Coagulation Factor XII (F12) Haploinsufficiency Is Protective Against Venous Thromboembolism

Background:

Inhibition of factor XII (FXII, gene: F12) represents a potentially transformative approach to antithrombotic therapy, as severe congenital deficiency in humans is not associated with bleeding while deletion or inhibition of FXII in preclinical models consistently protects against thrombosis. However, the impact of FXII deficiency on venous thromboembolism (VTE) risk in humans is controversial. Prior studies suffered from significant limitations such as small sample size, ascertainment bias, and heterogenous case/control groups. An accurate, population-level assessment of the clinical impacts associated with FXII deficiency could significantly impact ongoing efforts to develop novel anti-FXII therapies. Here, we combine analysis of large-scale genomic datasets with in vitro and in vivo translational studies to probe the role of FXII in VTE.

Methods:

We employed data from the UK Biobank (UKB, N=415,921) and NIH All of Us program (AoU, N=191,686), which contain paired genetic and clinical data. We identified participants with rare (MAF ≤ 1%) loss-of-function (LOF) variants in F12 (frameshift, essential splice site, and nonsense mutations). We used Firth’s penalized likelihood Cox regression modeling to evaluate the association between VTE and loss-of-function variants in F12 through single-gene burden analysis. Results from UKB and AoU were meta-analyzed using a random-effects model. Circulating FXII levels were measured in a subset of UKB participants using the Olink® Explore 3072 plasma proteomics panel (N=44,864). We directly measured plasma FXII levels on samples from mice and the Mass General Brigham (MGB) Biobank by ELISA. Kaolin-initiated thrombin generation assays were used to evaluate FXII-reconstituted plasma samples. The femoral vein electrolytic injury model was used to compare thrombus formation in wild-type (N=10), F12 heterozygous (N=9), and F12 homozygous knockout (N=8) mice.

Results:

Across the UKB and AoU datasets, 1920 participants had qualifying variants in F12 (99.9% heterozygous). Variant carrier status was significantly associated with protection against VTE (HR=0.65, 95% CI: 0.49-0.86, P=0.002). Modeling replicated known risk factors for VTE such as age, BMI, and blood group, thereby confirming that clinical annotation of the datasets was accurate. A leave-one-variant-out analysis showed that no single variant was responsible for the protective signal, and sensitivity analyses incorporating a polygenic risk score for VTE did not change the effect-size estimate for F12 LOF, indicating that F12 likely functions independently of known common genetic risk factors for VTE. F12 variant carriers had significantly lower circulating FXII levels (mean: 46.1%, P <0.0001) compared to non-carriers (mean 100% L-NPX, P=8.8 x 10^-8). These findings were corroborated by ELISA performed on plasma samples from 29 F12 variant carriers and matched 29 controls (12.68 mcg/ml in carriers vs. 25.17 mcg/ml in controls, P=4.8 x 10^-6). Thrombin generation assays performed on reconstituted FXII-deficient plasma showed significantly reduced peak thrombin generation at 50% FXII plasma levels (P=0.0002). Using a mouse femoral vein electrolytic injury model, we observed significant reductions in both platelet accumulation (P=0.032) and fibrin formation (P=0.025) in F12 heterozygous knockout mice compared to wild-type.

Conclusions:

In this multi-modal study, we show that inherited FXII deficiency is protective against VTE. Our data are consistent with the growing recognition that haploinsufficient states in coagulation can have biological effects without requiring total loss of function. Taken together, these findings suggest that targeting FXII is likely to be a safe and effective antithrombotic strategy.