Keywords
contact system - blood coagulation - polyphosphate - inflammation - factor XII
Schlüsselwörter
Kontaktphasesystem - Blutgerinnung - Polyphosphat - Entzündung - Faktor XII
Proteins of the Plasma Contact System
Proteins of the Plasma Contact System
The factor XII (FXII)-driven plasma contact system is a proinflammatory and procoagulant
plasma protease cascade that is initiated by FXII, in a reaction involving high-molecular-weight
kininogen (HK) and plasma prekallikrein (PK; [Fig. 1]).[1]
[2]
[3] Upon binding to negatively charged surfaces, the FXII zymogen is autocatalytically
converted to the serine protease FXIIa, which cleaves PK to active plasma kallikrein
(PKa). FXIIa initiates the intrinsic pathway of coagulation and leads to the PKa-mediated
liberation of the proinflammatory peptide hormone bradykinin (BK) from the precursor
HK (kallikrein–kinin pathway) in vivo. BK, in turn, is rapidly metabolized to vasoactive
and eventually inactive peptides. Furthermore, FXIIa has the capacity to activate
the fibrinolytic and complement systems in vitro. However, a possible in vivo relevance
of FXII contact activation for these reactions remains to be demonstrated.[4] In contrast to other coagulation factors, deficiency in FXII is not associated with
obvious coagulation abnormalities in humans and mice. FXII-deficient individuals are
phenotypically normal and do not bleed excessively. Our laboratory originally found
that both arterial and venous thrombus formation is defective in FXII-deficient (F12−/−
) mice while their hemostatic capacity is within the normal range.[5] It has also been shown that genetic ablation of F12 in mice leads to impaired venous thrombus formation in an inferior vena cava (IVC)
stenosis and femoral vein electrolytic injury model, but not in an IVC stasis model.[6] Even so, challenging the dogma of a coagulation “balance,” F12−/−
animals have a normal hemostatic capacity but are largely protected from thromboembolic
diseases such as ischemic stroke,[7] pulmonary embolism,[8] cancer-driven thrombosis,[9] or sepsis-associated thrombosis.[10] In addition, pharmacological inhibition of FXIIa-driven coagulation provides safe
thromboprotection in large animals.[11]
[12] Peptides, proteins, and small-molecular-weight inhibitors targeting FXII and FXIIa
are discussed in detail by Davoine et al.[13] In addition to coagulation, FXIIa initiates an inflammatory reaction by generation
of the peptide hormone BK.[14]
[15] Further processing of BK by removal of the C-terminal arginine through carboxypeptidase
N results in des-Arg9-BK. Binding of BK and des-Arg9-BK to its G-protein-coupled kinin
B2 and B1 receptors (B2R and B1R) initiates intracellular Ca2+-dependent signaling pathways that induce a plethora of inflammatory processes[16]
[17] leading to cytoskeleton rearrangement in endothelial[18]
[19] and epithelial cells.[20] Deficiency in C1 esterase inhibitor (C1INH, the major endogenous inhibitor of FXIIa
and PKa) or FXII gain-of-function mutations are associated with a BK-mediated, life-threatening
inflammatory disorder, referred to as hereditary angioedema (HAE).[21]
[22]
[23]
Fig. 1 Contact system–triggered pathways. The contact system is initiated by factor XII
(FXII) activation mediated by contact with anionic surfaces such as polyphosphate
(polyP) nanoparticles, mast cell heparin, cell-derived extracellular vesicles, misfolded
protein aggregates, and various exogenous surfaces. Activated FXII (FXIIa) drives
proinflammatory and procoagulant pathways, such as (1) plasma kallikrein (PK)-stimulated
release of bradykinin (BK) and des-Arg9-BK from high-molecular-weight kininogen (HK,
kallikrein–kinin system), followed by binding of the released peptide hormones to
kinin B2 receptor (B2R) and kinin B1 receptor (B1R); (2) proteolytic cleavage of factor
XI (FXI) leading to initiation of the intrinsic coagulation cascade with subsequent
thrombin and fibrin formation and clotting; (3) urokinase-type plasminogen activator
receptor (uPAR)-mediated activation of plasminogen to the protease plasmin, which
degrades formed fibrin clots, resulting in the formation of D-dimers; and (4) activation
of the classical complement pathway via the proteases C1r and C1s. While the C1 esterase
inhibitor (C1INH) plays an important role in complement inhibition, it is also the
main inhibitor of FXIIa and activated PK in vivo.
FXII: Domains and Their Functions
FXII: Domains and Their Functions
The plasma protein FXII is secreted by the liver into the circulatory system as an
inactive, single-chain enzyme precursor (zymogen). In addition, some blood cells such
as neutrophils express FXII[2]
[24] (https://www.proteinatlas.org/ENSG00000131187-F12/tissue). An N-terminal 19-residue signal sequence mediates secretion of the FXII zymogen.
In healthy individuals, the 80 kDa glycoprotein (596 amino acids) circulates in plasma
at a concentration of 30 to 35 µg/mL (375 nmol/L).[25] Proteolysis of FXII zymogen occurs at the single peptide bond Arg353–Val354 leading
to formation of the active protease FXIIa. Formation of FXIIa proceeds via two principal
mechanisms: (1) binding (“contact”) to negatively charged surfaces that induce a conformational
change and limited enzymatic activity (autoactivation) or (2) by proteases including
PKa and plasmin (fluid-phase activation). FXIIa is composed of two chains termed the
“heavy chain” (353 residues) and the “light chain” (243 residues). Both chains are
linked by an intramolecular disulfide bond spanning cysteine residues 340 and 467.
