First, all reads were

44 mapped against the human GRCh38 reference genome with TopHat (v.2.1.0) , followed by transcriptome assembly with Cufflinks (v.2.2.1)45. Subsequently, a read

46 count matrix was generated using HTSeqcount (v.0.6.1p1) . Trimmed mean of M- values normalization and differential gene expression were assessed with the Limma

47 software package . The statistical significance of each observed change in expression was tested between the patient group (two patient samples, two replicates per patient) and the control group (four control samples) with a false discovery rate threshold of

5%.

Immunohistochemistry

Slides with thicknesses of 4 μm were prepared from formalin-fixed (10%), paraffin- embedded tissue and pretreated for 5 min with 3% H2O2 to block endogenous peroxidase. After heat-induced antigen retrieval at pH 9, slides were blocked for biotin

(SP-2001, Vector Laboratories, Burlingame, CA). After overnight incubation at 4 °C with SOX9 (dilution 1/250, AF3075, R&D systems), FOXL2 (dilution 1/350, a gift from

Marc Fellous), and DDX4 primary antibody (dilution 1/100, Ab13840, Abcam), detection was performed by 30-min incubation at room temperature with a biotin- labeled secondary antibody, followed by another 30-min incubation with horseradish peroxidase (HRP)-conjugated Avidin-Biotin complex. Visualization was performed using 3,3-diaminobenzidine (DAB)/H202, (SOX9 and DDX4), and aminoethyl carbazole (AEC)/H202 (FOXL2); counterstaining was performed with hematoxylin.

General morphology was assessed on adjacent slides after hematoxylin and eosin staining.

195

SUPPLEMENTARY MATERIAL

Supplementary material is linked to the online version of the paper at http://www.nature.com/gim/journal/vaop/ncurrent/full/gim2016118a.html#suppleme ntary-information

ACKNOWLEDGMENTS

We are very grateful to the families who participated in this study. This work was supported by a grant from the Ghent University Special Research Fund

(BOF15/GOA/011) to E.D.B., by Belspo IAP project P7/43 (Belgian Medical Genomics

Initiative: BeMGI) to E.D.B., by the Hercules Foundation (AUGE/13/023 to E.D.B.), by a grant from the Ghent University Special Research Fund (BOF Starting Grant) to M.C., and by grant G0D6713N from the Research Foundation Flanders (FWO) to E.D.B. and

M.C. E.D.B. and M.C. are Senior Clinical Investigators of the FWO. F.C. is senior postdoctoral fellow of the FWO. We thank Francis Poulat for providing the TESCO- luc construct and Jan Tavernier for providing the NR5A1 ORF.

DISCLOSURE

The authors declare no conflict of interest.

REFERENCES

1. Ono, M. & Harley, V. R. Disorders of sex development: new genes, new concepts. Nat. Rev. Endocrinol. 9, 79–91 (2013). 2. Li, T.-F. et al. 46,XX testicular disorder of sexual development with SRY-negative caused by some unidentified mechanisms: a case report and review of the literature. BMC Urol. 14, 104 (2014). 3. Huang, B., Wang, S., Ning, Y., Lamb, a N. & Bartley, J. Autosomal XX sex reversal caused by duplication of SOX9. Am. J. Med. Genet. 87, 349–53 (1999). 4. Benko, S. et al. Disruption of a long distance regulatory region upstream of SOX9 in isolated disorders of sex development. J. Med. Genet. 48, 825–30 (2011). 5. Sutton, E. et al. Identification of SOX3 as an XX male sex reversal gene in mice and humans. Jounal Clin. Investig. 121, 328–341 (2011). 6. Haines, B. et al. Interchromosomal insertional translocation at Xq26.3 alters SOX3 expression in an individual with XX male sex reversal. J. Clin. Endocrinol. Metab. 100, E815–E820 (2015). 7. Seeherunvong, T. et al. 46,XX sex reversal with partial duplication of chromosome arm 22q. Am. J. Med. Genet. A 127A, 149–51 (2004). 8. Aleck, K. A., Argueso, L., Stone, J., Hackel, J. G. & Erickson, R. P. True Hermaphroditism With Partial Duplication of Chromosome 22 and Without SRY. Am. J. Hum. Genet. 4, 2–4 (1999).

196

9. Polanco, J. C., Wilhelm, D., Davidson, T.-L., Knight, D. & Koopman, P. Sox10 gain-of-function causes XX sex reversal in mice: implications for human 22q-linked disorders of sex development. Hum. Mol. Genet. 19, 506–16 (2010). 10. Parma, P. et al. R-spondin1 is essential in sex determination, skin differentiation and malignancy. Nat. Genet. 38, 1304–9 (2006). 11. Tomaselli, S. et al. Syndromic True Hermaphroditism Due to an R-spondin1 ( RSPO1 ) Homozygous Mutation. Hum. Mutat. 29, 220–226 (2008). 12. Slaney, S. F., Chalmers, I. J., Affara, N. A. & Chitty, L. S. An autosomal or X linked mutation results in true hermaphrodites and 46 , XX males in the same family. J Med Genet 35, 17–22 (1998). 13. Kuhnle, U., Schwarz, H. P., Li, U., Cleve, H. & Braun, A. Familial true hermaphroditism : paternal and maternal transmission of true hermaphroditism ( 46 , XX ) and XX maleness in the absence of Y-chromosomal sequences. Hum. Genet. 92, 571–576 (1993). 14. Ramos, E. S. et al. SRY-negative true hermaphrodites and an XX male in two generations of the same family. Hum. Genet. 97, 596–598 (1996). 15. Parker, K. L. & Schimmer, B. P. Steroidogenic Factor 1 : A Key Determinant of Endocrine Development and Function. Endocr. Rev. 18, 361–377 (1997). 16. Achermann, J. et al. Gonadal Determination and Adrenal Development Are Steroidogenic Factor-1 , in a Dose-Dependent Manner. J. Clin. Endocrinol. Metab. 87, 1829–1833 (1999). 17. Guerois, R., Nielsen, J. E. & Serrano, L. Predicting changes in the stability of proteins and protein complexes: A study of more than 1000 mutations. J. Mol. Biol. 320, 369–387 (2002). 18. Little, T. H. et al. Sequence-specific deoxyribonucleic acid (DNA) recognition by steroidogenic factor 1: a helix at the carboxy terminus of the DNA binding domain is necessary for complex stability. Mol. Endocrinol. 20, 831–43 (2006). 19. Uhlenhaut, N. H. et al. Somatic sex reprogramming of adult ovaries to testes by FOXL2 ablation. Cell 139, 1130–42 (2009). 20. Fukami, M. et al. Mastermind-like Domain-containing 1 (MAMLD1 or CXorf6) Transactivates the Hes3 Promoter, Augments Testosterone Production, and Contains the SF1 Target Sequence. J. Biol. Chem. 283, 5525–5532 (2007). 21. Kalfa, N. et al. Screening of mamld1 mutations in 70 children with 46,xy dsd: Identification and functional analysis of two new mutations. PLoS One 7, 1–7 (2012). 22. Guran, T. et al. Rare causes of primary adrenal insufficiency: Genetic and clinical characterization of a large nationwide cohort. J. Clin. Endocrinol. Metab. 101, 284–292 (2016). 23. Köhler, B. et al. Five Novel Mutations in Steroidogenic With Severe Underandrogenization But Without Adrenal Insufficiency. Hum. Mutat. 29, 59–64 (2008). 24. Barbaro, M., Cools, M., Looijenga, L. H. J., Drop, S. L. S. & Wedell, a. Partial deletion of the NR5A1 (SF1) gene detected by synthetic probe MLPA in a patient with XY gonadal disorder of sex development. Sex Dev. 5, 181–7 (2011). 25. Bashamboo, A. et al. Human male infertility associated with mutations in NR5A1 encoding steroidogenic factor 1. Am. J. Hum. Genet. 87, 505–12 (2010). 26. Köhler, B. et al. The spectrum of phenotypes associated with mutations in steroidogenic factor 1 (SF-1, NR5A1, Ad4BP) includes severe penoscrotal hypospadias in 46,XY males without adrenal insufficiency. Eur. J. Endocrinol. 161, 237–42 (2009). 27. Lourenco, D. et al. Mutations in NR5A1 associated with ovarian insufficiency. N. Engl. J. Med. 360, 1200–1210 (2009). 28. Hanley, N. A. et al. Expression of steroidogenic factor 1 and Wilms’ tumour 1 during early human gonadal development and sex determination. Mech. Dev. 87, 175–180 (1999). 29. Cools, M. et al. Pubertal androgenization and gonadal histology in two 46,XY adolescents with NR5A1 mutations and predominantly female phenotype at birth. Eur. J. Endocrinol. 166, 341–9 (2012).

197

30. Sekido, R. & Lovell-Badge, R. Sex determination involves synergistic action of SRY and SF1 on a specific Sox9 enhancer. Nature 453, 930–934 (2008). 31. Shen, W., Moore, C., Ikeda, Y., Parker, K. & Ingraham, H. Nuclear receptor steroidogenic factor 1 regulates the mullerian inhibiting substance gene: a link to the sex determination cascade. Cell 77, 651–661 (1994). 32. Luo, X., Ikeda, Y. & Parker, K. L. A cell-specific nuclear receptor is essential for adrenal and gonadal development and sexual differentiation. Cell 77, 481–490 (1994). 33. Sadovsky, Y. et al. Mice deficient in the orphan receptor steroidogenic factor 1 lack adrenal glands and gonads but express P450 side-chain-cleavage enzyme in the placenta and have normal embryonic serum levels of corticosteroids. PNAS 92, 10939–10943 (1995). 34. Inui, M. et al. Rapid generation of mouse models with defined point mutations by the CRISPR/Cas9 system. Sci. Rep. 4, 5396 (2014). 35. Camats, N. et al. Ten novel mutations in the NR5A1 gene cause disordered sex development in 46,XY and ovarian insufficiency in 46,XX individuals. J. Clin. Endocrinol. Metab. 97, E1294-306 (2012). 36. Baetens, D. et al. Extensive clinical, hormonal and genetic screening in a large consecutive series of 46,XY neonates and infants with atypical sexual development. Orphanet J. Rare Dis. 9, 209 (2014). 37. Wilhelm, D. et al. Antagonism of the testis- and ovary-determining pathways during ovotestis development in mice. Mech. Dev. 126, 324–36 (2009). 38. Veitia, R. a. FOXL2 versus SOX9: a lifelong ‘battle of the sexes’. Bioessays 32, 375–80 (2010). 39. Combes, A. N. et al. Gonadal defects in Cited2-mutant mice indicate a role for SF1 in both testis and ovary differentiation. Int. J. Dev. Biol. 54, 683–9 (2010). 40. Xie, Q.-P. et al. Haploinsufficiency of SF-1 Causes Female to Male Sex Reversal in Nile Tilapia, Oreochromis niloticus. Endocrinology 157, en.2015-2049 (2016). 41. Boulanger, L. et al. FOXL2 Is a Female Sex-Determining Gene in the Goat. Curr. Biol. 24, 404– 408 (2014). 42. Krieger, E. et al. Improving physical realism, stereochemistry, and side-chain accuracy in homology modeling: Four approaches that performed well in CASP8. Proteins Struct. Funct. Bioinforma. 77, 114–122 (2009). 43. Pettersen, E. F. et al. UCSF Chimera - A visualization system for exploratory research and analysis. J. Comput. Chem. 25, 1605–1612 (2004). 44. Trapnell, C., Pachter, L. & Salzberg, S. L. TopHat: Discovering splice junctions with RNA-Seq. Bioinformatics 25, 1105–1111 (2009). 45. Trapnell, C. et al. Transcript assembly and quantification by RNA-Seq reveals unannotated transcripts and isoform switching during cell differentiation. Nat. Biotechnol. 28, 511–515 (2010). 46. Anders, S., Pyl, P. T. & Huber, W. HTSeq – A Python framework to work with high-throughput sequencing data HTSeq – A Python framework to work with high-throughput sequencing data. Bioinformatics. 31, 0–5 (2014). 47. Ritchie, M. E. et al. Limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res. 43, e47 (2015).

198

SUPPLEMENTAL DATA

Figure S1 Results of transcriptional activation assays. Transcriptional activation was assessed for two different mutations in HeLa and KGN cells. Four different promoter constructs were used, here only the results for TESCO are shown. A second luciferase gene (Renilla luciferase) was cotransfected and used to normalize the luciferase intensities, that are shown as relative light units (R.L.U.). Mean values ± STD are shown. (*) statistically significant with a p-value < 0.05.

TESCO-luc, HeLa cells * * * * 7 * * 6 * 5

0 NLS FOXL2 NLS FOXL2 NLS FOXL2 NLS FOXL2 Mock NR5A1 WT NR5A1 274 NR5A1 275

TESCO-luc, KGN cells * 140 * 120 * 100

0 NLS FOXL2 NLS FOXL2 NLS FOXL2 NLS FOXL2 Mock NR5A1 WT NR5A1 274 NR5A1 275

199

Figure S2 Subcellular localization of wild type and mutant NR5A1. Subcellular localization of wild type (WT) and mutant NR5A1 proteins (p.Arg92Trp and p.Arg92Gln) in HeLa cells. DAPI (40,6-Diamidino-2-phenylindole) counterstain (blue) in the left panels shows staining of the nucleus. The middle panel shows the nuclear localization of GFP-fused NR5A1 protein, and the right panel is a merge. WT as well as mutated NR5A1 localize to the nucleus, no mislocalization or aggregates can be observed.

Figure S3 H&E-staining of individual I:2 of family 2. Hematoxylin and eosin staining of the distorted ovary of the mother of case 2 (individual I:2 of family 2), removed at age 41 years, showing typical ovarian differentiation with age-related depletion of follicles.

200

Table S1 Summary of the phenotypic characteristics of Cases 4-9.

Case 4 Case 5 Case 6 Case 7 Case 8 Case 9

Clinical phenotype at Presented at the age Presented at the age Presented at birth birth / first of 10 with proximal of 3 with clitoris with clitoris presentation hypospadias, breast hypertrophy, left hypertrophy, single development (M3), impalpable and right urogenital perineal left impalpable and labioscrotal gonad, opening, uterus right inguinal gonad . hemivagina and hemi- uterus

Gonadal phenotype R: ovotestis R: testis R: ovary R: NA R: ovotestis R:

L: ovary L: ovary L: testis L: ovotestis L: ovary L:

Germ cells Germ cells present in Germ cells present in Germ cells present in Germ cells present in Germ cells present in ovarian part ovarian part ovarian part ovarian part ovarian and testicular part

Genetic testing 46,XXSRYneg 46,XXSRYneg 46,XXSRYneg

Array-CGH: (also in gonads) Array-CGH: duplication chr12 duplication chr3 (128440338- Normal female array- (72772687-72915275), 128689778) CGH maternally inherited

201

Table S2 List of DSD genes used for filtering exome and RNA-seq data

AKR1C1 CHD7 DHCR7 FGFR1 GPR54 LEP MAP3K1 PROP1 SRD5A2 AKR1C2 CTNNB1 DHH FKBP4 HESX1 LEPR NELF PSMC3IP SRD5A2 AKR1C4 CYB5 DMRT1 FOG2 HHAT LGR8 NOBOX RSPO1 SRY AMH CYP11A1 EAP1 FOG2 HOXA13 LHCGR NR5A1 SEMA3A STAR AMHR2 CYP11B1 EMX1 FOXL2 HSD17B1 LHX1 NR5A2 SOX2 TSPYL1 AR CYP17A1 EMX2 FSHR HSD17B3 LHX3 NR0B1 SOX3 WDR11 ARX CYP19A1 FAM58A FSHβ HSD3B2 LHX4 PBX1 SOX8 WNT4 ATRX CYP1B2 FBLN2 GATA4 INSL3 LHX9 POR SOX8 WT1 BMP15 CYP21A2 FGF8 GATA4 KAL1 LHβ PROK2 SOX9 WWOX CBX2 DAX1 FGF9 GnRHR KISS1 MAMLD1 PROKR2 SRD5A1

Table S3 Overview of variants found in DSD genes in case 1

Gene Coverage VAF cDNA notation Protein notation Comments ExAC

SRD5A2 101 0,97 NM_000348.3:c.90dup p.(Ser31Leufs*105) rs61748121 0.1667

mother HoZ

FAM58A 2 1 NM_152274.3:c.17dup p.(Gly7Argfs*51) rs373552919 1

mother HoZ

MAP3K1 812 0,375 NM_005921.1:c.2845_2847d p.(Thr949del) rs5868032 0.7393 el

CYP21A2 22 0,636 NM_000500.7:c.29_31del p.(Leu10del) rs138498156 0.2934

TSPYL 351 0,84 NM_003309.3:c.526_528du p.(Val176dup) rs78371471 0.7628 p mother HoZ

CYP21A2 33 0,576 NM_000500.7:c.308G>A p.(Arg103Lys) rs6474 0.3341

NR5A1 151 0,483 NM_004959.4:c.274C>T p.(Arg92Trp) not present in genomic databases

HHAT 86 1 NM_001170587.1:c.1248+19 / rs4297293 0.9847 G>C mother HoZ

Abbreviations used: VAF: variant allele frequency; HoZ: homozygous.

202

Table S4 Overview of variants found in DSD genes in case 2

Gene Coverage VAF cDNA notation Protein notation Comments ExAC

SRD5A2 132 0,97 NM_000348.3:c.90dup p.(Ser31Leufs*105 rs61748121 0.1667 )

FAM58A 9 100 NM_152274.3:c.17dup p.(Gly7Argfs*51) rs373552919 1

FAM58A 9 100 NM_152274.3:c.54dup p.(Gln19Alafs*39) rs371528168 0.998

CYP21A2 28 0,43 NM_000500.7:c.29_31del p.(Leu10del) rs138498156 0.2934

TSPYL 347 0,82 NM_003309.3:c.526_528du p.(Val176dup) rs78371471 0.7628 p

MAP3K1 65 0,51 NM_005921.1:c.764A>G p.(Asn255Ser) rs56069227 0.0203

CYP21A2 53 0,68 NM_000500.7:c.308G>A p.(Arg103Lys) rs6474 0.3341

NR5A1 129 0,472 NM_004959.4:c.274C>T p.(Arg92Trp) not present in genomic databases

MAMLD 438 0,47 NM_001177465.2:c.2486C> p.(Ala829Val) rs186833799 0.0483 1 T

HHAT 76 1 NM_001170587.1:c.1248+19 ? rs4297293 0.9847 G>C

STAR 325 1 NM_000349.2:c.465+20A>C ? rs2720050 0.0343 (1000 genom es)

AKR1C2 108 0,50 NM_205845.2:c.447+17G>A ? rs61856066 0.1263

WWOX 155 0,51 NM_016373.3:c.108-12G>T ? rs67493355 0.286

203

Table S5 Overview of variants found in DSD genes in case 3

Gene Coverage VAF cDNA notation Protein notation Comments ExAC

KISS1 46 0,95 NM_002256.3:c.417del p.(*139Trpfs*8) rs71745629 0.2983

SRD5A2 101 0,96 NM_000348.3:c.90dup p.(Ser31Leufs*105 rs61748121 0.1667 )

FAM58A 3 0,33 NM_152274.3:c.17dup p.(Gly7Argfs*51) rs373552919 1

FAM58A 2 1 NM_152274.3:c.54dup p.(Gln19Alafs*39) rs371528168 0.998

MAP3K1 555 0,81 NM_005921.1:c.2845_2847d p.(Thr949del) rs5868032 0.7393 el

TSPYL 255 0,83 NM_003309.3:c.526_528du p.(Val176dup) rs78371471 0.7628 p

MAP3K1 65 0,51 NM_005921.1:c.764A>G p.(Asn255Ser) rs56069227

CYP21A2 81 0,37 NM_000500.6:c.185A>T p.(His62Leu) rs9378252 100

NR5A1 124 0,444 NM_004959.4:c.274C>T p.(Arg92Trp) not present in genomic databases

AKR1C3 3 0,33 NM_001253908.1:c.754C>A p.(Pro252Thr) / /

BMP15 32 0,5 NM_005448.2:c.308A>G p.(Asn103Ser) rs41308602 0.0875

FAM58A 3 0,33 NM_152274.3:c.55C>G p.(Gln19Glu) 0.0017

FAM58A 3 0,33 NM_152274.3:c.50A>G p.(Glu17Gly) / /

HHAT 64 1 NM_001170587.1:c.1248+19 ? rs4297293 0.9847 G>C

STAR 297 1 NM_000349.2:c.465+20A>C ? rs2720050 0.0343 (1000 genom es)

204

Table S6 Overview of variants found in DSD genes in case 4

Gene Coverage VAF cDNA notation Protein notation Comments Exac

SRD5A2 40 1 NM_000348.3:c.90dup p.Ser31Leufs*105 / 0.1667 MAP3K1 505 0.601 NM_005921.1:c.2845_2847del p.Thr949del rs5868032 0.7393 BMP15 177 1 NM_005448.2:c.786_788dup p.Leu263dup rs531409392 0.06708 poly-Gly AR 6 0.50 NM_000044.3:c.1412_1420del p.Gly471_Gly473del / stretch SOX3 14 0.36 NM_005634.2:c.1032_1091del p.Ala345_Ala364del / 0.00004233 AMH 10 0.20 NM_000479.3:c.1094A>G p.Asp365Gly / / LHB 10 0.40 NM_000894.2:c.104T>C p.Ile35Thr rs34349826 / LHX4 121 0.51 NM_033343.3:c.1022T>C p.Ile341Thr rs372952844 0.00001647 DMRT1 127 0.99 NM_021951.2:c.1037A>G p.Lys346Arg rs35846503 0.006213 KAL1 137 0.47 NM_000216.2:c.1600G>A p.Val534Ile rs808119 0.5916 LHX3 13 0.30 NM_014564.3:c.46G>T p.Gly16Cys / / AKR1C3 146 0.55 NM_001253908.1:c.197G>A p.Arg66Gln rs35961894 0.006901 SOX3 14 0.21 NM_005634.2:c.1036G>T p.Ala346Ser / / ATRX 183 0.97 NM_000489.3:c.2785C>G p.Gln929Glu rs3088074 0.3924 CYP19A1 205 0.51 NM_000103.3:c.602C>T p.Thr201Met rs28757184 0.02425 AKR1C4 276 0.48 NM_001818.3:c.404G>A p.Gly135Glu rs11253043 0.008072 CHD7 284 0.49 NM_017780.3:c.3697G>A p.Gly1233Ser rs190548814 0.0007453 PSMC3IP 28 0.25 NM_013290.6:c.88C>A p.Gln30Lys / / SOX3 41 0.58 NM_005634.2:c.127G>A p.Ala43Thr rs73637709 0.01584 AR 6 0.33 NM_000044.3:c.173A>T p.Gln58Leu rs200185441 / AR 7 0.42 NM_000044.3:c.176A>T p.Gln59Leu / / SOX3 9 0.22 NM_005634.2:c.221C>T p.Pro74Leu / / Cryptic Donor STAR 176 0.97 NM_000349.2:c.465+20A>G p.? 0.9796 Strongly Activated New Donor WT1 24 0.41 NM_001198551.1:c.-8_-7del p.? / Site

