Proteomics-based clustering for higher risk of venous thromboembolism in patients with nephrotic syndrome.

BACKGROUND

Nephrotic syndrome (NS) carries a high and heterogeneous risk of venous thromboembolism (VTE). Hypercoagulability has been considered as key factors, but in-depth molecular mechanisms remain unrevealed.

OBJECTIVES

We aimed to identify patients with NS who are under higher risk for VTE and explore key pathways and molecules using proteomic profiling.

METHODS

Serum proteomic data from 49 patients with NS were analyzed via unsupervised consensus clustering (top 1000 variable proteins, K = 2) and gene set variation analysis. Differentially expressed proteins were identified, and a protein-protein interaction network was constructed to pinpoint hub proteins. The levels of hub proteins were compared between the clusters, and their correlation with clinical biomarkers of renal function were analyzed. The cumulative incidence of VTE over 2 years was the primary end point. Trajectory analysis was performed to reveal the dynamic changes of renal function of each cluster.

RESULTS

Unsupervised clustering stratified 2 clusters. Cluster 2 had numerically higher prevalence of both baseline and cumulative VTE than in cluster 1 (43% vs 11%; P = 0.13 for baseline VTE; and 58% vs 22%; P = .074 for cumulative VTE) and had enriched activity in extracellular matrix organization and cell-cell interaction pathways. Differentially expressed protein analysis found 145 upregulated proteins in cluster 2, and protein-protein interaction network analysis identified 9 candidate hub proteins (CTNNB1, VCAN, POSTN, SDC1, MMP2, SERPINC1, APOA1, PCSK9, and LIPC), all significantly elevated in cluster 2, and MMP2 and POSTN were associated with markers of tubular injury.

CONCLUSION

Serum proteomics analysis identified a group of patients with NS who were potentially at higher risk of developing VTE. Extracellular matrix and endothelial glycocalyx damage might play an important role in VTE development among patients with NS.