The light chain comprises the catalytic domain with the typical triad of a serine
protease (His393, Asp442, Ser544). The heavy chain is composed of individual domains,
starting with the N-terminal fibronectin type II (Fib-II) domain, followed by a first
epidermal growth factor-like (EGF-I) domain, a fibronectin type I (Fib-I) domain,
a second EGF-like (EGF-II) domain, a kringle domain, and a C-terminal proline-rich
(PR) region. In particular, the heavy chain mediates contact to other proteins and
surfaces with implications for zymogen activation ([Table 1]). A FXII fragment that completely lacks the heavy chain was unable to contact activate.[11] However, the site(s) required for contact activation has/have remained enigmatic
and different studies came to discrepant results.
Table 1
FXII domains/regions and their functions
Attribute
|
Fib-II
|
EGF-I
|
Fib-I
|
EGF-II
|
KR
|
PR
|
Catalytic domain
|
Potential surface-binding site
|
✓[26]
[27]
|
✓[33]
|
✓[29]
[30]
|
✓[32]
[102]
|
✓[32]
[102]
|
|
|
Putative Zn2+-binding site
|
✓[35]
[103]
|
✓[35]
|
|
✓[35]
|
|
|
|
FXI interaction
|
✓[28]
|
|
|
|
|
|
|
Cell attachment
|
✓[104]
|
|
|
|
|
|
|
Heparin-binding site
|
|
|
✓[105]
|
|
|
|
|
Fibrin-binding site
|
|
|
✓[105]
|
|
|
|
|
Catalytic triad (His393, Asp442, Ser544)
|
|
|
|
|
|
|
✓[106]
|
Abbreviations: Fib-II, fibronectin type II domain; EGF-I, first epidermal growth factor-like
domain; Fib-I, fibronectin type I domain; EGF-II, second epidermal growth factor-like
domain; KR, kringle domain; PR, proline-rich region.
Anti-FXII antibodies that interfere with FXII contact activation have been used to
identify residues involved in surface binding. Monoclonal antibodies P5-2-1 and B7C9
identified a region within the Fib-II domain as a FXII surface binding site.[26]
[27] Challenging a role of residues 1–28 in the Fib-II domain for FXII contact activation,
a FXII deletion mutant lacking residues 3–19 was readily contact activatable by dextran
sulfate (DXS).[28] The monoclonal antibody B7C9 identified amino acids 134–153 as a part of the Fib-I
domain as potential polyanionic surface binding site.[29] Furthermore, anti-FXII antibody KOK5 that recognizes a discontinuous epitope in
the Fib-II domain involving residues 30–33, 40–47, and 57–60 inhibits kaolin-triggered
FXII contact system activation.[30] However, a 319-amino-acids-deleted FXII mutant, rFXII.lpc, that consists of the
light chain and the C-terminal part of the PR region but lacks the entire KOK5 epitope,
binds to kaolin and is activated by the anionic surface.[31] A FXII mutant spanning the EGF-II domain, kringle domain, and PR region, as well
as the light chain (rFXII-U-like) binds to surfaces and undergoes contact activation.[32] Moreover, a recent study based on recombinant N-terminal FXII deletion mutants indicated
that truncated FXII variants lacking the Fib-II domain (FXII Δ1–71) are more prone
for autoactivation.[33] This suggests that the Fib-II domain is involved in zymogen contact activation by
shielding its activation site. Using various deletion mutants, it was further shown
that the Fib-II domain is dispensable for FXII binding to kaolin or polyphosphate
(polyP), whereas the EGF-I domain appeared as indispensable.[33] This result is demonstrated in pull-down assays with recombinant FXII variants lacking
the EGF-I domain (FXII Δ1–112) that are defective in binding to kaolin and polyP.[33] Together, these studies reveal the complexity in activation of the FXII zymogen
and suggest a systematic approach to identify the regions of FXII involved in surface-triggered
autoactivation. Functions related to the kringle domain and PR region are not well
characterized. Based on in vitro experiments, these domains influence the conformation
of FXII and accordingly the exposure of its cleavage site at Arg353-Val354. A recently
described FXII point mutation (Trp268Arg) located in the kringle domain next to the
PR region disturbs zymogen quiescence and accelerates FXII activation leading to rash
and urticaria[34]; however, the underlying FXII-dependent mechanism remains to be characterized.
FXII: Zinc Binding
The divalent cation zinc (Zn2+) is involved in the binding and activation of FXII on negatively charged surfaces.
In particular, Zn2+ promotes the stabilization of the transition states formed during the conformational
change of FXII upon its autocatalytic cleavage. Putative Zn2+ binding sites within the Fib-II domain (residues His40–His44 and His78–His82) and
the EGF-like domains (residues His94–His131 and His174–His176) were described for
FXII.[35] Zn2+ circulates in plasma bound to proteins and is released by activated platelets. In
vitro studies have also shown that binding of FXII-stimulated platelets is enhanced
in the presence of Zn2+.[36] Moreover, Zn2+ facilitates FXII binding to endothelial and smooth muscle cells promoting mitogenic
activity in these cell types.[37] FXII binds urokinase-type plasminogen activator receptor (uPAR) and induces proliferation
via β1 integrin and EGF receptor (EGFR) signaling that leads to phosphorylation of
pERK1/2 and Akt.[38] In addition to uPAR, the globular C1q receptor (gC1qR) provides an attachment site
for contact system proteins on the cell surface. The interaction of C1q with gC1qR
is known to activate the classical complement cascade[39]; however, gC1qR also binds to FXII and HK and promotes their assembly on cell surfaces.[40] Complex formation with contact system proteins is facilitated by anionic loops of
gC1qR, whereas a cell-binding site allows attachment of gC1qR to plasma membrane proteins,
such as integrins.[41] Recent structural analysis revealed that the N-terminal Fib-II domain of FXII requires
the presence of Zn2+ for interaction with gC1qR.[42] By clustering FXII and HK into a planar ternary complex with a molecular weight
of 500 kDa, cell surface–bound gC1qR may enable initiation of the FXII-driven intrinsic
coagulation and kallikrein-mediated BK liberation.[42] However, other studies have shown that gC1qR has a mitochondrial targeting sequence
sufficient to localize the protein to intracellular compartments, challenging its
proposed role as a cell surface receptor for plasma proteins.[43]
FXII: Glycosylation
FXII is a glycoprotein of the β2 globulin fraction of serum electrophoresis separation.
