Báo cáo khoa học: Binding of activated Factor XII to endothelial cells affects its inactivation by the C1-esterase inhibitor pptx

Binding of activated Factor XII to endothelial cells affects
its inactivation by the C1-esterase inhibitor
Inger Schousboe
Department of Medical Biochemistry & Genetics, The Panum Institute, University of Copenhagen, Denmark
It is well known that activated Factor XII (FXIIa) and
kallikrein are rapidly inactivated in plasma as a result of
reaction with endogenous inhibitors. The purpose of this
may be to prevent uncontrolled deleterious spreading and
activation of target zymogens. Both FXII and the complex
plasma prekallikrein/high molecular mass kininogen
become activated when they bind, in a Zn
2+
-dependent
manner, to receptors on human umbilical vein endothelial
cells (HUVEC). The C1-esterase inhibitor (C1-INH) is by
far the most efficient inhibitor of FXIIa. In the present study
it has been investigated whether binding of FXIIa to
HUVEC might offer protection against inactivation by
C1-INH. It appeared that the relative amidolytic activity of
purified FXIIa bound to the surface of HUVEC decreased
according to the concentration of C1-INH in medium;
however, the decrease was smaller than that measured for
inactivation of FXIIa in solution. The secondary rate
constant for the inactivation was 3–10-fold lower for cell-
bound than for soluble FXIIa. The inactivation was found
to be caused by C1-INH binding to cell-bound FXIIa.
Accordingly, the amidolytic activity of saturated amounts of
cell-bound FXIIa was reduced in the presence of C1-INH
and was theoretically nonexistent at physiological C1-INH
concentrations. Amidolytic activity was, however, present
on HUVEC incubated with plasma indicating that the

endogenous C1-INH did not completely abolish the activity
of FXIIa generated during the incubation period. This
supports the hypothesis that binding to endothelial cells
protects the activated FXII against inactivation by its major
endogenous inhibitor. Hence, the function of FXII may be
localized at cellular surfaces.
Keywords: C1-esterase inhibitor; endothelial cells; Factor
XII.
Factor XII (FXII) and plasma prekallikrein (PPK) are
zymogens of proteolytically active factor XIIa (FXIIa) and
kallikrein, which constitute a linked activation system.
Contact with an artificial, negatively charged surface
activates the system in the presence of Zn
2+
and high
molecular mass kininogen (HK) [1]. Hence, this system was
named the contact activation system. Initiation of activation
of the system has been found to be a result of a slow
autodigestion and autoactivation of surface bound FXII by
FXIIa [2]. Binding of FXII to a negatively charged surface
induces a Zn
2+
-dependent conformational change in FXII
enhancing its susceptibility to become activated [3–5]. The
activated FXII then activates prekallikrein, which recipro-
cally enhances the activation of FXII and thereby the entire
system [6]. In blood plasma the proteolytic efficiency of the
proteinases is attenuated by endogenous inactivation of
FXIIa and kallikrein by serine proteinase inactivators. The
C1-esterase inhibitor (C1-INH), belonging to the super-

family of serine proteinase inhibitors (serpins), binds
covalently to the active site of its target [7]. It is the major
FXIIa inactivator [8] and accounts for > 90% of the
inactivation of FXIIa in plasma [9,10]. Its plasma concen-
tration is  2.5 l
M
[11]. Homozygous deficiency of C1-INH
is not known. Patients with ineffective C1-INH caused by
either a heterozygous deficiency, dysfunctional protein or
the presence of autoantibodies suffer episodes of painless
local swelling of the soft tissue resulting from increased
vasopermeability correlated with an increased plasma
concentration of activated FXII and cleavage products of
HK during attacks [12–14].
During the past decade convincing results have shown
that FXII and PPK bind to receptors on endothelial cells in
aZn
2+
-dependent manner. FXII binds directly to the cells
[15,16], whereas PPK binds via a bridge generated by the
ability of HK to bind to the cells as well as to PPK [17,18].
The binding results in activation of FXII and PPK by a so
far controversial mechanism [15,19–21]. Whatever this
mechanism might be, a prerequisite for function is that
the cell-bound and -generated FXIIa and kallikrein are
protected from inactivation by endogenous inhibitors in the
plasma. Activation of PPK plays a primary role in initiating
and maintaining the vascular response to injury. By
interacting with HK, kallikrein cleaves off the vasoactive
peptide, bradykinin [21–23]. Bradykinin stimulates endo-

thelial cell prostaglandin synthesis, leading to inactivation of
platelet function [24,25], and increased superoxide [26] and
nitric oxide formation [27]. Consequently bradykinin
is linked with the pathophysiological processes that
Correspondence to I. Schousboe, Department of Medical Biochemistry
& Genetics, The Panum Institute, University of Copenhagen,
Blegdamsvej 3C, DK-2200 Copenhagen N, Denmark.
Fax: + 45 35367980, Tel.: + 45 35327800,
E-mail:
Abbreviations: FXII, Factor XII; FXIIa, activated FXII; PPK,
plasma prekallikrein; HK, high molecular mass kininogen; C1-INH,
C1-esterase inhibitor; HUVEC, human umbilical vein endothelial
cells; CTI, corn trypsin inhibitor; SBTI, soybean trypsin inhibitor;
S-2302, H-
D
-Pro-Phe-Arg-para-nitroaniline.
(Received 3 September 2002, revised 1 November 2002,
accepted 14 November 2002)
Eur. J. Biochem. 270, 111–118 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03367.x
accompany inflammation and tissue damage which is
thought to play a major role in the symptomatology of
acute attacks in patients with angiodema [12,28].
Binding of FXIIa to different negatively charged artificial
surfaces has previously been shown to offer protection
against inactivation of FXIIa by C1-INH in a surface- and
C1-INH concentration-dependent manner [29]. Binding
and subsequent activation of FXII in the vascular system
may likewise offer protection of FXIIa activity against
inactivation by the C1-INH in plasma. FXII and FXIIa
bind indifferently to human umbilical vein endothelial cells

