Factor XI Deficiency
Factor XI (FXI) deficiency results in a mild bleeding disorder that was first recognized in 19531 when an American physician reported a Jewish family with abnormal bleeding after tonsillectomy and dental extractions. As the abnormal bleeding was manifested in both sexes (two sisters and their maternal uncle), and the clinical features were not consistent with hemophilia A or B, it was called hemophilia C.2 While spontaneous bleeding is very rarely reported in individuals with FXI deficiency (even for individuals with no measurable FXI levels), bleeding may occur after accidents or surgery, particularly in areas of high fibrinolytic activity (e.g. the oropharynx or the genitourinary tract).
FXI is a serine protease that is produced in the liver, and is unique among other coagulation factors as this enzyme circulates in the bloodstream as a homodimer complexed with high molecular weight kininogen (HK). FXI contains four homologous tandem repeats (apple domains) and a catalytic serine protease domain. The crystal structure of FXI shows a circle of apple domains arranged as a disc at the base of the catalytic domain.3 Activation of FXI is associated with a structural shift that exposes the FIX binding site. Activation of FIX reinforces the intrinsic pathway of coagulation.
The role of FXI in the coagulation pathway has been clarified in recent years. Although the traditional view was that FXI was activated by the ‘contact pathway’ by activated factor XII (FXII), the relationship of such FXI activation to physiologic hemostatic mechanisms was unclear. The discovery that FXI is activated by thrombin led to a revised view of the coagulation system in which FXI is not critical in initiation of coagulation, but rather reinforces the intrinsic pathway by activation of factor IX after its triggering by thrombin.4,5 FXI activation can occur by two different systems: through the calcium and phospholipid-dependent coagulation pathway or through the contact pathway, and so lies at a ‘bidirectional interface’ between these two systems.6 Combined deficiency of tissue factor (TF) and factor XI in mice is lethal in embryos, but factor XII deficiency together with TF deficiency results in normal development. Experimental studies suggests a critical but unexpected role for FXI activation via thrombin for placental hemostasis.7
The FXII-contact pathway, and thereby FXI activation by FXIIa, is important in the development of thrombosis. Absence of FXII is not associated with a bleeding disorder, and in fact is protective against thrombosis. FXII links hemostasis with the kinin and complement systems. Animal experiments show that mouse gene knockouts for either FXI or FXII are protected against thrombosis and do not have a bleeding disorder.8 These observations have led to the development of antithrombotic agents by inhibition of FXI (see below).
Thrombin activation of FXI is triggered by polyphosphate release from activated platelets. Polyphosphates (polymers of 60-100 phosphate units in length) are ubiquitous in nature, and act as cofactors that provide a template for assembly of FXI and FIX.5,9 Such findings suggest that physiological FXI activation in hemostasis is promoted by platelet polyphosphates, a model that explains why the contact pathway is not required for hemostasis. Short-chain polyphosphates also act as cofactors for TF pathway inhibitor inactivation by activated FXI.10
The FXI gene was elucidated in 198711 and is located on chromosome 4. Two mutations are responsible for the majority of cases of FXI deficiency in the Jewish population:12 a stop codon in apple 2 (Glu117Stop) and a missense mutation in apple 3 (Phe283Leu), which results in reduced secretion of FXI. Many other mutations (mainly point missense changes) have also been reported, and 220 are detailed on the factor XI website (http://www.factorxi.org/stats.php; last update in 2015). A third of the reported mutations are found in the serine protease domain, and others are spread across the 4 apple domains; 13% of the reported mutations are intragenic.
Individuals with two mutations (either homozygous or compound heterozygous) usually have FXI levels less than 15 IU/dL. Homozygotes for the Glu117Stop mutation produce no FXI and have a higher bleeding risk than homozygotes for Phe283Leu whose baseline level approximates 10 IU/dL. Compound heterozygotes have levels intermediate between these two. Some heterozygous individuals have a lower than expected FXI level due to inhibition of release of normal FXI dimers. This inhibition occurs where the mutant FXI interferes with and binds to wild-type FXI in the cell, a so-called dominant-negative effect. A number of different mutations with this complication have been described.13,14
FXI deficiency has been classified as ‘severe’ with FXI levels of 15-20 IU/dL and below, and ‘mild’ or partial deficiency where FXI levels exceed 15-20 IU/dL. The classification divides heterozygotes from homozygotes or compound heterozygotes but does not reflect bleeding risk since many individuals with very low FXI levels do not bleed. This terminology has been updated to ‘major’ (levels < 15 IU/dL) and ‘partial’ (levels > 15 IU/dL).15
A collective review of the classification of the rare bleeding disorders based on published evidence and information from four national and international databases noted that there is either a poor or no relationship between the FXI level and observed bleeding severity.16 Individuals with no detectable FXI activity may never experience bleeding. The classification of hemophilia A and B into severe, moderate and mild as defined by factor levels and clinical features is not applicable or appropriate in FXI deficiency. Recent experiments demonstrate that a pool of non-circulating FXI associated with the endothelium in mice can be released by polyanions (e.g. polyphosphates).17 This phenomenon may contribute to the reason why plasma FXI level correlates so poorly to the bleeding risk.
Measurement of thrombin generation in a low TF environment may relate more closely to bleeding tendency18 and may help elucidate the relationship between FXI levels and bleeding tendency. This has been confirmed by a recent study,19 which illustrates the difference between chemical (coagulation factors measurements) and more physiological measures of the role in thrombin generation.20 Ninety-seven patients were studied (with 50 controls). Patients were divided into those with a bleeding history (n=50) and those without (n=24). Thrombin generation in platelet-rich plasma with low TF concentration, and inhibition of contact activation, differentiated bleeders and non-bleeders. In addition, recent work demonstrates that bleeders have clot stability deficiency and increased sensitivity to fibrinolysis. Turbidity-based fibrinolysis assays may help to predict bleeding in FXI deficiency.21 In future, such measurements may predict more accurately which individuals are at risk of bleeding after surgery.
A further example of a lack of correlation between the laboratory tests and the bleeding tendency is provided by a mutation in the platelet binding domain (Ser248Asn), which is associated with bleeding but a normal activated partial thromboplastin time (aPTT).22