Identification of novel deep-intronic SERPINC1 variants by multiple genomic and transcriptomic approaches in a large cohort of families with hereditary antithrombin deficiency
Antithrombin deficiency is a rare hereditary predisposition to venous thromboembolism caused by variants in the SERPINC1 gene. Even in families exhibiting a clear autosomal dominant inheritance pattern, in up to 20% of cases no gene defects can be identified. Because distinct SERPINC1 variants are associated with different subtypes of antithrombin deficiency and corresponding risks of venous thromboembolism, a molecular diagnosis is important for counselling. We aimed to uncover novel SERPINC1 variants by using whole genome sequencing to cover the full SERPINC1 region.
In this single-centre cohort study, we included antithrombin deficient families in which no SERPINC1 variant had been previously identified. For each family, we included at least one deficient and one non-deficient member. We performed whole genome sequencing and used in-silicio tools to predict effects of novel variants. Variants were classified according to SERPINC1-specific ACMG criteria (v1.1.0). For variants predicted to lead to splicing abnormalities, we isolated RNA from peripheral blood of patients and performed reverse transcriptase PCR and Sanger sequencing. In addition, we introduced the variant in the genome of HepG2 hepatocytes using CRISPR-Cas9 to assess its effect on splicing.
We included 42 individuals from ten families. In four families comprising 15 individuals, we identified two perfectly cosegregating deep-intronic variants: (NM_000488.4) c.42-209T>G and c.1153+429C>T. Pedigree analysis revealed a common ancestor. With long-read sequencing we confirmed that both variants reside on the same allele. The c.42-209T>G variant is absent in the GnomAD (v4.1.0) and is located in an alternative exon upstream of the 5’UTR of a low expressed alternative transcript. In-silicio prediction tools suggested the introduction of a novel acceptor splice site at c.42-209, resulting in the inclusion of a pseudo-exon and a premature stopcodon due to a frameshift. We confirmed this aberrant splicing prediction by RT-PCR and Sanger sequencing of the resulting PCR products from CRISPR-Cas9 edited HepG2 cells. The variant leads to a reduction of SERPINC1 transcript levels. The c.1153+429C>T variant is prevalent in the GnomAD and is not predicted to have a pathogenic effect, leading to a likely benign classification. For the other six families, further analyses are ongoing.
We report the first pathogenic deep-intronic SERPINC1 variant causing antithrombin deficiency through aberrant splicing and resulting lower SERPINC1 transcript levels. This finding improves our understanding of the genetic aetiology and clinical practice for antithrombin deficiency.