(HUVEC) [16]. The purpose of the present investigation
was to analyse whether binding of FXIIa to the surface of a
confluent layer of HUVEC might lead to protection against
inactivation by C1-INH.
Experimental procedures
Materials
FXII was from obtained Haematologic Technologies as a
high concentration solution in 50% glycerol and stored at
)20 °C as recommended by the manufacturer. It migrated
as a single band with an M
r
of 80 kDa on reduced SDS/
PAGE and was unable to cleave amidolytically the FXIIa
substrate, S-2302 (see below). Two-chain FXIIa was
obtained lyophilized from Enzyme Research Laboratory.
It was dissolved as described by the manufacturer and
stored in aliquots at )80 °C. Once thawed it was never
refrozen. None of the proteins, including C1-INH, corn
trypsin inhibitor (CTI; Calbiochem) or soybean trypsin
inhibitor (SBTI; Sigma) were preserved with Zn
2+
binding
substances. Plasma was isolated from normal blood anti-
coagulated with hirudin (50 mgÆL
)1
)byusingastandard
blood bank procedure. H-
D
-Pro-Phe-Arg-para-nitroaniline
(S-2302) was from Chromogenix, and Sigmacoat (chlorin-

ated organo-polysiloxane in heptane) and gelatine from
porcine skin (G-8150) were from Sigma. 1,3,4,6-Tetra-
chloro-3a,6a-diphenyl glycoluril (Iodogene) was from
Pierce, and Na
125
I was from Amersham/Pharmacia. All
other reagents were of the purest grade commercially
available. Siliconized test tubes were used for all solutions
and dilutions.
Endothelial cell culture
Cryopreserved primary cultures of HUVEC delivered from
Clonetics were plated at a density of 2.5 · 10
3
cellsÆcm
)2
in
tissue culture flasks containing endothelial cell growth
medium supplemented with human recombinant endothel-
ial cell growth factor, fibroblast growth factor and insulin-
like growth factor, vascular endothelial cell growth factor,
ascorbic acid, hydrocortisone, heparin, gentamicin, ampho-
tericin and 2% fetal bovine serum according to the
recommendation of Clonetics. The cells were incubated at
37 °C in a humidified 95% air : 5% CO
2
atmosphere. The
medium was changed the day after plating and subsequently
every second day. At confluence (2.8 · 10
4
cellsÆcm

)2
)the
cells were subcultured. After seven to eight generations (two
passages) the cells were cryopreserved. During each passage,
the number of living cells was tested by Trypan blue
exclusion. For experiments the cryopreserved cells were
subcultured in microtiter plates at a density of 10
4
cellsÆcm
)2
.
Medium was changed every second day and the cells were
used 4–6 days after plating.
Inactivation of amidolytic activity of cell-bound FXIIa
The medium was aspirated from 12-well microtiter plates
and the cells were washed twice over a period of 20 min with
Locke’s buffer (154 m
M
NaCl, 5.6 m
M
KCl, 3.6 m
M
NaHCO
3
,2.3m
M
CaCl
2
,1.0m
M

MgCl
2
,5.6m
M
glucose,
5m
M
Hepes pH 7.4) containing 15 l
M
ZnCl
2
(wash buffer)
followed by a 10-min incubation with 1 mL 0.1% (w/v)
gelatine in wash buffer (block buffer). The washing proce-
dure, as judged visually by microscopic inspection, neither
removed cells from the surface nor changed the morphology
of the cells. The block buffer was aspirated and the cells
preincubated at room temperature with 400 lL FXIIa
diluted at least 300-fold in block buffer giving a final
concentration of 40 n
M
. This preincubation period allowed
FXIIa to bind to the cells without being affected by the
inhibitor. After a 60 min preincubation period, 40 lLof
varying concentrations of C1-INH or 0.15
M
NaCl were
added and mixed by horizontal circular movements and the
incubation continued for a further 60 min. To measure the
degree of inactivation of FXIIa in the medium 250-lL

aliquots of the medium above the cells were transferred to
and mixed with 250 lL0.8m
M
S-2302 in EDTA buffer
(50 m
M
Tris/HCl, 12 m
M
NaCl, 10 m
M
EDTA, pH 7.8).
The remaining medium was aspirated and the cells washed
continuously for 5 s in a mild stream of wash buffer.
Subsequently, 400 lL0.4m
M
S-2302 in EDTA buffer were
added. The presence of EDTA destroys the cells and
releases FXIIa into solution. Both series of samples with
S-2302 were incubated for a maximum of 4 h at room temp
and the concentration of p-nitroaniline generated during the
chosen incubation period was measured at 405 nm in
cuvettes with a 1-cm light path after addition of acetic acid
to a final concentration of 4% (v/v).
Comparable experiments were run in parallel. As deter-
mined by microscopic inspection, the cells were intact until
addition of the chromogenic substrate. No degradation or
change in specific amidolytic activity of FXIIa was observed
within 6 h of incubation with the chromogenic substrate.
Amidolytic activity on cells incubated with plasma
After washing the cells in 12-well microtiter plates as