205

Table S7 Overview of variants found in DSD genes in case 5

Gene Coverage VAF cDNA notation Protein notation Comments Exac

NR0B1 28 0.32 NM_000475.4:c.207C>A p.Cys69* / / SRD5A2 28 1 NM_000348.3:c.90dup p.Ser31Leufs*105 / 0.1667 MAP3K1 164 0.43 NM_005921.1:c.2845_2847del p.Thr949del rs5868032 0.7393 AR 30 1 NM_000044.3:c.234_239del p.Gln79_Gln80del / / AR 22 1 NM_000044.3:c.1409_1420del p.Gly470_Gly473del / 0.01489 MAMLD1 94 0.62 NM_001177465.2:c.1428_1430dup p.Gln477dup / 0.001151 POR 32 0.47 NM_000941.2:c.376A>G p.Ser126Gly / / CYP11B1 18 0.28 ENST00000377675.3:c.1370C>T p.Ala457Val rs4541 / SOX8 10 0.3 NM_014587.3:c.979T>G p.Ser327Ala / / WWOX 29 0.45 NM_001291997.1:c.415C>G p.Pro139Ala rs75559202 / AR 32 0.4 NM_000044.3:c.1960G>A p.Glu654Lys rs200737258 0.002680 ATRX 53 0.42 NM_000489.3:c.5773G>A p.Asp1925Asn / /

Table S8 Overview of variants found in DSD genes in case 6

Gene Coverage VAF cDNA notation Protein notation Comments Exac

SRD5A2 40 1 NM_000348.3:c.90dup p.Ser31Leufs*105 / 0.1667 New AR 59 0.22 NM_000044.3:c.2418_2442dup p.Ser815Valfs*23 Acceptor / Splice Site MAP3K1 170 1 NM_005921.1:c.2845_2847del p.Thr949del rs5868032 0.7393 AR 27 0.96 NM_000044.3:c.234_239dup p.Gln79_Gln80dup / / FBLN2 54 0.52 NM_001004019.1:c.244C>T p.Arg82Cys rs575900458 / TSPYL1 73 0.41 NM_003309.3:c.1098C>A p.Phe366Leu rs140756663 0.00367 WWOX 64 0.41 NM_130791.3:c.532C>T p.Pro178Ser / / SOX3 7 0.29 NM_005634.2:c.1114T>G p.Ser372Ala / / FAM58A 25 0.4 NM_152274.3:c.338G>A p.Arg113His rs150562029 /

206

Table S9 Overview of variants found in DSD genes in case 7

Gene Coverage VAF cDNA notation Protein notation Comment Exac

SRD5A2 30 1 NM_000348.3:c.90dup p.Ser31Leufs*105 / 0.1667 FAM58A 4 1 NM_152274.3:c.17dup p.Gly7Argfs*51 / / MAP3K1 222 1 NM_005921.1:c.2845_2847del p.Thr949del rs5868032 0.7393 AR 14 0.36 NM_000044.3:c.1418_1420dup p.Gly473dup / / DMRT1 27 0.44 NM_021951.2:c.442A>C p.Lys148Gln / / LHX3 27 0.22 NM_014564.3:c.982G>C p.Ala328Pro / / LHX3 24 0.25 NM_014564.3:c.979G>C p.Ala327Pro / / EMX2 23 0.26 NM_004098.3:c.238A>C p.Asn80His / / EMX2 30 0.2 NM_004098.3:c.244G>C p.Ala82Pro / / SOX8 12 0.25 NM_014587.4:c.685A>C p.Thr229Pro / / SOX8 13 0.23 NM_014587.4:c.688A>C p.Thr230Pro / / SOX9 39 0.31 NM_000346.3:c.715A>C p.Thr239Pro / / SOX9 42 0.26 NM_000346.3:c.718A>C p.Thr240Pro / / KAL1 73 0.49 NM_000216.2:c.1600G>A p.Val534Ile rs808119 0.5916 ATRX 142 0.35 NM_000489.3:c.2785C>G p.Gln929Glu rs3088074 0.3924 Cryptic Donor STAR 117 1 NM_000349.2:c.465+20A>G p.? 0.9796 Strongly Activated

207

Table S10 Overview of variants found in DSD genes in case 8

Gene Coverage VAF cDNA notation Protein notation Comments Exac

SRD5A2 59 1 NM_000348.3:c.90dup p.Ser31Leufs*105 / 0.1667

FAM58A 3 1 NM_152274.3:c.54dup p.Gln19Alafs*39 rs371528168 /

FAM58A 4 1 NM_152274.3:c.17dup p.Gly7Argfs*51 / /

MAP3K1 322 0.42 NM_005921.1:c.2845_2847del p.Thr949del rs5868032 0.7393

BMP15 147 0.45 NM_005448.2:c.786_788dup p.Leu263dup rs531409392 0.06708

AR 45 0.75 NM_000044.3:c.231_239dup p.Gln78_Gln80dup rs4045402 0.00002187

HHAT 40 0.55 NM_001122834.3:c.997G>A p.Gly333Arg rs61744143 0.002738

MAP3K1 8 0.25 NM_005921.1:c.479C>A p.Ala160Asp rs372955296 /

POR 54 0.5 NM_000941.2:c.1709G>A p.Arg570His / 0.0003066

CYP11B1 36 0.5 ENST00000377675.3:c.1370C>T p.Ala457Val rs4541 0.07390

NR5A1 99 0.44 NM_004959.4:c.779C>T p.Ala260Val / /

SOX8 24 0.21 NM_014587.4:c.685A>C p.Thr229Pro / /

SOX8 23 0.22 NM_014587.4:c.688A>C p.Thr230Pro / /

LHB 13 1 NM_000894.2:c.104T>C p.Ile35Thr rs34349826 /

ATRX 274 0.55 NM_000489.3:c.2785C>G p.Gln929Glu rs3088074 0.3924

MAMLD1 158 0.54 NM_001177465.2:c.1439T>C p.Val480Ala rs61740566 0.01835 Cryptic Donor STAR 142 1 NM_000349.2:c.465+20A>G p.? 0.9796 Strongly Activated

208

Table S11 Overview of variants found in DSD genes in case 9

Gene Coverage VAF cDNA notation Protein notation Comments Exac

SRD5A2 53 1 NM_000348.3:c.90dup p.Ser31Leufs*105 / 0.1667 EMX2 35 0.26 NM_004098.3:c.233dup p.Pro79Alafs*62 / / MAP3K1 292 0.44 NM_005921.1:c.2845_2847del p.Thr949del rs5868032 0.7393 AR 20 1 NM_000044.3:c.1418_1420del p.Gly473del / 0.001472 HSD3B2 136 0.5 NM_000198.3:c.707T>C p.Leu236Ser rs35887327 0.003687 CYP1B1 89 0.45 NM_000104.3:c.1328C>G p.Ala443Gly rs4986888 0.005280 FBLN2 64 0.45 NM_001004019.1:c.2658C>A p.Asn886Lys rs61731214 0.001426 PROP1 82 0.59 NM_006261.4:c.152G>C p.Gly51Ala rs2233783 0.01191 HOXA13 27 0.41 NM_000522.4:c.496C>A p.Pro166Thr / 0.03422 NOBOX 48 0.54 NM_001080413.3:c.1826C>T p.Pro609Leu rs115882574 0.001870 FGFR1 77 0.54 NM_023110.2:c.2314C>T p.Pro772Ser rs56234888 0.003822 CYP11B1 25 0.52 ENST00000377675.3:c.1370C>T p.Ala457Val rs4541 0.07390 DMRT1 59 0.39 NM_021951.2:c.1037A>G p.Lys346Arg rs35846503 0.006213 EMX2 29 0.24 NM_004098.3:c.244G>C p.Ala82Pro / / SOX8 20 0.35 NM_014587.4:c.676A>C p.Thr226Pro / / SOX8 24 0.33 NM_014587.4:c.685A>C p.Thr229Pro / / SOX8 30 0.27 NM_014587.4:c.688A>C p.Thr230Pro / / WWOX 110 0.57 NM_001291997.1:c.274C>A p.His92Asn rs74860463 0.001764 SOX9 30 0.36 NM_000346.3:c.715A>C p.Thr239Pro / / SOX9 23 0.43 NM_000346.3:c.718A>C p.Thr240Pro / / PROKR2 86 0.47 NM_144773.2:c.151G>A p.Ala51Thr rs144994507 0.004192 ATRX 199 0.5 NM_000489.3:c.2785C>G p.Gln929Glu rs3088074 0.3924 MAMLD1 53 0.47 NM_001177465.2:c.966C>A p.His322Gln rs62641609 0.006742 MAMLD1 120 0.46 NM_001177465.2:c.1439T>C p.Val480Ala rs61740566 0.01835 Cryptic Donor STAR 119 1 NM_000349.2:c.465+20A>G p.? 0.9796 Strongly Activated

209

Table S12 Overview of gynaecological examination in unaffected female carriers

Family 2 Family 3

I:2 II:2 I:2

Current age 48 21 46

Age at menarche 16 15.5 12

Early/excessive pubarche No No No

Acne No No No

Hirsutism No No No

Regular menses Regular, especially after No: sometimes absent for Prolonged menstrual pregnancies several months, followed cycle (1,5-2 months), by menorrhagia long-term use of oral Oral anticonception since anticonception age 16 because of menorrhagia Spontaneous pregnancies Yes None First 2 pregnancies spontaneous, last pregnancy by ICSI. Diagnostic testing at that time suggested male factor subfertility

Age at menopause 45 NA NA

Enlarged clitoris No No No

General medical problems No No No

Other remarks At age 41: acute ovariectomy because of ovarian cyst with torsion Abbreviations used: ICSI: intracytoplasmic sperm injection; NA: not applicable.

210

Table S13 List of rare variants (MAF<0.01) for patient 1, variants inherited from mother excluded (added in separate Excel file)

Table S14 List of rare de novo variants (MAF<0.01) for patient 1 (added in separate Excel file)

Table S15 List of rare variants (MAF<0.01) for patient 2 (added in separate Excel file)

Table S16 List of rare variants (MAF<0.01) for patient 3 (added in separate Excel file)

Table S17 Overview of all significantly differentially expressed genes (added in separate Excel file)

Table S13-17 are available at: http://www.nature.com/gim/journal/vaop/ncurrent/full/gim2016118a.html#supplementary- information

Table S18 Overview of the significantly differentially expressed DSD related genes.

Default analysis average gene Log fold change T p-value adjusted p-value expression LEPR -0.6681 6.123 -6.636 0.0001102 0.007119 MAMLD1 1.637 3.073 5.335 0.0005240 0.01289 IRF2BPL -0.7721 5.442 -4.092 0.002894 0.03156 Analysis with correction for fold change average gene Log fold change T p-value adjusted p-value expression MAMLD1 1.637 3.073 4.478 0.0009225 0.03436 LEPR -0.6681 6.123 -4.023 0.001605 0.04671

Fold change = ratio control group/ patient group; T = test-statistic value to compute significance of the observed change; p value = uncorrected p-value of the test statistic; adjusted p-value = the false discovery rate adjusted p- value of the test statistic

211

212

FOG2FOG2/ZFPM2/ZFPM2 HAPLOINSUFFICIENCY HAPLOINSUFFICIENCY CAUSES CAUSES PRIMARY PRIMARY OVARIAN OVARIAN INSUFFICIENCYINSUFFICIENCY

In preparation for the British Journal of Medicine.

Dorien Baetens1, Anne Bachelot2, Niels De Vriese1, Francis de Zegher3, Nele

Reynaert3, Jérôme Dulon2, Petra De Sutter4, Philippe Touraine2, Elfride De Baere1*,

Martine Cools5*

1Center for Medical Genetics Ghent, Ghent University and Ghent University Hospital, Ghent, Belgium

2UPMC Université Pierre et Marie Curie, Université Paris, Paris, France

3Paediatric Endocrinology, University Hospital Gasthuisberg, University of Leuven, Leuven, Belgium

4Department of Obstetrics and Gynaecology, Ghent University and Ghent University Hospital, Ghent,

Belgium

5Department of Pediatric Endocrinology, Ghent University and Ghent University Hospital, Ghent, Belgium

213

214

ABSTRACT

Background: Premature ovarian insufficiency (POI) is a major cause of female infertility. The genetic causes of POI remain unknown in most cases. Here, we hypothesized that haploinsufficiency of Friend of GATA2 gene (FOG2/ZFPM2) could contribute to the pathogenesis of POI, based on the identification of a heterozygous

FOG2 deletion in a child with an isolated 46,XY disorder of sex development (DSD) and in his 25-year-old mother with low serum anti-Müllerian hormone (AMH) levels, suggestive of a decreased ovarian reserve. This hypothesis is supported by the role of the FOG2/GATA4 complex in mouse ovarian development.

Methods: Copy number analysis of FOG2 in the index case was performed with array-

CGH followed by qPCR. Targeted next-generation sequencing of FOG2 was conducted in 31 Belgian and 357 French POI cases. Functional validation of the identified FOG2 variants was performed by luciferase assays in HEK293T cells.

Results: A heterozygous FOG2 deletion was found in a child with isolated 46,XY DSD and in his mother with putative POI risk. Subsequently, four novel likely pathogenic

FOG2 variants were found in a heterozygous state in a large POI cohort (4/388, 1%), one frameshift duplication and three missense mutations. Luciferase assays showed significantly lowered transcriptional activation of the mutated FOG2/GATA4 complex for several of the identified variants.

Conclusion: We demonstrate that FOG2 haploinsufficiency can lead to POI, supporting a role of FOG2 in human ovarian maintenance and contributing to the molecular pathogenesis of POI. Finally, we expand the phenotypic spectrum of FOG2 mutations, thus far only associated with cardiac anomalies, congenital diaphragmatic hernia and 46,XY DSD.

215

216

INTRODUCTION sex steroid deficiency and follicle

Gonadal development and functioning stimulating hormone (FSH) serum is driven by strictly regulated levels above 40 IU/L on at least two spatiotemporal expression of several measurements and with an interval of at

9,10 genes. Deviations from this tightly least one month . Over the years, many controlled process are known to cause tests have been developed to assess disorders of sex development (DSD), a ovarian reserve, e.g. assessment of the group of rare, congenital conditions number of antral follicles by characterized by atypical development transvaginal ultrasound (antral follicle of chromosomal, gonadal or phenotypic count), and serum measurement of FSH, sex1. As many genes involved in inhibin-B and anti-Müllerian hormone gonadal development continue to be (AMH) concentrations. The advantage expressed in adult life, an emerging of AMH measurement is that it is an concept is that mutations in known DSD early marker of reduced ovarian reserve,

11-13 genes can underlie more frequent independent of cyclic variation . POI conditions such as male infertility or can be the result of either follicle primary ovarian insufficiency (POI). An dysfunction or follicle depletion. Many exemplary gene in this context is cases are iatrogenic, following chemo-

Steroidogenic factor-1 (SF-1, NR5A1), or radiotherapy, or are due to

9 which was not only associated with autoimmune disorders . In some

46,XY DSD with or without adrenal patients POI is related to genetic defects failure2 but also with hypospadias male such as structural or numerical factor infertility3,4, POI and very recently abnormalities of the X chromosome (e.g. with 46,XX (ovo)testicular DSD5-8. Turner syndrome) or mutations in POI

POI affects approximately 1% of women genes (e.g. BMP15, FMR1, FOXL2,

9 and is characterized by a triad of NR5A1) . Although numerous genes symptoms occurring before the age of have been identified, the genetic causes

40: amenorrhea for at least four months, of POI remain unknown in the majority of cases. 217

In this paper we identified of a 46,XY DSD (chr 8: 104348705-107509840, heterozygous, maternally inherited GRCh37/hg19), encompassing several

Friend of GATA2 gene (FOG2/ZFPM2) genes including FOG2, previously

(MIM *603693) encompassing deletion implicated in the pathogenesis of 46,XY in a boy with isolated 46,XY DSD. FOG2 DSD and cardiac malformations (Figure is a cofactor of the GATA-binding 1A)17. The deletion was confirmed with protein 4 (GATA4) transcription factor, FOG2-specific qPCR, Sanger sequencing and mutations and structural variations of the entire FOG2 coding region in FOG2 and GATA4 have previously excluded the presence of additional been associated with 46,XY DSD and/or variants. Segregation analysis revealed cardiac anomalies14-17. Hormonal work- that the deletion is of maternal origin up was suggestive of gonadal (Figure 1B). Hormonal investigations in dysgenesis in the boy and reduced the 25-year-old mother revealed low ovarian reserve in the mother. Based on serum AMH (0.57 µg/L, reference range this, we hypothesized that before follicular phase, < 8.42 at age 25 haploinsufficiency of FOG2 could years, gradually decreasing to <0.3 at contribute to the pathogenesis of POI. age 46 years). Low AMH serum

Targeted next-generation sequencing of concentrations are known to be early

FOG2 in 31 Belgian and 357 French POI markers for decreased follicular reserve, cases revealed four heterozygous AMH values can be decreased or potentially pathogenic FOG2 variants, undetectable while FSH still reaches which were functionally characterized. normal levels. FSH and LH levels were RESULTS within the reference limits (0.84 U/L and

Identification of a heterozygous, 1.2 U/L respectively; reference range: maternally inherited FOG2 deletion in a FSH: 3-13 U/L, LH: 2-13 U/L); serum 46,XY boy with gonadal dysgenesis estradiol was below the detection limit of 12 ng/L (reference range: 12-166 A heterozygous deletion of 3.2 Mb was ng/µl). found in the index case with isolated

218

Figure 1: Identification of a FOG2 encompassing deletion in a patient with 46,XY DSD. (A) View of the UCSC genome browser (University of California, Santa Cruz) of the FOG2 deletion, chr 8: 104348705-107509840, GRCh37/hg19. (B) qPCR results confirming the FOG2 deletion in both the index patient and his mother. The patient’s father has two copies of the gene. Two reference genes were included for normalization. NC: negative control, SEM: standard error of the mean. (C) Schematic overview of the FOG2 protein and the location of the four novel variants identified in the Belgian/French POI cohort studied. 219

Table 1: Detailed overview of the identified FOG2 sequence variations.

Variant cDNA Protein Nucleotide Amino acid CADD- SIFT PolyPhen-2 Mutation Grantham Public conservation conservation score Taster distance databases

1 c.848_851dup p.(Glu285Valfs*5) / / / / / / / Not present

2 c.3086A>T p.(Lys1029Ile) Highly Highly 25.0 Deleterious Probably Disease 102 ExAC: All: conserved conserved: damaging causing T = 0.032% Tetraodon phyloP: 4.81 GnomAD: T= 0.022%

dbSNP: rs201729935

3 c.2501A>G p.(Lys834Arg) Highly Highly 24,9 Deleterious Probably Disease 26 ExAC: conserved conserved: Damaging causing ALL:G= Tetraodon 0.07222% phyloP: 5.13 GnomeAD: G= 0.0682%

dbSNP: rs113289249

4 c.934G>C p.(Ala312Pro) Moderately Highly 25.7 Deleterious Probably Disease 27 Not present conserved conserved: damaging causing Fruitfly phyloP: 4.08

220

Screening of a Belgian/French patient premature stop codon five amino acids cohort revealed four potentially later p.(Glu285Valfs*5), most likely pathogenic FOG2 variants undergoing nonsense mediated decay.

Targeted resequencing in a The missense variants (V2-4) were

Belgian/French cohort (nBelgian=31, nFrench= predicted to have a deleterious effect on

357) revealed four heterozygous protein function according to several sequence variants. All of them were prediction algorithms like SIFT, absent in publicly available exome PolyPhen-2 and Mutation Taster. The databases like ExAC and 1000 genomes, affected amino acids were moderately to or had a population frequency below highly conserved. Details about these

1%. Variant 1 (V1) is a four nucleotide variants are shown in Figure 1C and are duplication c.848_851dup, leading to a listed in Table 1. frameshift and the introduction of a

Figure 2: Results of the luciferase transcriptional activation experiments. Blue bars: transcriptional regulation of the AMH promoter by FOG2. Orange bars: transcriptional regulation of the AMH promoter by FOG2 variants with cotransfection of wild type (WT) GATA4 and NR5A1. Results are reported as relative light intensity units (R.L.U). The error bars represent the standard deviation of the data. *: significant difference compared to the corresponding WT condition at the 0.05 level. **: significant difference compared to the corresponding WT condition at the 0.01 level.

221

Luciferase assays show altered septum defect and 46,XY gonadal transcriptional activation. dysgenesis (GD), provided a first

Different experimental set-ups were argument for its involvement in human used to assess the transcriptional gonadal development as well. activation capacity of the FOG2/GATA4 Expression analysis of neighboring transcription complex on a human AMH genes showed no significant changes, promoter when FOG2 is mutated. Based relating the observed phenotype to on interactions described in literature, FOG2 disruption20. A second we co-transfected these FOG2 variants t(8;18)(q22;q21) translocation, associated with wild type (WT) GATA4 and with a 16 Mb FOG2 encompassing

NR5A1. Transfection of FOG2 alone deletion, was found in a syndromic case revealed discrete but significant displaying a combined congenital heart

18 differences for V1-V3 (pv1=0.0008, defect and 46,XY gonadal dysgenesis . pv2=0.0088, pv3=0.0006). Co-transfection In 2014, the first FOG2 point mutations with WT GATA4 and NR5A1 vector on associated with 46,XY gonadal the other hand revealed a significantly dysgenesis and without cardiac lowered transcriptional activation for all abnormalities were reported. variants except for V4 (pv1, pv2<0.0001, Segregation analysis revealed two pv3=0.0154) (Figure 2). asymptomatic female carriers.

