Factor XIII: driving (cross-)links in hemostasis, thrombosis, and disease

Abstract

Blood clots are complex structures composed of blood cells and proteins held together by a structural framework provided by an insoluble fibrin network. Factor (F) XIII is a protransglutaminase essential for stabilizing the fibrin network. Activated FXIII(a) introduces novel covalent cross-links within and between fibrin and other plasma and cellular proteins and thereby promotes fibrin biochemical and mechanical integrity. These irreversible modifications are also major determinants of clot composition and functional properties. As such, FXIII has central roles in hemostasis and wound healing, thrombosis, and many proinflammatory diseases associated with coagulation activation. The biochemical properties of FXIII are as interesting as its biology is unusual, giving rise to unique and still undefined mechanisms. Here, we review features underlying FXIII biology, biochemical function, biophysical impact, and (patho)physiologic implications in hemostasis, thrombosis, and disease.

Introduction

Vascular injury and exposure of extravascular cells and proteins to blood initiate proteolytic events that culminate in thrombin generation. Thrombin-mediated cleavage of soluble fibrinogen leads to the formation of an insoluble fibrin network.^1,2 Fibrin traps circulating cells, including leukocytes, platelets, and red blood cells (RBCs), and provides the structural backbone required to hold these components together. However, clinical observations, supported by work with animal models and biochemical and biophysical measurements of clots, reveal that fibrin formation alone is insufficient to achieve the structural integrity needed to fully support hemostasis and wound healing.^3–8

Factor (F)XIII, previously termed “fibrin stabilizing factor,” was the last of the clotting factors to be identified.⁹ Despite its essential role in hemostasis, FXIII remains one of the least understood players in clot formation. FXIII is one member of a superfamily of (pro)transglutaminases (transglutaminases 1–7, erythrocyte band 4.2, and FXIII) that exhibit diverse biological functions. All transglutaminases except erythrocyte band 4.2 catalyze the formation of ε-N-(γ-glutamyl)-lysyl cross-links, producing new intramolecular and intermolecular isopeptide bonds. Interestingly, microbial transglutaminases with similar enzymatic activity are used in the food industry and are informally but aptly referred to as “meat glue.”¹⁰ The ability to create, rather than cleave, covalent bonds distinguishes FXIII from other coagulation enzymes and enables it to fulfill unique functions in hemostasis, wound healing, tissue repair, pregnancy maintenance, and bone metabolism. FXIII contributes to arterial thrombosis, venous thrombosis (VT), and diseases marked by activation of coagulation (eg, liver failure, inflammatory bowel disease [IBD]). These situations highlight potential opportunities to target FXIII with novel therapeutics.

FXIII biology and biochemistry

FXIII protein is present in the plasma and cells. Plasma and cellular FXIII possess unique structural and functional properties and, therefore, have overlapping and non-overlapping roles.

Plasma FXIII-A₂B₂

The plasma form of FXIII circulates as 2 (pro)active site-containing A subunits (FXIII-A₂, encoded by F13A1) and 2 regulatory B subunits (FXIII-B₂, encoded by F13B) tightly associated (K_D ~10⁻¹⁰ M) in a noncovalent heterotetramer (FXIII-A₂B₂, ~70 nM).¹¹ Each FXIII-A subunit consists of a β-sandwich domain, a catalytic core, 2 β-barrel domains, and an N-terminal activation peptide.¹² Each FXIII-B subunit consists of 10 tandem repeat “sushi” domains similar to domains found in complement factor H (20 sushi domains) and β₂GP1 (5 sushi domains).