described above, the cells were incubated at room tempera-
ture with 400 lL hirudin anticoagulated plasma enriched
with 1/10 vol. 250 l
M
Zn
2+
in Locke’s buffer containing
either nothing, SBTI, CTI or both SBTI and CTI. After
incubation for 120 min the remaining medium was aspir-
ated and the cells washed continuously for 5 s in a mild
stream of wash buffer. Subsequently, the cells were incuba-
ted with S-2302 in EDTA buffer as described above. No
amidolytic activity could be measured when extracts of cells
were added to purified PPK and FXII.
The minimal concentration of CTI for total inhibition of
amidolytic activity of 40 n
M
FXIIa and the maximal
concentration of SBTI which could be applied without
affecting the activity of FXIIa at these concentrations were
determined by titration. The change in absorption over time
112 I. Schousboe (Eur. J. Biochem. 270) Ó FEBS 2003
was recorded continuously at 405 nm in mixtures of S-2302
(0.4 m
M
in EDTA buffer), containing FXIIa and varying
concentrations of CTI or SBTI. Complete inhibition
was observed at CTI concentrations > 0.005 mgÆmL
)1
( 0.5 l

M
) and no inhibition was recorded at SBTI
concentrations < 0.1 mgÆmL
)1
.
Radiolabelling of proteins
FXIIa was labelled with Na
125
I by the Iodogen method as
described by Salacinski et al. [30]. Following iodination,
125
I-labelled proteins were separated from
125
I
2
by gel-siev-
ing on a Sephadex G-25 column equilibrated with NaCl/Pi
(50 m
M
phosphate buffer, pH 7.4 containing 0.5
M
NaCl).
Specific labelling ranged from 10 · 10
9
to
19 · 10
9
c.p.m.Æmg protein
)1
.

Binding of
125
I-labelled proteins to HUVEC
and solubilization of the cells
The confluent layer of cells was incubated with
125
I-labelled
FXIIa and C1-INH added as described above. At the end of
the incubation period the cells were washed and solubilized
as described previously [16]. Amidolytic activity and radio-
activity were measured in 250-lL aliquots of the solubilized
cells. Activity was measured as the change in absorbance at
405 nm after a 4-h incubation with S-2302 (0.4 m
M
final
concentration).
Statistics
Student’s t-test was used to determine the level of signifi-
cance.
Results
Inactivation of the amidolytic activity of factor XIIa
bound to endothelial cells
To analyse if the presence of C1-INH would inactivate the
amidolytic activity of cell-bound FXIIa in a manner
comparable to that of inactivation of soluble FXIIa, the
following series of experiments was performed. HUVEC
were preincubed with purified FXIIa for 60 min, allowing
FXIIa to bind to the cells. Then C1-INH was added to the
incubation medium and the incubation continued for an
additional 60 min. This showed that the ability of C1-INH

to inactivate the amidolytic activity of FXIIa was markedly
reduced when FXIIa was bound to the cells compared to
the inactivation of FXIIa in the incubation medium. The
concentration of C1-INH which inactivated 50% of the
amount of FXIIa bound to the cells from a 40-n
M
solution
was approximately fivefold higher than the than that which
inactivated the activity in the 40 n
M
solution of FXIIa in
which the cells had been preincubated (Fig. 1). This
difference may be even larger if taking into account the
difference in amounts of FXIIa bound to the cells and that
in solution.
It was next analysed if amidolytic activity of FXIIa
generated on the cells from cell-bound FXII was likewise
protected from inactivation by C1-INH. As autoactivation
of HUVEC-bound FXII has been widely debated [15,19–
21] it was first analysed if FXII bound to HUVEC was able
to autoactivate. In the presence of 15 l
M
Zn
2+
, a concen-
tration previously shown to be optimal for binding of
purified FXII/FXIIa to HUVEC [16], a slightly but
significantly higher (P < 0.005) amidolytic activity could
be measured on the cells incubated in the presence than in
the absence of FXII (Table 1). During the period of

incubation with FXII, amidolytic activity had been gener-
ated also in the incubation medium. The total activity of this
was, however, lower than the total activity on the cells, and
in the presence of Zn
2+
fourfold higher than the activity of
FXIIa generated in the medium without cells (results not
shown). The difference in activity of FXIIa in the incubation
medium containing FXII may be due to equilibrium
binding of FXII/FXIIa to the cells. It is noticeable that
the presence of Zn
2+
alone increased significantly
(P < 0.005) the cell-generated S-2302 amidolytic activity.
This indicates that the cells themselves might have expressed
FXIIa- and/or kallikrein-like activity during cultivation. As
this activity could not be inhibited by either CTI, SBTI or
C1-INH (data not shown) it is unlikely that it was due to
either FXIIa or kallikrein adhering to the cells during
cultivation. It was subsequently shown that FXIIa gener-
ated by activation of FXII bound to cells responded to
increasing concentrations of the C1-INH in the incubation
medium in a manner similar to that of cell-bound FXIIa
(Fig. 1).
By following the progress of the inactivation it was
observed that the reaction between C1-INH and the cell-
bound FXIIa proceeded more slowly than the reaction
between C1-INH and FXIIa in the incubation medium
above the cells. The inactivation of cell-bound FXIIa was
Fig. 1. The effect of increasing concentrations of C1-INH on the activity

