Decreased in vitro thrombin generation and clot stability in
human FXII-null blood and plasma
Human clotting factor XII (FXII) deficiency is not associated
with bleeding problems, but the assumption that FXII is
irrelevant to physiological clotting has been challenged by
recent studies with FXII-null mice (Renne et al, 2005;
Kleinschnitz et al, 2006; Stoll et al, 2008). Thrombin genera-
tion, clot formation and clot stability have been examined
in vitro in human blood and plasma where FXII activity has
been artificially depleted or inhibited (Luddington & Baglin,
2004; Nielsen et al, 2005; Nielsen, 2009), revealing some
technical limitations to simulating FXII deficiency. We have
extended these studies using samples from a patient lacking
constitutive FXII expression, representing a genuine human
The patient was a healthy 18-year-old male with no prior
bleeding history, evaluated for an incidental finding of elevated
activated partial thromboplastin time (APTT) >180 s. Labo-
ratory testing reported undetectable plasma FXII (<0Æ01),
confirmed by immunoblot analysis. Normal range values were
reported for fibrinogen and other plasma proteins, coagulation
and Von Willebrand factor (VWF) levels [iu/ml: FVIII 1Æ52,
FIX 1Æ16, FXI 1Æ39, VWF antigen (VWF:Ag) 0Æ73, ristocetin
cofactor (VWF:RCo) 1Æ03; blood group O positive], complete
blood count, bleeding time and International Normalized
Ratio. Anticoagulated (sodium citrate 3Æ2%) blood was
collected by venepuncture from the FXII-null patient and
from normal control subjects with or without the addition of
corn trypsin inhibitor (CTI; Haematologic Technologies Inc.,
Essex Junction, VT, USA) at 20 lg/ml. Plasma was prepared by
Thrombin generation was measured via calibrated auto-
mated thrombography (CAT) assays performed with a
Thrombinoscope system (Synapse BV, Maastricht, Nether-
Consistent with previous studies (Luddington & Baglin,
2004), assays of normal plasma with and without CTI (Fig 1)
using low concentration tissue factor (TF; 1 pmol/l) activation
(n = 5) showed significantly decreased (P > 0Æ01) values for
mean (±standard deviation) peak thrombin (165 ± 68 nmol/l.
vs. 267 ± 65) and longer time to peak (7Æ0 ± 1Æ0 vs.
9Æ0 ± 1Æ8 min) for CTI-treated samples. Compared to normal
plasma with CTI (Fig 1), CAT assays of FXII-null patient
plasma samples (n = 3) showed significantly lower (P > 0Æ01)
peak thrombin(69Æ5 ± 9Æ1 nmol/l)
(11Æ7 ± 0Æ7 min). CAT assays of FXII immunodepleted normal
plasma gave similar results to CTI-treated normal samples.
Addition of purified FXII at a physiological concentration of
50 lg/ml to FXII-null plasma increased thrombin generation
to a level equivalent to CTI-treated normal plasma (Fig 1).
Assays using high TF (5 pmol/l) activation showed no signif-
icant differences among all sample groups, presumably due to
rapid induction of FVIIa-mediated thrombin generation.
Clotting was measured via thromboelastography (TEG)
assays performed with Haemoscope 5000 analysers (Haemo-
netics Corp., Braintree, MA, USA). As expected from clinical
laboratory results and previous studies (Nielsen et al, 2005),
unactivated FXII-null blood samples did not clot within
120 min. Clotting was observed with kaolin contact-pathway
activation but, compared to the mean for normal samples
(n = 17), FXII-null blood had a much lower maximum rate of
thrombus generation (MRTG; 3Æ7 mm/min vs. 10Æ7 ± 1Æ8
normal mean) and considerably extended time until clotting
onset (43Æ2 min vs. 6Æ0 ± 1Æ0). Clotting kinetics were also
decreased in kaolin-activated CTI-treated normal blood, but
not nearly to the extent observed in FXII-null samples.
A similar pattern of clotting kinetics was observed in TEG
assays using activation with low concentration TF (140 fmol/l),
but at 1Æ4 pmol/l TF no differences were observed between
FXII-null and normal blood.
Clot stability was tested via TEG fibrinolysis assays where
tissue-type plasminogen activator (tPA; Cathflo Activase?,
alteplase, Genentech) was added to samples at a final
concentration of 0Æ56 ng/ml. CTI has been shown to directly
inhibit tPA activity in TEG fibrinolytic assays (Nielsen, 2009),
thus CTI-treated normal samples cannot be reliably tested for
Fig 1. Thrombin generation in normal donor plasma without (Nor-
mal) or with corn trypsin inhibitor (Normal + CTI), and FXII-null
plasma without (FXII-null) or with added FXII (FXII-null + FXII).
The samples were activated with 1 pM TF.
