Causal and Candidate Gene Variants in a Large Cohort of Women With Primary Ovarian Insufficiency

Abstract

Context: A genetic etiology likely accounts for the majority of unexplained primary ovarian insufficiency (POI).

Objective: We hypothesized that heterozygous rare variants and variants in enhanced categories are associated with POI.

Design: The study was an observational study.

Setting: Subjects were recruited at academic institutions.

Patients: Subjects from Boston (n = 98), the National Institutes of Health and Washington University (n = 98), Pittsburgh (n = 20), Italy (n = 43), and France (n = 32) were diagnosed with POI (amenorrhea with an elevated follicle-stimulating hormone level). Controls were recruited for health in old age or were from the 1000 Genomes Project (total n = 233).

Intervention: We performed whole exome sequencing (WES), and data were analyzed using a rare variant scoring method and a Bayes factor-based framework for identifying genes harboring pathogenic variants. We performed functional studies on identified genes that were not previously implicated in POI in a D. melanogaster model.

Main outcome: Genes with rare pathogenic variants and gene sets with increased burden of deleterious variants were identified.

Results: Candidate heterozygous variants were identified in known genes and genes with functional evidence. Gene sets with increased burden of deleterious alleles included the categories transcription and translation, DNA damage and repair, meiosis and cell division. Variants were found in novel genes from the enhanced categories. Functional evidence supported 7 new risk genes for POI (USP36, VCP, WDR33, PIWIL3, NPM2, LLGL1, and BOD1L1).

Conclusions: Candidate causative variants were identified through WES in women with POI. Aggregating clinical data and genetic risk with a categorical approach may expand the genetic architecture of heterozygous rare gene variants causing risk for POI.

Keywords: genetics; menopause; sporadic; whole exome sequencing.

Figure 1.

Three examples of enriched pathways in the POI data set as determined by the permutation tests in cases (upper panels) and controls (lower panels). Enriched pathways that encompassed novel POI genes (Table 5) included (A) transcription/translation/DNA binding, (B) meiosis/DNA repair/homologous recombination, and (C) cell division/meiosis compared to (D) housekeeping genes. The number of damaged genes from the target gene list in the pathways of interest (red arrow) is compared to the distribution of damaged genes in random gene lists of equal number to the lists of interest (gray bars), burden-matched control genes (pink arrows), and housekeeping genes (green arrows). The burden-matched genes and housekeeping genes are not significantly enriched for any gene set. P-values are controlled for the false discovery rate.

Figure 2.

Drosophila melanogaster ovarian phenotype. Representative images of ovaries from RNA interference (RNAi) knockdowns that produced atrophic ovaries and a control. The affected gene is indicated. All other RNAi knockdowns that produced normal ovaries appear identical to the control and are not shown.

Figure 3.

Candidate genes in women with primary ovarian insufficiency. Variants in a number of genes involved in chromosome pairing and DNA damage and repair are involved in meiosis. The figure depicts candidate genes that are involved in chromosome movement, double-strand breaks, end resection, double-strand break repair, crossovers and dissociation, and resolution of Holliday junctions. Members of the nuclear pore complex (NUP43) play a role in chromosome movement and organization. After DNA replication (ORC6), the synaptonemal complex pairs homologous chromosomes (PSMC3IP) loaded with condensin and cohesion complex proteins (STAG3, REC8, NIPBL) and connects the synaptonemal complex to DNA repair proteins (SYCE1). During recombination, double-strand breaks form (ATM, ANKRD3, PIF1), ends are resected (BRCA1, SAMHD1, BOD1L1), and crossovers occur (HFM1) through strand invasion (PSMC3IP, MND1, RAD51). Subsequently, DNA double-strand break repair (CHD1L, POLG, POLK, MSH6, PCNA, NUPR1, APLF, NBN, RAD50, RUVBL2, MRE11), DNA repair (CDK7, MLH3, PRMT6, HELQ, TONSL), strand annealing (RECQL4), and repair via homologous recombination (BRCA2, BRIP1, FANCD2, HELQ, FANCM, FANCF, BLM, MCM9, USP36) take place. Kinetochore/chromosome assembly, orientation, and segregation (HAUS6, CENPF, NUP43, NCAPG2, LLGL1, NINL, ATRX) follow recombination.