Characterization of CNVs in a Large Cohort of VWD Patients Reveals Relationship between Disrupted Regions, Clinical Variables and Disease Type

BACKGROUND: Copy-number variants (CNVs) have been associated to all VWD types. Although many of the underlying pathological mechanisms have not been investigated, in-frame deletions have been shown to cause defects in VWF biosynthesis and secretion. While there are previous studies describing CNVs in VWD patients, these only discuss a small number of CNVs each, and are in varying populations, making it difficult to compare between populations as the differences could be population specific.

METHODS: A VWF high-resolution aCGH array was custom designed using long oligonucleotide probes ~60 bp in length to avoid multiple sequence variations, repeated sequences and the pseudogene. Data were manually inspected and analyzed with CytoSure Interpret Software. A multiplex PCR approach was used to map deletion breakpoints and to confirm the presence and zygosity of the deletions in the index case as well as family members. CNVs were mapped and t-tests performed on clinical laboratory values within each VWD type for those with and without a CNV. aCGH was performed in two datasets on low-VWF and VWD index cases with no known pathogenic VWF sequence variant for a total of 205 patients.

RESULTS: Of the 205 patients, CNVs were found in 25 (12%). No CNVs were found in low-VWF patients and the only duplication occurred in a Type 2 (n=1) patient. The remaining CNVs were deletions in VWD Type 1+1S (n=12), 1C (n=5), and 3 (n=7) patients. All Type 1C patients (n=5) had the same exon 33-34 deletion. All but 1 of the Type 1+1S patients (n=11) had the same exon 4-5 deletion, while the Type 3 patients (n=7) had deletions spread across the gene. Further, within the Type 1+1S patients, VWFpp/VWF:Ag ratio was significantly lower in those with a CNV than those without (two-tailed t-test, p=0.02). Lastly, ristocetin cofactor was significantly higher in the Type 1C patients with a CNV versus those without a CNV (two-tailed t-test, p=8.2x10^-5).

DISCUSSION: Here we have analyzed a large dataset of CNVs defined by aCGH in VWD patients with no known pathogenic VWF sequence variant. The location of the deletion in all Type 1C patients exactly matches the D4N module of the D4 domain of the mature VWF protein, a region where mutations and deletions are known to affect clearance. The D4 domain is the only D domain to also include an N module. Although previous work has shown that deletions of exons 33-34 potentially cause a mild reduction in VWF:Ag levels and only a moderate impact on secretion, we observed significantly higher VWF:RCo levels in those Type 1C patients with a CNV when compared to those without. As expected, Type 1C patients had a significantly higher VWFpp/VWF:Ag ratio than Type 1+1S patients overall (p=2.11x10^-15). However, the Type 1+1S patient with a CNV had significantly lower VWFpp/VWF:Ag ratio than those without a CNV. This work elucidates some of the underlying genetic mechanisms of CNVs in these patients.