of soluble and cell-bound FXIIa. In the presence of Zn
2+
, the cells were
preincubated with FXIIa for 1 h. Then C1-INH of varying concen-
trations was added in 1/10 incubation vol. After incubation for another
1 h, the amidolytic activity was measured on the cells (d)andinthe
medium (s) as described. Cells were also incubated for 2 h with a
mixture of 40 n
M
FXII and C1-INH (m). The medium was aspirated
and the amidolytic activity measured on the cells as described. The
medium and the cells were incubated with S-2302 in EDTA buffer for
1 h and 4 h, respectively. The final concentration of the C1-INH
is shown on the abscissa. The activity measured in the absence of
C1-INH has been given the index 100. The results are corrected for
activity measured on cells incubated in the absence of FXII/FXIIa.
Ó FEBS 2003 Factor XIIa activity on endothelial cells (Eur. J. Biochem. 270) 113
almost maximal after a 20-min incubation in the presence of
a high concentration of C1-INH (Fig. 2). The second-order
rate constant was calculated and in five individual experi-
ments found to be in the range 0.4)0.8 · 10
5
M
)1
Æmin
)1
.
This is 3–10-fold lower than the constant determined
analogously for the inactivation of FXIIa in the medium
(2.3)4.3 · 10

5
M
)1
Æmin
)1
). The decrease could be explained
by either one of two mechanisms: (a) the first is based on the
assumption that the cell-bound FXIIa is being pulled off the
cells by the binding of soluble FXIIa to the inactivator as
illustrated in Fig. 3A; (b) the second mechanism assumes
that the inactivator binds to FXIIa, while still bound to the
cell surface (Fig. 3B).
To analyse which of these mechnisms was followed, the
cells were preincubated with radioactively labelled FXIIa,
and the amidolytic activity as well as the amount of cell-
bound FXIIa were measured in identical aliquots of
lysates of the cells after incubation with varying concentra-
tions of C1-INH. This showed that the inactivation
only slightly affected the binding while the activity decreased
with increasing concentrations of the C1-INH to the level
of activity measured in the absence of FXIIa (see
Table 1) (Fig. 4). This suggests that C1-INH inactivates
Table 1. Amidolytic activity measured on the cells and in the incubation
medium after incubation of HUVEC with and without FXII in the
presence and absence of Zn
2+
. Confluent layers of HUVEC were
incubated at room temperature with block buffer in the absence and
presence of 15 l
M

Zn
2+
and 40 n
M
factor XII [to minimize the risk of
interference from contaminating FXIIa, the stock solution of FXII
(50 l
M
) was diluted twice with 0.2 m
M
phenyl-methyl-sulfuryl-fluoride
in 0.15
M
NaClandkeptat4°C for 20 h before start of the experi-
ment]. After 120 min incubation the medium was transferred to and
diluted (1 : 1) with 0.8 m
M
S-2302 and incubated for 4 h. As rapidly as
possible the rest of the medium was aspirated and the cells quickly
washed as described in Experimental procedures. Then 0.4 m
M
S-2302
was added to the cells and the incubation continued for the same
period of time as that used for the incubation medium (4 h). An ali-
quot of 0.4 m
M
S-2302 incubated under identical conditions was used
as blank. The results are obtained from six determinations in three
individual subcultures of confluent cells and given as means ± SD.
Medium composition

Total S-2302 amidolytic activity,
DOD/time period
On cells In medium
FXII + Zn
2+
0.263 ± 0.028 0.100 ± 0.0136
FXII ) Zn
2+
0.135 ± 0.015 0.038 ± 0.0031
No FXII + Zn
2+
0.199 ± 0.022*
a
0.000 ± 0.0015
No FXII ) Zn
2+
0.129 ± 0.026*
b
0.000 ± 0.0013
* Statistically significant differences (Student’s t-test) between
a
FXII + Zn
2+
and no FXII + Zn
2+
, and
b
no FXII + Zn
2+
and no FXII ) Zn

2+
at P < 0.005.
Fig. 2. The progress of inactivating cell-bound FXIIa by C1-INH. Cells
were incubated with 40 n
M
FXIIa for 1 h. Then C1-INH was added at
a final concentration of 100 n
M
(.) and in two separate experiments at
a final concentration of 400 n
M
(s, d), respectively, and the incuba-
tion was continued for the period of time indicated. The medium was
then aspirated, cells were washed and the remaining amidolytic activity
was measured as D of the samples after a 4-h incubation with S-2302 in
EDTA buffer.
Fig. 3. Possible mechanisms for C1-INH inactivation of cell-bound
FXIIa. (A) The C1-INH disturbs the binding of FXIIa to HUVEC by
binding to soluble FXIIa. (B) The C1-INH binds to receptor bound
FXIIa.
Fig. 4. Correlation between amidolytic activity and the amount of FXIIa
bound to the cells. The cells were incubated for 2 h with 40 n
M
radio-
actively labelled FXIIa and the radioactivity (s) and amidolytic
activity (d) were measured in the same aliquots of solubilized cells.
Following washing as described, the cells were solubilized by 10-min
incubation in 0.1
M
NH