Multiple lines of evidence document a hematopoietic origin for FXIII-A₂. Clinical observations reveal that FXIII-A polymorphisms track with bone marrow transplantation,¹³ and transcriptional analysis and studies using Cre/Lox mouse models suggest that aortic resident tissue macrophages are a primary site of plasma FXIII-A₂ production.¹⁴ It is unclear how FXIII-A₂ is secreted from nucleated cells. The FXIII-A polypeptide lacks a canonical endoplasmic reticulum signal sequence and, therefore, is not released through a classical secretion pathway. FXIII-A₂ secretion may involve subcellular trafficking in association with the Golgi apparatus.¹⁵ In contrast to FXIII-A₂, FXIII-B₂ polymorphisms track with liver transplantation, and liver-targeted lipid nanoparticles effectively reduce plasma FXIII-B₂.^13,16 Thus, the FXIII-A₂ and -B₂ homodimers are synthesized and secreted into plasma from separate tissues, and the FXIII-A₂B₂ complex forms in circulation (Figure 1). Interestingly, humans and mice with genetic loss of either FXIII-A₂ or FXIII-B₂ production have reduced circulating levels of both FXIII-A₂ and FXIII-B₂, and infusion to replace the missing protein increases the levels of both subunits.^17–19 Despite some potential overlap in transcriptional control,²⁰ compared with F13a1^+/+ mice, neither F13a1^−/− mice nor F13a1^−/− mice treated with recombinant FXIII-A₂ have altered hepatic F13b expression,²¹ suggesting regulation of these genes is distinct. Pharmacokinetic and pharmacodynamic studies suggest that the FXIII-A and -B subunits exhibit reciprocal stabilization essential for maintaining FXIII-A₂B₂ levels in the plasma.^17–19,21 Orthogonal approaches suggest that the FXIII-A₂ dimer binds asymmetrically to the FXIII-B₂ dimer and binding is achieved by interactions between FXIII-B₂ sushi domains 1 and 2 and the FXIII-A₂ core domains.^22,23 There is also approximately twofold higher total FXIII-B₂ relative to FXIII-A₂ in the plasma.¹⁸ The molecular economics or physiologic advantage of this stoichiometry is not known. Excess FXIII-B₂ may ensure that FXIII-A₂ is integrated into a stable FXIII-A₂B₂ complex and/or support rapid expansion of FXIII-A₂B₂ plasma levels in certain disease states.

Figure 1. FXIII-A₂ and FXIII-B₂ subunits are synthesized in separate tissues.

FXIII-B₂ and fibrinogen are independently synthesized and secreted by hepatocytes. FXIII-A₂ is synthesized by bone marrow-derived cells. Tissue macrophages thought to reside in the aorta release FXIII-A₂ into the plasma. Assembly of the FXIII-A₂B₂ heterotetramer and binding of FXIII-A₂B₂ and (free) FXIII-B₂ to fibrinogen occur in the plasma. The interaction between fibrinogen and FXIII-B₂ is mediated by fibrinogen γ-chain residues 390 to 396 and likely involves FXIII-B₂ sushi domains 8 to 10, although the FXIII-B₂ binding motif has not been identified. Each of these species (fibrinogen, fibrinogen plus FXIII-B₂, and fibrinogen plus FXIII-A₂B₂) is found in the plasma. In addition, hematopoietic progenitor cells package FXIII-A₂ into monocytes, mast cells, and other cells. Megakaryocytes package FXIII-A₂ into the platelets. Figure created with BioRender.com.

By virtue of their relative plasma concentrations, ~1% of fibrinogen is bound to FXIII-A₂B₂, but essentially all FXIII-A₂B₂ and (free) FXIII-B₂ circulates bound to fibrinogen (Figure 1), and both complexes coprecipitate with fibrinogen.²⁴ Initial studies suggested that FXIII-A₂B₂ binds to the fibrinogen alternatively spliced γ′ extension²⁵ or αC domain (residues 371–425, and specifically AαE396).²⁶ However, later studies revealed that fibrinogen residues γ390–396 (a highly conserved motif in humans [NRLTIGE], mice [NRLSIGE], and other animals) fulfill the major FXIII-A₂B₂ and (free) FXIII-B₂ carrier function, with only minor potential contributions of the γ′ extension or αC domain (Figure 2).^27–29 Indeed, FXIII-A₂B₂ binding to fibrinogen is mediated by FXIII-B₂ and likely involves sushi domains 8 to 10,^28,30,31 although the binding motif on FXIII-B₂ has not been identified. Interestingly, FXIII-B₂ sushi domains 6 to 10 were poorly resolved in a recent cryoelectron microscopy structure of FXIII-A₂B₂,³² suggesting this region is highly mobile. Flexibility within this region may facilitate the interaction between FXIII-B₂ and fibrinogen in the plasma. Importantly, FXIII-A₂B₂ levels are normal in afibrinogenemic humans³³ and mice,²¹ indicating FXIII-A₂B₂ binding to fibrinogen is not required to stabilize the heterotetramer or preserve its half-life in circulation. N-glycosylation increases the FXIII-B₂ half-life in circulation,³⁴ but definitive clearance mechanisms remain unknown.

Figure 2. Plasma FXIII-A₂B₂ binding to fibrinogen accelerates FXIII-A₂B₂ activation.

