Genotype Analysis of Adolescents with Heavy Menstrual Bleeding and Low Von Willebrand Activity - Report of a Multi-Center Study

Introduction: Low von Willebrand factor (VWF) activity is prevalent in adolescents with heavy menstrual bleeding (HMB). There is a need to better genetically characterize these patients and improve our understanding of the pathophysiology of their bleeding risk.

Methods: One of the main objectives of this multi-center, single arm, observational cohort study was to genotype adolescent females with HMB and low VWF (≥ 30 and ≤ 50 IU/dL) by means of whole exome sequencing (WES), to identify variants throughout the exome that may modulate risk for bleeding. Post-menarchal females < 21 years with HMB (defined as PBAC score >100) and low VWF were eligible for the study. All patients were enrolled by participating centers where blood samples were collected. Exome data for 86 cases and 900 unrelated pediatric (<21 years) controls taken from a genetic study on Chiari 1 Malformation (CM1) were aligned and sorted, and variants called using the Sentieon software package. Annotation including allele frequencies, function, amino acid change and Clinvar rating among others was obtained using ANNOVAR. Variants were retained if there was a genotype call rate of >0.8, GQ>20, AB between 0.3-0.7 and min DP>8. Case/control analyses included common and rare SNP associations, gene burden analysis of both rare nonsynonymous variants and ClinVar 'pathogenic' variants, and gene-set burden analyses. Variants were considered rare if they had a minor allele frequency of <1% in the Genome Aggregation Database (gnomAD). Count differences between cases/controls were determined by Fisher's Exact test.

Results: Of the 113 subjects enrolled, 86 had sufficient blood samples collected for WES. The median age was 16.2 years (range: 11.5-19.6). 36% of cases showed variants in VWF vs. 26% of controls (p=0.14). After multiple test correction, WES revealed a significant common variant association in FERMT2 with an LD block with a frequency of 24% in cases and 6% in controls (p=7.5x10^-7). Rare variant analysis showed a significant association with an intronic variant in ABCA13 (p=1.6x10^-7). Among the gene burden analysis using rare nonsynonymous variants, 2 genes of interest passed multiple test correction: IL12B and DTNBP1. When using ClinVar 'pathogenic' variants as the input for the gene burden analysis, 4 genes of interest passed multiple test correction: HBM, MYLK, RUNX1 and CD36. Gene-set burden analysis revealed 5 significant pathways of interest, including platelet degranulation, platelet alpha granule lumen and erythrocyte differentiation. We then focused a subset of known risk genes and compared the number of missense, nonsense and ClinVar 'pathogenic' variants between cases and controls. GP6, and MTHFR had significantly more missense variants in controls vs. cases (p=0.003 and 0.01, respectively), and F13B approached significance, with more missense variants in controls than cases (p=0.07).

Conclusion: We found VWF variants in 36% of subjects, in accordance with previous reports. We found novel SNP associations with variants in FERMT2 and ABCA13.FERMT2 encodes for the integrin, Kindlin 2, which has been shown to be critical for supporting vascular integrity. Several other relevant genes passed multiple test correction in gene burden analyses, including genes known to cause Hermansky-Pudlak syndrome (DTNBP1), familial platelet disorder (RUNX1), platelet glycoprotein IV deficiency (CD36) and aortic aneurysm (MYLK). This association with platelet disorders was strengthened by our gene-set burden results, which implicate several platelet-related pathways. These data suggest that while a subset of HMB patients may have their bleeding explained by variants in VWF, there is a role for other hemostasis, platelet biology and vascular integrity related gene variants which can contribute to the variation in bleeding severity in adolescents with low-VWF related HMB.

Study supported by an investigator-initiated research grant from Shire US Inc., now part of Takeda