DISCUSSION Functional validation of these variants showed reduced biological activity of FOG2 mutations were first identified in the mutant proteins but the pathogenic patients with structural congenital heart mechanism was not revealed17. defects such as septum defects and We first identified a FOG2 tetralogy of Fallot15,18,19. However encompassing deletion in a boy with identification of a balanced isolated 46,XY GD and his mother with t(8;10)(q23.1;q21.1) translocation, of low AMH values, suggesting POI risk. which the chromosome 8 breakpoint Previously no genomic rearrangements disrupts FOG2, in a boy with an atrial affecting FOG2 have been found in 222 patients without a cardiac phenotype. phases of gonadal development. Both

The subsequent identification of four knock-outs result in a similar phenotype novel, heterozygous FOG2 mutations in in which testicular cord formation is a French-Belgian POI cohort confirmed abrogated. Early (i.e. pre-testicular) our hypothesis that FOG2 mutations male specific genes like Nr5a1 and Wt1 may underlie POI. No cardiac anomalies are no longer expressed, Sry expression were reported in the medical history of is significantly reduced and these cases. The variants identified in downstream genes of Sry, such as Sox9 this study are distributed over the entire and Amh are undetectable (Figure 3A)21-

FOG2 gene, which is, together with the 24. deletion, suggestive of Interestingly, several lines of evidence haploinsufficiency. As variants that implicate the Fog2/Gata4 complex in were previously identified in patients ovarian development and functioning. with cardiac malformations are also Antonnen et al. have shown Fog2/Gata4 scattered over the gene, it is not possible expression in mice granulosa cells in to establish reliable genotype- different stages of follicle phenotype correlations. Luciferase development25. Later it was shown that assays revealed a significant effect of the the absence of a functional Fog2/Gata4

GATA4/FOG2 complex on complex results in ectopic expression of transcriptional activation of a human Dkk1, a known Wnt signaling inhibitor,

AMH-promoter for variants V1-V3. resulting in loss of expression of several

Although these functional data are in genes of the Wnt pathway. In addition, favor of a pathogenic effect of the loss of the Fog2/Gata4 complex leads to identified mutations, the precise downregulation of Foxl2, which is one of pathogenic mechanism remains to be the earliest markers of ovarian elucidated. development (Figure 3A)26. Early

Studies in knock-out mice propose a embryonic death of both Gata4ki/ki - a function for the Fog2/Gata4 knock-in model that abrogates the transcription complex in the earliest Fog2/Gata4 interaction and survives 223 longer than Gata4-/- mice - and Fog2-/- control of pituitary FSH secretion29,30. In mice precludes evaluation of adult vitro evidence demonstrates that the ovaries, however several sources of in Fog2/Gata4 complex can play a vitro evidence implicate both factors in regulatory role in steroidogenesis as adult ovarian functioning. Together Gata4 binding sites have been found in with Nr5a1, the Fog2/Gata4 complex is the promoters of different genes involved in transcriptional regulation of encoding steroidogenic enzymes21. The the Amh gene (Figure 3B). During involvement of the Fog2/Gata4 complex follicular maturation, Amh is in both Amh regulation and responsible for inhibiting excessive steroidogenesis is reminiscent of the role follicle recruitment by FSH27,28. Fog2 of Nr5a1, which has also been appears to have a coordinating role in implicated in 46,XY DSD and POI, sex this pathway as it can act as a repressor development and adult ovarian of Gata4-induced Amh activation functioning31. In agreement with

(Figure 3B)25. Besides Amh, the inhibin-α NR5A1-associated POI, FOG2- gene (Inha) is a second member of the associated POI is also characterized by transforming growth factor-β (TGF-β) incomplete penetrance, as FOG2 family which is regulated by mutations previously identified in

Fog2/Gata4. Inhibins form heterodimer patients with 46,XY DSD, did not lead to glycoproteins that are involved in the POI in two female carriers17.

224

Figure 3: Suggested working mechanisms for the FOG2 mutations. Blue: male specific genes. Orange: female specific genes. (A) Mouse knock-out models suggest that the Fog2/Gat4 complex is implicated in upregulation of Sry, thereby triggering the male developmental pathways22-24. In females, the Fog2/Gata4 complex inhibits Dkk1, a known inhibitor of the Wnt- pathways26. (B) In vitro experiments are in favor of a postnatal role for the complex. Together with NR5A1, the FOG2/GATA4 complex can activate AMH, which is known to regulate FSH in order to prevent excessive follicle maturation25,28. The green crosses indicate that FOG2 mutations can possibly inhibit the FOG2/GATA4 actions that are visualized by the black arrows.

CONCLUSION

We identified a heterozygous FOG2 encompassing deletion in a boy with 46,XY DSD and his mother with low AMH values, suggestive of a role of FOG2 haploinsufficiency in POI pathogenesis. Screening of a Belgian-French POI cohort revealed four potentially pathogenic FOG2 variants, with an effect on transcriptional activation.

Based on our findings, the incidence of FOG2-associated POI is 1% (4/388). Our study

225 shows that FOG2 haploinsuffiency is implicated in a spectrum of conditions ranging from 46,XY gonadal dysgenesis, whether or not associated with cardiac malformations, to 46,XX POI. This phenotypic spectrum is reminiscent of the one previously identified for NR5A1 mutations and underlines the importance of correct functioning of sex developmental genes throughout adult life.

ACKNOWLEDGEMENTS

We are most grateful to the families who participated in this study. This work was supported by grants from the Research Foundation Flanders (FWO) (Bilateral

Scientific Cooperation Brazil NXT-DSD to E.D.B. and M.C.), by Ghent University

Special Research Fund (BOF15/GOA/011), by Belspo IAP project P7/43 (Belgian

Medical Genomics Initiative: BeMGI) to E.D.B, by Ghent University Special Research

Fund (BOF Starting Grant) to M.C. E.D.B. and M.C. are Senior Clinical Investigators of the FWO. PATIENTS AND METHODS

Patients

The index case, currently four years old, was born at 38 weeks of gestation after an uneventful, spontaneous pregnancy. His birth weight was 2,575 kg and he measured

47 cm. The boy displayed proximal penile hypospadias, a fused hypoplastic scrotum and impalpable gonads. Hormonal work-up at 3 months showed low serum AMH (3

µg/L, reference range: 55.3-187 µg/L) and borderline increased FSH (12 U/L, reference range 1-12 U/L). Luteinizing hormone (LH) was within the reference range (1.2 U/L, reference range 1-9 U/L). At ten months, stretched penile length was 42 mm (mean of

46.5 mm ± 4.7 mm), and gonads were still not palpable. Discrete dysmorphic features were observed: flat nasal bridge, low set ears, downslanting palpebral fissures.

Psychomotor development was normal. Structural cardiac defects were excluded by cardiac ultrasonography, gonads were in an abdominal position. Gonadal biopsies, taken at the age of 13 months revealed dysgenetic testicular tissue on the left side, containing some spermatogonia, as indicated by positive DDX4 and negative OCT3/4

226 staining. On the right side no atypical testicularization was observed. Hypospadias was corrected in a two-stage procedure. Gonads could not be relocated in the scrotum and ultimately, gonadectomy was performed at the age of 2 years and 6 months in view of the increased gonadal germ cell cancer risk. Family history was negative for reproductive disorders. The mother had menarche at age 14 years and a normal menstrual cycle. She had two spontaneous pregnancies at age 23 and 25 (index case).

The father is known with asthma, hypercholesterolemia and epilepsy.

Belgian POI patients (n=31) were recruited in the fertility department of Ghent

University Hospital (Ghent, Belgium), in a group who had successful pregnancies after oocyte donation between 2004-2013. Women with iatrogenic or auto-immune disorders were excluded. All of them signed informed consent. DNA samples from

357 French POI patients were collected via the Service d’Endocrinologie et Médecine de la Reproduction, Hôpital Pitié – Salpêtriere (Paris, France). The study was approved by the ethical committee of the Ghent University Hospital (EC/2014/0592).

Array-CGH and qPCR

Microarray-based comparative genomic hybridization (array CGH) was performed with genome-wide arrays (180K, Agilent), with an overall probe spacing of 14 kb in general, 11 kb when only taking the RefSeq genes into account. Hybridization was performed as recommended by the company and results were visualized in ViVar32.

Quantitative PCR (qPCR) for copy number detection was performed as described33.

Primers were designed using Primer3plus and validated in silico and with a standard dilution series33. The experiment was done with the SsoAdvanced SYBR supermix

(Bio-Rad), and data-analysis was done with qbase+ software (Biogazelle).

Targeted resequencing

The coding exons and intron-exon boundaries of FOG2 were sequenced on a MiSeq instrument, using a standardized approach described by De Leeneer et al34. FOG2 primers were designed with primer3plus (http://www.bioinformatics.nl/cgi- bin/primer3plus/primer3plus.cgi). Sanger sequencing was used to confirm potentially

227 pathogenic variants and was performed on the ABI 3730XL DNA Analyzer (Applied

Biosystems) using the BigDye Terminator v3.1 Cycle Sequencing Kit (Applied

Biosystems), followed by data analysis with SeqScape v2.5 (Life Technologies).

Plasmid construction

We constructed a FOG2 and GATA4 wild type plasmid by using a Gateway LR recombination reaction (Invitrogen, Life Technologies), starting from gene specific pShuttle Gateway PLUS ORF clones (GC-U1317 and GC-Z5705, GeneCopoeia) and the

Gateway pcDNA-DEST47 vector (Invitrogen, Life Technologies). Mutagenesis was performed using the Q5 site-directed mutagenesis kit (New England Biolabs), and mutation-specific primers were designed with the NEBaseChanger software (New

England Biolabs). Mutation positive plasmids were grown up in One Shot TOP10 competent cells (Invitrogen, Life Technologies) and subsequently isolated with the

NucleoBond Xtra Midi/Maxi kit (Macherey-Nagel). All vectors were sequenced to ensure correct sequence. The luciferase reporter construct for AMH was a kind gift from dr. Nathalie Diclemente.

Cell culture and luciferase assays

HEK293T cells were cultured in DMEM supplemented with 10% FCS and 1% penicillin/streptavidin. One day before transfection 5000 cells/ well were seeded in a

96 well plate. Cells were transfected with a total of 200 ng/ well using Dharmafect kb transfection reagent (GE healthcare) and washed with phosphate-buffered-saline

(PBS) 24h later. After 48h, cells were washed a second time with PBS followed by cell lysis and assessment of the luciferase activities following the Dual-Glo Luciferase

Assay System (Promega) protocol. Results are reported as relative light units (R.L.U., i.e., the ratio of the Renilla luciferase to firefly luciferase activities). Each value is the mean of six independent experiments, and the error bars represents the standard standard deviation of the data.

228

REFERENCES

1. Hughes IA, Houk C, Ahmed SF, Lee PA. Consensus statement on management of intersex disorders. Endocrine. 2006;(91):554-562. doi:10.1136. 2. Achermann J, Masafumi I, Ito M, Hindmarsh P, Jameson J. A mutation in the gene encoding steriodogenic factor -1 causes XY sex reversal and adrenal failure in humans. Nat ge. 1999;22(june):1-2. 3. Bashamboo A, Ferraz-de-Souza B, Lourenço D, et al. Human male infertility associated with mutations in NR5A1 encoding steroidogenic factor 1. Am J Hum Genet. 2010;87(4):505-12. doi:10.1016/j.ajhg.2010.09.009. 4. Baetens D, Mladenov W, Delle Chiaie B, et al. Extensive clinical, hormonal and genetic screening in a large consecutive series of 46,XY neonates and infants with atypical sexual development. Orphanet J Rare Dis. 2014;9:209. doi:10.1186/s13023-014-0209-2. 5. Lourenco D, Brauner R, Lin L, et al. Mutations in NR5A1 Associated with Ovarian Insufficiency. N Engl J Med. 2009;360(12):1200-1210. 6. Baetens D, Stoop H, Peelman F, et al. NR5A1 is a novel disease gene for 46,XX testicular and ovotesticular disorders of sex development. Genet Med. 2016;(April). doi:10.1038/gim.2016.118. 7. Bashamboo A, Donohoue P a, Vilain E, et al. A recurrent p.Arg92Trp variant in steroidogenic factor-1 (NR5A1) can act as a molecular switch in human sex development. Hum Mol Genet. 2016;25(16):3446-3453. doi:10.1093/hmg/ddw186. 8. Igarashi M, Takasawa K, Hakoda A, et al. Identical NR5A1 Missense Mutations in Two Unrelated 46,XX Individuals with Testicular Tissues. Hum Mutat. 2016:1-4. doi:10.1002/humu.23116. 9. De Vos M, Devroey P, Fauser BCJM. Primary ovarian insufficiency. Lancet. 2010;376(9744):911- 21. doi:10.1016/S0140-6736(10)60355-8. 10. Tucker EJ, Grover SR, Bachelot A, Touraine P, Sinclair AH. Premature Ovarian Insufficiency: New Perspectives on Genetic Cause and Phenotypic Spectrum. Endocr Rev. 2016;37(6):609-635. doi:10.1210/er.2016-1047. 11. Broekmans FJ, Kwee J, Hendriks DJ, Mol BW, Lambalk CB. A systematic review of tests predicting ovarian reserve and IVF outcome. Hum Reprod Update. 2006;12(6):685-718. doi:10.1093/humupd/dml034. 12. Knauff EAH, Eijkemans MJC, Lambalk CB, et al. Anti-M??llerian hormone, inhibin b, and antral follicle count in young women with ovarian failure. J Clin Endocrinol Metab. 2009;94(3):786- 792. doi:10.1210/jc.2008-1818. 13. Hehenkamp WJK, Looman CWN, Themmen APN, De Jong FH, Te Velde ER, Broekmans FJM. Anti-Müllerian hormone levels in the spontaneous menstrual cycle do not show substantial fluctuation. J Clin Endocrinol Metab. 2006;91(10):4057-4063. doi:10.1210/jc.2006-0331. 14. Garg V, Kathiriya IS, Barnes R, et al. GATA4 mutations cause human congenital heart defects and reveal an interaction with TBX5. 2003;21:443-447. doi:10.1038/nature01820.1. 15. Pizzuti A, Sarkozy A, Newton AL, et al. Mutations of ZFPM2/FOG2 gene in sporadic cases of tetralogy of Fallot. Hum Mutat. 2003;22(5):372-7. doi:10.1002/humu.10261. 16. Lourenço D, Brauner R, Rybczynska M, Nihoul-Fékété C, McElreavey K, Bashamboo A. Loss- of-function mutation in GATA4 causes anomalies of human testicular development. Proc Natl Acad Sci U S A. 2011;108(4):1597-602. doi:10.1073/pnas.1010257108.

229

17. Bashamboo A, Brauner R, Bignon-Topalovic J, et al. Mutations in the FOG2/ZFPM2 gene are associated with anomalies of human testis determination. Hum Mol Genet. 2014;(6):1-9. doi:10.1093/hmg/ddu074. 18. Tan Z-P, Huang C, Xu Z-B, Yang J-F, Yang Y-F. Novel ZFPM2/FOG2 variants in patients with double outlet right ventricle. Clin Genet. 2012;82(5):466-71. doi:10.1111/j.1399-0004.2011.01787.x. 19. Huang X, Niu W, Zhang Z, et al. Identification of novel significant variants of ZFPM2/FOG2 in non-syndromic Tetralogy of Fallot and double outlet right ventricle in a Chinese Han population. Mol Biol Rep. 2014;41(4):2671-7. doi:10.1007/s11033-014-3126-5. 20. Finelli P, Pincelli a I, Russo S, et al. Disruption of friend of GATA 2 gene (FOG-2) by a de novo t(8;10) chromosomal translocation is associated with heart defects and gonadal dysgenesis. Clin Genet. 2007;71(3):195-204. doi:10.1111/j.1399-0004.2007.00752.x. 21. Zaytouni T, Efimenko EE, Tevosian SG. GATA Transcription Factors in the Developing Reproductive System. 1st ed. Elsevier Inc.; 2011. doi:10.1016/B978-0-12-386481-9.00004-3. 22. Tevosian SG, Albrecht KH, Crispino JD, Fujiwara Y, Eicher EM, Orkin SH. Gonadal differentiation, sex determination and normal Sry expression in mice require direct interaction between transcription partners GATA4 and FOG2. Development. 2002;129(19):4627-34. Available at: http://www.ncbi.nlm.nih.gov/pubmed/12223418. 23. Tevosian SG, Deconinck a E, Tanaka M, et al. FOG-2, a cofactor for GATA transcription factors, is essential for heart morphogenesis and development of coronary vessels from epicardium. Cell. 2000;101(7):729-39. Available at: http://www.ncbi.nlm.nih.gov/pubmed/10892744. 24. Crispino JD, Lodish MB, Thurberg BL, et al. Proper coronary vascular development and heart morphogenesis depend on interaction of GATA-4 with FOG cofactors. Genes Dev. 2001;15(7):839-44. doi:10.1101/gad.875201. 25. Anttonen M, Ketola I, Parviainen H, Pusa A-K, Heikinheimo M. FOG-2 and GATA-4 Are coexpressed in the mouse ovary and can modulate mullerian-inhibiting substance expression. Biol Reprod. 2003;68(4):1333-40. doi:10.1095/biolreprod.102.008599. 26. Manuylov NL, Smagulova FO, Leach L, Tevosian SG. Ovarian development in mice requires the GATA4-FOG2 transcription complex. Development. 2008;135(22):3731-43. doi:10.1242/dev.024653. 27. Visser J a, de Jong FH, Laven JSE, Themmen APN. Anti-Müllerian hormone: a new marker for ovarian function. Reproduction. 2006;131(1):1-9. doi:10.1530/rep.1.00529. 28. Tremblay JJ, Viger RS. Transcription Factor GATA-4 Enhances Müllerian Inhibiting Substance Gene Transcription through a Direct Interaction with the Nuclear Receptor SF-1. Mol Endocrinol. 1999;1(January):1388-1401. 29. Anttonen M, Parviainen H, Kyrönlahti a, et al. GATA-4 is a granulosa cell factor employed in inhibin-alpha activation by the TGF-beta pathway. J Mol Endocrinol. 2006;36(3):557-68. doi:10.1677/jme.1.01962. 30. Luisi S, Florio P, Reis FM, Petraglia F. Inhibins in female and male reproductive physiology: role in gametogenesis, conception, implantation and early pregnancy. Hum Reprod Update. 2005;11(2):123-35. doi:10.1093/humupd/dmh057. 31. Ferraz-de-Souza B, Lin L, Achermann JC. Steroidogenic factor-1 (SF-1, NR5A1) and human disease. Mol Cell Endocrinol. 2011;336(1-2):198-205. doi:10.1016/j.mce.2010.11.006. 32. Sante T, Vergult S, Volders PJ, et al. ViVar: A comprehensive platform for the analysis and visualization of structural genomic variation. PLoS One. 2014;9(12):1-12. doi:10.1371/journal.pone.0113800. 230

33. D’haene B, Vandesompele J, Hellemans J. Accurate and objective copy number profiling using real-time quantitative PCR. Methods. 2010;50(4):262-270. doi:10.1016/j.ymeth.2009.12.007. 34. De Leeneer K, Hellemans J, Steyaert W, et al. Flexible, scalable, and efficient targeted resequencing on a benchtop sequencer for variant detection in clinical practice. Hum Mutat. 2015;36(3):379-87. doi:10.1002/humu.22739.

231

232

CHAPTER 5:

CHAPTER 5: NON-CODING VARIATION IN DSD

Non-coding variation in DSD

233

234

NON-CODING VARIATION IN DISORDERS OF SEX DEVELOPMENT NON-CODING VARIATION IN DISORDERS OF SEX DEVELOPMENT

Clinical Genetics 2017

Dorien Baetens1, Berenice B. Mendonca2, Hannah Verdin1, Martine Cools³, Elfride De Baere1

1Center for Medical Genetics Ghent, Ghent University and Ghent University Hospital, Ghent, Belgium

2Disciplina de Endocrinologia, Laboratorio de Hormonios e Genetica Molecular LIM/42, Unidade de

Adrenal, Disc. de Endocrinologia e Metabologia, Hospital das Clínicas, Faculdade de Medicina da

Universidade de São Paulo, São Paulo, Brazil

3Department of Pediatric Endocrinology, Ghent University and Ghent University Hospital, Ghent,

Belgium

235

236

ABSTRACT

Genetic studies in Disorders of Sex Development (DSD), representing a wide spectrum of developmental or functional conditions of the gonad, have mainly been oriented towards the coding genome. Application of genomic technologies, such as whole- exome sequencing, result in a molecular genetic diagnosis in ∼50% of cases with DSD.

Many of the genes mutated in DSD encode transcription factors such as SRY, SOX9,

NR5A1, and FOXL2, characterized by a strictly regulated spatiotemporal expression.

Hence, it can be hypothesized that at least part of the missing genetic variation in DSD can be explained by non-coding mutations in regulatory elements that alter gene expression, either by reduced, mis- or overexpression of their target genes. In addition, structural variations such as translocations, deletions, duplications or inversions can affect the normal chromatin conformation by different mechanisms. Here, we review non-coding defects in human DSD phenotypes and in animal models. The wide variety of non-coding defects found in DSD emphasizes that the regulatory landscape of known and to be discovered DSD genes has to be taken into consideration when investigating the molecular pathogenesis of DSD.

237

238

INTRODUCTION factors (such as SRY, SOX9, NR5A1,

Sex development, whereby the initially FOXL2) requiring a rigorously regulated undifferentiated gonad is evolving to a spatiotemporal expression, it can be testicular or ovarian fate, is undoubtedly expected that part of the unsolved cases one of the most intriguing processes of are caused by a disturbed early embryonic life. Given the transcriptional or translational complexity of these processes, it is not regulation of known or yet to be surprising that they are prone to errors, discovered DSD genes, resulting from leading to so-called Disorders of Sex non-coding genetic defects. Indeed,

Development (DSD)1. With a prevalence studies of developmental conditions of around one in 2000–4500 births, they such as DSD have uncovered a wide are among the most common birth variety of non-coding mutations that defects2. In recent years, we have come affect regulatory elements and that to a better understanding of the result in reduced, mis- or molecular pathways involved in these overexpression of their target genes. complex processes and with the advent Structural variations such as of next-generation sequencing translocations, deletions, duplications technologies new genes and or inversions can alter the normal mechanisms have been identified3. chromatin conformation and can

Nevertheless, diagnostic approaches reposition enhancer elements which can using targeted resequencing of DSD lead to misexpression of their target gene panels and whole-exome gene. Such defects have been shown to sequencing leave a majority of cases be important in the pathogenesis of 5-7 unsolved4. Genetic studies in DSD have developmental conditions in general . mainly been oriented towards the Here, we review regulatory defects that coding portion of the genome. As many are known to be implicated in DSD, both of the genes involved in sex in human as well as in other mammalian development encode transcription model organisms.