Carbohydrates constitute for 14.2 kDa (17.8%) of the apparent molecular protein mass
of 80 kDa. In human FXII, two N- and seven putative O-glycosylation sites have been identified ([Table 2]). Most of the glycosylation sites are located within the PR region. Glycosylation
is crucial for the functionality and the physical properties of FXII. Consistent with
this, a nonglycosylated FXII variant expressed in Escherichia coli has very low or no solubility (unpublished observation of the authors). While glycosylation
is critical for solubility of the plasma protein, carbohydrates also regulate FXII
contact activation. For example, contact activation of a hypoglycosylated FXII variant
with a missense mutation at position 309 (Thr309Lys/Arg; identical to position 328
counting the 19 amino acid signal peptide) is largely accelerated.[44] Lack of a single O-linked FXII glycosylation leads to a gain-of-autoactivation.[45] Furthermore, the mutation produces an additional site for activation of the zymogen
by plasmin.[46] Synergistically, gain-of-contact activation and an extra cleavage site lead to excessive
production of BK. BK-signaling increases microvascular permeability leading to the
inherited swelling disease HAE. Three distinct types of HAE can be differentiated:
mutations in the gene SERPING1 encoding for C1INH are causative for HAE type I and HAE type II. Patients suffering
from these forms of HAE exhibit a decreased plasma concentration and impaired functionality
of C1INH, respectively. In contrast, the third type represents HAE with normal C1INH
(HAE-nC1INH) associated with mutations of genes other than SERPING1, such as KNG1,[47]
ANGPT1,[48]
PLG,[49] or F12.[50] A molecular aberration in F12 leads to the aforementioned Thr309Lys/Arg FXII mutation and causes HAE-FXII due to
the increased susceptibility of FXII for contact activation.[45]
Table 2
Glycosylation sites of FXII
FXII domain/region
|
Type of glycosylation
|
Amino acid
|
Amino acid position
|
Reference
|
Related findings
|
Reference
|
EGF-I
|
O-linked
|
Thr
|
90
|
[107]
|
|
|
Kringle
|
N-linked
|
Asn
|
230
|
[102]
[108]
|
Asn230Lys mutation (recombinant model, Schneider 2 (S2) cells): lack of glycosylation results in increased intracellular FXII levels
and defective secretion
|
[109]
|
Proline-rich
|
O-linked
|
Thr
|
280
|
[102]
|
|
|
O-linked
|
Thr
|
286
|
[102]
|
|
|
O-linked
|
Ser
|
289
|
[102]
|
|
|
O-linked
|
Thr
|
309
|
[102]
|
Thr309Lys/Arg mutation: loss of glycosylation leads to increased FXII activity in
HAE-FXII patient plasma
|
[45]
|
O-linked
|
Thr
|
310
|
[102]
|
|
|
O-linked
|
Thr
|
318
|
[102]
|
|
|
Catalytic domain
|
N-linked
|
Asn
|
414
|
[108]
[110]
|
Hypoglycosylated type of FXII in plasma of PMM2-CDG patients: FXII contact activation
is unaffected
|
[109]
|
Abbreviation: EGF-I, first epidermal growth factor-like domain.
FXII: Activation
Endogenous Activators
Over the years, an array of physiological structures with anionic surfaces have been
identified as putative FXII contact activators. Here, we focus on some of them, including
mast cell heparin, misfolded protein aggregates, polyP, and cell-derived extracellular
vesicles. A complete overview on FXII contact activators can be found here.[1]
Heparin, a polysaccharide with a high degree of sulfation and acetylation, is released
from mast cells stimulated by IgE/antigen. Mast cell heparin initiates FXII contact
activation in human plasma and in mice.[23] Subsequently, BK is released by the kallikrein–kinin system,[15] inducing vasodilation and an increase in permeability with implications for anaphylaxis[51] and edema.[23] In contrast, deficiency of FXII or B2R that is associated with defective BK signaling
interferes with mast cell heparin-driven acute swelling events in vivo.[23]
The amyloid β (Aβ) peptide is one of the pathological hallmarks of Alzheimer's disease
(AD) and, in its aggregated form, has the capacity to initiate FXII contact activation.
In particular, the Aβ42 isoform induces thrombin generation via the intrinsic coagulation
cascade in a FXIIa-mediated manner[52] and induces inflammatory reactions by the activated kallikrein–kinin system. Elevated
levels of FXIIa are found in the plasma of AD patients,[53] consistent with a role of the contact system in procoagulant and proinflammatory
states associated with AD.