4
OH, 1% (v/v) Triton X-100, and subsequently
neutralized by the addition of acetic acid. Aliquots of the solubilized
cells were counted in a c-counter and mixed 1 : 1 with 0.8 m
M
S-2302
inEDTAbuffer,incubatedfor4handacidifiedwithaceticacidbefore
reading the absorption at 405 nm. The results are shown as
means ± SD of six individual experiments using two individual sub-
cultures of cells. SD is shown by vertical bars, when extending beyond
the symbols.
114 I. Schousboe (Eur. J. Biochem. 270) Ó FEBS 2003
the amidolytic activity of the cell-bound FXIIa by binding
to FXIIa while still bound to the cellular receptor.
This mechanism was confirmed by analysing the effect of
C1-INH on the kinetics of FXIIa binding to the cells by
determining the S-2302 amidolytic activity of the bound
FXIIa. In the absence of C1-INH the activity increased
hyperbolically with the concentration of FXIIa in the
medium (Fig. 5A). The concentration of FXIIa leading to
half maximal activity – and thus half maximal binding –
was found to be 90 n
M
, and a value that did not change in
the presence of C1-INH. Increased C1-INH concentrations
resulted, however, in a decreased ÔmaximalÕ activity of the
bound FXIIa (Fig. 5B). This result complies with a
mechanism by which C1-INH binds to the cell-bound
FXIIa (compare Fig. 3B) and thus confirms the results
shown in Fig. 4. It is, however, not compatible with a

mechanism by which the bound FXIIa is being pulled off
the cells (compare Fig. 3A).
Zn
2+
-dependent amidolytic activity generated
on the surface of cells incubated with plasma
The experiments performed with purified FXIIa indicated
that a C1-INH concentration identical to that in plasma
would be expected to inactivate completely the expression of
FXIIa amidolytic activity on the cells. To verify this,
HUVEC were incubated with hirudin anticoagulated
plasma. Hirudin anticoagulation was chosen over that
based on citrate as the addition of hirudin was assumed not
to disturb the normal physiological concentration of
electrolytes and minerals, as does the addition of citrate.
Thus citrate would disturb the endogenous balance of the
concentration of free Zn
2+
, which is of importance for the
binding of not only FXII but also the HK/PPK complex to
HUVEC. Cells were incubated with the plasma for 2 h,
during which FXII was allowed to bind to the cells and to
become activated. Prior to the incubation, the plasma was
enriched with Zn
2+
increasing its concentration by 25 l
M
.
This concentration of Zn
2+

was found to enhance maxi-
mally the total amidolytic activity on the cells after
incubation with plasma. After incubation, the plasma was
aspirated, the cells washed and amidolytic activity was
measured according to the procedures described previously.
As S-2302 is a substrate not only for FXIIa but also for
kallikrein, the measured activity could be caused by an
FXIIa-independent activation of PPK [20]. In order to
separate the amidolytic activity of FXIIa from that of
kallikrein, the nonphysiological reversible inhibitors SBTI
and CTI were added to the plasma. CTI is a specific
inhibitor of FXIIa activity [31] and SBTI inhibits the
activity of several proteinases including kallikrein but not
FXIIa [32]. The advantage of using SBTI and CTI was that
they were both removed from the surface of the cells during
the washing procedure and thus did not interfere with the
subsequent measurement of amidolytic activity. Combined
addition of SBTI and CTI was found to inhibit 82% of the
totally amidolytic activity generated on the cells (Fig. 6).
The remaining 18% was identical to the activity generated
on the cells incubated with block buffer in the absence of
plasma as well as to the activity generated on the cells after
incubation with plasma to which no extra Zn
2+
had been
added. To analyse the relative magnitude of the SBTI- and
CTI-insensitive activity separately, multiple sets of HUVEC
were incubated in comparable series with plasma to which
the inhibitors were added separately. Of the 82% inhibited
by the presence of both SBTI and CTI, 65% was inhibited

by CTI alone and 91% by SBTI (Fig. 6). No CTI-
insensitive activity could be measured when SBTI
(0.01 mgÆmL
)1
) was added to the chromogenic substrate.
This indicates that the activity inhibited by the presence of
CTI may be due to reciprocal activation of FXII and PPK.
The difference between the activity generated in the presence
of CTI and that in the presence of SBTI, may be due to
FXIIa-independent PPK activation and the difference
between the activity generated in the presence of SBTI
and that generated in the presence of both SBTI and CTI
Fig. 5. Inhibiting effect of the C1-INH on the activity of varying
amounts of FXIIa bound to the cells. (A) Cells were incubated for 1 h
with the concentrations of FXIIa shown on the abscissa, and 1/10 vol.
0.15
M
NaCl (n) or C1-INH, respectively, was added leading to a final
C1-INH concentration of 206 n
M
(s)or82n
M
(d). After an addi-
tional 1-h incubation, the amidolytic activity on the cells was measured
as described. The results are shown as means ± SD of determinations
using three individual subcultures. (B) Reciprocal plots of the data
from (A) showing the amidolytic activity of bound FXIIa in the
absence (n) and presence (s, d) of the C1-INH (the slope of the lines
is 1/V
max

; V
max
is the maximal activity of the bound FXIIa). The
intercept with the abscissa is K
D
.
Ó FEBS 2003 Factor XIIa activity on endothelial cells (Eur. J. Biochem. 270) 115
may reflect a kallikrein-independent activation of FXII.
Thus both FXII and PPK became activated on the
endothelial cells incubated with plasma containing physio-
logically normal concentrations of the C1-INH but the
majority of the S-2302 amidolytic activity measured after a
2-h incubation was accounted for by kallikrein. No
amidolytic activity could be measured in the plasma which
had been incubated with HUVEC.
Discussion
The present study shows for the first time that the
amidolytic activity of FXIIa bound to HUVEC is
protected from inactivation by C1-INH. The protection
could be explained by a decreased second-order rate
constant for inactivation of the bound FXIIa compared
to inactivating FXIIa in the medium. The inactivation
only slightly affected the dissociation of FXIIa from
HUVEC indicating that C1-INH inactivates cell-bound
FXIIa by binding to FXIIa on the surface of the cells.
Accordingly, C1-INH reduced the activity of saturated
amounts of bound FXIIa without changing the K
D
for
binding of FXIIa to the cells.

