Studies of Adsorption, Activation, and Inhibition of Factor XII on Immobilized Heparin

)

THROMBOSIS RESEARCH

Pergamon ThrombosisResearch89 (1998)41-50

REGULAR ARTICLE

Studiesof Adsorption,Activation, and Inhibitionof Factor XII on ImmobilizedHe~arin A Javier Sanchezl, Graciela Elgue2, Johan Riesenfeld3 and Per Olssonl IDepartment of Surgical Sciences, Division of Experimental Surgery, Thoracic Clinics, Karolinska Hospital, Stockholm, Sweden; 2Department of Clinical Immunology and Transfusion Medicine, University Hospital, Uppsala, Sweden; 3Carmeda AB, Stockholm, Sweden. (Received 21 August 1997 by Editor B. Osterud; revised/accepted 3 December 1997)

Abstract The aim of the present investigation was to clarify whether immobilized heparin does, as previously suggested, prevent triggering of the plasma contact activation system. Purified FXII in the absence or presence of antithrombin and/or Cl esterase inhibitor as well as plasma was exposed for 1 to 600 seconds to a surface modified by end-point immobilization of heparin. With purified reagents, a process including surface adsorption and activation of FXII occurred within 1 second. In the presence of antithrombin, the resulting surface-bound a-FXHa was inhibited within that time. Likewise, the adsorption of native FXII from plasma was a rapid process. However, the inhibition of surface-bound a-FXIIa was slightlyslower than with purified components. Nevertheless, neither &FXHa nor FXIa were found in the plasma phase. Exposure of a surface prepared from heparin molecules, lacking antithrombin binding properties, to plasma resulted in surface-bound a-FXIIa within 1 second. In the liquid phase, ~-FXHa was detected after 2.5 seconds and, 12 seconds later, FXIIa and FXIa in complex with the Cl esterase inhibitor appeared. Abbreviations: FXII, factor XII; FXI, factor XI; AT, antithrombin;

Cl INH, Cl esterase inhibitor; ELISA, enzyme-linked immmunosorbent assay; PBS, phosphate-buffered saline; BSA, bovine serum albumin; LA, low affinity heparin surface. Corresponding author: J. Sanchez,Divisionof ExperimentalSurgery,ThoracicClinics,KarolinskaHospital,S-17176 Stockholm, Sweden;Tel: +46 8517 73274;Fax: +46 851773557. E-mail:

.

Addition of heparin to plasma prior to surface exposure did not prevent activation of surface-bound FXH, nor did it increase the ~-FXHa inhibition rate and prevent FXI activation in plasma, although &FXIIa and FXIa-AT complex formation occurred. It is concluded that surface-immobilized heparin, unlike heparin in solution, effectively inhibits the initial contact activation enzymes by an antithrombin-mediated mechanism, thereby suppressing the triggering of the intrinsic plasma coagulation pathway. @1998 Elsevier Science Ltd. KeyWords: Immobilized heparin; FXII, FXI; Antithrombin; Cl esterase inhibitor

A

technology for improving of the blood compatibility of artificial materials by surface immobilization of heparin, without compromising the antithrombin (AT) -binding properties of the heparin molecule, results in a surface with thrombo-resistant properties [1-4]. The immobilized heparin molecules catalyze the rapid inhibition of coagulation factors Xa and thrombin, and it has been suggested that the neutralization of these enzymes may account for the thromboresistant properties of the present heparin surface [4]. Recently it was shown that end-point immobilized heparin suppressed the activation of the plasma contact activation system by mediating rapid inhibition by AT of FXHa, adsorbed to and activated by the negatively charged surface [5]. In the present investigation, we studied the signifi-

0049-3848/98$19.00 + .00 @ 1998 Elsevier Science Ltd. Printed in the USA. All rights reserved