(A) Fibrinogen consisted of 2 Aα- (red), 2 Bβ- (blue), and 2 γ-chains (green) arranged in a trinodular structure with distal D domains and a central E domain. The AαC domain (αC) extends beyond the D domain and wraps back toward the E domain. FXIII-A₂B₂ circulates bound to fibrinogen γ-chain residues 390–396 through the FXIII-B subunits. During coagulation, thrombin cleaves N-terminal fibrinopeptides from the Aα- and Bβ-chains located within the E domain. (B) End-to-end polymerization of fibrin monomers brings γ-chain–bound FXIII-A₂B₂ near the thrombin at the D:E:D interface to form a ternary complex. This complex facilitates thrombin-mediated cleavage of the FXIII-A₂ activation peptides. Calcium-mediated dissociation of FXIII-B₂ from the FXIII-A subunits yields FXIIIa. (C) FXIIIa cross-links (X) the nearby γ-chains yielding γ-γ dimers. This γ-chain cross-linking also further promotes FXIII-B₂ dissociation from the clot. (D) FXIIIa translocates from the γ-chain to the αC region at or near α-chain residue E396 and catalyzes cross-link formation between fibrin α-chains and between fibrin and other plasma proteins (eg, α₂-AP). Figure created with BioRender.com.

Plasma FXIII-A₂B₂ activation and inactivation

The interaction between plasma FXIII-A₂B₂ and fibrinogen accelerates thrombin cleavage of the FXIII-A subunits.^28,30 Briefly, thrombin-mediated proteolysis of fibrinogen to fibrin and subsequent formation of half-staggered fibrin protofibrils bring C-terminal fibrin γ-chain regions (D domains) near the central fibrin E domain (Figure 2). By virtue of FXIII-A₂B₂ binding to residues γ390–396 within the D domain, protofibril formation also brings FXIII-A₂B₂ near thrombin bound to the central fibrin E domain. This proximity facilitates cleavage of the 37-amino acid N-terminal activation peptide from the FXIII-A subunits (Figure 2). Activation peptide release is accompanied by calcium ion binding to the FXIII-A subunits, causing a conformation change that releases FXIII-B₂ and converts the FXIII-A₂ dimer to an open monomeric conformation (FXIII-A*, FXIIIa) permitting substrate access to the activation cleft.^35,36 Plasmin or enzymes released by granulocytes may cleave and inactivate FXIIIa.^37,38

Plasma FXIIIa function

Processes that determine the composition and structural stability of hemostatic and thrombotic clots are primarily mediated by FXIIIa derived from plasma FXIII-A₂B₂. Proteomics studies have identified >100 substrates for FXIIIa, many of which become cross-linked to the clot.³⁹ Among these substrates are coagulation and adhesive proteins, including fibronectin and von Willebrand factor.^40,41 The primary substrate of plasma FXIIIa is fibrin. Cross-linking within and between fibrin γ- and α-chains generates γ-γ dimers and high molecular weight γ-α– and α-α–rich species. Cross-linking has relatively minor effects on overall fibrin network structure.⁴² However, cross-linking compacts individual fibers⁴³ and increases the stiffness of individual fibers and the fibrin network.^7,8 Non–cross-linked fibrin is highly elastic (ie, stretchy), and this is thought to result from monomer unfolding, straightening, and/or protofibril sliding.^8,44,45 Cross-linking prevents sliding, reduces extensibility, and enhances fibrin mechanical integrity during platelet-mediated clot contraction and wound healing. The ability of plasma FXIIIa to rapidly cross-link fibrin in concert with its polymerization also promotes RBC retention in contracting clots.⁴⁶ FXIIIa does not directly cross-link RBCs to the clot.⁴⁶ Rather, this function is specifically associated with the formation of α-chain–rich species^46,47 and may stem from increased stiffness imparted by the introduction of α-chain cross-links. FXIIIa-mediated cross-linking of α₂-antiplasmin (α₂-AP) to fibrin localizes α₂-AP to fibrin in flow and during clot contraction, and this mechanism protects clots against plasmin-mediated degradation (fibrinolysis).^48,49

Cellular FXIII-A₂

Many cells of the bone marrow and mesenchymal lineage, including monocytes/macrophages, megakaryocytes and platelets, osteoblasts, and other cells have FXIII-A₂. Although cellular FXIII-A₂ has identical active site-containing subunits as plasma FXIII-A₂B₂, the role(s) of cellular FXIII-A₂, including platelet FXIII-A₂, are less clear and more controversial.