The analysis of the protecting effect that the cells might
exert on FXIIa against inactivation by the C1-INH was
originally thought to be performed on HUVEC incubated
with FXII. The advantage of using this set up instead of cells
preincubated with FXIIa was that the C1-INH would affect
FXIIa when generated on the cell surface while its influence
on the putative low concentrations of FXIIa in the medium
would be insignificant. However, although the activity of
FXIIageneratedonthecellsbyactivationofFXIIinthe
presence of Zn
2+
was significantly higher than that
generated in the absence of Zn
2+
, the fraction of bound
FXII that became activated was judged to be too low for
further investigation. A single series of experiments per-
formed on cells incubated with FXII indicated, however,
that the degree of inactivation of cell-bound FXIIa was
independent of the cells being incubated with FXII or
FXIIa. That the amidolytic activity of FXIIa was protected
against inactivation when bound to a cellular surface is in
accordance with a previous investigation showing that
incubation of FXIIa with kaolin and sulfatide protects the
amidolytic activity of FXIIa against inactivation [29]. Three
mechanisms of protection of FXIIa activity against inacti-
vation in the presence of kaolin and sulfatide have been
suggested. One of these was that C1-INH disturbs the
reversible association of FXIIa with the binding surface
either by binding to the surface or to the soluble FXIIa. This

mechanism was found not to be involved in the inactivation
of cell-bound FXIIa. The other two mechanisms, explaining
decreased inactivation, were based on either steric hindrance
induced by FXIIa binding to the surface or to conforma-
tional changes of FXIIa occurring when bound to the
surface. Both of these mechanisms may be involved in the
protection against C1-INH inactivation of FXIIa bound to
HUVEC. Factor XIIa has been shown to change confor-
mation in the presence of Zn
2+
[3,5] and binding of FXIIa
to the cells is Zn
2+
-dependent [16]. A conformational
change of FXIIa could, however, not by itself explain the
protection against inactivation, as the presence of Zn
2+
in
the incubation medium only slightly protected soluble
FXIIa from inactivation by the C1-INH (I. Schousboe,
unpublished data).
It may be argued that C1-INH could bind to HUVEC
in the absence of FXII thus reducing the effective dose of
the inhibitor available for inactivation, which might at least
partly explain the protection. However, a reduction in the
effective dose of C1-INH would affect not only inactiva-
tion of the cell-bound FXIIa but also the inactivation of
thefreeFXIIainthemediumabovethecells.The
consequence of this would have been that the secondary
rate constant for binding the inhibitor to FXIIa would

decrease but this decrease would be identical for the cell-
bound FXIIa and the FXIIa in the medium above the cells
and would not change the ratio between the values. Thus
FXIIa-independent binding of C1-INH to the cells would
not have changed the overall results showing that cell-
bound FXIIa is protected from inactivation by C1-INH by
a factor of 3–10. Comparison of the rate constant for
inactivation of FXIIa in the medium with that previously
published for inactivation of FXIIa in solution [7] indicates
furthermore that it is unlikely that C1-INH should bind to
HUVEC. The second-order rate constant for inactivation
of soluble FXIIa was found to be unaffected by the
presence of HUVEC.
The gradual decrease in amidolytic activity of cell-bound
FXIIa with increasing concentrations of C1-INH was not
correlated with a decreased binding of radioactively labelled
FXIIa. This suggests that C1-INH binds to FXIIa on the
cell surface consistent with a C1-INH concentration-
dependent decreased amidolytic activity of saturated
amounts of bound FXIIa, which did not affect the K
D
for
binding FXIIa to the cells.
Fig. 6. S-2302 amidolytic activity measured on HUVEC after incuba-
tion with hirudin anticoagulated plasma in the presence and absence of
reversible FXIIa and kallikrein inhibitors. Cells were incubated for 2 h
with plasma diluted 0.9-fold with either 250 l
M
Zn
2+

in Locke’s buffer
or with this Zn
2+
-solution containing either SBTI, CTI or both [final
concentrations: SBTI of 0.1 mgÆmL
)1
;CTI,0.05mgÆmL
)1
(4 l
M
].
Then the plasma was aspirated and the cells washed rapidly with
Locke’s buffer containing 25 l
M
Zn
2+
. Subsequently, S-2302 in
EDTA buffer was added for determination of cell-bound activity. The
hatched parts of the columns show the measured activity as
mean ± SD of five individual subcultures. The blank parts of the
columns show the calculated inhibited activity and the P-values indi-
cate the level of significance (Student’s t-test) of the change in activity
in the absence and presence of the inhibitors.
116 I. Schousboe (Eur. J. Biochem. 270) Ó FEBS 2003
In plasma the generation of FXIIa by activation of
FXII is closely connected to activation of PPK. The
S-2302 amidolytic activity measured on cells incubated
with plasma may thus be due to both FXIIa and
kallikrein activity. The S-2302 amidolytic activity meas-
ured on the cells after a 2-h incubation with Zn

