A Long contiguous stretches of homozygosity disclosed a novel biallelic pathogenic variant in STAG3 causing Primary Ovarian Insufficiency


 Background: Primary ovarian insufficiency (POI) refers to an etiologically heterogeneous disorder characterized by hypergonadotropic hypogonadism that represents a major cause of infertility in women under 40 years. Most cases of isolated POI still appear sporadically, but ∼10–15% have an affected first-degree relative, indicating a significant genetic etiology. Several genes implicated in development, meiosis, hormonal signaling and metabolism are involved in the genetic form of the disorder in both syndromic and isolated POI. However, most cases of POI remain unsolved even after exhaustive investigation. Results: Here is reported a 19-year-old Senegalese female affected by non-syndromic POI showing primary amenorrhoea, who well answered to the hormonal induction of puberty reaching a complete sexual maturation in two years. In order to investigate the presence of a genetic defect, aCGH-SNP analysis was performed. A 13.5 Mb long contiguous stretch of homozygosity (LCSH) region on chromosome 7q21.13-q22.1 was identified where the exome sequencing revealed a novel homozygous 4-bp deletion (c.3381_3384delAGAA) in STAG3 . Pathogenic variants in this gene, encoding for a meiosis-specific protein, have been previously reported as cause POI in only seven families and recently as cause of infertility in a male. The here identified mutation leads to the truncation of the last 55 aminoacids, confirming the important role in meiosis of the STAG3 C-terminal domain. Conclusions: In conclusion we identified a loss of function variant in STAG3 in a Senegalese woman with POI reinforcing the role the cohesin complex in the genetic etiology of this disorder. This gene should be included in the screening of POI to offer a better genetic counseling and long term follow-up considering the risk of ovarian tumor in woman carrying pathogenic variants in genes involved in germ cell differentiation.


Background
Primary ovarian insufficiency (POI) is characterized by depletion of ovary follicles, leading to hypoestrogenism and hypergonadotropic hypogonadism, infertility and amenorrhea in women younger than 40 years (1,2). This disorder represents one of the main causes of infertility affecting approximately 1% of women (3,4) with a worldwide prevalence of 3.7% (5). Women with POI may show a wide range of clinical phenotypes, from primary (PA) to secondary amenorrhea (SA) and other congenital or acquired abnormalities and present either syndromic or isolated forms of the disease (6,7). PA is usually diagnosed in adolescence in patients showing delayed puberty and absence of secondary sex characteristic development, while SA, which represents the most frequent POI phenotype, may occur at any age after menarche and is characterized by normal pubertal development (8,9). Although POI is a heterogeneous disorder caused by iatrogenic, viral, or autoimmune factors, more than 70% of cases remain idiopathic (1). Chromosomal aberrations have long been recognized as a cause of POI and at least 10-13% of the syndromic cases present anomalies revealed by standard karyotype (10). About 10-15% of women with POI have an affected first-degree relative and different modes of inheritance can be observed in families, suggesting the presence of several monogenic causes in the etiology of the disease (7). To date the use of different genomic approaches, including linkage studies, sequencing of candidate genes and whole exome sequencing (WES) allowed the identification of pathogenic alterations in more than 60 genes implicated in both syndromic and isolated POI (1,10). Additionally, submicroscopic copy number variations (CNVs) encompassing genes potentially implicated in reproductive function have emerged as an important genetic determinant in POI (11). The genes altered in POI encode for protein involved in DNA repair and meiosis (1) and pathogenic variants also predispose to different forms of cancer (6,12,13,14). Thus the early diagnosis of the molecular mechanism underlying POI is essential to develop strategies for preventing the irreversible consequences on fertility, to improve clinical management and to perform genetic counseling with a long-term follow-up also considering tumor susceptibility.
Mutations in STAG3 (stromal antigen 3), which encodes a subunit of the cohesin complex participating to sister chromatid pairing during meiosis, have been identified as a rare POI monogenic cause. To date STAG3 biallelic variants have been reported in seven families worldwide, five of which were consanguineous pedigree (6,(14)(15)(16)(17)(18)(19). All the affected women had isolated POI except for a patient belonging to a Palestinian family (6) who presented simultaneous bilateral ovarian tumors.
Here we report the identification of a novel biallelic pathogenic variation in STAG3 through a combined approach of CGH-SNP microarray and Clinical Exome Sequencing (CES) in a young Senegalese woman affected by non-syndromic POI.