ª 2010 Blackwell Publishing Ltd, British Journal of Haematology, 152, 108–121
clot stability in vitro. TF activation (‡1Æ4 pmol/l) was used to
produce consistent initial clot formation kinetics in normal
and FXII-null samples, and assays were run until the amplitude
returned to the baseline value of 2 mm to yield a value for the
clot lysis time (CLT, min). We observed that clot stability was
notably decreased in FXII-null blood compared to normal
(Fig 2), as indicated by a CLT of 47 min versus the normal
mean (n = 5) of 83 ± 22Æ6 min. Similar assays were performed
using plasma samples, which also showed significantly lower
(P > 0Æ01) mean CLT for FXII-null samples (63Æ2 ± 11Æ2 min;
n = 8) compared to normal plasma samples (92Æ0 ± 17Æ6;
n = 7). Results for FXII-depleted normal plasma were within
the normal range.
In summary, we observed that using low levels of TF
activation in vitro plasma thrombin generation was depressed
in FXII-null samples compared to both normal samples and
those treated to inactivate FXII (with CTI) or deplete it. FXII-
null blood also showed depressed clot formation kinetics at
low levels of TF. When sufficient TF was used to produce
uniform clotting kinetics in blood and plasma, FXII-null clots
were observed to be less stable as assessed by tPA-mediated
fibrinolysis than those formed in normal and FXII-depleted
samples (CTI-treated samples cannot be reliably tested for clot
Clot structure is influenced by thrombin generation (Weisel,
2007; Wolberg & Campbell, 2008) and TF activity in whole
blood has been estimated to be 20 fmol/l or less (Butenas et al,
2008). Within this physiological range our results indicate that
FXII may positively influence the kinetics of clot formation.
Our results are also in line with studies indicating that lack of
FXII activity in FXII-null mice (Renne et al, 2005; Kleinschnitz
et al, 2006; Stoll et al, 2008) protects mice from arterial and
venous thrombosis without affecting bleeding. In humans,
decreased levels of FXII have been observed to have a negative
effect on survival, but not severe deficiency (Endler et al,
2007). Given the evidence for a physiological role for FXII in
clotting and the differences we have observed between assays of
genuine human FXII-null samples and those simulating FXII
deficiency, it would be worthwhile to extend this analysis to
other cases of constitutive lack of FXII expression.
Fred G. Pluthero1,2
Leonardo R. Branda ˜o1
Walter H.A. Kahr1,2
1Department of Paediatrics, Division of Haematology/Oncology, The
Hospital for Sick Children, and University of Toronto,2Program in Cell
Biology, Research Institute of The Hospital for Sick Children, Toronto,
ON, Canada, and3Department of Haematology, Mercy University
Hospital, Cork, Ireland.
Butenas, S., Orfeo, T. & Mann, K.G. (2008) Tissue factor activity and
function in blood coagulation. Thrombosis Research, 122(Suppl 1),
Endler, G., Marsik, C., Jilma, B., Schickbauer, T., Quehenberger, P. &
Mannhalter, C. (2007) Evidence of a U-shaped association between
factor XII activity and overall survival. Journal of Thrombosis and
Haemostasis, 5, 1143–1148.
Kleinschnitz, C., Stoll, G., Bendszus, M., Schuh, K., Pauer, H.U.,
Burfeind, P., Renne, C., Gailani, D., Nieswandt, B. & Renne, T.
(2006) Targeting coagulation factor XII provides protection from
pathological thrombosis in cerebral ischemia without interfering
with hemostasis. Journal of Experimental Medicine, 203, 513–518.
Luddington, R. & Baglin, T. (2004) Clinical measurement of thrombin
generation by calibrated automated thrombography requires
contact factor inhibition. Journal of Thrombosis and Haemostasis, 2,
Nielsen, V.G. (2009) Corn trypsin inhibitor decreases tissue-type
plasminogen activator-mediated fibrinolysis of human plasma.
Blood Coagulation and Fibrinolysis, 20, 191–196.
Nielsen, V.G., Cohen, B.M. & Cohen, E. (2005) Effects of coagulation
factor deficiency on plasma coagulation kinetics determined via
thrombelastography: critical roles of fibrinogen and factors II, VII, X
and XII. Acta Anaesthesiologica Scandinavica, 49, 222–231.
Renne, T., Pozgajova, M., Gruner, S., Schuh, K., Pauer, H.U., Burfeind,
P., Gailani, D. & Nieswandt, B. (2005) Defective thrombus forma-
tion in mice lacking coagulation factor XII. Journal of Experimental
Medicine, 202, 271–281.
Stoll, G., Kleinschnitz, C. & Nieswandt, B. (2008) Molecular mecha-
nisms of thrombus formation in ischemic stroke: novel insights and
targets for treatment. Blood, 112, 3555–3562.
Weisel, J.W. (2007) Structure of fibrin: impact on clot stability. Journal
of Thrombosis and Haemostasis, 5(Suppl 1), 116–124.
Wolberg, A.S. & Campbell, R.A. (2008) Thrombin generation, fibrin
clot formation and hemostasis. Transfusion and Apheresis Science,
Keywords: blood coagulation, coagulation factors, fibrinolysis,
First published online 29 September 2010
Fig 2. TEG fibrinolysis assays of FXII-null and representative normal
donor whole blood. Clotting was activated with TF and fibrinolysis
stimulated with tPA. Clot breakdown is demonstrated by the conver-
gence of the amplitude curve back to baseline as indicated by the clot
lysis time (CLT).
ª 2010 Blackwell Publishing Ltd, British Journal of Haematology, 152, 108–121