239

REGULATORY DEFECTS IN DSD 2008, a more extensive phylogenetic

SRY footprinting study showed four large conserved regions based on multiple Two decades ago several individuals sequence alignments. Two of the were described who had regulatory previously identified SRY regulatory defects in Sex determining region Y (SRY) defects8,11,12 proved to be located within associated with 46,XY DSD; however for one of these regions, supporting their none of them the exact working association with gonadal dysgenesis. mechanism could be elucidated8-11. Intriguingly, it was noticed that mouse Concomitantly, comparative sequence 5′ sequences, as well as Sry coding analysis revealed highly conserved sequences were different from those of motifs surrounding SRY, suggesting a other species, suggesting different function in transcriptional regulation12- regulatory mechanisms. This is reflected 13. The first irrefutable evidence for the by a unique murine expression pattern importance of correct SRY regulation in of Sry15. humans was provided by Assumpção et al. who identified a three basepair (bp) SOX9 deletion located in the SRY promoter in The most extensively studied regulatory a patient with 46,XY complete gonadal region involved in DSD is undoubtedly dysgenesis and the father, who had the SRY-box 9 (SOX9) region. Loss-of- undergone several reconstructive function coding mutations and surgeries for severe hypospadias in regulatory defects including childhood. This deletion partially translocations, deletions and inversions removes a Sp1 binding site from the SRY can cause campomelic dysplasia (CD), a promoter. Previous research had semi-lethal skeletal malformation showed that Sp1 is one of the syndrome with 46,XY DSD in 75% of transcription factors (together with male cases, while both coding and

NR5A1 and WT1) necessary for regulatory duplications are associated

14 transcriptional activation of SRY . In with 46,XX testicular DSD (called

240

RevSex)16,17. In addition, disruptions of regions, although there are no clear-cut specific long-range cis-regulatory genotype–phenotype correlations. The elements are a known cause of Pierre first and most upstream interval (−1230

Robin sequence (PRS), leading to to 1030 kb upstream of SOX9) is craniofacial anomalies18,19. The situation associated with PRS, the second (−932 to is even more complex: SOX9 regulatory 601 kb upstream of SOX9) with milder duplications that overlap with RevSex forms of CD and acampomelic duplications but extend to the campomelic dysplasia (ACD) and the neighbouring KCJN genes do not result third (−375 to 50 kb upstream of SOX9) in female-to-male sex reversal but in with more severe CD, respectively (Fig.

Cooks syndrome, a condition 1). The shortest region of overlap (SRO) characterized by brachydactyly and of the duplications found in 46,XX DSD, anonychia20. A comprehensive overview a 74-kb interval (−517 to −595 kb of of reported regulatory rearrangements SOX9), represents the critical region for of the SOX9 region are listed in Table 1 this phenotype, called RevSex24. This and depicted in Fig. 1. In an attempt to region was further refined by Kim et al. unravel the complex regulation of SOX9 to a 68-kb interval from 516 to 584 kb expression, several studies searched for upstream of SOX9, called the XX sex conserved non-coding elements (CNEs) reversal region (XXSR). Similarly, the that might act as tissue-specific authors identified the SRO of 46,XY DSD enhancers16,21-24 (Fig. 1). Different tissue- causing deletions, the XY sex reversal specific enhancers have been identified region (XYSR) of 32.5 kb, located 607.1–

(involved in neurodevelopment, 639.6 kb upstream of SOX926. Almost chondrogenesis and craniofacial simultaneously, Hyon et al. reduced this development); however the testis- critical region to 41 kb27 (Fig. 1). specific SOX9 regulation remains Despite several efforts to identify enigmatic25. In general, SOX9 regulatory gonadal-specific enhancers, the rearrangements that are associated with repositioning of which might cause sex specific phenotypes cluster in four main 241 reversal, no such elements have been More insights into the pathogenic found to date. In this respect, a mechanisms underlying different particular duplication that causes 46,XX classes of SOX9 regulatory duplications

DSD is of special interest. Benko et al. were recently provided by Franke et al.29 identified a 148 kb duplication in a by chromosome conformation capture

46,XX ovotesticular DSD patient that studies (capture Hi-C and 4C-seq). As was also present in the father and more mentioned above, small regulatory surprisingly in the paternal, unaffected duplications encompassing the RevSex grandmother. It was hypothesized that region lead to 46,XX DSD, while larger changes in the regulatory region that duplications including this region but harbors CNEs, could affect the extending to the KCNJ genes cause chromatin state of the SOX9 promoter Cooks syndrome. In addition, a third and more specifically of the testis-specific class of duplications was described, enhancer of SOX9 (TESCO) element, by containing the RevSex region and the alternative interactions between a gene desert between the SOX9 and specific regulatory element and TESCO. KCNJ genes but not the genes

Chromatin profiling in fibroblasts themselves. Duplications of this type do showed that both the father and son had not lead to a specific phenotype (Fig. 1). a more repressive signature in the Franke et al. attribute these differential region of the duplication and a more phenotypic expressions to different open chromatin state starting from the disruptions of topologically associated

TESCO enhancer towards SOX9, domains (TADs)29 that represent higher compared with female controls, order chromatin structures in which confirming this hypothesis. These interactions take place. They are epigenetic changes can be seen as a separated from each other by genetic background-based, boundaries, restricting the contacts that superimposing regulatory element that are established between enhancers and can contribute to the observed their target genes, and allowing intra- incomplete penetrance28. TAD but no inter-TAD interactions30,31. 242

The first class of duplications, leading to SOX9 regulatory elements and that is female-to-male sex reversal, affects a misexpressed, hence resulting in a single TAD and are called intra-TAD developmental limb phenotype. As duplications. The other two classes SOX9 is located in another TAD in the affect two neighboring TADs and their latter situation, there is no sex boundary (inter-TAD duplications). developmental phenotype29. The

Intra- and inter-TADs have a different concepts of this study will have impact impact on chromatin organization, on the interpretation of copy number leading to a distinct phenotypic variations that are found by routine expression. The intra-TAD duplication genetic testing in DSD-related does not alter the original TAD structure phenotypes. but leads to an increased number of SOX3 interactions in the duplicated region SOX regulatory rearrangements in DSD itself and with SOX9, causing increased patients are not restricted to the SOX9 SOX9 activity and thus testis locus, but have also been described in development in an XX individual. The the SOX3 region. Comparative sequence inter-TAD duplications have an impact analysis and cytogenomic analyses led on chromatin structure and create an to the hypothesis that SRY arose from additional TAD (neo-TAD). The this gene32. Loss-of function mutations duplication found in the phenotypically do not lead to sex determination defects, normal family forms a neo-TAD that is suggesting that SOX3 has no critical isolated from the rest of the genome and function in normal sex determination in that does not interact with the humans. However, it was hypothesized neighboring SOX9 and KCNJ genes, that certain gain-of-function variants which are not included in the leading to ectopic SOX3 activation, duplication. Duplications found in could lead to female-to-male sex Cooks syndrome lead to a neo-TAD reversal. In 2011, Sutton et al. described containing the KCNJ2 gene that has three patients displaying 46,XX ectopic contacts with the duplicated 243 testicular DSD in which distinct SOX3 congenital hypotonia and a spectrum of regulatory rearrangements were craniofacial abnormalities34. The region found32. In addition, a transgenic mouse associated with sex development model in which Sox3 is ectopically encompasses 1 Mb starting from the expressed in the developing XX gonads, telomeric end and spanning the DMRT is characterized by female-to-male sex gene cluster. On the basis of its reversal. These results show that expression pattern, haploinsufficiency gonadal SOX3 expression can act as a of DMRT1 was hypothesized to be the trigger for male sex development in the underlying mechanism35,36. Only one absence of SRY32. More recently an telomeric 9p deletion that does not insertional translocation was reported in encompass the DMRT genes has been an individual with 46,XX ovotest/test described in association with sex

DSD33, which was shown to result in reversal37,38. According to a previous robust SOX3 expression in patient- study, the critical DMRT1 regulatory derived lymphoblasts, while it was region reaches out to 6.5 kb upstream of absent in XX and XY control the gene39; however the deletion individuals33. described by Calvari et al. has its

breakpoint ∼40 kb upstream of DMRT1 DMRT137,38. The authors hypothesize Deletions of the short arm of that the more distal location, in closer chromosome 9 (9p) are associated with proximity to the telomere, could also two distinct phenotypes. Telomeric affect DMRT1 expression despite the deletions of 9p24.3 result in a genital intact proximal regulatory region38. phenotype ranging from isolated Moreover, it cannot be ruled out that hypospadias to gonadal dysgenesis in this deletion affects additional more XY individuals. More proximal remote regulatory elements. However, a interstitial deletions cause 9p- more complex mechanism in which malformation syndrome, that is DMRT1 dysregulation together with characterized by intellectual disability, other effects would lead to gonadal 244 dysgenesis, should also be considered. cis-acting regulatory elements. The

A combined dosage threshold effect absence of adrenal insufficiency in this could also explain the incomplete patient suggests that the deletion does penetrance that was observed in several not result in a loss-of-function effect. It pedigrees38. was suggested that the deletion affects

silencing cis-acting elements, thereby DAX1 leading to increased activation of DAX1 The dosage sensitive Nuclear Receptor and thus 46,XY sex reversal42. Other subfamily 0 group B member 1 (NR0B1, possible mechanisms such as the loss of DAX1) gene is known to cause an insulator or the juxtaposition of a congenital adrenal hypoplasia (AHC) long-range, gonadal ridge-specific associated with hypogonadotropic enhancer cannot be excluded; however hypogonadism when mutated or the net result should be increased DAX1 deleted, and 46,XY complete or partial expression42. Interestingly, a large gonadal (testicular) dysgenesis when regulatory inversion, located about 4 kb duplicated40,41. DAX1 downregulates the upstream of DAX1, was identified in a Anti-Müllerian Hormone (AMH) gene patient with AHC43. This inversion during female development, therefore relocates an SF1-binding site that was increased DAX1 expression can cause previously associated with murine sex AMH repression during male development. These findings represent development and lead to gonadal another example of defects that, by dysgenesis and an undervirilized affecting different regulatory elements, phenotype in 46,XY individuals41,42. can cause distinct phenotypes, as was Smyk et al. reported a 46,XY female with also illustrated for the SOX9 region (see a maternally inherited deletion of ∼250 above). kb upstream of DAX142. In silico analysis of the deleted region identified several AR putative SF1-binding sites, Additional interesting regulatory evolutionarily conserved elements and mechanisms were added to the

245 molecular pathogenesis of DSD, more showed expression of a short specifically in the AR gene implicated in polypeptide, reduced levels of normal androgen insensitivity syndrome (AIS). AR protein, while the mRNA levels

First, a novel and recurrent 5’′ remain unchanged. Overall, this is a untranslated region (5′UTR) mutation novel mechanism explaining the was identified, confirming a 20-year-old complete AIS phenotype in these hypothesis44. Indeed Mizokami and patients44. Very recently, Känsäkoski et

Chang had previously shown that al. identified a deep intronic AR variant certain regions of the AR 5’ UTR (c.2450-118A>G) in two 46,XY sisters influence induction of translation with complete AIS by whole-genome without having an effect on sequencing. This variant creates a de transcription. It was hypothesized that novo 5’ splice site and a putative exonic mutations in this region might cause splicing enhancer motif, leading to the

AIS, because they lead to a lower level of preferential formation of two aberrantly

AR, without changing the transcript spliced transcripts. The de novo 5’ splice levels45. The identified variant site leads to the activation of a 3’ splice introduces a new start codon to the AR site 84 bp upstream of the variant. In

5’ UTR, and together with an in-frame ∼50% of the crypticly spliced transcripts stop codon 183 nucleotides a 202 nucleotide fragment downstream downstream, this creates a short of the 3’ splice site is included. Both upstream open reading frame (uORF). transcripts are predicted to cause a

The short alternative transcripts premature stop codon, predisposing resulting from this can reduce them to nonsense mediated decay translation levels of the downstream (NMD). Quantification of the normal AR translated gene by inhibiting ribosome transcript in patients’ genital fibroblasts progression or by complicating the showed 10% levels compared with binding of pre-initiation complexes to controls. Despite normal amounts of the original translation start site44. total AR mRNA, the AR protein was

Functional validation of this variant undetectable in patients’ genital 246

Figure 1: Schematic overview of 5’ regulatory defects of the SOX9 region. UCSC genome browser, build GRCh37/hg19. At the top, in black; conserved non-coding elements (CNEs) based on sequence homology between different species16,21-24. Identified inversions, deletions, duplications and translocations are depicted in orange, red, green and blue respectively. In general four regulatory intervals can be distinguished, shown in grey at the top of the picture (references in Table 1). The first interval, from 1030-1230 kb upstream of SOX9 is associated with Pierre Robin Sequence (PRS); the second (601-932 kb upstream of SOX9) is associated with acampomelic campomelic dysplasia (ACD) and isolated 46,XY DSD; the third (50-375 kb upstream of SOX9) is associated with more severe campomelic dysplasia (CD). The RevSex region of 74 kb is the shortest region of overlap between the duplications identified in 46,XX DSD patients24. Recently, the RevSex region was even further refined to a 62 kb region (XXSR, pink vertical rectangle), at the same time a 46,XY DSD defining region was proposed (blue vertical rectangle)26. Hyon et al. reduced the critical region for XX DSD to 41 kb27.

247 fibroblasts, suggesting that its stability hybridization showed Sox9 expression is compromised, even if aberrant in the sex reversed XX gonads. In silico mRNA is translated46. This study shows analyses did not provide evidence for the power of integrating whole-genome the involvement of a new sex sequencing data with cDNA analyses to development gene, but suggested that study deep intronic mutations in AIS the phenotype was caused by altered patients. Previously, one other intronic regulation of Sox9 expression during mutation had been identified which embryogenesis either by affecting the had an impact on dihydrotestosterone surrounding chromatin structure or by binding, no effect on mRNA splicing rearrangement of tissue-specific, long- was reported for this mutation47. range, cis-acting enhancer/repressor

elements48. However, two follow-up Taken together, the above studies papers showed that the underlying emphasize the importance of non- mechanism was more complex and that coding AR mutations in the perturbed Sox9 regulation alone was pathogenesis of AIS in case no coding insufficient to cause the sex reversal AR mutation is found phenotype49,50. The first paper

ANIMAL MODELS illustrates that the sex reversal

SOX9: odd sex mouse and CanRevSex dog phenotype is depending on the genetic

In 2000, Bishop et al. created an background of the transgenic mouse, as insertional transgenic mouse (Odd sex, crossing with an A/J strain resulted in

Ods) with the insertion of two normal female mice. Interestingly, tyrosinase copies under the control of a amodifier locus on chromosome 18,

Dct promoter, accompanied by a 150 kb Odsm, was found. Homozygosity for deletion approximately 1 Mb upstream the allele from the original mouse strain of Sox948. This rearrangement manifests strongly favors Ods sex reversal, while itself in sex reversal in XX mice. the A/J Odsm allele inhibits Ods sex

49 Histological analysis and in situ reversal . The second paper shows that upregulation of Sox9 was not solely 248 depending on the removal of a potential detected in both affected dogs and in regulatory element, but that it was healthy controls. For CNVR1, an provoked by the presence of the Dct association between high copy promoter, which had a long-range numbers and XX DSD was noticed, activating effect50. suggesting the presence of another

CNV or genetic variation in this region In dogs, XX DSD is quite a common that is associated with the phenotype; phenotype, although no molecular no such association was observed for causes were identified until recently. CNVs in CNVR2 however; although Rossi et al. were the first to identify this is part of the CanRevSex region52. SOX9 encompassing duplications in XX Recently, the assembly of the canine sex reversed dogs51. Given the fact that SOX9 regulatory region was revised, in humans both SOX9 encompassing revealing that the CanRev- Sex region is and regulatory duplications are also located upstream of SOX9. A new associated with 46,XX DSD, in silico analysis shows complete Marcinkowska-Swojak et al. explored homology between the human and the canine SOX9 regulatory region. canine region53. The existence of this Comparative sequence analysis new assembly suggests that CNVs revealed that the canine homologue found in dogs should be reassessed to (CanRevSex) of the human RevSex gain better understanding of their region (HumRevsex), which is located involvement in canine XX DSD. ∼0.5 Mb upstream of SOX9, can be found about 9 Mb downstream of FOXL2: regulatory defects in Polled

SOX9. Customized high-resolution Intersex goats and inpiggyBac mice copy number screening revealed two A third regulatory animal model with highly variable regions CNVR1 and 2, sex reversal is the Polled Intersex (PIS) respectively upstream and downstream goat, characterized by dominant of SOX9; however copy number polledness and recessive XX sex variants (CNV) in these regions were reversal. Linkage analysis located the

249 causal defect to chromosome band 1q43 also result in 46,XX sex reversal,

(the orthologue of the 3q23 locus in however no coding or regulatory humans) and via positional cloning the mutations of FOXL2 have been found in causal genetic defect was identified: a 46,XX (ovo)testicular DSD cases to

11.7 kb deletion affecting the date61,62. transcription of Pis-regulated transcript 1 A second Foxl2 associated regulatory (Pisrt1), a long non-coding RNA, and animal model is the piggyBac (PB) Forkhead box L2 (Foxl2)54. The exact mouse63. Insertional mutagenesis led to mechanism behind the sex reversal the identification of a mouse with a PB phenotype was unclear at that time. insertion approximately 160 kb Generation of homozygous Foxl2 upstream of Foxl2. This mutation leads knockout mice via zinc-finger nuclease to reduced Foxl2 expression and causes mutagenesis proved that isolated loss a phenotype reminiscent of BPES. of Foxl2 without affecting lncRNA Additional experiments showed expression is sufficient for developing interaction of the mutated region with the sex reversal phenotype in mice55. In the Foxl2 transcription start site in humans, coding FOXL2 mutations and developing craniofacial tissues, deletions have been described in showing that the piggyBac mouse is a blepharophimosis, ptosis, epicanthus good animal model to study Foxl2 inversus syndrome (BPES), a regulation and the mechanisms developmental eye condition underlying human BPES63. associated (type I) or not (type II) with primary ovarian insufficiency (POI), FUTURE PERSPECTIVES AND and rarely associated with isolated CONCLUSIONS

POI56-58. Interestingly, several cis-acting The discoveries discussed in this review regulatory deletions have been emphasize the importance of correct described in human cases with BPES58- spatiotemporal gene expression during

60. It has been suggested that sex development and highlight an homozygous FOXL2 mutations might increasing number of non-coding 250 defects in human and animal sex and gene expression. Less frequent reversed phenotypes. While in most mechanisms, however, like uORFs in

DSDs cases coding variants in known AR44 and chromatin alterations in the and yet to be discovered disease genes SOX9 region28 will evolve from future are to be expected, it is conceivable that studies. Finally, the increasing number at least part of the missing genetic of non-coding defects in DSD illustrates variation can be explained by non- that non-coding regions such as 5′UTRs, coding mutations in regulatory intronic regions, proximal and distant elements. A common action mechanism cis-regulatory elements, mostly ignored for such regulatory defects is removal in a routine context, should be or malpositioning of tissue-specific examined in DSD patients with normal enhancer or silencer elements, leading coding regions. to perturbed transcriptional activation

Table 1: Overview of SOX9 regulatory rearrangements

References Phenotype Karyotype Breakpoint locations

Inversions

Maraia et al. Severe CD 46,XX,inv(17)(q12;q25) /° 199164

Mansour et al. Mild CD + complete 46,XY,inv(17)(q11.2;q24.3- 75-350 kb upstream of SOX916 199565 gonadal dysgenesis q25.1)

Deletions

Pop et al. Case 1: mild CD 46,XY 379 -1869 kb upstream of SOX9. 200422 + 46,XY DSD Length: 1.49 Mb

Benko et al. Family 4: PRS 1.38 Mb upstream of SOX9 200918 (chr17:68,663,405-68,738,405)

Length: 75 kb

Case 1: PRS 1.58 Mb upstream of SOX9

251

(chr17:68,219,155-68,538,005)

Length: >319 kb

Case 2: PRS 1.56 Mb downstream of SOX9

(chr17:71,641,405-71,677,405)

Length: 36 kb°

Lecointre et al. Familial ACD Child: 46,XY 517 kb-1.477 Mb upstream of SOX9 200966 (+ 46,XY DSD in Mother: 46,XX Length: 960 kb child)

Benko et al. Family 4: 46,XY DSD 46,XY 405-645 kb upstream of SOX9 201117

White et al. 46,XY DSD, cleft 46,XY 235 kb upstream of SOX9 201167 palate, short stature Length: 1.19 Mb

Baghavath et 46,XY DSD 46,XY 353 kb upstream of SOX9 al. 201468 Length: 349 kb

Kim et al. Case 1: 46,XY DSD 46,XY 403.3–639.6 kb upstream of SOX9 201526 Length: 236 kb

Case 2: 46,XY DSD 46,XY 607.1–672.5 kb upstream of SOX9

Length: 65 kb

Case 3: 46,XY DSD 46,XY 510–646 kb upstream of SOX9

Length: 136 kb

Case 4: 46,XY DSD 46,XY 132.1–709.0 kb upstream of SOX9

Length: 577 kb

Duplications

Kurth et al. Family 1: symmetric Familial ~97 kb upstream of SOX9 2009 brachydactyly, anonychia (chr 17: 68,053,486-70,019,661)

Length: 1.96 Mb

Family 2: symmetric Familial /° brachydactyly, anonychia Length: 1.91 Mb

Family 3: symmetric 46,XX /° brachydactyly, Length: 1.52 Mb anonychia Familial /°

252

Family 4: symmetric Length: 1.21 Mb brachydactyly, anonychia

Benko et al. Family 1: 46,XX DSD 46,XX ~353 kb upstream of SOX9 (max. 201117 size chr17:69,069,079-69,764,059, min. size chr17: 69,107,506- 69,712,726)

Length: 605-694 kb

Family 2: 46,XX DSD* 46,XX 447-595 kb upstream of SOX9 (chr17:69521863-69670036)

Length: 148 kb

~508 kb upstream of SOX9 (max. Family 3: 46,XX DSD 46,XX size chr17:68829028-69609453 and min. size coordinates chr17:68838024-69599915)

Length: 761-780 kb

Cox et al. 46,XX DSD 46,XX 600 kb upstream of SOX9 201169 Length: 178 kb

Vetro et al.70 46,XX DSD 46,XX 514 kb upstream of SOX9

Length: 96 kb

Xiao et al. 46,XX DSD 46,XX 510-584 kb upstream of SOX9 201371 Length: 74 kb

(chr17:69,533,305-69,606,825)

Hyon et al. 46,XX DSD 46,XX ~600 kb upstream of SOX9 201527 Length: 83.8 kb

(chr17: 69,491,366 -69,575,195)

Kim et al. Case 5: 46,XX DSD 46,XX 516– 659 kb upstream of SOX9 201526 Length: 143 kb

Case 6: 46,XX DSD 46,XX 259–703 kb upstream of SOX9

Length: 444 kb

Case 7: 46,XX DSD 46,XX 264–744 kb upstream of SOX9 (min. size chr17:69 371 938– 69 852 771;

253

max. size chr17:69 370 916–69 855 932)

Length: 480 kb

Vetro et al. Case 1: 46,XX DSD 46,XX 294 kb–658 kb upstream of SOX9 201572 (min. size chr17:69 458 883 - 69823311; max. size chr17:69 401 099 – 69 878 197)

Length: 364 kb

Case 2: 46,XX DSD 46,XX 431-572 kb upstream of SOX9

(min. size chr17:69 544 737-69 686 379; max size chr17: 69 510 367- 69 764 059)