Patient description
The patient was a 19-year-old female from Senegal, referring to the Emergency Room Department for dyspareunia and pelvic pain after her first sexual intercourse. She was also suffering from primary amenorrhea and anosmia. Familiar anamnesis revealed the absence of parental consanguinity. Computerized tomography (CT) investigation showed vaginal stenosis and failure to display ovaries. Physical examination showed absence of breast and pubic hair development, consistent with Tanner stage 1.
After hormonal and radiological diagnostic work-up, a puberty induction therapy using transdermal 17ß-oestradiol (oestradiol hemyhidrate 25 µg/3 days for 2 months, then 50 µg/3 days) was started.   Then VCF files were annotated with the wANNOVAR tool. Finally the variants were filtered and prioritized using a personalized bioinformatics pipeline: variants with a read coverage of less than 5x and a Q score below 20 were filtered out; allelic frequency in public databases variants <0.1% (1000 Genomes, the ESP cohort data set, GnomAD, Exome Aggregation Consortium) were considered; synonymous variants were excluded; for the missense variants, at least four in-silico prediction tools (SIFT, CADD, Polyphen, MutationTaster) were used.

Results
An array CGH-SNP assay was used as first tier test to investigate the presence of both DNA copy number variations (CNVs) and Loss of heterozygosity (LOH) regions. No chromosomal unbalance was identified. However the aCGH-SNP revealed a LCSH (Long contiguous stretches of homozygosity) region of 13.5 Mb (from rs2374083 to rs1990167) on chromosome 7q21.13-q22.1 (Fig 2A). This region contained 127 OMIM genes among which the best candidate to explain the patient's phenotype was STAG3, whose mutations had been previously detected in patients with autosomal recessive POI (6,(14)(15)(16)(17)(18)(19).  (20). We did not find any other mutations in genes known to cause POI and consequently STAG3 p.E1128Mfs*42 was considered the best disease-causing candidate variant in this patient. The best candidate in the LCSH region was STAG3, that has been previously associated with a recessive form of POI consistent with the patient's phenotype. STAG3 encodes a subunit of the multiprotein cohesin complex required during meiosis I for homologous chromosome pairing, correct synapsis and segregation of chromosomes, proper recombination and DNA repair (23). In mammals, meiosis-specific cohesin subunits includes a SMC1 subunit (SMC1β), two additional αkleisins (RAD21L and REC8) and a stromal antigen protein (STAG3) that are specifically 9 expressed in meiosis. Moreover, Stag3 specifically localizes on the spindle apparatus and is required for microtubule stability and spindle assembly to preserve the euploidy in the mouse eggs (24).

Discussion
Females Stag3 -/mice lack ovarian follicles indicating a severe ovarian dysgenesis (25) and also develop ovarian tumors as observed in humans (6). It is likely that homozygous or compound heterozygous variants in STAG3 may also affect male fertility due to the early prophase I arrest and apoptosis in both sex germ cells. Indeed, all Stag3 -/male mice described to date are infertile (26,27) and a biallelic loss-of-function STAG3 variant has been recently reported in an infertile male (28) affected by complete bilateral meiotic arrest.
Up to date, only ten variants in STAG3 have been reported as genetic causes of POI in seven pedigrees (Fig 3) (6,(14)(15)(16)(17)(18)(19) with a predicted loss of function effect in most cases. The here identified STAG3 frameshift variant is a 4-bp deletion in exon 30 and the resulting transcript would either undergo nonsense-mediated decay (29) or result in a truncated protein devoided of the Cterminal domain similarly to a previously described frameshift mutation truncating the last 206 C-terminal residues, in exon 28 (14). The C-teminal domain is specific of STAG3 as it is not conserved neither in STAG1 nor in STAG2, that belong to the same family but are involved in somatic cell division (30). The findings of two POI patients with pathogenic variants truncating the C-terminal support an important specific role of this STAG3 domain during meiosis.
In the present patient, hormone replacement therapy (HRT) allowed to induce puberty and achieve complete secondary sexual characteristics, adequate growth, uterine development, and menstrual fluxes. HRT currently represent the mainstay therapy for women with POI and plays a pivotal role in reducing long-term comorbidities (e.g. osteoporosis and cardiovascular disease) and improving sexual health by restoring normal serum estrogen concentrations according to age (3,8,31).
Taken together our observations strength the importance of a correct diagnosis and clinical management in women with POI. The diagnostic workflow should include the molecular test with multi-gene panel including STAG3 to allow an appropriate ovarian monitoring in the long term follow-up, considering the high risk to develop ovarian tumors.

Conflict of Interest
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Availability of data and materials
All data generated in this study are included in the published article