A novel homozygous growth differentiation factor 9 variant associated with premature ovarian insufficiency: A case report

Abstract

Background

Premature ovarian insufficiency (POI) is a condition marked by reduced ovarian function. The variants in several genes have been identified in association with POI. Growth differentiation factor 9 (GDF9) is one of these genes, which encodes a secreted protein with an essential role in follicular development.

Case Presentation

Our study investigates the cause of early secondary amenorrhea in 2 affected sisters with POI from a consanguineous Iranian family. Exome sequencing identified a novel homozygous GDF9 variant (c.275T $>$ C; p.Leu92Pro) in the probands. Based on family segregation analysis, the variant was detected in a heterozygous state in the parents of the affected sisters and their grandmother. Notably, neither the mother nor the grandmother showed symptoms of the disease. Bioinformatics analysis and protein structural alteration due to the substitution of leucine with proline also supported the deleterious potential impact of this variant.

Conclusion

In summary, our findings highlight the role of GDF9 biallelic variants in the etiology of POI and suggest a more cautious interpretation of heterozygous variants in the context of disease.

1. Introduction

Premature ovarian insufficiency (POI) is a condition characterized by altered ovarian function, usually diagnosed in women during the second or third decade of life (1). Recent research has provided valuable information regarding its etiology, clinical features, management options, and impact on quality of life. The prevalence of POI is approximately 1–3% among women under the age of 40, but the condition's etiology remains complex and mostly unknown (2).

The major clinical manifestations of POI are amenorrhea for at least 4 months, increased levels of gonadotropins, most particularly follicle-stimulating hormone (FSH), and hypoestrogenism.

Individuals with POI often experience symptoms such as hot flashes and night sweats (1, 2). The diagnosis of POI usually requires clinical evaluation and laboratory tests to provide an initial differential diagnosis from hypogonadotropic hypogonadism, polycystic ovary syndrome, and related disorders. Transvaginal ultrasonography may also reveal small and inactive ovaries or a depleted follicle reserve (3, 4).

Autoimmune disorders and genetic abnormalities, such as turner and fragile X syndromes, are the well-known causes of POI (5). Additionally, variants in other genes involved in follicle development, such asgrowth differentiation factor 9 (GDF9), have been associated with POI. GDF9 encodes a member of the transforming growth factor-beta superfamily secreted by oocytes, which plays a key role in folliculogenesis. Initial studies indicated autosomal dominant inheritance for GDF9-related POI, whereas recent evidence suggested an autosomal recessive pattern (6). Our study investigated the genetic basis of POI in an Iranian family with 2 affected sisters using whole-exome sequencing (WES). Our findings highlight the potential autosomal recessive inheritance of POI caused by GDF9 variants, and further expand the mutational spectrum of this gene associated with the disease.

2. Case Presentation

We investigated 2 POI-affected sisters (IV-1 and IV-2) (Figure 1), who were referred to Shahid Mohammadi hospital, Bandar Abbas, Iran.

They experienced menarche at ages 14 and 13 and subsequently developed secondary amenorrhea at ages 17 and 15, respectively. The physical examination revealed normal external genitalia and secondary sexual characteristics. Hormonal evaluations demonstrated elevated gonadotropins in participants IV-1 (FSH = 61.3 IU/L, luteinizing hormone [LH] = 37.4 IU/L) and IV-2 (FSH = 77.1 IU/L, LH = 43.8 IU/L), with low estradiol levels and absent anti-Müllerian hormone levels, confirming a hypergonadotropic hypogonadism profile. The ultrasonography revealed small ovaries in both cases. In case IV-1, the left ovary measured 11 $\times$ 12 $\times$ 4 mm and the right ovary 16 $\times$ 14 $\times$ 6 mm. In case IV-2, the left ovary measured 16 $\times$ 11 $\times$ 4 mm and the right ovary 18 $\times$ 15 $\times$ 5 mm. Antral follicles were absent in both probands, while uterine volumes remained within normal limits. Additional laboratory investigations excluded other common causes of POI. The laboratory evaluations of thyroid function, together with adrenal hormone profiles, demonstrated normal results, and cytogenetic analysis revealed normal female karyotypes (46, XX) in the participants. The molecular examination for the fragile X mental retardation-1 premutation also has a negative outcome. Hormone therapy was initiated for them, consisting of 10 mg medroxyprogesterone acetate for 14 days per month and 1.25 mg conjugated equine estrogen for 25 days per month. This regimen, together with vitamin D and calcium supplementation, effectively managed their symptoms.