Length: 141 kb

Translocations

Young et al. Severe CD + 46,XY 46,XY, t(2;17)(q35;q23-24) 88 kb upstream of SOX975 199273 and DSD Foster et al. 199474

Tommerup et Case 1: severe CD + 46,XY,t(7;17)(q32;q24.2) or 50 kb upstream of SOX9 al. 199376 and 46,XY DSD (q34;q25.1) Wagner et al. 199477 134- Case 2: ACD + 46,XY 46,XY,t(13;17)(q22;q25.1) 142 kb upstream of SOX975,78 DSD

Case 3: severe CD 46,XX,t(1;17)(q42.13;q24.2) 173-179 kb upstream of SOX975,78 or (q25.1)

Mansour et al. Mild CD/ACD 46,XX,t(4;17)(q21.3;q23.3) /° 199565

Ninomiya et ACD + 46,XY DSD 46,XY,t(12;17)(q21.32;q24.2) 74-88 kb upstream of SOX975 al. 199679 or (q25.1)

Wirth et al. Mild CD+ PRS + 46,XY,t(6;17)(q11;q24) 212-224 kb upstream of SOX975,78 199678 46,XY DSD

Wagner et al. ACD 46,XY,t(17;22)(q25.1;p11.2) 900.2 kb upstream of SOX975,81 199780 (chr17:69,216,927-69,217,135)

Wunderle et Severe CD 46,XY,t(9;17)(p13;q23.3- 110-140 kb upstream of SOX9 al. 199816 24.1)

254

Savarirayan Mild CD 46,XX,t(5;17)(q15;q25.1) /° and Bankier 199882

Pfeiffer et al. Case 1: severe CD 46,XY,t(10;17)(q24;q23) 228-229 kb upstream of SOX9 199975

Case 2: severe CD 46,XX,t(5;17)(q13.3;q24.2) 288-319 kb upstream of SOX9

Hill-Harfe et Mild CD Familial t(13;17) 931.8 kb upstream of SOX9 al. 200581 (chr17:69,185,496)

Velagaleti et Case 1: ACD + PRS 46,XX,t(4;17)(q28.3;q24.3) ~900 kb upstream of SOX9 al. 200523 (chr17:69,217,932)

~1.3 Mb downstream of SOX9 Case 2: ACD + 46,XY 46,XY,t(4;7;8;17)(4qterr4p15 DSD .1::17q25.1r17qter;7qterr7p1 5.3::4p15.1r4pter;8pterr8q1 2.1::

7p15.3r7pter;17pterr17q25. 1::8q12.1r8qter)

Leipoldt et al. Case 1: mild CD + 46,XY,t(1;17)(q42.1;q24.3) 375 kb upstream of SOX9 (chr17: 200783 46,XY DSD 69,741,445)

Case 2: ACD + 46,XY 789 kb upstream of SOX9 DSD 46,X,t(Y;17)(q11.2;q24.3)

Jakobsen et al. PRS 46,XX,t(2;17)( q23.3;q24.3) 1.13 Mb upstream of SOX9 200719

Benko et al. Family 1: PRS t(2;17)(q32;q24) 1.16 Mb upstream of SOX9 200918 (chr17:68,663,405-68,738,405)

Family 2: PRS t(5;17)(q15;q24) 1.03 Mb upstream of SOX9

(chr17:68,219,155-68,538,005)

Family 3: PRS t(2;17)(q24.1;q24.3) 1.23 Mb upstream of SOX9

(chr17:71,641,405)

Refai et al. 46,XX testicular DSD 46,XX,t(12;17)(14.3;24.3) 776-811 kb upstream of SOX9 201084 (chr17:69,399,629)

Jakubiczka et ACD + 46,XY DSD 46,XY,t(7;17)(q33;q24).ish 492 kb upstream of SOX9 al. 201085 t(7;17) (wcp7+,wcp17+;wcp7+wcp

255

17+) with a deletion of approximately 4.2 Mb

Vetro et al. 46,XX DSD 46, XX, t(11;17) (p13;q24.3) 929 kb upstream of SOX9 201572

Abbreviations used: ACD: acampomelic dysplasia, CD: campomelic dysplasia, DSD: disorder of sex development, kb: kilobase, Mb: megabase, PRS: Pierre Robin sequence. Genomic build GRCh37/hg19 was used for all genomic coordinates. *: duplication inherited from paternal grandmother, characterized by incomplete penetrance. °: the genomic coordinates for these rearrangements were not provided in the cited references , therefore they are not shown in Figure 1.

ACKNOWLEDGEMENTS

This work was supported by the Ghent University Special Research Fund

(BOF15/GOA/011) to E. D. B., by the Ghent University Special Research Fund (BOF

Starting Grant) to M. C., and by a grant G0D6713N from the Research Foundation

Flanders (FWO) to E.D. B., B. B. M. andM. C., E. D. B. andM. C. are Senior Clinical

Investigators of the FWO. H. V. is postdoctoral fellow of the FWO.

REFERENCES

1. Biason-lauber A. Control of sex development. Best Pract Res Clin Endocrinol Metab. 2010;24(2):163-186. 2. I.A.Hughes. Early Management and Gender Assignment. Congenit Endocrinopathies New Insights into Endocr Dis Diabetes Endocr Dev Basel, Karger, 2007, vol 11, pp 47–57 Early. 2007;11:47-57. 3. Bashamboo a, Ledig S, Wieacker P, Achermann JC, Achermann J, McElreavey K. New technologies for the identification of novel genetic markers of disorders of sex development (DSD). Sex Dev. 2010;4(4-5):213-224. 4. Ohnesorg T, Croft B, Tan J, Sinclair AH. Using ROADMAP Data to Identify Enhancers Associated with Disorders of Sex Development. Sex Dev. 2016;3052:59-65. 5. Kleinjan DA, Van Heyningen V. Long-Range Control of Gene Expression: Emerging Mechanisms and Disruption in Disease. Am J Hum Genet. 2005;76:8-32. 6. Spielmann M, Mundlos S. Looking beyond the genes: the role of non-coding variants in human disease. Hum Mol Genet. 2016;0(0):ddw205. 7. Zhang F, Lupski JR. Non-coding genetic variants in human disease. Hum Mol Genet. 2015;24(R1):R102-R110. 8. McElreavy K, Vilain E, Abbas N, et al. XY sex reversal associated with a deletion 5’ to the SRY “HMG box” in the testis-determining region. Proc Natl Acad Sci U S A. 1992;89(22):11016-11020. 9. McElreavey K, Vilain E, Barbaux S, et al. Loss of sequences 3’ to the testis-determining gene, SRY, including the Y pseudoautosomal boundary associated with partial testicular determination. Proc Natl Acad Sci U S A. 1996;93(16):8590-8594. 10. Kwok C, Tyler-Smith C, Mendonca BB, et al. Mutation analysis of the 2 kb 5’ to SRY in XY females and XY intersex subjects. J Med Genet. 1996;33(6):465-468.

256

11. Poulat F, Desclozeaux M, Tuffery S, Jay P, Boizet B, Berta P. Mutation in the 5’ noncoding region of the SRY gene in an XY sex-reversed patient. Hum Mutat. 1998;Suppl. 1:S192-S194. 12. Veitia RA, Fellous M, McElreavey K. Conservation of Y chromosome-specific sequences immediately 5’ to the testis determining gene in primates. Gene. 1997;199(1-2):63-70. 13. Margarit E, Guillén a, Rebordosa C, et al. Identification of conserved potentially regulatory sequences of the SRY gene from 10 different species of mammals. Biochem Biophys Res Commun. 1998;245(2):370-377. 14. Assumpção J, Caldas Ferraz L, Benedetti C, et al. A naturally occuring deletion in the SRY promoter region affecting the Sp1 binding site is associated with sex reversal. J Endocrinol Invest. 2005;28:651-656. 15. Ross DGF, Bowles J, Koopman P, Lehnert S. New insights into SRY regulation through identification of 5’ conserved sequences. BMC Mol Biol. 2008;9:85. 16. Wunderle VM, Critcher R, Hastie N, Goodfellow PN, Schedl A. Deletion of long-range regulatory elements upstream of SOX9 causes campomelic dysplasia. Proc Natl Acad Sci U S A. 1998;95(18):10649-10654. 17. Benko S, Gordon CT, Mallet D, et al. Disruption of a long distance regulatory region upstream of SOX9 in isolated disorders of sex development. J Med Genet. 2011;48(12):825-830. 18. Benko S, Fantes J a, Amiel J, et al. Highly conserved non-coding elements on either side of SOX9 associated with Pierre Robin sequence. Nat Genet. 2009;41(3):359-364. 19. Jakobsen LP, Ullmann R, Christensen SB, et al. Pierre Robin sequence may be caused by dysregulation of SOX9 and KCNJ2. J Med Genet. 2007;44(6):381-386. 20. Kurth I, Klopocki E, Stricker S, et al. Duplications of noncoding elements 5’ of SOX9 are associated with brachydactyly-anonychia. Nat Genet. 2009;41(8):862-863. 21. Bagheri-Fam S, Ferraz C, Demaille J, Scherer G, Pfeifer D. Comparative genomics of the SOX9 region in human and Fugu rubripes: conservation of short regulatory sequence elements within large intergenic regions. Genomics. 2001;78(1-2):73-82. 22. Pop R. Screening of the 1 Mb SOX9 5’ control region by array CGH identifies a large deletion in a case of campomelic dysplasia with XY sex reversal. J Med Genet. 2004;41(4):e47-e47. 23. Velagaleti GVN, Bien-Willner GA, Northup JK, et al. Report Position Effects Due to Chromosome Breakpoints that Map ∼900 Kb Upstream and ∼1.3 Mb Downstream of SOX9 in Two Patients with Campomelic Dysplasia. Am J Hum Genet. 2005;76:652-662. 24. Smyk M, Szafranski P, Startek M, Gambin A, Stankiewicz P. Chromosome conformation capture-on-chip analysis of long-range cis-interactions of the SOX9 promoter. Chromosom Res. 2013;21(8):781-788. 25. Bagheri-Fam S, Barrionuevo F, Dohrmann U, et al. Long-range upstream and downstream enhancers control distinct subsets of the complex spatiotemporal Sox9 expression pattern. Dev Biol. 2006;291(2):382-397. 26. Kim G-J, Sock E, Buchberger A, et al. Copy number variation of two separate regulatory regions upstream of SOX9 causes isolated 46,XY or 46,XX disorder of sex development. J Med Genet. 2015;52(4):240-247. 27. Hyon C, Chantot-Bastaraud S, Harbuz R, et al. Refining the regulatory region upstream of SOX9 associated with 46,XX testicular disorders of Sex Development (DSD). Am J Med Genet A. 2015;167A(8):1851-1858. 28. Lybæk H, Bruijn D De, Dijk AH a DE, et al. RevSex duplication-induced and sex-related differences in the SOX9 regulatory region chromatin landscape in human fibroblasts. Epigenetics. 2014;9(3):416-427. 29. Franke M, Ibrahim DM, Andrey G, et al. Formation of new chromatin domains determines pathogenicity of genomic duplications. Nature. 2016;(Idi):1-15. 30. Dixon JR, Selvaraj S, Yue F, et al. Topological domains in mammalian genomes identified by analysis of chromatin interactions. Nature. 2012;485(7398):376-380.

257

31. Lupiáñez DG, Kraft K, Heinrich V, et al. Disruptions of Topological Chromatin Domains Cause Pathogenic Rewiring of Gene-Enhancer Interactions. Cell. 2016;161(5):1012-1025. 32. Sutton E, Hughes J, White S, et al. Identification of SOX3 as an XX male sex reversal gene in mice and humans. Jounal Clin Investig. 2011;121(1):328-341. 33. Haines B, Hughes J, Corbett M, et al. Interchromosomal insertional translocation at Xq26.3 alters SOX3 expression in an individual with XX male sex reversal. J Clin Endocrinol Metab. 2015;100(5):E815-E820. 34. Onesimo R, Orteschi D, Scalzone M, et al. Chromosome 9p deletion syndrome and sex reversal: novel findings and redefinition of the critically deleted regions. Am J Med Genet A. 2012;158A(9):2266-2271. 35. Raymond CS, Parker ED, Kettlewell JR, et al. A region of human chromosome 9p required for testis development contains two genes related to known sexual regulators. Hum Mol Genet. 1999;8(6):989-996. 36. Raymond CS, Kettlewell JR, Hirsch B, Bardwell VJ, Zarkower D. Expression of Dmrt1 in the Genital Ridge of Mouse and Chicken Embryos Suggests a Role in Vertebrate Sexual Development. Dev Biol. 1999;215(2):208-220. 37. Calvari V, Bertini V, De Grandi a, et al. A new submicroscopic deletion that refines the 9p region for sex reversal. Genomics. 2000;65(3):203-212. 38. Barbaro M, Balsamo A, Anderlid BM, et al. Characterization of deletions at 9p affecting the candidate regions for sex reversal and deletion 9p syndrome by MLPA. Eur J Hum Genet. 2009;17(11):1439-1447. 39. Lerer I, Sagi M, Meiner V, Cohen T, Zlotogora J, Abeliovich D. Deletion of the ANKRD15 gene at 9p24.3 causes parent-of-origin-dependent inheritance of familial cerebral palsy. Hum Mol Genet. 2005;14(24):3911-3920. 40. Muscatelli F, Strom TM, Walker a P, et al. Mutations in the DAX-1 gene give rise to both X- linked adrenal hypoplasia congenita and hypogonadotropic hypogonadism. Nature. 1994;372:672-676. 41. Bardoni B, Zanaria E, Guioli S, et al. A dosage sensitive locus at chromosome Xp21 is involved in male to female sex reversal. Nat Genet. 1994;7(4):497-501. 42. Smyk M, Berg JS, Pursley A, et al. Male-to-female sex reversal associated with an approximately 250 kb deletion upstream of NR0B1 (DAX1). Hum Genet. 2007;122(1):63-70. 43. Skinningsrud B, Husebye ES, Gilfillan GD, et al. X-linked congenital adrenal hypoplasia with hypogonadotropic hypogonadism caused by an inversion disrupting a conserved noncoding element upstream of the NR0B1 (DAX1) gene. J Clin Endocrinol Metab. 2009;94(10):4086-4093. 44. Hornig NC, de Beaufort C, Denzer F, et al. A Recurrent Germline Mutation in the 5’UTR of the Androgen Receptor Causes Complete Androgen Insensitivity by Activating Aberrant uORF Translation. PLoS One. 2016;11(4):e0154158. 45. Mizokami A, Chang C. Induction of translation by the 5’-untranslated region of human androgen receptor mRNA. J Biol Chem. 1994;269(41):25655-25659. 46. Känsäkoski J, Jääskeläinen J, Jääskeläinen T, et al. Complete androgen insensitivity syndrome caused by a deep intronic pseudoexon-activating mutation in the androgen receptor gene. Sci Rep. 2016;6(May):32819. 47. Audi L, Fernández-Cancio M, Carrascosa A, et al. Novel (60%) and recurrent (40%) androgen receptor gene mutations in a series of 59 patients with a 46,XY disorder of sex development. J Clin Endocrinol Metab. 2010;95(4):1876-1888. 48. Bishop CE, Whitworth DJ, Qin Y, et al. A transgenic insertion upstream of sox9 is associated with dominant XX sex reversal in the mouse. Nat Genet. 2000;26(4):490-494. 49. Qin Y, Poirier C, Truong C, Schumacher A, Agoulnik AI, Bishop CE. A major locus on mouse chromosome 18 controls XX sex reversal in Odd sex (Ods) mice. Hum Mol Genet. 2003;12(5):509- 515.

258

50. Qin Y, Kong LK, Poirier C, Truong C, Overbeek PA, Bishop CE. Long-range activation of Sox9 in Odd Sex (Ods) mice. Hum Mol Genet. 2004;13(12):1213-1218. 51. Rossi E, Radi O, De Lorenzi L, et al. Sox9 duplications are a relevant cause of SRY-negative XX sex reversal dogs. PLoS One. 2014;9(7). 52. Marcinkowska-Swojak M, Szczerbal I, Pausch H, et al. Copy number variation in the region harboring SOX9 gene in dogs with testicular/ovotesticular disorder of sex development (78,XX; SRY-negative). Sci Rep. 2015;5(February):14696. 53. Rossi E, Radi O, Lorenzi L De. A Revised Genome Assembly of the Region 5′ to Canine SOX9 Includes the RevSex Orthologous Region. Sex Dev 2015. 54. Pailhoux E, Vigier B, Chaffaux S, et al. A 11.7-kb deletion triggers intersexuality and polledness in goats. Nat Genet. 2001;29(4):453-458. 55. Boulanger L, Pannetier M, Gall L, et al. FOXL2 Is a Female Sex-Determining Gene in the Goat. Curr Biol. 2014;24(4):404-408. 56. Baere E De, Dixon MJ, Small KW, et al. Spectrum of FOXL2 gene mutations in blepharophimosis-ptosis-epicanthus inversus ( BPES ) families demonstrates a genotype – phenotype correlation. Hum Mol Genet. 2001;10(15):1591-1600. 57. De Baere E, Beysen D, Oley C, et al. FOXL2 and BPES: mutational hotspots, phenotypic variability, and revision of the genotype-phenotype correlation. Am J Hum Genet. 2003;72(2):478-487. 58. Beysen D, Raes J, Leroy BP, et al. Deletions involving long-range conserved nongenic sequences upstream and downstream of FOXL2 as a novel disease-causing mechanism in blepharophimosis syndrome. Am J Hum Genet. 2005;77(2):205-218. 59. D’haene B, Attanasio C, Beysen D, et al. Disease-causing 7.4 kb cis-regulatory deletion disrupting conserved non-coding sequences and their interaction with the FOXL2 promotor: implications for mutation screening. PLoS Genet. 2009;5(6):e1000522. 60. Verdin H, D’haene B, Beysen D, et al. Microhomology-Mediated Mechanisms Underlie Non- Recurrent Disease-Causing Microdeletions of the FOXL2 Gene or Its Regulatory Domain. Spinner NB, ed. PLoS Genet. 2013;9(3):e1003358. 61. De Baere E, Fellous M, Veitia R a. The transcription factor FOXL2 in ovarian function and dysfunction. Folia Histochem Cytobiol. 2009;47(5):S43-S49. 62. De Baere E, Lemercier B, Christin-Maitre S, et al. FOXL2 mutation screening in a large panel of POF patients and XX males. J Med Genet. 2002;(39):7-10. 63. Shi F, Ding S, Zhao S, et al. A piggyBac insertion disrupts Foxl2 expression that mimics BPES syndrome in mice. Hum Mol Genet. 2014;23(14):3792-3880. 64. Maraia R, Saal M, Wangsa D. inversion in a patient with campomelic dysplasia ; case report and etiologic hypothesis. Clin Genet. 1991;39:401-408. 65. Mansour S, Hall CM, Pembrey ME, Young ID. A clinical and genetic study of campomelic dysplasia. J Med Genet. 1995;32:415-420. 66. Lecointre C, Pichon O, Hamel A, et al. Familial acampomelic form of campomelic dysplasia caused by a 960 kb deletion upstream of SOX9. Am J Med Genet Part A. 2009;149(6):1183-1189. 67. White S, Ohnesorg T, Notini A, et al. Copy number variation in patients with disorders of sex development due to 46,XY gonadal dysgenesis. PLoS One. 2011;6(3):e17793. 68. Bhagavath B, Layman LC, Ullmann R, et al. Familial 46,XY sex reversal without campomelic dysplasia caused by a deletion upstream of the SOX9 gene. Mol Cell Endocrinol. 2014;393(1- 2):1-7. 69. Cox JJ, Willatt L, Homfray T, Woods CG. A SOX9 duplication and familial 46,XX developmental testicular disorder. N Engl J Med. 2011;364(1):91-93. 70. Vetro A, Ciccone R, Giorda R, et al. XX males SRY negative: a confirmed cause of infertility. J Med Genet. 2011;48(10):710-712. 71. Xiao B, Ji X, Xing Y, Chen Y wei, Tao J. A rare case of 46, XX SRY-negative male with a ~74-kb duplication in a region upstream of SOX9. Eur J Med Genet. 2013;56(12):695-698.

259

72. Vetro A, Dehghani MR, Kraoua L, et al. Testis development in the absence of SRY: chromosomal rearrangements at SOX9 and SOX3. Eur J Hum Genet. 2014;23:1-8. 73. Young ID, Zuccollo JM, Maltby EL, Broderick NJ. Campomelic dysplasia associated with a de novo 2q;17q reciprocal translocation. J Med Genet. 1992;29(4):251-252. 74. Foster JW, Dominguez-Steglich M a, Guioli S, et al. Campomelic dysplasia and autosomal sex reversal caused by mutations in an SRY-related gene. Nature. 1994;372(6506):525-530. 75. Pfeifer D, Kist R, Dewar K, et al. Campomelic dysplasia translocation breakpoints are scattered over 1 Mb proximal to SOX9: evidence for an extended control region. Am J Hum Genet. 1999;65(1):111-124. 76. Tommerup N, Schempp W, Meinecke P, et al. Assignment of an autosomal sex reversal locus (SRA1) and campomelic dysplasia (CMPD1) to 17q24.3-q25.1. Nat Genet. 1993;4(2):170-174. 77. Wagner T, Wirth J, Meyer J, et al. Autosomal sex reversal and campomelic dysplasia are caused by mutations in and around the SRY-related gene SOX9. Cell. 1994;79(6):1111-1120. 78. Wirth J, Wagner T, Meyer J, et al. Translocation breakpoints in three patients with campomelic dysplasia and autosomal sex reversal map more than 130 kb from SOX9. Hum Genet. 1996;97(2):186-193. 79. Ninomiya S, Isomura M, Narahara K, Seino Y, Nakamura Y. Isolation of a testis-specific cDNA on chromosome 17q from a region adjacent to the breakpoint of t(12;17) observed in a patient with acampomelic campomelic dysplasia and sex reversal. Hum Mol Genet. 1996;5(1):69-72. 80. Wagner T, Tommerup N, Wirth J, et al. A somatic cell hybrid panel for distal 17q: GDIA maps to 17q25.3. Cytogenet Cell Genet. 1997;76:172-175. 81. Hill-Harfe KL, Kaplan L, Stalker HJ, et al. Fine mapping of chromosome 17 translocation breakpoints > or = 900 Kb upstream of SOX9 in acampomelic campomelic dysplasia and a mild, familial skeletal dysplasia. Am J Hum Genet. 2005;76(4):663-671. 82. Savarirayan R, Bankier A. Acampomelic campomelic dysplasia with de novo 5q;17q reciprocal translocation and severe phenotype. J Med Genet. 1998;35(6):597-599. 83. Leipoldt M, Erdel M, Bien-Willner GA, et al. Two novel translocation breakpoints upstream of SOX9 define borders of the proximal and distal breakpoint cluster region in campomelic dysplasia. Clin Genet. 2007;71(1):67-75. 84. Refai O, Friedman A, Terry L, et al. De Novo 12;17 translocation upstream of sox9 resulting in 46,xx testicular disorder of sex development. Am J Med Genet Part A. 2010;152(2):422-426. 85. Jakubiczka S, Schröder C, Ullmann R, et al. Translocation and deletion around SOX9 in a patient with acampomelic campomelic dysplasia and sex reversal. Sex Dev. 2010;4(3):143-149.

