Genetics of venous thromboembolism revised

In this issue of Blood, Lindström and colleagues¹  report genomic and transcriptomic association of 16 novel susceptibility loci for venous thromboembolism (VTE). Moreover, Mendelian randomization causally linked blood traits to thrombosis.

Familial aggregation of VTE was recognized in 1905 by Briggs,²  but it was not until 1965 that Egeberg identified the first genetic risk factor for VTE (ie, inherited deficiency of antithrombin [SERPINC1]) (see figure).³  In 1981 and 1984, inherited deficiencies of protein C (PROC) and protein S (PROS1) were recognized as risk factors for VTE. However, it was not until the discovery of resistance to activated protein C (APC resistance) by Dahlbäck et al in 1993 that it became evident that genetic factors are common risk factors of VTE.⁴  APC resistance was linked to a mutation in the factor V gene (F5), factor V Leiden (rs6025).⁵  In 1996, Poort et al reported a common genetic variation (rs1799963) in the 3′-untranslated region of the prothrombin gene (F2) is associated with elevated plasma prothrombin levels and VTE risk.⁶  These 5 inherited defects are called major thrombophilias, and all involve coagulation or anticoagulant genes. No new major thrombophilia has been discovered since 1996. Instead, genome-wide association studies (GWASs) have discovered a number of VTE-associated loci, reviewed by Trégouët and Morange.⁷  The GWAS-discovered risk variants are weak risk factors for VTE but are prevalent in the population. Before the present study, all genes linked to VTE were directly or indirectly linked to the coagulation system (FGG, F2, F5, F8, F11, KNG1, VWF, ABO), anticoagulation pathways (SERPINC1, PROC, PROS1, PROCR, THBD), or platelets (VWF, GP6, ABO, STXBP5, ZFPM2), although some loci (SLC44A2, TSPAN15, LRP4) had no known biological link to VTE.

The present study by Lindström et al is the largest meta-analysis of GWAS data for VTE, including 18 studies with 30 234 VTE cases and 172 122 controls. The study is the first trans-ancestry GWAS of VTE. It is also the first transcriptome-wide association study (TWAS). The GWAS identified 11 newly associated genetic loci (C1orf198, PLEK, OSMR-AS1, NUGGC/SCARA5, GRK5, MPHOSPH9, ARID4A, PLCG2, SMG6, EIF5A, and STX10). The TWAS identified 5 additional genetic loci using imputed gene expression in whole blood (SH2B3, SPSB1, RP11-747H7.3, RP4-737E23.2) and in liver (ERAP1). Some previous associations could not be confirmed in the present GWAS, such as the THBD loci.⁷  The present study gives important contributions to the functions of genes involved in the pathogenesis of VTE (see figure). In the figure, the genes are grouped after potential links to VTE. The strongest genetic risk factors for VTE (ie, classical thrombophilia) are all related to coagulation (F2 and F5) or the anticoagulation system (SERPINC1, PROC, PROS1). However, the present study adds several loci all involved directly or indirectly in platelet biology but also loci involved in erythrocyte biology or inflammation. The present study also used Mendelian randomization and showed that blood traits, especially related to platelets, are causally associated with VTE risk. Thus, platelets that before have been implicated mostly in arterial thrombosis are also of importance for VTE. Thus, the present study suggests that the formed elements in the blood need more attention from researchers in the field of VTE.

Inflammatory conditions, such as infection and immune-mediated diseases, have been linked to VTE,⁸  but the present study is the first to link several genes related to inflammation to VTE underlining the importance of inflammation in VTE pathobiology. Whether inflammation per se is a driver of thrombosis or whether inflammation is due to its effects on coagulation, anticoagulation, and platelets remains to be determined.

The present study also tested a new genetic risk score (GRS) based on 37 loci. The GRS constructed from these variants will become a useful research tool due to its strong impact on VTE risk, although its clinical usefulness remains to be determined. For instance, a recent study has shown that family history of VTE affects the risk of both VTE and major bleeding following hip and knee replacement surgical procedures.⁹  It could therefore be worthwhile to determine the value of the new GRS for prediction of both VTE and bleeding.

The heritability on the observed scale due to genotyped variants was 23%. Transforming this to the liability scale assuming a disease prevalence of 0.5% resulted in a heritability of 15%. This is much lower than the estimated heritability of 40% to 60% from studies of families, twins, siblings, and half-siblings.²  The cause of the remaining heritability is unclear and might be due to rare variants, although a recent exome study by Lindström et al of 8332 cases and 16 087 controls of European ancestry and 382 cases and 1476 controls of African American ancestry found no significant novel rare variants.¹⁰  However, the design of exome studies needs to be developed because burden analysis could not confirm that the 3 major thrombophilia genes (SERPINC1, PROS1, and PROC) are associated with VTE.¹⁰  Only the Ser501Pro variant of PROS1 (PS Herleen mutation; rs121918472) with a minor allele frequency of 0.005 could be confirmed. Family studies using new strategies might be an option. In the present study, variants in novel loci were all intronic or intergenic. Maybe-whole genome sequencing will be an important way for mapping the missing heritability of VTE. As in the present study, using expression data from Genotype-Tissue Expression might be another option for the detection of functional variants.

With the present milestone study by Lindström et al, the list of VTE-associated loci has grown considerably. Genes have been identified that are linked not only to coagulation and anticoagulation but also to platelets, erythrocytes, inflammation, or genes without any so far known biological link with thrombosis (see figure). The results will be important not only for future genetic and clinical studies but also for directing future function-structure studies of the molecular cause of VTE.