Hematopoietic cell and leukocyte FXIII-A₂

In addition to secreting FXIII-A₂ into the plasma (Figure 1), monocytes/macrophages maintain an intracellular pool of FXIII-A₂. FXIII-A₂ is also expressed in B-cell progenitor lymphoblasts in acute lymphoblastic leukemia and has been implicated as a molecular marker and pathologic mechanism in myeloid leukemias.⁵⁰ FXIII-A₂ stores may help define cellular functional responses. For example, FXIII-A expression is up-regulated in “M2” polarized macrophages⁵¹ and FXIII-A–deficient macrophages display differential gene induction in response to treatment with the M2-polarizing cytokine interleukin-4.⁵² Activated monocytes and macrophages can also externalize FXIII-A₂ to their surface.^53,54 Exposed FXIII-A₂ may cross-link fibrin, leukocytes, and extracellular matrix to protect thrombi against fibrinolytic degradation, contribute to matrix remodeling and tissue repair, and alter macrophage effector function in diseases, including fibrosis, infection, and cancer.⁵⁵ Mast cells contain high levels of FXIII-A,⁵⁶ which may facilitate control of certain bacterial infections.⁵⁷ Uncovering specific functional role(s) of cellular FXIII-A₂ in the face of abundant plasma FXIII-A₂B₂ may be difficult but perhaps enabled by new tools capable of selectively ablating the plasma pool of FXIII-A₂B₂.¹⁶

Platelet FXIII-A₂

FXIII-A₂ is the only transglutaminase in human megakaryocytes and platelets^58,59 and the major transglutaminase in mouse platelets.^60,61 In humans, approximately half of circulating FXIII-A₂ is in the platelets, where it is present at ~83 000 copies per platelet (~60 fg) and comprises ~3% of the platelet proteome.^62,63 Because platelets take up very little FXIII-A₂B₂ from the plasma, most platelet FXIII-A₂ is synthesized and packaged into the platelets by megakaryocytes during thrombopoiesis (Figure 1).^64,65 Platelets also contain F13A1 transcripts, and ribosome occupancy of platelet F13A1 increases after the platelets are stimulated by thrombin, suggesting platelets are capable of de novo FXIII-A synthesis.⁶⁶ Situations in which de novo FXIII synthesis contributes to platelet function remain unclear.

Platelet FXIII-A° activation and exposure

In contrast to plasma FXIII-A₂B₂, cellular FXIII-A₂ is activated nonproteolytically to FXIII-A° by high calcium concentrations achieved within the cell; calcium binding causes a reversible⁶⁷ conformational change that exposes the active sites⁶⁸ and produces a structure that is functionally distinct from FXIII-A*.⁶⁹ Interestingly, whereas platelets carry most of their soluble coagulation proteins in α-granules, most platelet FXIII-A₂ is present in the cytoplasm.⁷⁰ Mechanisms mediating the mobilization and exposure of platelet FXIII-A₂ during platelet activation are not well understood. Platelet FXIII-A₂ is not released after stimulation of platelets with single agonists (eg, thrombin, collagen, plasmin). Dual stimulation with thrombin plus collagen/convulxin to simultaneously activate both protease-activated receptors and glycoprotein VI enables externalization of a small amount of platelet FXIII-A₂ that can be detected in a focal region on the surface of phosphatidylserine-positive platelets termed the platelet “body” or “cap,”⁷¹ including on platelet-derived extracellular vesicles.⁷² However, even after dual stimulation with high agonist concentrations, most FXIII-A is retained within the platelet cytoplasm, where it is protected from thrombin- or plasmin-mediated proteolysis.⁷³ Thus, despite its abundance, during hemostasis, platelet FXIII-A° has only a minor role in cross-linking plasma proteins and does not promote RBC retention in the clots.^61,74 These data suggest that platelet FXIII-A₂ is a protected pool with biological role(s) that differ from plasma FXIII-A₂B₂. This pool might be released as platelets fragment or disintegrate during the resolution of hemostatic clots or intravascular thrombi. Delayed exposure of transglutaminase activity may help preserve the structural integrity of the clot and/or promote other functions related to resident cell migration, proliferation, or tissue remodeling.