260

CHAPTER 6: DISCUSSION & PERSPECTIVES

CHAPTER 6: Discussion &

261 Perspectives

262

DISCUSSION

With a prevalence of approximately one in 2 000 to 4 500 births, DSD conditions are among the most common birth defects. Taking also milder phenotypes like hypospadias and micropenis into account, the prevalence approximates one in 200 births1,2. The birth of a child with manifest genital atypicalities, especially if gender assignment is not straight forward, is very distressing for the parents. Care by an expert multidisciplinary team, including specialists on the different aspects of DSD, is therefore required. Over the years, considerable progress has been made in the diagnostic process and availability of diagnostic techniques for, amongst others, DSD conditions3,4. However, despite the advent of next-generation sequencing (NGS) technologies and subsequent identification of new factors involved in the molecular pathogenesis of DSD, an underlying genetic defect can be identified in only ~50% of cases5,6. Accurate genetic diagnosis and better understanding of the involved molecular pathways can provide information on expected prognostic factors, guide management and facilitate acceptance and genetic counselling6. Furthermore, elucidation of the molecular pathways involved in DSD, can also shed a light on the pathogenesis of related reproductive phenotypes such as male infertility or primary ovarian insufficiency (POI), which occur at much higher frequency.

The Ghent multidisciplinary DSD team has been offering centralized multidisciplinary care for families affected by a DSD condition for many years and is internationally recognized for its expertise. Over the years, DNA has been collected in a substantial number of DSD cases and their parents, enabling our team to contribute to the unravelling of the underlying molecular pathways. Accordingly, the main aim of this

PhD project was to identify new disease genes and mechanisms for some of these intriguing conditions. As a first step, we aimed to critically assess the appropriateness of the, at that time, available genetic testing in small children with atypical sexual development.

263

46,XY neonates and infants with atypical genitalia: who to investigate?

Before the advent of NGS approaches, genetic work-up in 46,XY children with undervirilized genitalia had a diagnostic yield of ~20-40%7,8 and was relatively time- consuming and expensive. Milder phenotypes had often been related to environmental factors, low birth weight and multiple genetic factors rather than to a monogenic cause; and therefore, the question arose whether hormonal and genetic work-up is truly beneficial in these cases9-11. However, mutations in genes such as AR,

NR5A1 and WT1, generally associated with more severe forms of genital ambiguity had also been identified in milder undervirilization phenotypes12-15. In study 1, we performed a systematic, extensive genetic work-up in a cohort of 32 male-assigned neonates and infants with a 46,XY karyotype and a variable degree of genital undervirilization. External masculinization scores (EMS) ranged from 2/12 to 12/12, allowing us to compare the diagnostic yield of our approach in the group with low

EMS scores (EMS<7) as compared to the group with higher EMS scores (EMS≥7).

Our data revealed no important differences in diagnostic yield in cases with low versus higher EMS scores (17.6 versus 26.5% respectively), suggesting that the decision to perform a thorough diagnostic work-up in 46,XY boys with atypical genitalia should not be based on the severity of the phenotype alone. A comparable study by Su et al. does state that EMS-scores below three indicate a higher likelihood for identification of an underlying genetic defect and that this criterion justifies genetic screening.

However our data do not support this statement as we have a slightly higher diagnostic yield in the group with higher EMS scores16. In our cohort, array-CGH proved to be a valuable technique for syndromic cases, demonstrated by a 33% success rate in this group, but its clinical detection rate drops when including non-syndromic cases, which is in line with previous data17. Sequential sequencing of known DSD genes in this cohort could only reveal three NR5A1 mutations, and no pathogenic changes in AR, SRY or WT1. In accordance with our results, this sequential first-line approach has been replaced in our center by a NGS-based DSD gene panel,

264 implemented in the automated workflow described by De Leeneer et al18.This allows a more flexible and cost-effective analysis of the most important DSD genes in a clinical setting. A very successful example of targeted NGS sequencing in DSD diagnostics has been reported by Eggers et al. Using Haloplex enrichment, they sequenced 64 known

DSD genes and 967 candidate gene in a large cohort of DSD cases. Their strategy lead to an overall diagnostic yield of 43% in 46,XY DSD cases and 17% in 46,XX DSD cases19.

Despite these promising numbers, a large portion of the investigated cases remains unexplained with a targeted approach. Therefore, we anticipate that targeted whole exome sequencing (WES), in which WES data-analysis initially focusses on a targeted gene panel, or WES in general will gain importance in the diagnostic work-up of DSD conditions (both 46,XY and 46,XX DSD), not only in a research setting, but in a clinical context as well. Of interest, several algorithms have become available meanwhile, that allow direct copy number variant (CNV) detection on WES data20-22. Although these tools have high specificity and sensitivity, their results still need to be validated with more established techniques like array-CGH. Given the lowering costs and technical improvements, it can be anticipated that whole genome sequencing (WGS) will be introduced as the “one-for-all” diagnostic tool in the near future23. The implementation of these exome- or genome-wide techniques in a diagnostic setting is however accompanied by ethical concerns as they may result in the identification of secondary and incidental finding that are unrelated to the initial clinical question. To address these concerns the American College of Medical Genetics appointed a working group on incidental findings to make recommendations on the management of these secondary findings. The specific recommendations were reported by Green et al. and include the screening of genes related to 59 medically actionable conditions24.

Comprehensive pre- and post-test counseling are of utmost importance and should be available to all patients undergoing these tests.

265

Unravelling the molecular pathogenesis of DSD phenotypes

The main goal of this PhD project was the identification of novel genetic causes or mechanisms for DSD conditions and related phenotypes. In our center, first line diagnostic testing for DSD currently consists of combined copy number screening using genome-wide array-CGH, intersex MLPA (a commercially available MLPA kit containing probes against common DSD genes, P185, MRC Holland) and targeted resequencing of a DSD gene panel. For those cases in which this sequential approach did not reveal a molecular diagnosis, we used WES to identify the underlying molecular defect. The main advantage of WES, exploring the entire coding sequence of the patient’s DNA, accounts for its main disadvantage as well. In each patient thousands of variants are found, but in most cases only one of them is of clinical importance, making the quest for the pathogenic variant like looking for a “needle in a haystack”. To reduce the size of this metaphorical haystack, two different WES strategies were applied. In case of consanguinity, we combined WES with homozygosity mapping, as previously described for DSD and other conditions25-30. In this project identification of the homozygous regions was done using genome-wide

SNP genotyping. WES data were subsequently filtered based on these regions. Today, several algorithms have been developed that extract these homozygous regions directly from WES data, overruling the need of previous SNP-chip genotyping31,32. For non-consanguineous families, we opted for a trio-sequencing strategy to reduce the number of possibly pathogenic variants. The rationale behind this strategy is that unaffected parents do not carry the pathogenic mutation and that only de novo variants can be related to the phenotype. Despite disappointing results of this strategy in a large

DSD cohort19, this strategy has also proven to be very useful in other conditions like autism spectrum disorders and intellectual disability phenotypes33-35. However, for

DSD phenotypes this de novo paradigm has to be used with caution. For instance, 46,XY

DSD phenotypes can be caused by mutations in pro-testis genes. The presence of a mutation in such a gene in the mothers of a child with 46,XY DSD could be ‘tolerated’,

266 as defects in testicular development do not apply to them. Therefore, only variants inherited from the parent with the same sex chromosomes can be disregarded when filtering WES data for DSD phenotypes. Besides this DSD-specific consideration, the possibility of incomplete penetrance should be kept in mind.

Identification of ESR2 as a novel candicate disease gene for 46,XY DSD

So far, mutations in several nuclear receptor genes have been implicated in DSD pathogenesis: Nuclear Receptor subfamily 5 group A member 1 (NR5A1, also known as

Steroidogenic Factor 1 or SF1)36, Androgen Receptor (AR, NR3C4)37, Nuclear Receptor

Subfamily 0 Group B Member 1 (NR0B1, also known as DAX1)38. In the past, ESR2 polymorphisms had already been associated with subtle differences of sex development such as hypospadias and micropenis and a heterozygous missense variant has been described in a girl with primary amenorrhea10,39-42. In study 3, we identified biallelic and monoallelic ESR2 mutations in one syndromic and two non- syndromic 46,XY DSD cases respectively, putting forward ESR2 as a novel disease gene for 46,XY DSD and expanding the spectrum of nuclear receptors implicated in

46,XY DSD. Genetic and structural data were in favor of a potential pathogenic effect, however the search for a pathogenic mechanism is not straight forward.

Nuclear receptors, and hence ER-β, are known to exert their functions via different signaling pathways. The most common signaling mode for ERs involves ligand- dependent receptor dimerization, inducing transcriptional regulation of ER target genes43,44. Alternatively, ERs can also interact with other transcription factors and modify their transcription responses, and non-genomic pathways signaling via second messengers have been described43,44. Furthermore, ligand-independent ER-signaling has been described43,44. Interestingly, one of these alternative pathways involves

MAPK signaling. Unfortunately, only scarce experimental data concerning these pathways is available and these interactions are highly cell specific, making gonadal specific validation challenging44.

267

The mitogen-activated protein kinase (MAPK) signaling pathway has previously been implicated in sex developmental processes; gain-of-function MAP3K1 mutations have been found in several 46,XY DSD cases and Map3k4 loss-of-function variants in mice result in male-to-female sex reversal45,46, although the mechanisms by which the MAPK signaling functions differ between the two species. In humans, MAPK signaling is essential for tilting the gene expression balance towards SOX9 expression; gain-of- function mutations of MAP3K1 lead to a higher phosphorylation of downstream targets, causing a decreased expression of SOX9 and an increased expression of female-specific genes like β-catenin45,47,48. As a result, the ovary-determining pathways can overrule the induction of male-specific SRY expression47,48. In mice, the Mapk pathway acts through direct upregulation of Sry; homozygous loss-of-function variants of Map3k4 are reflected in a male-to-female sex reversal phenotype46,48. We hypothesize that ESR2 gain-of-function variants can cause increased activation of the

MAPK pathway via a non-genomic working mechanism, thereby tilting the expression balance from SOX9 to β-catenin signaling, and resulting in male-to-female sex reversal.

Further experimental confirmation of this hypothesis is recommended but will be challenging as no relevant experimental set-ups are commonly available (see future perspectives).

Identification of NR5A1 as a novel disease gene for 46,XX testicular and ovotesticular DSD

While many genes have been implicated in the molecular pathogenesis of 46,XY DSD, far less is known about genes and mutations underlying 46,XX DSD. Known etiologies include SRY translocations, copy number variations and regulatory defects of different

SOX genes (SOX9, SOX3, SOX10) and inactivating RSPO1 mutations49-57. In study 3, we have screened a cohort of 13 cases with 46,XX testicular or OT DSD to identify the new molecular causes for these phenotypes. To our surprise, we identified the same

NR5A1 variant in three of them, (c.274C>T, p.(Arg92Trp)). NR5A1 mutations had previously been associated with different 46,XY DSD phenotypes, male infertility and

POI and the gene has many known functions in both developing and adult

268 gonads36,58,59. Genetic and structural arguments were both in favor of pathogenicity although an exact working mechanism could not be demonstrated.

Initially, we hypothesized that this mutation, located in the Ftz-F1 domain which is implicated in DNA-binding stability and specificity, would result in a more stable binding of NR5A1 to the SOX9 promoter, thereby directly counteracting the FOXL2- mediated SOX9 inhibition that normally occurs during female development60.

However, this hypothesis could not be confirmed by luciferase transcriptional activation experiments, as a decreased transcriptional activation of a SOX9 TESCO- reporter construct was observed. Furthermore, transcriptome analysis using RNA-seq on patient-derived lymphocytes did not show upregulation of other early male- specific genes, which is not surprising given the limited biological relevance of the used material. A second, more attractive hypothesis is that NR5A1 also plays a role in triggering the activation of early ovary-specific genes like WNT4. Two animal models are in favor of these female-specific NR5A1 actions; namely the Cited2 knock-out mouse and the Nr5a1 deficient Nile Tilapia61,62. In addition, Bashamboo et al. have demonstrated NR5A1 expression in the early human embryonic ovary, suggesting its functional importance63. It is conceivable that this specific NR5A1 mutation affects the early ovary-promoting function of NR5A1, ultimately resulting in insufficient/delayed expression of female-specific factors and as a result, escape from their testis- suppressing activity. This subtle balance of testis- and ovary-specific NR5A1 functions might also explain the observed phenotypic heterogeneity and incomplete penetrance.

The final phenotype is dependent on whether certain thresholds of male and female specific factors have been reached in a timely fashion.

Remarkably, at the same time two other studies reporting on the same NR5A1 mutation in 46,XX DSD cases were published63,64. In the study of Bashamboo et al. the p.Arg92Trp variant was identified in a boy with testicular DSD and his sister with

46,XY partial gonadal dysgenesis. Transcriptional activation assays showed a reduced biological activity on AMH and SOX9 TESCO reporter constructs (in line with our

269 findings), explaining the 46,XY DSD phenotype. Furthermore, they assessed the capability of mutated NR5A1 (in combination with β-catenin) to activate

DAX1/NR0B1; the p.Arg92Trp variant showed selective loss of this synergistic activation, possibly accounting for the 46,XX DSD phenotype63. A previously identified

NR5A1 mutation affecting the same codon, p.Arg92Gln, had no effect on DAX1 activation, which can explain why this variant does not lead to a gonadal phenotype in a 46,XX individual with adrenal insuffiency65. This mutation had previously been identified in a case of 46,XY gonadal dysgenesis and adrenal insufficiency and it was already established to lead to reduced SOX9 activity66. The third study concerning the p.Arg92Trp mutation adds two additional cases with (ovo)testicular DSD and puts forward an alternative working mechanism. The DAX1-mediated anti-testis action has previously been attributed to its antagonizing effect on NR5A1-mediated activation of

SOX9. Here, luciferase data showed that mutated NR5A1 is less susceptible to DAX1- induced suppression on SOX9-TESCO. The authors anticipate that unresponsiveness of the mutated NR5A1 to DAX1 suppression would alter the ovarian developmental processes64. Overall, our p.Arg92Trp study and the two others agree on the fact that this mutation exerts its effect rather by hampering NR5A1-mediated pro-ovarian functions than by upregulation of pro-testicular NR5A1 activity. Additional functional studies will further elucidate the exact underlying mechanism.

Interestingly a mutant mouse was generated using CRISPR/Cas9 for mutation p.Arg92Trp67. Phenotypical characterization revealed both similarities and discrepancies between humans and mice. All Nr5a1R92W/R92W mice had typical female external genitalia and no adrenals were present, while in humans the p.Arg92Trp mutant has so far not been associated with an adrenal phenotype. Nr5a1R92W/R92W XY mice had intra-abdominal hypoplastic gonads and histological examination revealed the absence of testicular cords at 13.5-18.5 dpc, reminiscent of the 46,XY DSD case described by Bashamboo et al63,68. In Nr5a1R92W/R92W XX mice however the gonadal phenotype is not in accordance with the human situation; examination of XX mice

270 revealed normal or slightly hypoplastic gonads that were histologically indistinguishable from wild type counterparts68. These discrepancies between mouse and human ovarian phenotypes are reflected in the NR5A1/Nr5a1 expression pattern; in humans NR5A1 mRNA expression in the fetal ovary starts at 6-10 weeks of gestation while in mice, Nr5a1 transcripts are barely detectable in the sex determining period63,69.

The anti-testis role for NR5A1 that we suggest is therefore most likely of no importance during murine development.

In conclusion, these new NR5A1 data provide a new, apparently frequent genetic cause for 46,XX testicular and OT DSD. Furthermore, these data contribute to the discussion whether 46,XX and 46,XY DSD are two different entities or whether they are two end points of the spectrum of one underlying condition. The fact that this recurrent p.Arg92Trp mutation can result in both 46,XX and 46,XY DSD is strongly in favor of a phenotypic spectrum. Intrafamilial phenotypic variability has also been described in other families in which the underlying molecular defect remains enigmatic70-72. These findings can have an impact on genetic counseling for NR5A1 mutations. Previously,

NR5A1 mutations had only been related to POI in 46,XX individuals, while our data show that other, more severe phenotypes can occur in XX individuals and that within families, phenotypic variability can range from 46,XY gonadal dysgenesis to 46,XX

(ovo)testicular DSD. However, so far only the p.Arg92Trp mutation has been associated with 46,XX DSD.

Identification of FOG2 as a novel disease gene for POI

Many genes involved in gonadal development remain expressed throughout adulthood and may play a role in gonadal maintenance and functioning. Therefore, unraveling the molecular pathways involved in DSD can eventually uncover genetic mechanisms behind more frequently occurring reproductive conditions such as male infertility or POI, as previously mentioned.

271

In the beginning of this PhD project we identified a heterozygous, maternally inherited, FOG2 encompassing deletion in a boy with 46,XY gonadal dysgenesis without cardiac defects. At that time FOG2 affecting rearrangements had only been associated in patients with congenital cardiac anomalies, whether or not in combination with 46,XY DSD73,74. In 2014, a paper reported on the first FOG2 missense mutations found in two cases with isolated 46,XY DSD. Segregation analysis in their families revealed two symptom-free, female carriers75.

In view of a reported mouse phenotype in Fog2 null mice76, suggesting a role for Fog2 in the ovary, hormonal investigations in the index’ mother revealed a low AMH level, putting her at risk for POI. These combined data led us to hypothesize that FOG2 mutations could be an underlying cause of POI. Targeted resequencing of the FOG2 coding region in a French-Belgian POI cohort, convincingly confirmed this hypothesis as it resulted in the identification of four additional heterozygous FOG2 mutations

(1%). Luciferase assays showed a significantly reduced transcriptional activation for three out of four of these variants, supporting this hypothesis.

Given the early prenatal death of Fog2-/- mice, this model is not suitable to study a possible Fog2 role in mature ovarian function; however, in vitro evidence has implicated the Fog2/Gata4 transcription complex in follicular maturation, via Amh induction, and in steroidogenesis77. This situation is reminiscent of the role Nr5a1 plays during both sex development and adult ovarian functioning. NR5A1-associated POI is also characterized by incomplete penetrance, as is the case for FOG2 mutations and

POI. Accordingly, it can be anticipated that in concordance with NR5A1, FOG2 might be involved in other conditions along the broad DSD spectrum as well.

272

LIMITATIONS OF OUR RESEARCH AND FUTURE PERSPECTIVES

More DSD genes and mechanisms to discover

The past four and a half years of work have resulted in the identification of one new disease gene for 46,XY DSD and expansion of the phenotypic spectrum of two known

DSD genes. It goes without saying that the many unresolved cases will extend the DSD disease gene list even further. Given the low prevalence of DSD conditions and the small size of DSD pedigrees, collaboration between different research groups will be key as each of the newly discovered DSD genes will only account for a small portion of DSD cases. In that light, the recent foundation of the SDgenematch system (in the context of the Cooperation in Science and Technology (COST) action “DSD-net”), where research groups will safely disclose their candidate genes, can only be welcomed warmly. Such collaborative initiatives will avoid parallel efforts on similar projects and will allow the identification of additional patients with defects in the same gene. A similar initiative in the field of inherited retinopathies, the European Retinal

Disease Consortium, has already proven to be very useful (http://www.erdc.info/)78.

While the search for a genetic cause is still mostly focused on the coding regions of the human genome, many examples of non-coding defects have been described in various developmental disorders over the last decades79,80. DSDs form a good paradigm for these non-coding mechanisms, as gonadal development is strictly regulated by the temporospatial expression of different transcription factors. In study 5, we have reviewed the currently reported regulatory defects in DSD pathogenesis. Lowering costs and technical improvements will soon allow the application of whole genome sequencing in a clinical setting, allowing to detect coding and non-coding variation and structural defects with one single test. However today one big obstacle remains: with the advent of NGS technologies the bottleneck has shifted from variant identification itself to variant interpretation. This drawback is even more important for non-coding variations as the regulatory code remains poorly understood79. Closing of

273 this “interpretive gap” forms one of the main challenges in the human genetics field80.

The Encyclopaedia of DNA Elements (ENCODE) consortium has provided massive experimental data concerning the function of regulatory elements, and it can contribute to the variant interpretation process although it cannot be seen as the

Rosetta stone for our genetic code80,81. Futhermore, the Roadmap Epigenomics project aims to provide a collection of reference epigenomes and to provide a reference for comparison and integration of regulatory findings82. The major drawback of both

ENCODE and Roadmap data, is the fact that the available information is generated on differentiated cell types, making them less applicable for DSD variant interpretation.

Analogous to the development of different algorithms for coding variant annotation, several comparable tools have been created for non-coding variants, e.g. Genome-

Wide Annotation of VAriants (GWAVA) and Combined Annotation-Dependent

Depletion (CADD)83,84.

New tools for functional validation of candidate genes

Besides these computational approaches for variant interpretation, reliable approaches for in vitro and in vivo functional validation are becoming more important. The introduction of RNA-seq was a revolutionary step forward for transcriptome analysis and the development of different chromosome conformation capture technologies (3C,

4C-seq, Hi-C, Capture-C, …) has enabled the assessment of chromatin interactions and genomic architecture85,86. Although these techniques have proven their utility, they have one disadvantage in common, making them less attractive for the functional validation of candidate DSD genes; all of these techniques require the use of relevant cell types. Even though several gonadal cell types are available, they are all differentiated cell lines87,88. Ideally, functional validation of DSD candidate genes should be performed in cell lines resembling the cellular environment of the urogenital ridge or the bipotential gonad. The same argument applies to patient-derived gonadal tissues, if available; at the time of sampling the gonad is already differentiated.

274

Two promising techniques might be helpful to overcome these issues: induced pluripotent stem cells (iPSC) and ex vivo culturing of entire murine gonads. The iPSC technology allows reprogramming of differentiated adult cells to pluripotent cells and subsequent differentiation to a cell type of choice. Bugamim et al. have described the generation of murine induced Sertoli cells89 and several reports on the generation of human iPSC-derived germ cells have been published90-93. The next step could be the generation of an iPSC-derived cell line that resembles the bipotential gonad. Another opportunity can be found in the ex vivo culturing of whole murine gonads94,95. The main advantage of this technique is that it preserves gonadal tissue integrity. This way cell contacts are maintained and the intercellular signaling between somatic cells and germ cells is unchanged. However, despite optimizations of the culturing protocol, the surrounding molecular environment will inevitably differ from the in vivo situation94,95.