Polyphosphate
It is known for decades that activated platelets promote coagulation in a FXII-dependent
manner,[54]
[55]
[56]
[57]
[58]
[59] but the platelet-derived FXII activator had remained unknown. Platelets store high
concentrations of the inorganic polymer polyP in their dense granules.[60] PolyP is a linear polymer composed of orthophosphate [P]i monomers that are linked by energy-rich phosphoanhydride bonds. PolyP complexed with
calcium ions (Ca2+) is packed in dense granules in high concentrations (up to the molar range) leading
to characteristic dark spots in transmission electron microscopy images. Activated
platelet-derived polyP potently initiates FXII contact activation[8] providing the long-sought link of primary and secondary hemostasis.[61] PolyP is instable in plasma and is degraded by endogenous phosphatases.[62] Targeting polyP using synthetic and recombinant binding proteins or exopolyphosphatases,
enzymes that degrade polyP in bacteria and yeast, interferes with the activity of
the polymer to induce FXII contact activation.[63] Only a small portion of platelet polyP (<5%) is released into the cell supernatant
upon activation,[63] but the vast majority remains attached to the plasma membrane, indicating that the
polymer operates on cell surfaces. In contrast to synthetic polyP, naturally occurring
polymers are bound to divalent metal cations such as Ca2+ (and possibly others, e.g., Zn2+). The resulting aggregate formation leads to insoluble polyP nanoparticles that are
retained on the platelet membrane.[64] Intravital microscopy has visualized polyP nanoparticles on the surfaces of activated
platelets and within platelet-rich thrombi confirming that the majority of platelet
polyP is retained on the plasma membrane.[65]
The insoluble Ca2+ and Zn2+-rich polyP nanoparticles found in nature challenge the hypothesis that size of individual
polymers would determine polyP function in coagulation. While chain length of synthetic
polyP determines its activity on coagulation reactions in plasma,[66] natural platelet polyP forms insoluble Ca2+-rich nanoparticles. These particles function in coagulation reactions independently of the chain length of the individual polyP molecule.[65] Furthermore, these particulate polyP form triggers FXII activation on plasma membranes
of platelets, megakaryocytes, various cancer cells, and exosomes/microparticles derived
from these cells, with implications for thrombosis in murine models.[9] In contrast to insoluble particulate polyP, a potential in vivo function of soluble
polyP remains to be shown. Notably, soluble polyP may have anticoagulant effects since
Ca2+-free synthetic short-chain polyP acts as a chelator for Ca2+ ions in plasma.[66]
Platelets release small amounts of short chain polyP (around 80 monomers) that is
soluble and found in the supernatant. However, and similar to other cells the majority
of platelet polyP comprises long chain insoluble polymers with a chain length > 200
[P]i subunits. Methods to isolate polyP from cells and tissues have been established.
The phenol/chloroform extraction method (Werner's protocol[67]) selects for water soluble (short chain) polyP, while anion exchanger is based on
purification of both soluble (short chain) and insoluble (long chain) polyP.[68] Despite that polyP is found in every cell in nature, the biosynthesis of the polymer
is a topic of ongoing research. So far, mainly prokaryotes but also yeast have served
as model organisms for the analysis of enzymes that synthesize polyP with different
chain lengths.[68] In yeast, intracellular levels of polyP and inositol pyrophosphate are interdependent
from each other and inositol pyrophosphate modulates the cellular [P]i influx.[69] The concentrations of both molecular species are regulated by inositol hexakisphosphate
kinase (IP6K1) activity. In Saccharomyces cerevisiae and mice, genetic ablation of Ip6k1 severely reduces polyP and inositol pyrophosphate levels.[70] While reduced inositol pyrophosphate levels do not significantly affect secondary
hemostasis, decreased platelet polyP impairs activation of the coagulation cascade
in a FXII-dependent manner.[71] Consequently, IP6K1 deficiency protects genetically modified mice from lethal venous
thromboembolic events. The xenotropic and polytropic receptor 1 (XPR1) that was originally
desired as [P]i exporter represents another polyP-regulating protein. Genetic and pharmacologic targeting
of XPR1 interferes with [P]i export and results in polyP accumulation in platelets.[72] Accordingly, surface exposure of polyP from stimulated XPR1-deficient platelets
is increased. Excess platelet polyP has no effect on hemostasis but accelerates arterial
and venous thrombosis in a FXII-dependent manner. Therefore, modulating proteins that
regulate polyP content in platelets may be useful for the prevention of thrombotic
disorders.
Cell-Derived Extracellular Vesicles
Extracellular vesicles (EV) modulate blood coagulation under physiological and pathological
conditions. Originally, platelet-derived EV termed “microparticles” have been shown
to initiate FXII contact activation, while red blood cell–derived EV were inactive
in producing FXIIa.[73] Flow cytometric and biochemical analyses with EV isolated from platelet concentrates
confirmed that EV support thrombin generation in an FXII-dependent manner. However,
in that specific study, red blood cells were also found to contribute to EV-driven
FXIIa production,[74] suggesting that distinct populations or age-dependent effects on the coagulation
system exist in red blood cells. Red blood cell–derived EV participate in activation
of the intrinsic coagulation pathway by two mechanisms: (1) the classical contact
activation and (2) a FXII-independent pathway involving PK activation followed by
direct stimulation of factor IX (FIX) by a yet unknown protease.[75] Recently, and in line with the previous study, PKa-mediated FIX activation was confirmed
in plasma ex vivo.[76] The clinical consequence of red blood cell–triggered coagulation may explain the
procoagulant state in patients after blood transfusions.