2+
-
enriched human plasma was found primarily to be due to
kallikrein (Fig. 6). A considerable amount of the total
amidolytic activity of kallikrein was, however, generated
by an FXIIa enhanced activation mechanism, which was
inhibited by CTI. That only a negligible amount of the
cell-bound S-2302 amidolytic activity could be measured
as FXIIa activity is in accordance with the suggested
mechanism for FXIIa inactivation on cell membranes. By
binding the C1-INH to the receptor/FXIIa complex no
free receptors will be available for binding soluble FXII
from the plasma. The reason why no FXIIa activity could
be measured on the surface of HUVEC after incubation
with plasma may therefore be that FXIIa had been
gradually inactivated during the 2-h incubation by binding
C1-INH to the receptor bound FXIIa. All of the receptors
on the cells are saturated with FXII at the normal FXII
concentration (350 n
M
), which is 4–6-fold higher than the
dissociation constant for binding to HUVEC. The normal
concentration of C1-INH is 2.5 l
M
[11] and more than
sufficient to completely inactivate any FXIIa amidolytic
activity generated on the cells. The process of binding
FXII to the cellular receptor, activating the bound FXII
and inactivating the cell bound FXIIa by binding C1-INH
is a complicated time-dependent series of events. It may

appear to play a significant role in regulating the function
of cell-bound FXII. Disturbing the regulation by reducing
the concentration of C1-INH is known to have serious
consequences. Patients with a diminished concentration of
C1-INH, having antibodies towards C1-INH or a dys-
functional C1-INH suffer from angiodema, a disease
resulting in episodic swelling of subcutaneous tissues,
bowel walls and upper airways. The vasopermeability
mediator responsible for the swelling may be bradykinin,
a cleavage product of HK generated by proteolytic attack
by kallikrein generated by activation of PPK as a
consequense of vessel injury. The activation of FXII
and PPK/HK may start with a local increase in the
concentration of Zn
2+
as a result of leakage from
disturbed cells. This leakage may lead to binding of FXII
and PKK/HK to respective receptors on endothelial cells.
Localized at the site of injury, FXII and PPK become
activated and start a series of events involving generation
of bradykinin and activation of plasminogen. The Zn
2+
-
dependent enhanced FXIIa/kallikrein-like amidolytic
activity on HUVEC measured in the absence of FXIIa
may be the initiating factor. Binding of C1-INH to
activated FXII may prevent spreading of activation and
be of importance in the subsequent elimination of the
disturbed cells by macrophage invation.
Acknowledgements

I thank Jørn Dalsgaard Nielsen (of The University Hospital Gentofte)
for helping to deliver the hirudin anticoagulated plasma and B. Harder
for technical assistance. This study was supported by grants from the
Danish Health Research Council and the Novo Nordic Foundation.
References
1. Schousboe, I. (1990) The inositol-phospholipid-accelerated acti-
vation of prekallikrein by activated factor XII at physiological
ionic strength requires zinc ions and high-M
r
kininogen. Eur. J.
Biochem. 193, 495–499.
2. Silverberg, M., Dunn, J.T., Garen, L. & Kaplan, A.P. (1980)
Autoactivation of human Hageman factor. Demonstration uti-
lizing a synthetic substrate. J. Biol. Chem. 255, 7281–7286.
3. Rojkjaer, R. & Schousboe, I. (1997) The surface-dependent
autoactivation mechanism of FXIIa. Eur. J. Biochem. 243,
160–166.
4. Citarella, F., Wuillemin, W.A., Lubbers, Y.T. & Hack, C.E.
(1997) Initiation of contact activation in plasma is dependent on
factor XII autoactivation and not on enhanced susceptibility of
FXII for kallikrein cleavage. Br. J. Haematol. 99, 197–205.
5.Bernardo,M.M.,Day,D.E.,Halvorson,H.R.,Olson,S.T.&
Shore, J.D. (1993) Surface-independent acceleration of factor XII
activation by Zinc ions. II. Direct binding and fluorescence stud-
ies. J. Biol. Chem. 268, 12477–12483.
6. Griffin, J.H. (1978) Role of surface-dependent activation of
Hageman factor (blood coagulation factor XII). Proc. Natl Acad.
Sci. USA 75, 1998–2002.
7. Patston, P.A., Gettins, P., Beechem, J. & Schapira, M. (1991)
Mechanism of serpin action: evidence that C1 inhibitor functions

as a suicide substrate. Biochemistry 30, 8876–8882.
8. Pixley, R.A., Schapira, M. & Colman, R.W. (1985) The regulation
of human factor XIIa by plasma proteinase inhibitors. J. Biol.
Chem. 260, 1723–1729.
9. Schreiber, A.D., Kaplan, A.D. & Austen, F.K. (1973) Inhibition by
C1 inhibitor of Hageman factor fragment activation of coagula-
tion, fibrinolysis and kinin-generation. J. Clin. Invest. 52, 1402.
10. De Agostini, A., Lijnen, H.R., Pixley, P.A., Colman, R.W. &
Schapira, M. (1984) Inactivation of factor XII active fragment in
normal plasma: predominant role of C1 inhibitor. J. Clin. Invest.
73, 1542–1549.
11. Spa
¨
th, P.J. (1997) C1 inhibitor. In Human Protein Data (Haeberli,
A., ed.). Wiley-VCH, Verlag Gmbh, Weinheim.
12. Agostoni, A. & Cicardi, M. (1992) Hereditary and acquired
C1-inhibitor deficiency: Biological and clinical characteristics in
235 patients. Medicine 71, 206–215.
13. Cicardi, M., Bergamaschini, L., Cugno, M., Beretta, A., Zingale,
L.C., Colombo, M. & Agostoni, A. (1998) Pathogenetic and
clinical aspects of C1inhibitor deficiency. Immunobiology 199, 366–
376.
14. Cugno, M., Cicardi, M., Coppola, R. & Agostoni, A. (1996)
Activation of factor XII and cleavage of high molecular weight
kininogen during acute attacks in hereditary and aquired C1
inhibitor deficiencies. Immunopharmacology 33, 361–364.
15. Reddigari, S.R., Shibayama, Y., Brunnee, T. & Kaplan, A.P.
(1993) Human Hageman factor (factor XII) and high molecular
weight kininogen compete for the same binding site on human
umbilical vein endothelial cells. J. Biol. Chem. 268, 11982–11987.