Furthermore, this technique uses murine tissues and although mice are the predominant animal model used in sex developmental research, differences between murine and human sex development might be bigger than initially anticipated. From the known differences in SRY expression patterns, we can expect that these differences start early on in sex development: in mice Sry is only transiently expressed starting at

10.5 dpc and fading away two days later while in humans SRY expression is maintained throughout adult life96,97. If such an important difference is seen that early during sex development, one can only speculate about the deviations further down these pathways.

In an ideal world…

Working in a DSD multidisciplinary team for the past four and a half years has given me the opportunity to witness the extensive efforts of all team members to offer the best possible care to children with a DSD condition and their families. It has also allowed me to reflect on DSD management from different viewpoints.

During this PhD fellowship, I have learned the importance of shared clinical decision making. Joint consultations attended by an endocrinologist, a urologist and a 275 psychologist, where individuals with a DSD condition and their parents get information concerning different aspects of the condition, can only be beneficial to the patients and their families. However, to make these collaborative efforts even more effective, structural support for DSD teams should be available.

A standardized approach consisting of a thorough clinical examination and hormonal evaluation can form a handhold to guide the genetic diagnostic process. If no financial considerations had to be taken into account, WES or eventually WGS would be available for all individuals with a DSD condition. Based on the results of the clinical and hormonal work-up an initial targeted search could be performed, followed by an exome/genome-wide evaluation of the data. However, the implementation of these applications to the routine clinical setting would require better interpretation algorithms and a DSD genetics team for functional validation.

Additional to these clinical and genetic evolutions, I also believe that education is key on several levels. First of all, education of health professionals: too often, wrong or misleading information is given immediately after the birth of a child with atypical genitalia, placing the parents in a very stressful situation. Given that DSDs are rare conditions, local obstetricians and pediatricians only see few cases. Earlier and better referral to a specialized center, would reduce the stress encountered by new parents.

However, local practitioners may not even be aware of the existence of an expertise center for these conditions. In that light, it would be beneficial for DSD management, and by extension, for other rare conditions, if governmental steps would be taken to officially appoint centers of expertise, as is often the case in other countries. The establishment of European networks of reference for rare diseases (ERN), guided by the European Commission, is an excellent opportunity in this matter. These ERNs aim to achieve international cooperation concerning treatment and research options for rare diseases. However, today the Belgian government has not yet appointed official national centers of expertise. In that light and to enhance rare diseases related research and care, the Flemish government has founded the Flemish network for rare disorders,

276 although this project is still in its infancy. Second, general education concerning sex and gender should be adjusted. Before engaging in this PhD fellowship, I had never heard of DSD conditions and I assume this is also the case for most parents at the time of diagnosis. Courageous ‘outings’ by people with a DSD condition can help to raise awareness but if we really want children with DSD to feel confident, we should make sure that variations of sex development are perceived as normally occurring variations of nature, such as variations in eye color, length, sexual orientation, …

REFERENCES

1. Thyen, U., Lanz, K., Holterhus, P.-M. & Hiort, O. Epidemiology and initial management of ambiguous genitalia at birth in Germany. Horm. Res. 66, 195–203 (2006). 2. Ahmed, S. F. Prevalence of hypospadias and other genital anomalies among singleton births, 1988-1997, in Scotland. Arch. Dis. Child. - Fetal Neonatal Ed. 89, 149F–151 (2004). 3. I.A.Hughes, Department of Paediatrics, Addenbrooke’s Hospital, University of Cambridge, Cambridge, U. & Abstract. Early Management and Gender Assignment. Congenit. Endocrinopathies. New Insights into Endocr. Dis. Diabetes. Endocr Dev. Basel, Karger, 2007, vol 11, pp 47–57 Early 11, 47–57 (2007). 4. Hiort, O. et al. Management of disorders of sex development. Nat. Rev. Endocrinol. 10, 520–9 (2014). 5. Bashamboo, A., Ledig, S., Wieacker, P., Achermann, J. & McElreavey, K. New Technologies for the Identification of Novel Genetic Markers of Disorders of Sex Development ( DSD ). Sex. Dev. 15, 213–224 (2010). 6. Baxter, R. M. et al. Exome sequencing for the diagnosis of 46, XY disorders of sex development. J. Clin. Endocrinol. Metab. 100, E333–E344 (2015). 7. Ono, M. & Harley, V. R. Disorders of sex development: new genes, new concepts. Nat. Rev. Endocrinol. 9, 79–91 (2013). 8. Köhler, B. et al. Analysis of the Wilms’ tumor suppressor gene (WT1) in patients 46,XY disorders of sex development. J. Clin. Endocrinol. Metab. 96, E1131-6 (2011). 9. Toppari, J., Virtanen, H. E., Main, K. M. & Skakkebaek, N. E. Cryptorchidism and hypospadias as a sign of testicular dysgenesis syndrome (TDS): environmental connection. Birth Defects Res. A. Clin. Mol. Teratol. 88, 910–9 (2010). 10. van der Zanden, L. F. M. et al. Aetiology of hypospadias: a systematic review of genes and environment. Hum. Reprod. Update 18, 260–83 (2012). 11. Ahmed, S. F. & Rodie, M. Investigation and initial management of ambiguous genitalia. Best Pract. Res. Clin. Endocrinol. Metab. 24, 197–218 (2010). 12. Köhler, B. et al. The spectrum of phenotypes associated with mutations in steroidogenic factor 1 (SF-1, NR5A1, Ad4BP) includes severe penoscrotal hypospadias in 46,XY males without adrenal insufficiency. Eur. J. Endocrinol. 161, 237–42 (2009).

277

13. Allali, S. et al. Mutation analysis of NR5A1 encoding steroidogenic factor 1 in 77 patients with 46, XY disorders of sex development (DSD) including hypospadias. PLoS One 6, e24117 (2011). 14. Brauner, R. et al. Clinical, biological and genetic analysis of anorchia in 26 boys. PLoS One 6, e23292 (2011). 15. Köhler, B., Schumacher, V., l’Allemand, D., Royer-Pokora, B. & Grüters, a. Germline Wilms tumor suppressor gene (WT1) mutation leading to isolated genital malformation without Wilms tumor or nephropathy. J. Pediatr. 138, 421–4 (2001). 16. Su, R., Adam, M. P., Ramsdell, L. & Fechner, P. Y. Can the external masculinization score predict the success of genetic testing in 46 , XY DSD ? AIMS Genet. 2, 163–172 (2015). 17. Ledig, S. et al. Array-CGH analysis in patients with syndromic and non-syndromic XY gonadal dysgenesis: evaluation of array CGH as diagnostic tool and search for new candidate loci. Hum. Reprod. 25, 2637–46 (2010). 18. De Leeneer, K. et al. Flexible, scalable, and efficient targeted resequencing on a benchtop sequencer for variant detection in clinical practice. Hum. Mutat. 36, 379–87 (2015). 19. Eggers, S. et al. Disorders of sex development: insights from targeted gene sequencing of a large international patient cohort. Genome Biol. 17, 243 (2016). 20. Tattini, L., D’Aurizio, R. & Magi, A. Detection of Genomic Structural Variants from Next- Generation Sequencing Data. Front. Bioeng. Biotechnol. 3, 92 (2015). 21. Tan, R. et al. An Evaluation of Copy Number Variation Detection Tools from Whole-Exome Sequencing Data. Hum. Mutat. 35, 899–907 (2014). 22. Johansson, L. F. et al. CoNVaDING: Single Exon Variation Detection in Targeted NGS Data. Hum. Mutat. 37, 457–464 (2016). 23. Carss, K. J. et al. Comprehensive Rare Variant Analysis via Whole-Genome Sequencing to Determine the Molecular Pathology of Inherited Retinal Disease. Am. J. Hum. Genet. 75–90 (2016). doi:10.1016/j.ajhg.2016.12.003 24. Green, R. C. et al. ACMG recommendations for reporting of incidental findings in clinical exome and genome sequencing. Genet. Med. 15, 565–74 (2013). 25. Alkuraya, F. S. The application of next-generation sequencing in the autozygosity mapping of human recessive diseases. Hum. Genet. (2013). doi:10.1007/s00439-013-1344-x 26. Abu-safieh, L. et al. Autozygome-guided exome sequencing in retinal dystrophy patients reveals pathogenetic mutations and novel candidate disease genes. 236–247 (2013). doi:10.1101/gr.144105.112.236 27. Coppieters, F. et al. Identity-by-descent-guided mutation analysis and exome sequencing in consanguineous families reveals unusual clinical and molecular findings in retinal dystrophy. Genet. Med. 16, 671–80 (2014). 28. Zangen, D. et al. Testicular differentiation factor SF-1 is required for human spleen development. J. Clin. Invest. 124, 2071–2075 (2014). 29. Walsh, T. et al. Whole exome sequencing and homozygosity mapping identify mutation in the cell polarity protein GPSM2 as the cause of nonsyndromic hearing loss DFNB82. Am. J. Hum. Genet. 87, 90–4 (2010). 30. Al-Romaih, K. I. et al. Genetic diagnosis in consanguineous families with kidney disease by homozygosity mapping coupled with whole-exome sequencing. Am. J. Kidney Dis. 58, 186–95 (2011). 31. Pippucci, T. et al. EX-HOM (EXome HOMozygosity): a proof of principle. Hum. Hered. 72, 45– 53 (2011). 278

32. Görmez, Z., Bakir-Gungor, B. & Sagiroglu, M. S. HomSI: a homozygous stretch identifier from next-generation sequencing data. Bioinformatics 30, 445–7 (2014). 33. O’Roak, B. J. et al. Exome sequencing in sporadic autism spectrum disorders identifies severe de novo mutations. Nat. Genet. 43, 585–9 (2011). 34. Vissers, L. E. L. M. et al. A de novo paradigm for mental retardation. Nat. Genet. 42, 1109–12 (2010). 35. de Ligt, J. et al. Diagnostic exome sequencing in persons with severe intellectual disability. N. Engl. J. Med. 367, 1921–9 (2012). 36. Ferraz-de-Souza, B., Lin, L. & Achermann, J. C. Steroidogenic factor-1 (SF-1, NR5A1) and human disease. Mol. Cell. Endocrinol. 336, 198–205 (2011). 37. Mongan, N., Tadokoro-Cuccaro, R., Bunch, T. & Hughes, I. Androgen insensitivity syndrome. Best Pract. Res. Clin. Endocrinol. Metab. 29, 569–580 (2015). 38. Orekhova, a. S. & Rubtsov, P. M. DAX1, an unusual member of the nuclear receptor superfamily with diverse functions. Mol. Biol. 49, 65–76 (2015). 39. Ban, S. et al. Genetic polymorphisms of ESR1 and ESR2 that may influence estrogen activity and the risk of hypospadias. Hum. Reprod. 23, 1466–71 (2008). 40. Beleza-Meireles, A. et al. Polymorphisms of estrogen receptor β gene are associated with hypospadias. J. Endocrinol. Invest. 29, 5–10 (2006). 41. van der Zanden, L. F. M. et al. Genetics of hypospadias: are single-nucleotide polymorphisms in SRD5A2, ESR1, ESR2, and ATF3 really associated with the malformation? J. Clin. Endocrinol. Metab. 95, 2384–90 (2010). 42. Asadi, M., Ghafouri-Fard, S., Zare-Abdollahi, D., Ebrahim-Habibi, a & Matin, N. Estrogen receptor mutation in a girl with primary amenorrhea. Clin. Genet. 83, 497–8 (2013). 43. Pettersson, K. & Gustafsson, J. A. Role of Estrogn Receptor Beta in Estrogen Action. Annu. Rev. Physiol. 64, 65–92 (2001). 44. Marino, M., Galluzzo, P. & Ascenzi, P. Estrogen signaling multiple pathways to impact gene transcription. Curr. Genomics 7, 497–508 (2006). 45. Pearlman, A. et al. Mutations in MAP3K1 cause 46,XY disorders of sex development and implicate a common signal transduction pathway in human testis determination. Am. J. Hum. Genet. 87, 898–904 (2010). 46. Bogani, D. et al. Loss of mitogen-activated protein kinase kinase kinase 4 (MAP3K4) reveals a requirement for MAPK signalling in mouse sex determination. PLoS Biol. 7, e1000196 (2009). 47. Loke, J. et al. Mutations in MAP3K1 tilt the balance from SOX9/FGF9 to WNT/??-catenin signaling. Hum. Mol. Genet. 23, 1073–1083 (2014). 48. Ostrer, H. Disorders of Sex Development (DSDs): An Update. J. Clin. Endocrinol. Metab. 99, 1503–9 (2014). 49. Margarit, E. et al. SRY Gene Transferred to the Long Arm of the X Chromosome in a Y-Positive XX True Hermaphrodite. 28, 25–28 (2000). 50. Huang, B., Wang, S., Ning, Y., Lamb, a N. & Bartley, J. Autosomal XX sex reversal caused by duplication of SOX9. Am. J. Med. Genet. 87, 349–53 (1999). 51. Benko, S. et al. Highly conserved non-coding elements on either side of SOX9 associated with Pierre Robin sequence. Nat. Genet. 41, 359–364 (2009). 52. Sutton, E. et al. Identification of SOX3 as an XX male sex reversal gene in mice and humans. Jounal Clin. Investig. 121, 328–341 (2011).

279

53. Haines, B. et al. Interchromosomal insertional translocation at Xq26.3 alters SOX3 expression in an individual with XX male sex reversal. J. Clin. Endocrinol. Metab. 100, E815–E820 (2015). 54. Seeherunvong, T. et al. 46,XX sex reversal with partial duplication of chromosome arm 22q. Am. J. Med. Genet. A 127A, 149–51 (2004). 55. Aleck, K. A., Argueso, L., Stone, J., Hackel, J. G. & Erickson, R. P. True Hermaphroditism With Partial Duplication of Chromosome 22 and Without SRY. Am. J. Hum. Genet. 4, 2–4 (1999). 56. Parma, P. et al. R-spondin1 is essential in sex determination, skin differentiation and malignancy. Nat. Genet. 38, 1304–9 (2006). 57. Tomaselli, S. et al. Syndromic True Hermaphroditism Due to an R-spondin1 ( RSPO1 ) Homozygous Mutation. Hum. Mutat. 29, 220–226 (2008). 58. Bashamboo, a. & McElreavey, K. NR5A1/SF-1 and development and function of the ovary. Ann. Endocrinol. (Paris). 71, 177–182 (2010). 59. Hoivik, E. a, Lewis, A. E., Aumo, L. & Bakke, M. Molecular aspects of steroidogenic factor 1 (SF- 1). Mol. Cell. Endocrinol. 315, 27–39 (2010). 60. Uhlenhaut, N. H. et al. Somatic sex reprogramming of adult ovaries to testes by FOXL2 ablation. Cell 139, 1130–42 (2009). 61. Combes, A. N. et al. Gonadal defects in Cited2-mutant mice indicate a role for SF1 in both testis and ovary differentiation. Int. J. Dev. Biol. 54, 683–9 (2010). 62. Xie, Q.-P. et al. Haploinsufficiency of SF-1 Causes Female to Male Sex Reversal in Nile Tilapia, Oreochromis niloticus. Endocrinology 157, en.2015-2049 (2016). 63. Bashamboo, A. et al. A recurrent p.Arg92Trp variant in steroidogenic factor-1 (NR5A1) can act as a molecular switch in human sex development. Hum. Mol. Genet. 25, 3446–3453 (2016). 64. Igarashi, M. et al. Identical NR5A1 Missense Mutations in Two Unrelated 46,XX Individuals with Testicular Tissues. Hum. Mutat. 1–4 (2016). doi:10.1002/humu.23116 65. Guran, T. et al. Rare causes of primary adrenal insufficiency: Genetic and clinical characterization of a large nationwide cohort. J. Clin. Endocrinol. Metab. 101, 284–292 (2016). 66. Achermann, J. et al. Gonadal Determination and Adrenal Development Are Steroidogenic Factor-1 , in a Dose-Dependent Manner. J. Clin. Endocrinol. Metab. 87, 1829–1833 (1999). 67. Inui, M. et al. Rapid generation of mouse models with defined point mutations by the CRISPR/Cas9 system. Sci. Rep. 4, 5396 (2014). 68. Miyado, M. et al. The p.R92W variant of NR5A1/Nr5a1 induces testicular development of 46,XX gonads in humans, but not in mice: phenotypic comparison of human patients and mutation- induced mice. Biol. Sex Differ. 7, 56 (2016). 69. Takasawa, K. et al. FOXL2 transcriptionally represses Sf1 expression by antagonizing WT1 during ovarian development in mice. FASEB J. 28, 2020–2028 (2014). 70. Slaney, S. F., Chalmers, I. J., Affara, N. A. & Chitty, L. S. An autosomal or X linked mutation results in true hermaphrodites and 46 , XX males in the same family. J Med Genet 35, 17–22 (1998). 71. Kuhnle, U., Schwarz, H. P., Li, U., Cleve, H. & Braun, A. Familial true hermaphroditism : paternal and maternal transmission of true hermaphroditism ( 46 , XX ) and XX maleness in the absence of Y-chromosomal sequences. Hum. Genet. 92, 571–576 (1993). 72. Ramos, E. S. et al. SRY-negative true hermaphrodites and an XX male in two generations of the same family. Hum. Genet. 97, 596–598 (1996).

280

73. Finelli, P. et al. Disruption of friend of GATA 2 gene (FOG-2) by a de novo t(8;10) chromosomal translocation is associated with heart defects and gonadal dysgenesis. Clin. Genet. 71, 195–204 (2007). 74. Tan, Z.-P., Huang, C., Xu, Z.-B., Yang, J.-F. & Yang, Y.-F. Novel ZFPM2/FOG2 variants in patients with double outlet right ventricle. Clin. Genet. 82, 466–71 (2012). 75. Bashamboo, A. et al. Mutations in the FOG2/ZFPM2 gene are associated with anomalies of human testis determination. Hum. Mol. Genet. 1–9 (2014). doi:10.1093/hmg/ddu074 76. Anttonen, M., Ketola, I., Parviainen, H., Pusa, A.-K. & Heikinheimo, M. FOG-2 and GATA-4 Are coexpressed in the mouse ovary and can modulate mullerian-inhibiting substance expression. Biol. Reprod. 68, 1333–40 (2003). 77. Zaytouni, T., Efimenko, E. E. & Tevosian, S. G. GATA transcription factors in the developing reproductive system. Advances in genetics 76, (Elsevier Inc., 2011). 78. Coppieters, F. et al. Isolated and Syndromic Retinal Dystrophy Caused by Biallelic Mutations in RCBTB1, a Gene Implicated in Ubiquitination. Am. J. Hum. Genet. 99, 470–480 (2016). 79. Spielmann, M. & Mundlos, S. Looking beyond the genes: the role of non-coding variants in human disease. Hum. Mol. Genet. 0, ddw205 (2016). 80. Zhang, F. & Lupski, J. R. Non-coding genetic variants in human disease. Hum. Mol. Genet. 24, R102–R110 (2015). 81. Consortium", T. E. project. An integrated encyclopedia of DNA elements in the human genome. Nature 489, 57–74 (2012). 82. Consortium, R. E. et al. Integrative analysis of 111 reference human epigenomes. Nature 518, 317–330 (2015). 83. Ritchie, G. R. S., Dunham, I., Zeggini, E. & Flicek, P. Functional annotation of noncoding sequence variants. 11, (2014). 84. Kircher, M. et al. A general framework for estimating the relative pathogenicity of human genetic variants. Nat. Genet. 46, 310–315 (2014). 85. Wang, Z., Gerstein, M. & Snyder, M. RNA-seq: a revolutionary tool for transcriptomics. Nat. Rev. Genet. 10, 57–63 (2009). 86. Schmitt, A. D., Hu, M. & Ren, B. Genome-wide mapping and analysis of chromosome architecture. Nat. Publ. Gr. (2016). doi:10.1038/nrm.2016.104 87. Havelock, J. C., Rainey, W. E. & Carr, B. R. Ovarian granulosa cell lines. Mol. Cell. Endocrinol. 228, 67–78 (2004). 88. Rahman, N. a & Huhtaniemi, I. T. Testicular cell lines. Mol. Cell. Endocrinol. 228, 53–65 (2004). 89. Buganim, Y. et al. Direct reprogramming of fibroblasts into embryonic sertoli-like cells by defined factors. Cell Stem Cell 11, 373–386 (2012). 90. Ishii, T. Human iPS Cell-Derived Germ Cells : Current Status and. 1064–1083 (2014). doi:10.3390/jcm3041064 91. Mouka, A., Tachdjian, G., Dupont, J., Drévillon, L. & Tosca, L. In Vitro Gamete Differentiation from Pluripotent Stem Cells as a Promising Therapy for Infertility. Stem Cells Dev. 25, 509–21 (2016). 92. Hendriks, S., Dancet, E. A. F., Van Pelt, A. M. M., Hamer, G. & Repping, S. Artificial gametes: A systematic review of biological progress towards clinical application. Hum. Reprod. Update 21, 285–296 (2015). 93. Leng, L. et al. Differentiation of primordial germ cells from induced pluripotent stem cells of primary ovarian insufficiency. Hum. Reprod. 30, 737–748 (2015). 281

94. Mcclelland, K. S. & Bowles, J. Culturing murine embryonic organs : Pros , cons , tips and tricks. 1–7 (2016). doi:10.1016/j.diff.2016.01.008 95. Hogg, K. & Western, P. S. Differentiation of Fetal Male Germline and Gonadal Progenitor Cells Is Disrupted in Organ Cultures Containing Knockout Serum Replacement. Stem Cells Dev. 24, 2899–911 (2015). 96. Hacker, A., Capel, B., Goodfellow, P. & Lovell-badge, R. Expression of Sry , the mouse sex determining gene. 1614, 1603–1614 (1995). 97. Salas-Cortés, L. et al. The human SRY protein is present in fetal and adult Sertoli cells and germ cells. Int. J. Dev. Biol. 43, 135–140 (1999).

