Feasibility of characterizing von Willebrand disease using high-throughput affinity-based proteomics

Abstract

Abstract Background: Von Willebrand disease (VWD) is the most common inherited bleeding disorder, caused by quantitative or qualitative defects in von Willebrand factor (VWF), a key protein in hemostasis. Accurate diagnosis and subtype classification remain challenging due to clinical and laboratory heterogeneity. High-throughput affinity-based proteomic technologies, such as the Olink proximity extension assay (PEA), enable simultaneous quantification of thousands of proteins in plasma samples and may offer novel insights into VWD biology. However, the feasibility of applying Olink proteomics to characterize VWD and differentiate among subtypes has not been previously evaluated. We aimed to assess whether high-throughput proteomics data can reflect VWD-related phenotypic variation. Methods: Plasma samples from 316 well-characterized congenital VWD patients from the French reference center for VWD representing different subtypes, along with 43 healthy controls, were profiled using the Olink Explore HT platform, which quantifies over 5400 circulating proteins, including VWF. Laboratory measurements were performed centrally for VWD patients and controls. The following parameters were collected: Factor VIII activity using one-stage assay (FVIII:C), VWF antigen (VWF:Ag), VWF activity (VWF:Act), VWF collagen-binding assay (VWF:CB), VWF propeptide (VWFpp) and VWF multimers. The correlation between VWF levels measured by Olink and conventional VWF assays was assessed using Spearman's rank correlation. A linear model adjusted for age was used to assess differences in VWF levels between VWD subtypes and healthy controls. All statistical analyses were performed using R version 4.4.1 in RStudio. Results: The mean [IQR] age was 38.2 [20.9-52.1] years old and 22.8 [20.0-23.3] years old respectively for VWD patients and healthy controls. The sex ratio was 184/132 for VWD patients and 21/22 for healthy controls. The majority of VWD patients had type 2 VWD (n=213, 67.4%), followed by type 1 (n=74, 23.4%) and type 3 (n=29, 9.2%). Olink-derived VWF levels showed a strong correlation with conventional VWF:Ag measurements (Spearman's ρ = 0.74, p < 2.2 × 10⁻¹⁶). In the age-adjusted linear model, VWF levels were significantly lower in all VWD subtypes compared to healthy individuals. The largest reduction was observed in type 3 VWD (β = –6.02, p < 2 × 10⁻¹⁶), followed by type 2M (β = –2.23, p = 2.2 × 10⁻¹²), type 1 (β = –2.17, p = 5.6 × 10⁻¹²), type 2N (β = –0.88, p = 0.025), and type 2A (β = –0.73, p = 0.016). Conclusions: This study demonstrates the feasibility of using Olink PEA proteomics to characterize VWD from plasma samples. VWF levels derived from Olink data strongly correlate with standard clinical assays and reflect subtype-specific differences compared to healthy controls, even after adjusting for age. These findings support the utility of high-throughput proteomic approaches in VWD research, offering potential for novel biomarker discovery and improved disease stratification.