Platelet FXIII-A function

Several functions have been attributed to platelet FXIII-A₂; however, these are controversial. Because humans⁴ and mice^14,61 with genetic loss of FXIII-A₂ have a normal platelet count, FXIII-A₂ is not required for platelet production. FXIIIa can cross-link cytoskeletal proteins, including myosin, actin, filamin, and vinculin, and colocalize with heat shock protein 27 that mediates actin dynamics.^75–78 Some studies reported that interactions between FXIII-A₂, fibrin, and α_IIbβ₃ are required for platelet contraction.^79,80 However, cystamine used to inhibit FXIIIa in several of these studies also inhibits thrombin,⁸¹ which complicates the interpretations. Others have revealed that platelet binding to polymerizing fibrin requires activated α_IIbβ₃ but not fibrin cross-linking⁸² and observed normal or only slightly delayed contraction in platelet-rich plasma from FXIII-A–deficient humans and mice.^61,83,84 FXIII-A₂ can cross-link α-granule proteins to serotonin and has been implicated in the formation of highly active collagen- and thrombin-activated platelets.⁸⁵ However, F13a1^−/− mice have been reported to produce collagen- and thrombin-activated platelets,⁸⁶ suggesting platelet FXIII-A₂ is not required for this mechanism.

Contribution of FXIII to hemostasis, thrombosis, and related diseases

Given the ability of FXIII(a) to introduce new covalent bonds and stabilize proteins and protein complexes, it is not surprising that FXIII has essential roles in hemostasis and thrombosis. Mechanisms have been linked to the canonical role of FXIIIa in stabilizing fibrin matrices but may also reflect noncanonical roles. In addition, proinflammatory conditions, including liver disease, IBD, trauma, surgery, and sepsis, can lead to loss of FXIII through activation and consumption, and supplementation with plasma-derived FXIII-A₂B₂ or recombinant FXIII-A₂ has been reported to improve healing.^87–91 In these situations, it is unclear whether benefits are due to enhanced stabilization of clots within damaged or inflamed tissue and/or to non-hemostatic effects of FXIII(a) on these pathologies. Here, we highlight the role of FXIII in hemostasis and thrombosis, including liver disease and IBD, two settings in which FXIII(a) stabilization of fibrin contributes to disease etiology.

FXIII in hemostasis

FXIII(a)-mediated stabilization of fibrin is essential to protect clots against mechanical disruption and biochemical dissolution (Figure 3). Congenital FXIII deficiency is rare and associated with subcutaneous bleeding and menorrhagia, delayed rebleeding, poor wound healing, and spontaneous miscarriage in humans and mice.^3,4 Congenital FXIII deficiency is most often associated with mutations within F13A1; however, mutations within F13B are also reported.⁹² Partial or acquired FXIII deficiency often secondary to liver disease, IBD, trauma, surgery, and sepsis/disseminated intravascular coagulation may be common and underdiagnosed.⁹¹ Bleeding in FXIII deficiency stems from insufficient cross-linking between fibrin chains and between fibrin and α₂-AP. FXIII deficiency is also associated with the highest rate of intracranial hemorrhage compared with other clotting factor deficiencies, including fibrinogen deficiency,^5,6 suggesting FXIII has a fibrin(ogen)-independent role in maintaining vascular integrity in the brain. Individuals with F13A1 mutations that reduce FXIII-A antigen or activity may be treated with recombinant FXIII-A₂, which rapidly complexes with endogenous FXIII-B₂ to produce a functional heterotetramer. Because FXIII-B₂ stabilizes FXIII-A₂, individuals with F13B mutations lack both FXIII-B and plasma FXIII-A and must be treated with plasma-derived FXIII-A₂B₂. Restoration of FXIII-A₂B₂ in the plasma normalizes most hemostatic defects associated with FXIII deficiency, demonstrating a major role for plasma FXIII-A₂B₂ in hemostasis.⁹³

Figure 3. During hemostasis, FXIIIa-mediated cross-linking protects against mechanical disruption and fibrinolysis and enhances red blood cell retention in the clots.

Injury and exposure of blood components to extravascular tissues promote platelet deposition to form an initial plug, followed by activation of coagulation and thrombin generation. Thrombin cleaves fibrinogen to fibrin and FXIII-A₂B₂ to activated FXIIIa. Polymerized fibrin traps local blood cells. FXIIIa cross-linking (X) of fibrin mechanically stabilizes fibrin fibers and increases red blood cell retention during platelet-mediated clot contraction. FXIIIa cross-linking of α₂-AP to fibrin retains α₂-AP within the contracted clot and reduces fibrinolysis. Figure created with BioRender.com.