PII SO049-3848(97)O031 O-1

42

J. Sanchez et al./Thrombosis Research 89 (1998) 41-50

cance of this inhibitory mechanismfor the thromboresistant properties and also for the possibly improved compatibility with the fibrinolyticand complement systems of end-point immobilized heparin surfaces, that has been observed experimentally and clinically [6–10].An experimental model was developed to estimate the rates at which FXII, either purified or in plasma, is adsorbed, activated, and inhibited by the heparin surface. In addition to AT, the effect of Cl esterase inhibitor (Cl INH), thought to be a main inhibitor of active FXH [11], was studied. Besides measuring enzymatic activities on the surface or released to the fluid phase, the formation of FXIIa-AT and FXIIa-Cl INH enzyme inhibitor complexes was determined immunologically.Generation of FXIa, as a marker of intrinsic pathway coagulation activation, was also measured as a complex with either of the two inhibitors. 1. Materials Tris-buffer (50 mmol/1Tris-HCl, 12 mmol/1NaCl, pH 7.8) was used in the determinations of FXII and FXIIa. Human lyophilized coagulation Factor XII (FXII) was purchased from Enzyme Research Laboratories Inc. (IN). The powder was reconstituted in 0.46ml of distilled water to a final concentration of 1.1 mg/ml (activity of 28 U/ml). Aliquots of 30 pl were dispensed and stored at –70”C. Human antithrombin (AT) was obtained from Pharmacia & Upjohn (Uppsala, Sweden). The dry powder was dissolved in water to a stock solution of 50 IU/ml and stored in small portions at –70°C. Prior to use, the stock solution was diluted in Trisbuffer to a final concentration of 1U/ml, equivalent to 2800 nmol/1. Cl esterase inhibitor of human origin (Cl INH) was purchased from Immuno AG (Vienna, Austria). It was diluted to a stock solution of 50 IU/ ml and stored in small aliquots at –70”C. Before use, the stock solution was diluted in Tris-buffer to a final concentration of 1 U/ml, equivalent to 2.3 pmol/1. Human plasma was prepared from venous blood from 10healthy donors. The blood (9 volumes) was drawn into polystyrene centrifuge tubes containing 0.13 mol/1sodium citrate solution (1 volume) and centrifuged at room temperature for 20 minutes at

3000g to obtain platelet-poor plasma. The pooled plasma was dialyzed against 0.1 mol/1 NaCl, 1.0 mmol/1EDTA in 50 mmol/1Tris-HCl buffer, pH 7.4, and dispensed in small aliquots, which were stored at –70”C [5,12]. Human FXII-deficient plasma was obtained from Helena Laboratories (Beaumont, TX). Immunodepletion of AT and Cl INH from this plasma was performed as previously described [5]. The resulting plasma contained 0.8 units prokallikrein/ml and was used as substrate for FXIIa, in the following denoted FXIIa substrate plasma. Synthetic chromogenic peptide substrates for kallikrein (S-2302)and for FXI (S-2366)purchased from Chromogenic AB (Molndal, Sweden), were dissolved in water to final concentrations of 2.5 and 0.6 mmol/1,respectively, and stored in the dark at 4“C. Cephotest, an aqueous suspension of phospholipid (cephalin) and ellagic acid, was purchased from NycoMed A/S (Oslo, Norway). Heparin, specific activity 174 IU/mg, was obtained from Pharmacia-Upjohn (Uppsala, Sweden). Polyethylene tubing (1 mm id., Portex, Hythe, UK), surface-modified by end-point immobilization of partially depolymerized heparin, as described previously [5,12]. The capacity of the immobilizedheparin to bind AT at an ionic strength of 0.15 (total AT-binding capacity) was 19 pmol/ cm2and at an ionic of strength of 0.4 (binding to the specific AT binding sequence only), it was 4 pmol/cm2 [4,5,12,13]. A heparin surface lacking specific AT binding domains was also used and prepared as previuosly described [5,12]. Popcorn inhibitor was purchased from Unicorn Diagnostics Ltd. (London, UK). It was reconstituted in distilled water to a final concentration of 4 p,mol/1,dispensed in 200 @ aliquots and stored at –70”C. 1.1. ELISA assayfor FXIa and FXIIa Phosphate-buffered saline (PBS) (10 mmol/1phosphate, 145mmol/1NaCl), pH 7.4, was used as coupling buffer. The working buffer, was PBS containing l’%obovine serum albumin and 0.1?40Tween 20 and the washing buffer, PBS with O.lYOTween 20. Human coagulation Factor XIIa (FXIIa) and human coagulation Factor FXIa (FXIa) were purchased from Enzyme Research Laboratories Inc.

