ESPE Abstracts

Background: Sequencing of Primary Ovarian Insufficiency (POI) cohorts have identified variants in >100 “POI genes” in up to 50% of women. Establishing pathogenicity of these variants is challenging. Whether early-onset POI (EO-POI; adolescent-onset) has a unique genetic profile remains unknown.

Methods: We performed exome sequencing (Nonacus) in an EO-POI cohort (sporadic and familial) and unaffected family members. Filtering (QCI Qiagen) retained variants which were 1) rare/novel (minor allele frequency <0.01%), 2) predicted pathogenic in silico, 3) enriched in the cohort compared to gnomAD v4.0 controls; and 4) pathogenic/likely pathogenic (ACMG). Variants were further filtered using a tiered, evidence-based approach, considering three categories of variants: 1) variants in genes included on the Genomics England Primary Ovarian Insufficiency PanelApp; 2) variants in POI PanelApp genes following unexpected inheritance patterns or variants in other genes previously associated with POI; and 3) homozygous variants in novel candidate POI genes.

Results: A total of 149 women with EO-POI were recruited (31 familial POI from 17 kindreds (8 consanguineous); 118 sporadic POI). Of these, 81.2% (n = 121) had primary amenorrhea. Of 17 kindred with familial POI, 13 (76.5%) had a Category 1, 2, or 3 variant identified (n = 6 homozygous Category 1 variants in STAG3, MCM9, PSCM3IP, POLR2C, YTHDC2, and ZSWIM7 and n = 5 Category 2 variants in NLRP11 and PRKD1 (heterozygous), PLEC, IGSF10, and KMT2A (homozygous), PDE3A, POLR2H, MSH6, and CLPP (polygenic)). A total of 65.3% (n = 77) women with sporadic POI had a variant identified (20.3% (n = 24) Category 1 and 42.4% (n = 50) Category 2). Within the entire cohort, heterozygous, homozygous, and polygenic variants in genes previously associated with POI were found in 30.9%, 9.4%, and 21.8% respectively. Most heterozygous variants were present in gnomAD controls, albeit enriched in our cohort. The cohort was then screened for Category 3 variants, revealing 10 novel POI candidate genes in 8 women. These included genes with animal gonadal insufficiency models (PCIF1, DND1, MEF2A, TGFBR1), DNA repair genes (XRCC1, MMS22L, RXFP3), mitochondrial genes (PPAN, CLUH, COQ10B, MRPS14), and others (C4orf33, ARRB1).

Discussion: Using a tiered, evidence-based approach, we characterise genetic variants in POI-associated genes in, to our knowledge, the largest EO-POI cohort described. We suggest POI is a complex and often multigenic disorder and highlight the difficulty of establishing the pathogenicity of single heterozygote variants in this condition. Compared to previous POI cohorts, a higher proportion of young women with EO-POI have autosomal recessively inherited, monogenic POI. Lastly, we identify convincing novel candidate genes for POI.