Several studies have tested the hypothesis that supraphysiologic concentrations of FXIII that enhance fibrin cross-linking may promote clot stability and hemostasis in hemophilia. These studies reveal that FXIII supplementation accelerates FXIII activation and cross-linking of fibrin and α₂-AP and increases whole blood clot mass, strength, and resistance to fibrinolysis.^94–96 The use of FXIII concentrates may be particularly helpful in persons with acquired inhibitors and refractory bleeding despite hemostatic therapy. The stabilizing effect of FXIII in this setting depends on sufficient thrombin generation to activate FXIII.⁹⁶

FXIII in thrombosis

During VT, inappropriate activation of coagulation within the blood vessels can lead to exuberant and dysregulated thrombin generation and intravascular fibrin deposition (Figure 4).⁹⁷ Trapping of the blood cells by cross-linked fibrin and clot contraction produces a consolidated thrombus and venous occlusion (Figure 4). Anticoagulants that limit thrombin generation and/or its activity mitigate fibrin deposition and reduce VT. During arterial thrombosis, vascular damage leading to platelet accumulation and activation enables fibrin deposition. Thrombolytic therapy to facilitate endogenous fibrinolytic mechanisms dissolves fibrin and restores vessel patency. Consequently, the ability of FXIII to stabilize fibrin against mechanical disruption and biochemical dissolution has important implications for both venous and arterial thromboembolic diseases. Genetic studies and experimental models of thrombosis have identified mechanisms by which FXIII contributes to these events.

Figure 4. During VT, FXIIIa-mediated cross-linking increases fibrin mechanical stability, enhances red blood cell retention and thrombus mass, and reduces fibrinolysis.

VT is thought to initiate in venous valve pockets, where nonlaminar blood flow and inflammatory mediators activate endothelial cells and lead to expression of cell adhesion molecules. Leukocytes and platelets recruited to the endothelial surface facilitate thrombin generation and thrombin-mediated proteolytic conversion of fibrinogen to fibrin and FXIII-A₂B₂ to activated FXIIIa. Polymerized fibrin traps resident cells, leading to an occlusive thrombus. FXIIIa cross-linking (X) of fibrin stiffens fibrin fibers and increases thrombus mass by promoting red blood cell retention in the thrombi during platelet-mediated contraction. FXIIIa cross-linking of α₂-AP to fibrin retains α₂-AP within the contracted clot and reduces fibrinolysis. Figure created with BioRender.com.

FXIII V34L polymorphism in venous and arterial thromboses

Associations between polymorphisms encoded by F13A1 and F13B and venous and arterial thromboses suggest that FXIII-A₂B₂ contributes to these etiologies.⁹⁸ The most recognized polymorphism is a valine-to-leucine substitution in FXIII-A residue 34 present in ~25% of White individuals of European descent.⁹⁹ Although individual epidemiologic studies have not consistently associated the 34Leu allele with thrombosis, meta-analyses suggest that the 34Leu allele offers modest but significant protection against VT and coronary artery disease.⁹⁸ Characterization of mechanisms underlying this association revealed unexpected biochemistry and exposed an elegant relationship between fibrin formation and cross-linking. FXIII residue 34 is located in 3 amino acids N-terminal to the cleavage site that releases the FXIII-A activation peptide. The 34Leu allele facilitates thrombin-mediated release of the FXIII-A activation peptide, causing 2.5-fold earlier FXIII-A₂ activation, and therefore, faster fibrin cross-linking.¹⁰⁰ Interestingly, effects of the 34Leu allele on clot structure and stability depend on the local fibrinogen concentration.¹⁰¹ Notably, high fibrinogen concentrations accelerate fibrin formation and increase fibrin network density, strength, and stability, and elevated total circulating fibrinogen is an established risk factor for thrombosis.¹⁰² However, the 34Leu allele mitigates these prothrombotic effects at high fibrinogen concentrations by increasing fibrin fiber thickness and clot permeability and decreasing clot stability.¹⁰¹ The 34Leu allele also decreases RBC retention and clot mass in vitro, suggesting it also decreases thrombus burden in vivo.¹⁰³

FXIII in VT

Relatively few studies have investigated the relationship between circulating FXIII level and VT, and interpretation of retrospective studies has been limited by a consumptive loss of FXIII that occurs during thrombosis.^104,105 Studies in mice reveal that genetic loss of FXIII (F13a1^+/−, F13a1^−/−) delays FXIII activation and fibrin cross-linking and decreases RBC retention in venous thrombi, and, therefore, reduces VT mass ex vivo and in vivo.^27,61 This effect is specifically attributed to plasma FXIII-A₂B₂, but not platelet FXIII-A₂.⁶¹ Because FXIII-A₂B₂ binding to fibrinogen facilitates FXIII-A₂B₂ activation and fibrin cross-linking, loss of this interaction delays these events.^27,28 Accordingly, mice expressing mutated fibrinogen (Fgg^390−396A) that eliminates this FXIII binding motif exhibit delayed FXIII-A₂B₂ activation, and, therefore, delayed fibrin cross-linking.²⁷ Disruption in the coordination of these events allows clot contraction to occur before sufficient fibrin stabilization and permits RBC extrusion during clot contraction in vitro.²⁸ When subjected to VT models in vivo, Fgg^390−396A mice produce significantly smaller thrombi with reduced RBC content.²⁷ Collectively, these data suggest that the timing of FXIIIa-mediated fibrin cross-linking is critical to ensure fibrin fibers acquire viscoelastic properties before the generation of platelet contractile forces. Loss of this coordination compromises fibrin structural integrity during clot consolidation and results in the formation of smaller thrombi. Whether this mechanism may be targeted to reduce VT remains unknown.