J. Sanchez et al./Thrombosis Research 89 (1998) 41-50

(IN). The protein concentrations were 1.07mg/ml and 0.54 mg/ml, respectively. Goat IgG anti-human antibodies, were purchased by Affinity Biological Inc. (Chicago, IL), goat anti-human FXI antibodies by The Binding Site, (Birmingham, UK), rabbit anti-human AT and rabbit anti-human Cl INH by DAKO AIS (Glostrup, Denmark) and Streptavidin-conjugated horseradish peroxidase by Amersham Life Science (UK). Anti-human AT and anti-human Cl INH were biotinylated as described [14]. Biotinamidocaproate N-hydroxisuccinimideester was purchased from Sigma (St. Louis, MO). Chromogen: 1-2 phenylenediamine dihydrochloride and HZ02in citrate/phosphate-buffer (35 mmol/1and 67 mmol/1,pH 5.0). 2. Methods 2.1. ExperimentalProcedure Aliquots of 200 pJ of purified FXII; with or without addition of AT and/or Cl INH; in Tris-buffer or 200 p,lof plasma, were deposited at the distal end of a non-heparinized tubing segment, followed by insertion of a 200 PI air bubble and 400 pl of Trisbuffer. The segment was connected to a 60-cm piece of surface-heparinized tubing, through which the material was pumped at room temperature by using a motor-driven syringe. The flow velocity ranged from 0.05 to 12 ml/min, corresponding to contact times of 1, 2, 6, 12, 24, and 240 seconds between any point on the tubing surface and the 25.6-cmlong column of reagent solutions. The passage-time of the liquid column through the tubing segment ranged from 2.5 to 600 seconds and represents the time during which a hypothetical point on the column was in contact with the heparin surface. The flow was laminar at all the velocities with wall-shear stress values ranging from 6 to 1.5 and 9 to 2.3 N/cmzfor buffer and plasma, respectively [13]. 2.2. Measurements Following passage through the tubing segment, the aliquots containing FXII were collected and diluted 1 to 10 in Tris-buffer. A 200 p,] sample of this dilution was incubated with 10 pl of FXIIa sub-

43

strate plasma at 37”C,and, 2 minutes later, 100 pl of the chromogenic kallikrein substrate S-2302was added. The enzymatic reaction was stopped after 15minutes by addition of 100pl of citric acid (20% w/v), and the absorbance at 405 nm was read. The non-specificbackground absorbance achievedwhen the heparinized tubing segments had been replaced with untreated tubing segment (less than 0.08) was subtracted from the readings obtained. FXII adsorbed to the heparin surface was determined essentially as described before [5,12,15].Immediately after contact with the FXH-containing solution or plasma, the tubing segments were rinsed with Tris-buffer and incubated as closed rotating loops with a reaction mixture consisting of the FXII activating agent Cephotest (200 pJ), prokallikrein supplied in FXII-deficient plasma (FXIIa substrate plasma, 10 pi), and Tris-buffer (20 pJ) for 2 minutes (flow velocity 5.5 cm/s). The reaction mixture was collected and incubated with the chromogenic kallikrein substrate, S-2302,for 5 minutes at 37°C and the enzymatic reaction stopped by addition of 100 pl of citric acid (207. w/v). The resulting absorbance at 405 nm was read. Determination of surface-associated FXII which had undergone spontaneous conversion to enzymatically active FXII during contact with the surface was performed by the same assay procedure, however, omitting the FXII activating agent Cephotest, replacing it by FXH buffer. Under the assay conditions used, there was an excess of both enzyme substrates, i.e., prokallikrein (in FXIIa substrate plasma) and the chromogenic kallikrein substrate. The non-specificbackground (0.06)was subtracted from the absorbance readings obtained. Determination of FXIIa-AT, FXHa-Cl INH, FXIa-AT, and FXIa-Cl INH complexes in the aliquots having been exposed to the surfaces was performed using solid-phase enzyme-linked immunosorbent assays (ELISA). The test was based on the same principle as that used to determine thrombin-AT complex [16]. The wells of the microtiter plate were coated with goat polyclonal captureantibodies against coagulation factors XII or XI. Non-specific sites in the solid phase were blocked with 1‘-%. BSA in PBS. Enzyme-inhibitor complex as well as free FXH or free FXI in the sample solution (diluted 1/4 in PBS buffer IYo bovine serum albumin) bind to the corresponding specific solid phase antibodies. Subsequently, the com-

44

J. Sanchez et al./Thrombosis Research 89 (1998) 41-50

plexes bound to the surface react with biotinylated polyclonal anti-human AT or Cl INH which react with the inhibitor moiety of the complex taken up by the capture antibody. The complexes are detected by addition of streptavidin conjugated horseradish peroxidase, followed by the chromogen reagent. Standards for FXIIa-AT, FXIIa-Cl INH, FXIaAT, and FXIa-Cl INH were prepared by mixing well defined amounts of purified FXIIa or FXIa with a molar excess of AT or Cl INH and allowing them to react for 1 hour at 37°C in the presence of heparin (5 IU/ml). Residual FXIIa and FXIa activities were measured using the chromogenic substrates S-2302 and S-2366, respectively. The amounts of complex formed were regarded as equivalent to the consumption of FXIIa and FXIa. The standard solutions were diluted in normal plasma before the assay and then 1:4in PBS working buffer solution. The intra-assay coefficient of variation was determined and was found to be less than 4Y0in all four assays.Cross-reactions between antibodies to FXIa and FXIIa or antibodies against AT and Cl INH were less than 57.. 3. Results