Biogenesis of EV is accompanied by the translocation of the phospholipid phosphatidylserine
(PS) from the inner to the outer membrane leaflet.[77] In addition, exposure of PS on the cell surface is a hallmark of programmed cell
death. Apoptotic cells are procoagulant by presenting PS as an attachment site for
FXII.[78] A role of phospholipids in FXII-triggered contact activation is also known from
circulating triglyceride-rich particles. Thromboelastography and thrombin generation
assays revealed that very low density lipoproteins mediate FXII contact activation
in a phosphatidylethanolamine- but not PS-dependent manner.[79] In conclusion, further studies are required to analyze, whether phospholipids, especially
PS, have a supportive role in FXII-driven coagulation or whether they are capable
to directly induce FXII contact activation.
Exogenous Activators
The white clay material kaolin mainly contains kaolinite, a hydrated aluminum silicate.
Similar to Ca2+-polyP, Ca2+-kaolin provides a negatively charged surface for FXII binding and subsequent contact
activation.[80] Due to these procoagulant activities, hemostatic wound dressings (QuikClot Combat
Gauze) are coated with kaolin to rapidly control trauma-induced bleeding. However,
long-term exposure of the injured tissue to these hemostatic dressings increases the
thromboembolic risk, as kaolin can enter the circulation and trigger clot formation.[81] Wounds rarely remain clean and are more often contaminated with soil, especially
in terrestrial vertebrates. Soil is predominantly composed of silicate minerals and
induces activation of the intrinsic coagulation cascade in a FXII-dependent manner.[82] Consistent with previous studies showing that birds and some marine mammals are
deficient in FXII,[83]
[84] soil-stimulated clotting was defective in dolphin or chicken blood.[82] In a murine bleeding model, treatment of an injured vessel with a polyacrylic acid–coated
filter paper reduced underlying blood loss in a FXII-dependent manner, indicating
that the polymer functions as a contact activator.[85]
In addition to kaolin and other silica-rich materials, glass, sulfatides, and ellagic
acid (EA) also represent artificial surfaces that initiate FXII contact activation.[86]
[87] EA is a polyphenolic compound associated with antiproliferative and antioxidant
effects. Similarly with kaolin and polyP, EA forms complexes with divalent metal ions[88] and only these EA particles, but not soluble free EA molecules, are capable to induce
FXII contact activation.[88] Administration of an infusion containing EA caused thrombosis in patients by direct
activation of FXII or platelet stimulation.[89]
A well-known FXII contact activator is the sulfated polysaccharide DXS. High-molecular-weight
DXS (500 kDa) stimulates FXII contact activation, whereas short-chain DXS (5 kDa)
is inactive for zymogen activation.[90] Challenging pigs with DXS led to a BK-mediated drop in blood pressure[91] but did not induce thrombosis. This suggests that DXS specifically triggers the
kallikrein–kinin system leading to inflammation but is inactive for driving coagulation.
Therefore, injection of the polymer is useful to analyze BK-mediated edema independently
on coagulation in mouse models.[92] The synthetic polysaccharide oversulfated chondroitin sulfate (OSCS) acts similarly
with high-molecular-weight DXS. OSCS is a potent FXII contact activator. Infusion
of heparin contaminated with OSCS led to hypotonic shock[93] with fatal outcome due to BK-mediated hypersensitivity reactions.[94]
The structurally related glycosaminoglycan fucosylated chondroitin sulfate (FCS) also
activates the contact system in human plasma. Based on carbohydrate–protein interaction
studies, PK binds readily to FCS in the presence of the other contact system proteins,
leading to the hypothesis that FXIIa is generated mainly by PKa-mediated fluid-phase
activation.[95]
All of the described endogenous and exogenous FXII contact activators have a net negative
surface charge, which appears to be necessary for binding/activation of FXII. Vice
versa, the positively charged polyethylenimine and the inert polymer Teflon AF failed
to induce FXII contact activation human plasma.[96]
Relevance of FXII for Diagnostic Coagulation Tests
Relevance of FXII for Diagnostic Coagulation Tests
FXII contact activation is the mechanistic basis for one of the most commonly performed
diagnostic coagulation screening tests, the activated partial thromboplastin time
(aPTT). aPTT assays are performed more than 5 billion times annually and are used
for (1) preoperative screening; (2) monitoring of intravenous application of heparin;
(3) detection of lupus anticoagulants/antiphospholipid syndrome; and (4) identification
of deficiency in the contact system proteins including FXII, PK, and HK and the coagulation
factors FXI, FIX, and FVIII. Kaolin, micronized silica, or EA is commonly used to
trigger FXII-dependent clotting in aPTT assays. Initially, citrated plasma is incubated
with artificial surfaces for up to 90 seconds. Within that time a substantial amount
of FXII zymogen undergoes conversion to FXIIa that subsequently activates factor XI
(FXI) to FXIa. Upon recalcification, the coagulation cascade proceeds and eventually
forms fibrin. Accordingly, aPTT represents the time that elapses from recalcification
to the formation of a fibrin clot. FXII-deficient plasma has a prolonged/abnormal
aPTT, as clotting is impaired in vitro. Notably, extending the incubation time with
artificial surfaces normalizes aPTT in PK-deficient plasma due to surface-mediated
FXII contact activation that compensates for the absence of PKa-mediated fluid-phase
FXII activation.[97] aPTT assays could be discussed as a tool for diagnosing C1INH-HAE, as in one study,
approximately 73% of HAE patients with C1INH deficiency had a shortened aPTT compared
with HAE patients with functional C1INH.[98]
On the basis of abnormal prolonged aPTT results and sequencing analysis, several mutations
in the F12 gene have been identified that were not associated with a history of bleeding events
or pathological hemostatic capacity ([Table 3]
[Fig. 2]). This may be one reason why congenital FXII deficiency, as an autosomal recessive
inherited disorder, tends to be diagnosed incidentally during routine coagulation
testing, e.g., prior to surgery. Most of the reported F12 mutations result in a quantitative defect (cross-reacting material [CRM]-negative
FXII deficiency) represented by a markedly reduced FXII activity (FXII:C) and antigen
levels (FXII:Ag). However, a few genetic variations within the F12 gene are causative for a qualitative defect (CRM-positive FXII deficiency) indicated
by a decreased FXII:C but nearly normal FXII:Ag.