16. Schousboe, I. (2001) Rapid and cooperative binding of factor XII
to human umbilical vein endothelial cells. Eur. J. Biochem. 268,
3958–3963.
17. Schmaier, A.H., Kuo, A., Lundberg, D., Murray, S. & Cines, D.B.
(1988) The expression of high molecular weight kininogen on
human umbilical vein endothelial cells. J. Biol. Chem. 263, 16327–
16333.
18. Hasan, A.A., Cines, D.B., Ngaiza, J.R., Jaffe, E.A. & Schmaier,
A.H. (1995) High-molecular weight kininogen is exclusively
membrane bound on endothelial cells to influence activation of
vascular endothelium. Blood 85, 3134–3143.
19. Motta,G.,Rojkjaer,R.,Hasan,A.A.,Cines,D.B.&Schmaier,
A.H. (1998) High molecular weight kininogen regulates
Ó FEBS 2003 Factor XIIa activity on endothelial cells (Eur. J. Biochem. 270) 117
prekallikrein assembly and activation on endothelial cells: a novel
mechanism for contact activation. Blood 91, 516–528.
20. Rojkjaer,R.,Hasan,A.A.,Motta,G.,Schousboe,I.&Schmaier,
A.H. (1998) Factor XII does not initiate prekallikrein activation
on endothelial cells. Thromb. Haemost. 80, 74–81.
21. Joseph, K., Ghebrehiwet, B. & Kaplan, A.P. (2001) Activation of
the kinin-forming cascade on the surface of endothelial cells. Biol.
Chem. 382, 71–75.
22. Zhao, Y., Qui, Q., Mahdi, F., Shariat-Madar, Z., Rojkjaer, R. &
Schmaier, A.H. (2001) Assembly and activation of HK-PK com-
plex on endothelial cells results in bradykinin liberation and NO
formation. Am. J. Physiol. Heart. Circ. Physiol. 280, H1821–
H1829.
23. Nishikawa, K., Shibayama, Y., Kuna, P., Calcaterra, E., Kaplan,
A.P. & Reddigari, S.R. (1992) Generation of vasoactive peptide
bradykinin from human umbilical vein endothelium-bound high

molecular weight kininogen by plasma kallikrein. Blood 80, 1980–
1988.
24. Hong, S.L. (1980) Effect of bradykinin and thrombin on pro-
stacyclin synthesis in endothelial cells from calf and pig aorta and
human umbilical cord vein. Thromb. Res. 18, 787–795.
25. Crutchley, D.J., Ryan, J.W., Ryan, U.S. & Fischer, G.H. (1983)
Bradykinin-induced release of prostacyclin and thromboxanes
from bovine pulmonary artery endothelial cells. Studies with lower
homologs and calcium antagonists. Biochem. Biophys. Acta 751,
99–107.
26. Holland, J.A., Pritchard, K.A., Papolla, M.A., Wolin, M.S.,
Rogers, N.J. & Stemerman, M.B. (1990) Bradykinin induces
superoxide anion release from human endothelial cells. J. Cell.
Physiol. 143, 21–25.
27. Palmer, R.M.J., Ferrige, A.G. & Moncada, S. (1987) Nitric oxide
release accounts for the biological activity of endothelium-derived
relaxing factor. Nature 327, 524–526.
28. Nielsen, E.W., Johansen, H.T., Hogasen, K., Wuillemin, W.,
Hack, C.E. & Molness, T.E. (1996) Activation of the complement,
coagulation, fibrinolytic and kallikrein-kinin systems during
attacks of hereditary angiodema. Scand. J. Immunol. 44, 185–192.
29. Pixley, R.A., Schmaier, A. & Colman, R.W. (1987) Effect of
negatively charged activating compounds on inactivation of factor
XIIa by Cl inhibitor. Arch. Biochem. Biophys. 256, 490–498.
30. Salacinski, P.R., McLean, C., Sykes, J.E., Clement-Jones, V.V. &
Lowry, P.J. (1981) Iodination of proteins, glycoproteins, and
peptides using a solid-phase oxidizing agent, 1,3,4,6-tetrachloro-3
a,6 a-diphenyl glycoluril (Iodogen). Anal. Biochem. 117, 136–146.
31. Hojima, Y., Pierce, J.V. & Pisano, J.J. (1980) Hageman factor
fragment inhibitor in corn seeds. Purification and characterization.

Thromb. Res. 20, 149–162.
32.Tans,G.,Janssen-Claessen,T.,Rosing,J.&Griffin,J.H.
(1987) Studies on the effect of serine protease inhibitors on
activated contact factors. Application in amidolytic assays for
factor XIIa, plasma kallikrein and factor XIa. Eur. J. Biochem.
164, 637–642.
118 I. Schousboe (Eur. J. Biochem. 270) Ó FEBS 2003