WES was conducted for both affected sisters. The SureSelectQXT kit (Agilent) was used for library preparation, and paired-end sequencing with 100x coverage was performed on the NovaSeq 6000 (Illumina, San Diego, CA, USA). The FastQC tool (v.0.11.9) was used to evaluate the quality of the Fastq data, and genomic alignment against the hg19 assembly was performed using BWA-MEM (v.0.7.17). The variants were called GATK HaplotypeCaller (v.4.4) and then annotated with ANNOVAR. Variants were filtered based on minor allele frequency (MAF $<$ 1%) and American College of Medical Genetics and Genomics guidelines, with only high-confidence variants selected for final interpretation. The ClinVar, Franklin, Varsome, and InterVare databases were used to examine variant classification, and the OMIM, DisGeNET, and GeneCards databases were used to investigate the gene-disease associations.

WES identified a novel homozygous GDF9 gene variant (NM_005260.6: c.275T $>$ C; Leu92Pro) in the POI-affected sisters. PCR amplification (95 C for 5 min, followed by 28 cycles at 95 C for 30 sec, 58 C for 45 sec, 72 C for 30 sec, and 72 C for 5 min), Sanger sequencing, and subsequent segregation analysis were conducted using specifically designed forward 5 $\text{'}$ -TTTTCCTATTAGCCTTGGTTCTC-3 $\text{'}$ and reverse 5 $\text{'}$ -CACATACCTGTTACCTGGTCTCC-3 $\text{'}$ primers (Figures 1 and 2).

The variant was observed to be absent from the Iranome database (https://iranome.com) and Genome Aggregation Database v2.1.1 (gnomAD v2.1.1) (https://gnomad.broadinstitute.org). Based on American College of Medical Genetics and Genomics criteria, the c.275T $>$ C variant was categorized as a variant of uncertain significance, supported by PM1, PM2, and PP3 evidence. Analyses using various in-silico tools, such as I-mutant2.0 and MutationTaster, predicted the effect of the variant (Table I). Conservation assessment was performed using ConSurf and Clustal Omega (Figure 3). Structural prediction of the region containing Leu92 was generated using the PSIPRED database (https://bioinf.cs.ucl.ac.uk) (Figure 4), and 3-dimensional molecular modeling was conducted with Chimera software (Figure 5).

Figure 1.

Family pedigree and molecular genetic analysis. The arrows represent the probands. Filled symbols indicate affected individuals, and half-filled symbols indicate carriers.

Figure 2.

Chromatograms of the identified variant in the affected sisters (IV-1 and IV-2), their carrier parents (III-2 and III-3), and grandmother (II-5). The wild-type chromatograms (III-4, III-6) are not shown.

Figure 3.

Sequence alignment demonstrates high conservation of the affected residue (Leu92) across species.

Figure 4.

Leu92 is located within an α-helix region of the protein, a structural element important for maintaining local stability.

Figure 5.

Substitutions alter protein properties, with proline disrupting the α-helix and hydrogen bonding. Blue dashed lines indicate hydrogen bonds.

Table 1.

In silico predictions of the c.275T $>$ C variant

In silico tool	Complete form	Prediction	Score
MetaRNN	Meta recurrent neural network	Deleterious	0.95
REVEL	Rare exome variant ensemble learner	Deleterious	0.736
PolyPhen-2	Polymorphism phenotyping v2	Deleterious	0.999
CADD	Combined annotation dependent depletion	Pathogenic	26.1
MutPred	-	Pathogenic	0.908
DANN	Deleterious annotation of genetic variants using neural networks	Deleterious	0.999
DEOGEN2	-	Deleterious	0.964
AlphaMissense	-	Pathogenic	0.939
ClinPred	-	Pathogenic	0.997
PROVEAN	Protein variation effect analyzer	Deleterious	-5.81
FATHMM-XF	Functional analysis through hidden Markov models eXtended features	Deleterious	0.98
Mutationassessor	-	Medium	2.6
SIFT	Sorting intolerant from tolerant	Pathogenic	0
Mutation taster	-	Deleterious	0.8
DynaMut2	-	Decrease stability	-0.9 kcal/mol
I-Mutant2.0	-	Decrease stability	-1.59 kcal/mol
GERP	Genomic evolutionary rate profiling	High constraint	6.02

Ethical Considerations

Written informed consent was obtained from both cases in this study.