Table 3
F12 gene mutations associated with FXII deficiency (update of mutations reviewed by Naudin
et al[111])
Case no.
|
Exon/Intron
|
Genetic variation
|
Comments
|
Protein
|
Condition
|
c.-4 genotype
|
FXII:C
|
FXII:Ag
|
Reference
|
1
|
Exon 1
|
c.-13C > T
|
5′ UTR
|
–
|
Heterozygous FXII deficiency
|
C/T
|
22%
|
–
|
[112]
|
2
|
Exon 1
|
c.-13C > T
|
5′ UTR
|
–
|
Compound heterozygous FXII deficiency; CRM-negative
|
–
|
< 1%
|
< 1%
|
[113]
|
c.-8C > G
|
5′ UTR
|
–
|
3
|
Exon 1
|
c.-8C > G
|
5′ UTR
|
–
|
Homozygous FXII deficiency
|
T/T
|
< 1%
|
–
|
[112]
|
4
|
Exon 1
|
c.-8C > G
|
5′ UTR
|
–
|
Homozygous FXII deficiency; CRM negative
|
–
|
< 1%
|
1%
|
[114]
[115]
|
Intron B
|
c.116–224C > T
|
TaqI restriction site
|
–
|
5
|
Exon 1
|
c.-8C > G
|
5′ UTR
|
–
|
Compound heterozygous FXII deficiency; CRM-negative
|
–
|
< 1%
|
< 1%
|
[113]
|
Exon 13
|
c.1558T > C
|
Missense mutation
|
p.Gln520* (Gln501*)
|
6
|
Exon 1
|
c.-8C > G
|
5′ UTR
|
–
|
Compound heterozygous FXII deficiency; CRM-negative
|
–
|
< 1%
|
< 1%
|
[113]
|
Exon 14
|
c.1697C > T
|
Missense mutation
|
p.Pro566Leu (Pro547Leu)
|
7
|
Exon 3
|
c.158A > G
|
Missense mutation
|
p.Tyr53Cys (Tyr34Cys)
|
FXII Tenri, homozygous FXII deficiency; CRM-negative
|
–
|
3%
|
3%
|
[116]
|
8
|
Exon 4
|
c.218G > C
|
Missense mutation
|
p.Cys73Ser (Cys54Ser)
|
Heterozygous FXII deficiency
|
T/T
|
1%
|
–
|
[117]
[118]
|
9
|
Exon 4
|
c.249delG
|
Frameshift mutation
|
p.Gln83Hisfs*12 (Gln64Hisfs*12)
|
Compound heterozygous FXII deficiency
|
C/T
|
< 0.5%
|
–
|
[119]
|
Exon 6
|
c.405C > A
|
Nonsense mutation
|
p.Cys135* (Cys116*)
|
10
|
Exon 4
|
c.251G > C
|
Missense mutation
|
p.Arg84Pro (Arg65Pro)
|
Homozygous FXII deficiency
|
T/T
|
< 10%
|
<17%
|
[120]
|
11
|
Exon 5
|
c.303_304delCA
|
Frameshift mutation
|
p.His101Glufs*36 (His82Glufs*36)
|
Heterozygous carrier
|
C/T
|
39%
|
41.6%
|
[121]
|
12
|
Exon 5
|
c.303_304delCA
|
Frameshift mutation
|
p.His101Glufs*36 (His82Glufs*36)
|
Homozygous FXII deficiency
|
T/T
|
0%
|
1.2%
|
[121]
|
13
|
Exon 6
|
c.405C > A
|
Nonsense mutation
|
p.Cys135* (Cys116*)
|
Homozygous FXII deficiency
|
T/T
|
<0.5%
|
–
|
[119]
[122]
|
14
|
Exon 6
|
c.425G > C
|
Missense mutation
|
p.Arg142Pro (Arg123Pro)
|
Homozygous FXII deficiency; CRM-negative
|
T/T
|
<1%
|
<10%
|
[123]
|
15
|
Exon 7
|
c.583delC
|
Frameshift mutation
|
p.His195Thr*55 (His176Thr*55)
|
Heterozygous carrier
|
C/T
|
53.4%
|
–
|
[118]
|
16
|
Exon 7
|
c.583delC
|
Frameshift mutation
|
p.His195Thr*55 (His176Thr*55)
|
Compound heterozygous FXII deficiency
|
T/T
|
<1%
|
–
|
[118]
|
Exon 10
|
c.1092_1093insC
|
Frameshift mutation
|
p.Lys365fs*68 (Lys346fs*68)
|
17
|
Exon 8
|
c.721T > G
|
Missense mutation
|
p.Trp241Gly (Trp222Gly)
|
Heterozygous carrier
|
–
|
40%
|
55%
|
[124]
|
18
|
Exon 8
|
c.721T > G
|
Missense mutation
|
p.Trp241Gly (Trp222Gly)
|
Compound heterozygous FXII deficiency; CRM-negative
|
–
|
<3%
|
<10%
|
[124]
|
Exon 12
|
c.1396C > A
|
Missense mutation
|
p.Arg466Ser (Arg447Ser)
|
19
|
Exon 8
|
c.776G > A
|
Missense mutation
|
p. Gly259Glu (Gly240Glu)
|
Homozygous FXII deficiency; CRM-negative
|
C/C
|
0.6%
|
<1%
|
[125]
|
20
|
Exon 8
|
c.799C > G
|
Missense mutation
|
p.Arg267Gly (Arg248Gly)
|
Homozygous FXII deficiency; CRM-negative
|
C/C
|
1.5%
|
<1%
|
[125]
|
21
|
Exon 9
|
c.809_811delACA
|
Deletion mutation
|
p.Asn271del (Asn252del)
|
Compound heterozygous FXII deficiency
|
–
|
2.0%
|
5.2%
|
[126]
|
Exon 10
|
c.1078G > A
|
Missense mutation
|
p.Gly360Arg (Gly341Arg)
|
22
|
Exon 9
|
c.856_864delAGCTGGGAG
|
Frameshift mutation
|
p.Ser286_Glu288 (Ser267_Glu269)
|
Heterozygous FXII deficiency; CRM-negative
|
T/T
|
<1%
|
<1%
|
[127]
|
23
|
Exon 10
|
c.1027G > C
|
Missense mutation