282

CURRICULUM VITAE

______

Personalia ______

Name Dorien Baetens Adress Hoofdstraat 12 9200 Dendermonde Telephone +32/ 496686507 E-mail [email protected] Date of birth 15/10/1989 Place of birth Dendermonde, Belgium Nationality Belgian ______

Education ______

2010-2012 Master of Science in Biomedical Sciences magna cum laude Ghent University, Ghent

Thesis: Exploratie van het regulatorisch domein van FOXL2 in patiënten met syndromaal ovarieel falen en identificatie van een nieuw ziektegen voor 46,XY DSD door middel van homozygositeitsmapping Promotor: Prof. dr. Elfride De Baere Tutor: dr. Hannah Verdin

2007-2010 Bachelor of Science in Biomedical Sciences cum laude Ghent University, Ghent

2001-2007 Latin – Latin/Mathematics - Sciences/Mathematics (8h) Sint-Vincentius Instituut, Dendermonde

283

______

Professional activities ______

2012-present PhD candidate in the Center for Medical Genetics Ghent, CMGG and the Department of Pediatric Endocrinology, Ghent University

Thesis: SeX(X)Y genes: unraveling the molecular pathogenesis of disorders of sex development

Promotor: prof. dr. Martine Cools

Co-promotor: prof. dr. Elfride De Baere

______

Grants and awards ______

11-13/06/2015 Best long oral presentation, I-DSD symposium, Ghent, Belgium

20-24/06/2014 Outstanding abstract award, ENDO meeting, Chicago, Illinois, United States

31/05 – 03/06/2014 ESHG Young Investigator Award nominee, ESGH meeting, Milan, Italy 2014

______

Publications ______

Baetens D, De Vriese N, Dulon J, Bachelot A, Touraine P, De Sutter P, Cools M, De Baere E. FOG2 (ZFPM2) haploinsufficiency causes primary ovarian insufficiency. In preparation.

Baetens D, Guran T, De Cauwer L, Mendonca B, Van der Linden M, Stoop H, Looijenga L, De Bosscher K, Cools M, De Baere E. Biallelic and monoallelic ESR2 mutations are associated with 46,XY disorders of sex development. Under review in Genetics in Medicine.

284

Baetens D, Cools M, De Baere E (2016). Non-coding variation in Disorders of sex development. Clinical Genetics 2017: 91: 163–172. doi: 10.1111/cge.12911.

Baetens D, Stoop H, Peelman F, Todeschini AL, Rosseel T, Coppieters F, Veitia R, Looijenga L, De Baere E, Cools M (2016). NR5A1 is a novel disease gene for 46,XX testicular and ovotesticular disorders of sex development. Genetics in Medicine 2016: 19, 367–376. doi: 10.1038/gim.2016.118.

Cools M, Goemare S, Baetens D, Raes A, Desloovere A, Kaufman J, De Schepper J, Jans I, Vanderschueren D, Billen J, De Baere E, Fiers T, Bouillon R (2015). Calcium and bone homeostasis in heterozygous carriers of CYP24A1 mutations: a cross-sectional study. BONE 81 (2015), p.89-96.

Baetens D, Delle Chiaie B, Menten B, Desloovere A, Callewaert B, Van Laecke E, Hoebeke P, De Baere E, Cools M (2014). Extensive clinical, hormonal and genetic screening in a large consecutive series of 46,XY neonates and infants with atypical sexual development. Orphanet Journal of Rare Disease (2014), 9: 209.

Baetens D, Verdin H, Cools M, De Baere E (2013) Forkhead transcription factors in genetic disease. eLS. John Wiley & Sons, Ltd: Chichester.DOI: 10.1002/9780470015902.a0024256

______

Oral presentations ______

Baetens D, De Vriese N, Reynaert N, Dulon J, Bachelot A, Touraine P, De Zegher F, De Sutter P, Cools M, De Baere E. Z ZFMP2 (FOG2) is a new candidate disease gene for primary ovarian insufficiency. Research Day, Student Research Symposium, April 20, 2017, Ghent, Belgium

Baetens D, Todeshini A, Veitia R, Stoop H, Looijenga L, De Baere E, Cools M. A heterozygous NR5A1 missense mutation c.274C>T is a frequent cause of 46,XX (ovo)testicular DSD. 55th Annual ESPE Meeting, September 10-12, 2016, Paris, France

Baetens D, Todeshini A, Veitia R, Stoop H, Looijenga L, De Baere E, Cools M. A heterozygous NR5A1 missense mutation c.274C>T is a frequent cause of 46,XX (ovo)testicular DSD. 16th Annual Meeting of the BeSHG, February 4-5,2016, Leuven, Belgium

285

Baetens D, Güran T, De Cauwer L, Looijenga L, De Bosscher K, Cools M, De Baere E. Identification and functional characterization of ESR2, a new disease gene for 46,XY disorders of sex development (DSD). 5th I-DSD symposium, June 11-13, 2015, Ghent, Belgium

Baetens D, Mladenov V, Delle Chiaie B, Menten B, Desloovere A, Iotova V, Callewaert B, Van Laecke E, Hoebeke P, De Baere E, Cools M. Extensive clinical, hormonal and genetic screening in a large consecutive series of 46,XY neonates and infants with atypical sexual development. Science Day 2015, March 5, 2015, Ghent, Belgium

Baetens D, Güran T, De Cauwer L, Looijenga L, De Bosscher K, Cools M, De Baere E. Identification and functional characterization of ESR2, a new disease gene for 46,XY disorders of sex development (DSD). The European Society of Human Genetics (ESHG) conference 2014, May 31 – June 3, 2014, Milan, Italy

Baetens D, Güran T, De Cauwer L, Looijenga L, De Bosscher K, Cools M, De Baere E. Identification and functional characterization of ESR2, a new disease gene for 46,XY disorders of sex development (DSD). (Poster teaser) 14th Annual Meeting of BeSHG, February 7, 2014, Antwerp, Belgium.

Baetens D, Acke T, Van Cauwenbergh C, delle Chiaie B, Debeaufort C, Cools M, De Baere E. Identification of three NR5A1 mutations in a cohort of patients with 46,XY DSD and POF. (Poster teaser) FWO-WOG meeting, July 3, 2013, Leuven, Belgium

______

Poster presentations ______

Baetens D, De Vriese N, Reynaert N, Dulon J, Bachelot A, Touraine P, De Zegher F, De Sutter P, Cools M, De Baere E. ZFMP2 (FOG2) is a new candidate disease gene for primary ovarian insufficiency. 17th Annual Meeting BeSHG 2017, February 17, 2017, Louvain-La-Neuve, Belgium

Baetens D, Todeschini A, Peelman F, Stoop H, Coppieters F, Rosseel T, Van Laethem T, Looijenga L, Veitia R.A, De Baere E, Cools M. Dysregulation of NR5A1 is a novel and recurrent cause of 46,XX (ovo)testicular disorder of sex development. The European Society of Human Genetics (ESHG) conference 2016, May 21-24, 2016, Barcelona, Spain

286

Baetens D, Reynaert N, De Zegher F, Cools M, De Baere E. Identification of a 3.2 Mb deletion and a missense mutation in ZFPM2 in two 46,XY disorders of sex development patients. American Society of Human Genetics Annual Meeting 2015, October 6-10, 2015, Baltimore, Maryland, United states

Baetens D, Wilhelm M, Delle Chiaie B, Menten B, Desloovere A, Iotova V, Callewaert B, Van Laecke E, Hoebeke P, De Baere E, Cools M. Integrated screening in a large consecutive series of 46,XY neonates and infants with atypical sexual development. 43e jaarlijks congres van de Belgische Vereniging voor Kindergeneeskunde, March 12- 13, 2015, Liege, Belgium

Baetens D, Wilhelm M, Delle Chiaie B, Menten B, Desloovere A, Iotova V, Callewaert B, Van Laecke E, Hoebeke P, De Baere E, Cools M. Extensive clinical, hormonal and genetic screening in a large consecutive series of 46,XY neonates and infants with atypical sexual development. 15th Annual Meeting BeSHG 2015, March 6, 2015, Charleroi, Belgium

Baetens D, Wilhelm M, Delle Chiaie B, Menten B, Desloovere A, Iotova V, Callewaert B, Van Laecke E, Hoebeke P, De Baere E, Cools M. Extensive clinical, hormonal and genetic screening in a large consecutive series of 46,XY neonates and infants with atypical sexual development. 53th Annual ESPE Meeting, September 18-20, 2016, Dublin, Ireland

Baetens D, Güran T, De Cauwer L, Looijenga L, De Bosscher K, Cools M, De Baere E. Identification and functional characterization of ESR2, a new candidate gene for 46,XY Disorders of sex development. ENDO meeting, June 20-24, 2014, Chicago, Illinois, United States

Baetens D, Güran T, De Cauwer L, Looijenga L, De Bosscher K, Cools M, De Baere E. Identification and functional characterization of ESR2, a new disease gene for 46,XY disorders of sex development (DSD). 14th Annual Meeting of BeSHG, February 7, 2014, Antwerp, Belgium.

Baetens D, Güran T, De Cauwer L, Looijenga L, De Bosscher K, Cools M, De Baere E. Identification and functional characterization of ESR2, a new disease gene for 46,XY disorders of sex development (DSD). Keystone symposia: Nuclear Receptors: Biological Networks, Genome Dynamics and Disease, Januari 10-15, 2014, Taos, New Mexico, United States.

287

Baetens D, Acke T, Van Cauwenbergh C, delle Chiaie B, Debeaufort C, Cools M, De Baere E. Identification of three NR5A1 mutations in a cohort of patients with 46,XY DSD and POF. FWO-WOG meeting, July 3, 2013, Leuven, Belgium.

Baetens D, Acke T, Van Cauwenbergh C, delle Chiaie B, Debeaufort C, Cools M, De Baere E. Identification of three NR5A1 mutations in a cohort of patients with 46,XY DSD and POF. 13th Annual Meeting of BeSHG, March 15th, 2013, Brussels, Belgium.

______

Courses and Workshops

______

2010 – 2011 Certificate FELASA category C, Laboratory Animal Science, Ghent University, Ghent, Belgium

10 – 14/09/2012 The Omics revolution, FTNLS, Leuven, Belgium

17 – 21/09/2012 Workshop Next-Generation Sequencing, FTNLS, Leuven, Belgium

25/02 – 07/03/2013 Effective Oral presentations, Principiae, Doctoral Schools, Ghent, Belgium

07/06/2013 DSD Training Workshop for New Investigators, I-DSD, Glasgow, United Kingdom

1 – 4/07/2013 The Epigenetics Revolution: development and disease biology revisited, FTNLS, Leuven, Belgium

Oct 2013 – Jun 2014 Permanent Education Course in Human Genetics, BeSHG, Interuniversitary, Belgium

25/08 – 01/09/2014 Speed reading course, Doctoral Schools, Ghent, Belgium

22 – 23/10/2014 Biogazelle qPCR course: from qPCR experiment design to data analysis, Ghent, Belgium

8 – 10/06/2015 1st DSDnet Training School: holistic care and research in DSD, Ghent, Belgium

15 – 16/09/2016 f-TALES workshop “Light on the dark side of the genome”, Ghent, Belgium

288

22/09/2016 9th From PhD to Job Market: 'Believing in First Mover Advantages”, Doctoral School, Ghent, Belgium

______

Conferences ______

- 11th Annual Meeting of the BeSHG, March 4, 2011, Louvain-la-Neuve, Belgium. - 12th Annual Meeting BeSHG 2012, March 2, 2012, Liège, Belgium. - NCMLS New Frontiers symposium, December 3-5, 2012, Nijmegen, The Netherlands. - 13th Annual Meeting of BeSHG, March 15, 2013, Brussels, Belgium. - 4th I-DSD Symposium, June 7-9, 2013, Glasgow, United Kingdom. - FWO-WOG meeting, July 3, 2013, Leuven, Belgium. - ‘DSD: van embryo naar senoir’, symposiareeks van de kinderchirurgische groep Sophia 150 jaar, December 12, 2013, Rotterdam, The Netherlands. - Keystone symposia: Nuclear Receptors: Biological Networks, Genome Dynamics and Disease, Januari 10-15, 2014, Taos, New Mexico, United States. - 14th Annual Meeting of BeSHG, February 7, 2014, Antwerp, Belgium. - The European human genetics conference 2014, May 31 – June 3, 2014, Milan, Italy. - 16th International Congress of Endocrinology & the Endocrine Society, 96th Annual Meeting & Expo, June 20-24, 2014, Chicago, Illinois, United States. - Ghent University Science Day 2015, March 5, 2015, Ghent, Belgium. - 15th Annual Meeting BeSHG 2015, March 6, 2015, Charleroi, Belgium. - 5th I-DSD Symposium, June 11-13, 2015, Ghent, Belgium - American Society of Human Genetics Annual Meeting 2015, October 6-10, 2015, Baltimore, Maryland, United States. - 16th Annual Meeting of the BeSHG, February 4-5, 2016, Leuven, Belgium. - The European human genetics conference 2016, May 21-24, 2016, Barcelona, Spain. - 55th Annual ESPE Meeting, ESPE 2016, September 10-12, 2016, Paris, France.

______

Supervision of students ______

Master thesis

2013-2014 Ibran Mertens Master of Science in the Industrial Sciences: Biochemistry

289

Mutatiescreening van FOG2 in een 46,XY DSD patiëntencohorte en identificatie van nieuwe kandidaatgenen met behulp van IBD-mapping Promotor: Prof. dr. Elfride De Baere Co-promotor: Dr. Katrien Strubbe Tutor: Dorien Baetens

2014-2015 Malaïka Van der Linden Master of Sciences in Biomedical Sciences Homozygositeitsmapping en exoomsequenering voor de genetische karakterisatie van 46,XY disorders of sex development Promotor: Prof. dr. Elfride De Baere Tutor: Dorien Baetens

2015-2016 Santusha Prija Bihari Master of Sciences in Biomedical Sciences NXT-DSD: next-generation sequencing technologieën voor de genetische karakterisatie van disorders of sex development (DSD) and primaire ovariële insuffficiëntie Promotor: Prof. dr. Elfride De Baere Tutor: Dorien Baetens

2016-2017 Niels De Vriese Master of Sciences in Bio-engineering Promotor: Prof. dr. Elfride De Baere Tutor: Dorien Baetens

Z-lijn papers, 2nd Bachelor of Medicine

2012-2013 Kim Van der Eecken Personal genome sequencing: mogelijkheden en ethische implicaties. Promotor: Prof. dr. Elfride De Baere Supervision: Dorien Baetens

290

2013-2014 Alexander Barbary Incidentele vondsten bij next generation sequenicng: Melden of niet? Promotor: Prof. dr. Elfride De Baere Supervision: Dorien Baetens

2013-2014 Maurits De Wolf Personal genome sequencing: mogelijkheden en ethische implicaties. Promotor: Prof. dr. Elfride De Baere Supervision: Dorien Baetens

2014-2015 Ka Chien Choi Ethische implicaties van genome sequencing: incidental findings, vertellen of verzwijgen? Promotor: Prof. dr. Elfride De Baere Supervision: Dorien Baetens

2014-2015 Sören Verleysen Genoomsequenering: incidentele vondsten en ethische implicaties. Promotor: Prof. dr. Elfride De Baere Supervision: Dorien Baetens

2015-2016 Jelle Van Damme Ethische implicaties bij accidentele ontdekkingen in genoomwijd onderzoek. Promotor: Prof. dr. Elfride De Baere Supervision: Dorien Baetens

2015-2016 Tessa Seys Personal genome sequencing: ethische implicaties. Hoe omgaan met incidentele resultaten bij genome sequencing? Promotor: prof. dr. Elfride De Baere Supervision: Dorien Baetens

291

292

Dank je wel…

Dit boekje en het bijhorende werk is niet alleen mijn verdienste en is slechts tot stand kunnen komen dankzij de hulp van vele anderen. Aan het einde van dit doctoraatstraject wil ik uiteraard graag alle mensen bedanken die op welke manier dan ook hebben bijgedragen tot deze thesis. Er wordt vaak gezegd dat iets bloed, zweet en tranen heeft gekost, een zegswijze die hier wel zeer toepasbaar is.

Bloed... van de patiënten. Zonder hun toestemming en medewerking kan er geen onderzoek gevoerd worden, kunnen geen resultaten bekomen worden en kon deze thesis niet geschreven worden. Als onderzoeker staan we hier niet altijd bij stil, dus dank aan de patiënten en hun ouders.

Zweet... van zowel mezelf als mijn promotoren,

Martine en Elfride. Heel erg bedankt om mij onder jullie vleugels te nemen als jonge doctoraatsstudent en mij de afgelopen jaren zo enthousiast te begeleiden. Ik kan mij geen twee betere rolmodellen voorstellen om aan een (hopelijk) succesvolle loopbaan te beginnen. Niet alleen stonden jullie steeds paraat met jullie onuitputtelijk advies en uitgebreide expertise, jullie hebben mij ook gemotiveerd op momenten dat alles minder haalbaar leek en terecht gewezen wanneer ik nonchalant werd. Ook aan het einde van mijn doctoraatsperiode ben ik nog steeds verwonderd over hoe doortastend, perfectionistisch en gepassioneerd jullie zijn in jullie job. Jullie combineren met schijnbaar gemak al jullie taken en verantwoordelijkheden en hebben mij geleerd dat je met de nodige inzet erg veel kan verwezenlijken.

293

Toen ik als student eerste master startte aan mijn onderzoeksstage waren er slechts drie doctoraatsstudenten in het DNA-labo. Kim, Frauke en Hannah, jullie waren al ervaren rotten toen ik een groentje was en hebben me vaak verder kunnen helpen met jullie ervaring, bedankt! Ook voor mijn generatie: Caroline, Miriam, Basamat en

Kristof; een dikke merci. Samen hebben we het grootste deel van onze doctoraatsperiode doorgebracht in Blok B. Bedankt om er te zijn wanneer het even wat minder ging en om die vele uren die we in ons bureau doorbrachten op te vrolijken met grappige you-tube filmpjes, ridicule zelftests, koekjespauzes, de loofkes en crazy lunchbreaks maar bovenal met jullie aanwezigheid. Voor de nieuwere garde: Giulia,

Sarah, Stijn, Mattias, Jeroen, Toon, Greet, Marieke en Thalia. Het was fijn om de afgelopen jaren steeds nieuwe collega’s te krijgen en de groep te zien groeien.

Naast deze researchbende zijn er ook nog de laboranten, ook aan jullie een welgemeende dank-je-wel. Om één van mijn favoriete tv-quotes aan te halen en omdat ik het Elien beloofd had: ‘ ik ben blij dat jullie in mijn team zitten’ (Guido Pallemans -

Het Eiland). Bedankt voor de antwoorden op al mijn vragen en voor de babbels tijdens het PCR’en of sequenties bekijken. Het is ook mede dankzij jullie dat ik altijd (of op zijn minst meestal) met veel plezier kwam werken.

Daarnaast wil ik ook de andere leden van het DSD-team bedanken. Dankzij jullie verschillende achtergronden heb ik geleerd om ook verder te kijken dan de DNA- moleculen lang zijn. Bedankt om mij meer vertrouwd te maken met de klinische beelden op de DSD combi poli’s, om mij te laten mee volgen met een operatie, om mij ook in te leiden in de psychologische kant van de zaken. Aan ieder van jullie, bedankt om mij jullie kant van het DSD-verhaal te tonen.

Karolien en Lode, ook jullie verdienen hier een afzonderlijke paragraaf. Karolien, bedankt om zo hard met ons te willen meedenken in het ESR2 verhaal en om me te laten proeven van de sfeer in jullie labo. Lode, bedankt om mij op te leiden tot professioneel transfecteerder (sense the irony, het zal wel aan de HEBS gelegen

294 hebben). Bedankt voor de vele cellen die je voor mij uitzaaide, ik kwam steeds met veel plezier naar jullie labo.

Reiner, Anne-Laure and the other members of the Paris team. Thank you for welcoming me in your lab. Thank you for letting me be a part of the Parisian life, even though it was so suddenly shaken while I was there. Thank you for the experience and for your input in the NR5A1 project.

En tranen... van mijzelf bij mijn achterban (en natuurlijk ook veel gelach en plezier). In de laatste, maar zeker niet minst belangrijke plaats, wil ik mijn vrienden en familie bedanken. Een doctoraatsperiode durft al wel eens tot twijfels en de daarbij horende emoties leiden (vooral dan de laatste maanden). Dankzij elk van jullie, kon ik er steeds weer voor een tijdje tegenaan.

Voor mijn vriendinnen: de BMW-girls, de SVI-survivors en mijn mede-oudleiding bij de chiro aka ‘de cavakeelkes’. Dankzij de gezellige avonden met elk van jullie werden de laatste zware maanden toch wat draaglijker. Bedankt om mijn hoofd eens af en toe leeg te maken en mijn maagje daarbij meestal te vullen. Ik kijk al uit naar de volgende komen drinken/eten, weekendjes en andere dates.

Ook voor mijn familie is een bedankje op zijn plaats. Voor moeke, meter en peter, de tantes en nonkels, neven en nichten. Ondanks de grote hoeveelheden eten die we steeds samen verorberen, kan nog steeds iedereen door dezelfde deur, iets wat in veel families niet vanzelfsprekend is! Ook dat is van onschatbare waarde. Bedankt om regelmatig eens te vragen hoe het opschoot, wat ik nu weer net onderzocht, of ik al publicaties had en wanneer jullie nu eindelijk naar die lang verwachte receptie mochten komen 

Daarnaast wil ik ook graag mijn “schoonfamilie” bedanken. Jullie deur staat ook altijd open voor mij en het heeft niet heel lang geduurd voor ik mij ook in jullie huis thuis voelde. Roger, Martine, Thomas en Sylvia: bedankt om er te zijn!

295

Ook voor mijn zus Evelien, bedankt om mijn oudere zus te zijn. Om mij te vertellen dat alles altijd wel goed komt en om beter te kunnen relativeren dan ik. Bedankt ook voor Gaston, de rakker, mijn metekindje (Tim: hiervoor wil ik jou ook een klein beetje bedanken ). Jullie kleine, guitige uk tovert steeds weer een lach op mijn gezicht.

Enkele van de meest bijzondere woorden van dank gaan naar mijn ouders. Bedankt papa, om er telkens weer op te hameren hoe belangrijk studeren was en om steeds het onderste uit de kan te willen. Het ging er vroeger langs het ene oor in en het andere weer uit, maar aangezien ik hier nu sta, moet er toch een en ander blijven hangen zijn.

Bedankt mama, om er telkens voor ons te staan. Als een echt moeke was/ben je altijd in de weer voor jouw dochters. Geen vraag is jou te veel, geen boodschap te zwaar, geen omweg te ver. Kort samengevat: zonder de kansen die jullie ons gegeven hebben en het warme nest waarin we opgroeiden, had ik dit nooit kunnen doen.

Tot slot nog een grote dank-je-wel voor de hoofdrolspeler in mijn leven, diegene die de laatste maanden met al mijn grillen heeft moeten leven en die dat hopelijk nog lang wil doen (al zal ik de grillen een beetje proberen te beperken). Lieve Sam, bedankt om mij te doen lachen, mij op te peppen, mij te kalmeren, mij soms gewoon te laten zijn.

Kortom, bedankt voor ALLES. Ik hoop voor jou dezelfde steun en toeverlaat te kunnen zijn, terwijl jij jouw doctoraat afrondt.

296

297

298