FXIII in PE

Fibrin cross-linking by FXIII is essential to protect clots against pulmonary embolism (PE). After FeCl₃ application to the femoral vein, F13a1^−/− mice have increased clot embolization compared with F13a1^+/+ mice.¹⁰⁶ Furthermore, after FeCl₃ application to the femoral vein or inferior vena cava, mice expressing fibrinogen-bearing mutations in the γ-chain cross-linking sites (FGG3X) have increased embolic events and accumulation of fluorescent fibrin(ogen) within the lungs, suggesting γ-chain cross-linking protects against PE.⁴⁷ A study using a ligature/removal model that recapitulates aspects of PE pathophysiology in humans (VT with subsequent embolization) confirmed that complete deficiency increases clot embolization to the lungs, but also revealed that 50% of FXIII is sufficient to stabilize thrombi against embolization.¹⁰⁷ Interestingly, FXIII deficiency is only rarely associated with PE; a meta-analysis of PE prevalence among individuals with congenital coagulation deficiencies associated PE with reduced levels of clotting factors XI, IX, VIII, and VII, but not FXIII.¹⁰⁸ In total FXIII deficiency, increased susceptibility of non–cross-linked clots to fibrinolysis, a pathway especially prominent in lung vasculature, may mitigate PE risk.

FXIII in arterial thrombosis

Genetic loss of FXIII in mice does not delay FeCl₃-induced arterial occlusion.¹⁰⁹ However, studies in rabbits and dogs suggest pharmacologic inhibition of FXIIIa before arterial thrombus formation facilitates thrombolysis.^110,111 Unfortunately, pharmacologic limitations of existing FXIIIa inhibitors have prevented the translation of this observation to the clinical setting.¹¹²

FXIII in liver disease

Hepatic dysfunction in severe liver disease (eg, cirrhosis) is associated with reduced plasma FXIII.⁸⁷ Decreased levels are most likely explained by reduced FXIII-B₂ synthesis by the diseased liver. As such, hepatic dysfunction may contribute to coagulopathy by compromising clot cross-linking and stability. Interestingly, compelling experimental and clinical evidence suggests that this relationship is bidirectional, wherein coagulation factors, including FXIII, contribute to the development of liver disease and regeneration. For example, obesity promotes metabolic dysfunction-associated steatotic liver disease, but obese F13a1^−/− mice have less insulin resistance and decreased disease severity.^113,114 Intriguingly, when fed a high fat diet, Fgg^390−396A mice, which lack the plasma FXIII-A₂B₂ binding site, also develop less severe liver pathology.¹¹⁵ Notably, however, fibrinogen expressed by Fgg^390−396A mice also does not bind leukocyte integrin receptor α_Mβ₂,¹¹⁶ so the relative contributions of FXIII- and fibrin(ogen)-mediated functions are unclear.

Despite its pathologic role in steatotic liver disease, FXIII appears essential for liver repair/regeneration after bile duct injury¹¹⁷ and partial resection.¹¹⁸ Low preoperative plasma FXIII(a) activity is associated with development of post-hepatectomy liver failure,¹¹⁸ a major complication of this type of liver surgery. Studies in mice reveal that liver regeneration after partial resection is associated with intraoperative hepatic fibrin(ogen).¹¹⁹ Notably, after experimental partial hepatectomy, F13a1^−/− mice have dramatically reduced hepatocyte proliferation and reduced liver regeneration, revealing the importance of FXIIIa cross-linking of these intrahepatic fibrin(ogen) deposits.¹¹⁸