|
p.Ala343Pro (Ala324Pro)
|
Heterozygous carrier
|
T/T
|
29%
|
30%
|
[128]
|
24
|
Exon 10
|
c.1027G > C
|
Missense mutation
|
p.Ala343Pro (Ala324Pro)
|
Compound heterozygous FXII deficiency; CRM-negative (Ala324Pro) and CRM-positive (Gly531Glu)
|
T/C
|
35%
|
81%
|
[128]
|
Exon 13
|
c.1649G > A
|
Missense mutation
|
p.Gly550Glu (Gly531Glu)
|
25
|
Exon 10
|
c.1078G > A
|
Missense mutation
|
p.Gly360Arg (Gly341Arg)
|
Homozygous FXII deficiency
|
T/T
|
12%
|
10%
|
[129]
|
26
|
Exon 10
|
c.1078G > A
|
Missense mutation
|
p.Gly360Arg (Gly341Arg)
|
Heterozygous carrier
|
C/T
|
35%
|
38%
|
[129]
|
27
|
Exon 10
|
c.1078G > A
|
Missense mutation
|
p.Gly360Arg (Gly341Arg)
|
Compound heterozygous FXII deficiency
|
–
|
2.0%
|
1.0%
|
[126]
|
Exon 13
|
c.1561G > A
|
Missense mutation
|
p.Glu521Lys (Glu502Lys)
|
28
|
Exon 10
|
c.1093_1094insC
|
Frameshift mutation
|
p.Lys365fs*68 (Lys346fs*68)
|
Heterozygous carrier
|
T/T
|
17.1%
|
–
|
[118]
|
29
|
Exon 10
|
c.1092_1093insC
|
Frameshift mutation
|
p.Lys365Glnfs*69 (Lys346Glnfs*69)
|
Compound heterozygous FXII deficiency
|
C/T
|
<0.5%
|
–
|
[119]
[122]
|
Exon 14
|
c.1744G > A
|
Missense mutation
|
p.Gly582Ser (Gly563Ser)
|
30
|
Exon 10
|
c.1095G > C
|
Missense mutation
|
p.Lys365Asn (Lys346Asn)
|
FXII Ofunato, CRM-reduced FXII deficiency
|
T/T
|
5%
|
4.5%
|
[130]
|
31
|
Exon 10
|
c.1115G > C
|
Missense mutation
|
p.Arg372Pro (Arg353Pro)
|
FXII Locarno, CRM-positive FXII deficiency
|
–
|
<1%
|
46%
|
[131]
[132]
|
32
|
Exon 10
|
c.1231G > A
|
Missense mutation
|
p.Ala411Thr (Ala392Thr)
|
Factor XII Shizuoka, homozygous FXII deficiency; CRM-negative
|
T/T
|
<3%
|
<10%
|
[133]
|
33
|
Exon 10
|
c.1240C > A
|
Missense mutation
|
p.Leu414Met (Leu395Met)
|
Heterozygous carrier
|
–
|
<28%
|
34%
|
[115]
|
34
|
Exon 10
|
c.1240C > A
|
Missense mutation
|
p.Leu414Met (Leu395Met)
|
Compound heterozygous FXII deficiency; CRM-negative
|
–
|
<5%
|
5%
|
[115]
|
Intron M
|
c.1681–1G > A
|
Acceptor splice site mutation
|
–
|
35
|
Exon 11
|
c.1318C > A
|
Missense mutation
|
p.Gln440Lys (Gln421Lys)
|
Heterozygous carrier; CRM-negative
|
C/T
|
23%
|
28%
|
[123]
|
36
|
Exon 12
|
c.1515G > C
|
Missense mutation
|
p.Trp505Cys (Trp486Cys)
|
FXII Mie-1, homozygous FXII deficiency
|
T/T
|
<5%
|
<5%
|
[134]
|
37
|
Exon 13
|
c.1561G > A
|
Missense mutation
|
p.Glu521Lys (Glu502Lys)
|
Homozygous FXII deficiency; CRM-negative
|
C/C
|
4%
|
3.8%
|
[125]
|
38
|
Exon 13
|
c.1561G > A
|
Missense mutation
|
p.Glu521Lys (Glu502Lys)
|
Heterozygous carrier
|
C/C
|
34%
|
33%
|
[135]
|
39
|
Exon 13
|
c.1561G > A
|
Missense Mutation
|
p.Glu521Lys (Glu502Lys)
|
Compound heterozygous FXII deficiency; CRM-negative
|
C/T
|
4%
|
5%
|
[135]
|
c.1637T > C
|
Missense mutation
|
p.Met546Thr (Met527Thr)
|
40
|
Exon 13
|
c.1583C > T
|
Missense mutation
|
p.Ser528Phe (Ser509Phe)
|
Heterozygous carrier
|
T/T
|
29%
|
–
|
[122]
|
41
|
Exon 13
|
c.1583C > T
|
Missense mutation
|
p.Ser528Phe (Ser509Phe)
|
Compound heterozygous FXII deficiency
|
T/T
|
5%
|
–
|
[119]
[122]
|
Exon 14
|
c.1744G > A
|
Missense mutation
|
p.Gly582Ser (Gly563Ser)
|
42
|
Exon 13
|
c.1637T > C
|
Missense mutation
|
p.Met546Thr (Met527Thr)
|
Heterozygous carrier
|
C/T
|
30%
|
32%
|
[135]
|
43
|
Exon 13
|
c.1669G > A
|
Missense mutation
|
p.Asp557Asn (Asp538Asn)
|
Heterozygous FXII deficiency
|
–
|
5%
|
6.8%
|
[136]
|
44
|
Intron M
|
c.1681–1G > A
|
Acceptor splice site mutation
|
–
|
Heterozygous carrier
|
–
|
<26%
|
<40%
|
[115]
|
45
|
Intron M
|
c.1681–1G > A
|
Acceptor splice site mutation
|
–
|
Homozygous FXII deficiency; CRM-negative
|
–
|
<1%
|
<1%
|
[115]
[137]
|
46
|
Exon 14
|
c.1681G > A
|
Missense mutation
|
p.Gly561Ser (Gly542Ser)
|
Homozygous FXII deficiency