3. Discussion

POI manifests through ovarian dysfunction and menopause-like symptoms, while genetic defects and autoimmune disorders serve as the main causes for its development. However, the underlying reason remains unidentified in more than half of POI cases (7).

In this study, we describe a family where 2 sisters were affected by early secondary amenorrhea before age 20. WES analysis identified a homozygous GDF9 variant (c.275T $>$ C; Leu92Pro) in the probands, whereas their parents and grandmother were heterozygous carriers. Neither the carrier mother nor the grandmother exhibited clinical features of POI or premature menopause. Their mother and grandmother experienced menopause at ages 49 and 55, respectively. The variant was absent in population databases, and in silico analyses predicted its deleterious effects. According to the PSIPRED database, leucine at position 92 was located within an α-helix. Therefore, the Leu92Pro substitution was expected to disrupt local protein structure and α-helix formation.

GDF9 functions as an oocyte-produced growth factor that plays a crucial role in the initial stages of follicle development and in maintaining ovarian health. Different genetic variations in the GDF9 gene show association with both poor follicle development and diminished ovarian reserve, alongside cases of familial POI. Functional studies indicate that several GDF9 missense variants impair protein maturation and reduce the biological activity of conditioned media from HEK293F cells compared with the wild-type protein (8).

Initial cohort studies suggested that heterozygous GDF9 variants may cause POI in an autosomal dominant manner. Studies in French, Chinese, and Indian cohorts identified several heterozygous missense variants, including p.Ser186Tyr, p.Pro103Ser, p.Thr238Ala, p.Lys67Glu, and p.Val216Met in individuals with POI. Most variants were rare, and their pathogenicity remained uncertain. However, functional assays demonstrated the effect of missense variants, such as p.Lys67Glu and p.Pro103Ser, on protein function. In contrast, recent studies indicate that heterozygous variants have a milder effect. These studies suggested autosomal recessive inheritance of the disease due to GDF9 variants. They also claimed that heterozygous variants of this gene are associated with normal ovarian function or secondary amenorrhea at older ages (1, 8–11) (Table II).

The normal phenotypes of the mother and grandmother of our study proband further support the recessive mode of inheritance. It was observed that the clinical severity of POI is influenced by both the number of affected GDF9 alleles and the specific nature of the mutations. While a single heterozygous missense variant may play a contributory role in the onset of premature menopause, it may not necessarily lead to early menopause in all carriers. In contrast, compound heterozygous or homozygous variants that were experimentally confirmed or predicted to result in substantial impairment of protein function were more likely to cause POI or early amenorrhea.

Table 2.

GDF9 homozygous variants based on previous studies and the current study

References	Variant	Protein change	Phenotype	Notes
Present study	c.275T $>$ C	p.(Leu92Pro)	Secondary amenorrhea	Normal secondary sexual characteristics; small ovaries
Almatrafi et al ., 2024 (10)	c.1121C $>$ T	p.(Pro374Leu)	Secondary amenorrhea	Normal secondary sexual characteristics; menopause at 25 yr
Jordan et al. , 2024 (8)	c.1294del	p.(Ile432LeufsTer15)	Primary-secondary amenorrhea	Normal secondary sexual characteristics
Verma et al. , 2021 (1)	c.604C $>$ T	p.(Gln202*)	Primary amenorrhea	Normal secondary sexual characteristics
Yang et al. , 2019 (11)	c.1157G $>$ T	p.(Cys386Phe)	Primary amenorrhea	Normal puberty; small ovaries
França et al. , 2018 (9)	c.783delC	p.(Ser262Hisfs*2)	Primary amenorrhea	Breast development: tanner II; pubic hair: tanner III

4. Conclusion

In conclusion, we report a homozygous variant in GDF9 co-segregating with POI in affected sisters, providing support for a Mendelian gene-disease association. This finding also expands the known mutational spectrum of GDF9 and underscores its importance for genetic counseling and molecular diagnosis in families affected by POI.

Author Contributions

B. Haj Mohammad Hassani: Conceptualization, bioinformatics analysis, literature review, and article writing. N. Ghasemi: Literature review and article writing. K. Malekzadeh: Data collection, supervision, editing, and review.

Acknowledgments

The researchers sincerely appreciate both the cases and the health service staff of Shahid Mohammadi hospital, Bandar Abbas, Iran who contributed to this research. The authors declare that no funds, grants, or other support were received during the preparation of this manuscript. The authors of this article did not use artificial intelligence (AI) in any way.

Conflict of Interest

The authors declare that there is no conflict of interest.