Editorial 

March 3, 2025

Protein S Genomics and Proteomics Refine Thrombosis Risk

Margaret V. Ragni

JAMA. 2025;333(16):1401-1402. doi:10.1001/jama.2025.1883

Protein S is a vitamin K–dependent protein named after Seattle, where it was isolated and characterized. Subsequently recognized as a multifunctional enzyme that plays a critical role in the regulation of coagulation, protein S limits excessive clot formation through its action on clot initiation and clot propagation.^1,2 As a cofactor for tissue factor pathway inhibitor, protein S limits clot initiation by inactivating extrinsic factor Xa,³ and as a cofactor for activated protein C, protein S limits clot propagation by inactivating intrinsic factors Va and VIIIa.⁴ In individuals with protein S deficiency, regulation of coagulation is impaired and the risk of thrombosis is increased. A clinically important manifestation of protein S deficiency is neonatal purpura fulminans, a rare and potentially fatal disorder characterized by microvascular thrombosis and skin necrosis. The relevance of protein S regulation of coagulation is further underscored by the recent development of a silencing RNA that inhibits protein S as a novel approach to restore hemostasis and reduce bleeding in hemophilia.⁵

Protein S deficiency may be inherited or acquired. Hereditary protein S deficiency is a rare autosomal disorder caused by variants in the PROS1 gene, located on chromosome 3. Among those with hereditary protein S deficiency, most are heterozygous for a PROS1 variant; rare individuals with homozygous or compound heterozygous variants have more severe disease. In its mature form, protein S exists free (40%) or bound (60%) to complement C4b–binding protein, but it is free protein S that interacts with activated protein C and tissue factor pathway inhibitor and regulates coagulation and, when deficient, impairs coagulation regulation, thereby increasing the risk of thrombosis.⁶ Protein S deficiency is defined as plasma protein S antigen of less than 60% and is associated with a 2- to 5-fold increased relative risk of venous thromboembolism (VTE), which varies with the population studied.⁷

The prevalence of protein S deficiency and the risk of thrombosis remain challenging to define, as studies are limited by small kindreds, assay variability, and lack of defined diagnostic thresholds or genetic confirmation. In kindreds with protein S deficiency, thrombosis typically occurs at a young age and is increased by acquired factors such as hormone therapy, pregnancy, surgery, trauma, and immobilization. Although less common, thrombosis in protein S deficiency may occur in unusual sites, eg, axillary, mesenteric, or cerebral veins. Arterial thrombosis, including acute ischemic stroke and arterial occlusion, has been reported in some families with protein S deficiency, although controversy remains whether protein S deficiency is a risk for arterial thrombosis.⁸

In this issue of JAMA, Chaudhry and colleagues⁹ improve understanding of protein S deficiency by genomics and proteomics profiling of a large population-based study of more than 600 000 individuals. They performed cross-sectional analyses of multiple “omics” datasets, specifically genomic and proteomic data, from the UK Biobank (n = 426 436) and US National Institutes of Health (NIH) All of Us biorepository (n = 204 006), with individual-level demographic, laboratory, and clinical data available from both cohorts. A functional impact score ranging from 0 to 1, based on in silico prediction modeling, was calculated to determine the likelihood that an identified genetic variant would disrupt protein S activity in vivo. A functional impact score of 1 indicated variants most likely and a functional impact score of 0 those least likely to disrupt protein activity.¹⁰ Logistic regression and linear regression modeling adjusting for age, sex, and ancestry^10,11 were used to determine thrombosis risk across protein S levels and PROS1variants in these databases. Kaplan-Meier analysis of incident VTE events was used to evaluate ongoing risk among those with PROS1 variants.

In their analyses, the authors established that the highest thrombosis risk was found in individuals with inherited variants that disrupted protein S (PROS1), specifically caused by nonsense, frameshift, and splice site variants; these variants are rare (0.00091%) but associated with a strong risk for VTE (odds ratio, 14.01) that persists throughout life. Thrombosis risk in these individuals could be predicted by plasma proteomics data indicating the highest functional impact score of 1.0 and associated with a 50% reduction in free protein S levels. In contrast, individuals with protein S deficiency with less damaging missense variants are more common (0.22%) and have lower VTE risk (odds ratio, 1.98), more moderate reduction in plasma protein S antigen levels (48% of normal), and a lower functional impact score of 0.7 or greater.

Chaudry and colleagues also found that individuals with PROS1 variants, whether disruptive or not, have lower circulating protein S levels than those without PROS1 variants. Among those without PROS1 variants, individuals who have low protein S levels, whether acquired, environmental, or trans-acting genetic, are more likely than those with PROS1 variants to develop VTE and, in contrast to those with PROS1 variants, may develop arterial thrombosis, primarily peripheral arterial disease.

So, what is the clinical impact of the findings of this study on the diagnosis and management of individuals with protein S deficiency? What past beliefs can be safely discarded regarding thrombosis risk in protein S deficiency, what, if any, changes should be considered in our approach to clinical screening for and management of protein S deficiency, and what future studies should be pursued?

For patients with protein S deficiency who require lifelong anticoagulation, or for those with unprovoked VTE or recurrent or life-threatening VTE, the current screening guidance is to avoid protein S testing as testing will not change treatment¹²; this guidance is unlikely to change based on the current findings. Similarly, these findings are unlikely to change the current strategy of selective testing for protein S deficiency in a member of an affected kindred,¹² as counseling is essential to reduce VTE risk and to provide guidance on whether and how long to continue anticoagulation or thromboprophylaxis.

However, these findings may change the current management strategy¹² for those with protein S deficiency. Specifically, in those for whom protein S testing is indicated and in whom a low free and/or total protein S antigen is found, the PROS1 gene should be sequenced. This management change is important because detection of the disruptive variants associated with a high risk of thrombosis may prompt consideration of longer-term anticoagulation for acute thrombosis and longer-term thromboprophylaxis in at-risk settings,¹² eg, with transient risk factor for VTE, with nonsurgical major transient or hormonal risk factor, in the antepartum and postpartum period, or for ambulatory cancer management. These potential revisions to the management of protein S deficiency represent critical issues for review and debate by the American Society of Hematology’s VTE guidelines experts.¹²

The study findings should dispel the belief that protein S deficiency is associated only with mild thrombosis risk. Individuals with protein S deficiency who have inherited the rare but highly disruptive PROS1 variants (including nonsense, frameshift, and splice site variants) have a significantly increased risk of VTE (odds ratio, 14.01). Moreover, these individuals were also assigned the study’s highest clinical impact score of 1.0. As these functional impact scores were derived from clinical thrombosis risk factors analyzed by probability modeling and regression analyses, it would seem possible that a thrombosis prediction score, much like the 4Ts score for heparin-induced thrombocytopenia,¹³ could be developed, validated, and potentially implemented in clinical practice to facilitate and simplify thrombosis management of protein S deficiency. In the future, as additional PROS1 variants are identified, such a prediction score could optimally provide a thrombosis risk profile and indication for and duration of anticoagulation. Longitudinal studies in individuals with protein S deficiency will also be important to clarify clinical outcomes and, importantly, to incorporate genomics data to guide clinical care and optimize management.