|
–
|
<1%
|
<1%
|
[126]
|
47
|
Exon 14
|
c.1765G > C
|
Missense mutation
|
p.Gly589Arg (Gly570Arg)
|
Heterozygous carrier; CRM-positive
|
–
|
10%
|
74%
|
[115]
|
48
|
Exon 14
|
c.1768T > A
|
Missense mutation
|
p.Cys590Ser (Cys571Ser)
|
FXII Washington, CRM-positive FXII deficiency
|
–
|
<1%
|
80%
|
[138]
|
49
|
Exon 14
|
c.1744G > A
|
Missense mutation
|
p.Gly582Ser (Gly563Ser)
|
Heterozygous carrier
|
T/T
|
25%
|
–
|
[122]
|
Abbreviations: CRM, cross-reacting material; FXII:C, factor XII activity; FXII:Ag,
factor XII antigen levels.
Notes: Letter prefixes indicate the type of reference sequence: “c.”—coding DNA reference
sequence (NM_000505.4), “p.”—protein reference sequence (NP_000496.2). The protein
reference sequence contains an N-terminal signal peptide with a length of 19 amino
acids. Changes in the amino acid sequence of the mature FXII protein are described
in (parentheses). The c.-4 genotype indicates a C/T polymorphism localized in the
5′ untranslated region (5′ UTR) at nucleotide position four upstream of the translation
initiation codon ATG. Nucleotide as well as amino acid changes are depicted as follows:
“ > ”—substitution, “del”—deletion, “ins”—insertion, “fs”—frameshift, “*”—translation
termination codon.
Fig. 2 Genetic variations in the F12 gene associated with FXII deficiency. The genomic DNA encoding for human FXII comprises
14 exons (E1–E14) and 13 introns. E10, E13, and E14 are mainly affected by F12 mutations (indicated by orange lines) leading to reduced FXII activity and antigen
levels. The genetic variations c.116–224C > T and c.1681–1G > A are localized in introns
that are not represented by their actual size. Starting with the N-terminus, FXII
is characterized by a signal peptide (S), a fibronectin type II (Fib-II) domain, followed
by a first epidermal growth factor-like (EGF-I) domain, a fibronectin type I (Fib-I)
domain, a second EGF-like (EGF-II) domain, a kringle domain, a proline-rich (PR) region,
and a C-terminal catalytic domain with an overall length of 615 amino acids. Compared
with the other domains, amino acid changes are abundant in the catalytic domain of
FXII. Reference sequences: transcript—NM_000505.4, protein—NP_000496.2.
In addition to the F12 mutations described in [Table 3], a polymorphism in the 5′-untranslated region of F12 at nucleotide position four upstream of the translation initiation codon ATG (c.-4C > T,
referred to as 46 C/T)[99] has additional implications for FXII:C and FXII:Ag. Transcriptional and translational
analyses showed that this 46 T allele is associated with decreased translational efficiency
compared with 46 C, resulting in low FXII:C and FXII:Ag.[99]
Reduced FXII levels may also be associated with the transfusion of blood cell or platelet
concentrates.[100] Platelet activation and subsequent microparticle release occurs upon extended storage
times,[101] which then drives FXII activation by exposure to PS and platelet membrane-bound
polyP. Thus, reduced FXII levels appear as a consequence of zymogen “consumption”
in patients receiving platelet concentrates.
Conclusion
Because of its selective role in thrombosis-sparing hemostasis, FXII has gained considerable
interest as a target for safe anticoagulation over the last decades. In addition,
FXII is a promising target for interference with proinflammatory BK-mediated responses.
Contact activation of FXII is still incompletely understood, and identification of
the “FXII contact activation site” will represent a major step forward. Detailed insight
into the mechanism of FXII activation will help develop specific FXII contact inhibitors
for potential clinical and diagnostic applications.