FXIII in colitis

IBD (ie, Crohn disease and ulcerative colitis) is characterized by immune-driven damage to the intestinal epithelial lining. Prolonged intestinal inflammatory insult can result in the formation of chronic wounds covering a significant surface area of the intestinal lining, ulcerations, and bleeding within intestinal tissue and the lumen of the gastrointestinal tract. Accordingly, colitis is characterized by a prothrombotic state and linked to activation of coagulation within the inflamed intestinal tissue. Studies have consistently detected reduced circulating FXIII in humans with active IBD relative to those with inactive IBD or healthy controls,^88,89 as well as in wild-type mice subjected to the established dextran sodium sulfate (DSS) mouse model of colitis.¹²⁰ Compared with control wild-type mice, DSS-challenged F13a1^−/− mice exhibited increased and prolonged disease activity and ulceration, and treatment of F13a1^−/− mice with recombinant FXIII-A reverses the enhanced disease activity.¹²⁰ Interestingly, Fgg^390−396A mice that lack the plasma FXIII-A₂B₂ binding site are protected from DSS-induced inflammation and intestinal damage.¹²¹ Thus, FXIII(a) suppresses IBD without requiring fibrinogen cofactor activity for its activation.

FXIII(a) as a therapeutic target

Because FXIII fortifies the clot structural framework, FXIII is an intriguing target to promote hemostasis or reduce thrombosis and thrombotic disease.

Congenital or acquired FXIII deficiency

Administration of plasma-derived FXIII-A₂B₂ or recombinant FXIII-A₂ restores circulating FXIII in individuals with FXIII deficiency. FXIII-A₂B₂ also contaminates plasma-derived fibrinogen preparations,¹²² so fibrinogen supplementation may also enhance FXIII levels and activity. Partial FXIII deficiency is surprisingly common,⁹¹ but the implications of this for bleeding risk, particularly during acute disease and disease progression, are unclear. Whether FXIII administration in these settings may limit pathogenesis and improve outcomes warrants continued investigations in preclinical and clinical studies.

Thrombosis and thrombotic disease

All current antithrombotic therapeutics suppress thrombin generation or its activity. It is therefore not surprising that each of these therapies increases bleeding risk. Because plasma FXIII enhances thrombus mass and protects against lysis, reducing or inhibiting FXIII may increase the susceptibility of intravascular thrombi to proteolysis and mitigate complications associated with persistent occlusive thrombi. Although complete deficiency causes severe bleeding complications, partial deficiency of FXIII is not generally associated with severe bleeding, PE, or delayed wound healing in humans or mice, suggesting a potential therapeutic range in which FXIII antagonism may be effective and safe.^107,108 Moreover, in contrast to other available anticoagulants, targeting FXIII(a) would not reduce thrombin generation or activity or prevent fibrin deposition. Indeed, FXIII(a) antagonism may allow for combined therapy with lower dosing of traditional anticoagulants.

Several FXIII(a) inhibitors, including antibodies, peptides, and small molecules, are amenable for in vitro studies, but major barriers, including short half-life and insufficient specificity, have limited their use in in vivo studies.¹¹² Several strategies to suppress FXIII or its activity are under consideration. Administration of small interfering RNA to knock down FXIII-B₂, and therefore plasma FXIII-A₂B₂, reduces plasma FXIII in mice for several weeks and enhances carotid artery reperfusion after arterial thrombosis, but this is currently restricted to the research setting.¹⁶ Tridegin is a leech saliva-derived protein with good selectivity for FXIIIa over other transglutaminases, but it is structurally complex, which complicates large-scale synthesis.¹²³ Creative strategies are needed to develop FXIII inhibitors that maintain the selectivity and potency of natural peptides, such as tridegin, but with chemical modifications or drug delivery approaches to support the half-life desired for a durable therapy. Antibody-based strategies inspired by biochemical studies may also prove attractive. Importantly, because FXIII has multiple biological roles, studies are needed to understand potential negative side effects of FXIII inhibition on hemostasis, pregnancy maintenance, bone development, and other clinical settings.

Conclusions

FXIII is a major determinant of clot structure and stability, and the importance of FXIIIa-catalyzed covalent modifications is revealed by bleeding in severe FXIII deficiency. FXIII has also been implicated in the pathophysiology of thrombosis and other diseases and may be a therapeutic target for reducing these pathologies. Unsolved mysteries in FXIII biology, biochemistry, and (patho)physiology beckon to clinicians and basic/translational scientists. What is the evolutionary advantage of synthesizing FXIII-A₂ and FXIII-B₂ in separate tissues? How is this biology linked to mechanisms regulating expression of these subunits? How does cellular FXIII-A₂ contribute to platelet function, and under what settings does this occur? Answering these questions will fill fundamental gaps and uncover novel strategies for limiting hematologic and other diseases.