3.1. ExperimentswithPurifiedComponents The uptake and spontaneous activation of FXII on the heparin surface, in the absence or presence of Cl INH and/or AT, are shown in Figure 1.As seen, the transport and binding of FXII to the heparin surface were dependent on the time during which the surface was in contact with the FXII solution. Neither Cl INH nor AT had any effect on those processes (Figure 1A). The spontaneous conversion of adsorbed FXII to enzymatically active FXIIa was also shown to depend on the contact time (Figure IB). The fraction of spontaneously activated FXIIa in relation to the total FXII on the surface appeared smaller at shorter exposure times. This observation might indicate that activation, although rapid, requires more time than does the adsorption. Addition of Cl INH to the FXH solution did not reduce the degree of spontaneous activation of adsorbed FXH at short contact times and only marginally at longer times (Figure IC). In contrast, the addition of AT completely quench-

Abs 405 2.50

El

2.25

A

2.00 1.75 1.50

B

1.25

n

1.00

1!

0.75

~u~~

~~ 1

2

6

12

24

240

Exposuretime (See) Fig. 1. Uptake, activation, and inhibition ofadsorbed FXII on the heparin surface in the absence or presence of AT or Cl INH. (A) Total FXII taken up by the surface during exposure to a solution of purified FXII. Identical results were obtained in the presence of respective inhibitors (not shown). Surface-bound, spontaneously activated FXII after exposure to (B) FXII, (C) FXII in the presence of Cl INH, and (D) FXII in the presence of AT. Six determinations were made for each solution and exposure time. The data are presented as box plots (StatView@,Apple). Symbols indicate (beginning from below): circle, minimum value; horizontal line, IOth percentile; bottom of box, lower quartile; line inside box, median; top of box, upper quartile; horizontal line, 90th percentile; and circle, maximum value.

ed the generation of surface-bound enzymatic activity (Figure ID). Tubing segments pre-incubated with FXH solution for 6 minutes were incubated with 200 pl of a solution of popcorn inhibitor (4 pmol/1)as closed rotating loops for up to 15 minutes, which proved to have no effect on the surface-bound enzymatic activity. Popcorn inhibitor is known to inhibit the enzymaticallyactive degradation product of FXIIa, ~-FXHa, but not the primary activation product a-FXIIa [17–19].This finding therefore seems to indicate that the surface-associated activity mainly represents ct-FXIIa, in keeping with the fact that

J. Sanchez et al./Thrombosis Research 89 (1998) 41-50

nmol / 1

4.0 3.5 3.0 -

Eiil

2.0 : l!! ‘“5 .m

0.5 IEEa, -,, 2.5

I

c [[

L1

1.0

I

A

Ea”

2.5 -

o-.J,

❑

m

5

[R Ez?3,

15

n 30

,,

45

plexes were generated, indicating quantitative inhibition of surface-bound activated FXII by AT, followed by release of complex into the liquid phase (Figure 2C). In conclusion, the experiments with purified reagents seem to demonstrate that the end-point imm-obilized heparin surface, due to its negative charge density, rapidly adsorbs and activates FXII. Simultaneously, however, there is a rapid ATmediated inhibition of the initial FXII activation product. The immobilized, functionally active heparin molecules catalyze a rapid inhibition of all activated FXH, most certainly ~-FXIIa, by AT adsorbed to the heparin surface along with FXII. 3.2. ExperimentswithPlasma

‘, 60

~

,

B

600

Exposure time ( see) Fig. 2. Measurements of FXIIa-Cl INH and FXIIa-AT complexes in various reaction mixtures containg FXII after exposure to the heparin surface. (A) FXIIa-Cl INH complex formed in a FXII solution containing Cl INH. (B) FXIIa-Cl INH complex formed in a FXII solution containing Cl INH and AT. (C) FXIIa-AT complex formed in a FXII solution containing AT. Box plot. Each box corresponds to six determinations.

a-FXIIa, but not ~-FXIIa, contains a region with the ability to bind to negatively charged surfaces [11,20]. Addition of FXIIa substrate plasma to the FXH solution, following its exposure to the heparin surface, resulted in the development of an intense kallikrein activity (absorbance values ranging from 0.9 to 1.0), demonstrating that activation of FXII had occurred. In the presence of Cl INH, similar enzymaticactivityoccurred, and increasingamounts of FXIIa-Cl INH complexes were formed with time (Figure 2A). When AT was included in the mixture along with FXII and Cl INH, the capacity of the surface-exposed solution to generate kallikrein in the FXIIa substrate plasma was completely abolished and no FXIIa-Cl INH complexes were detected (Figure 2B). Instead, FXIIa-AT com-

In the plasma milieu, FXH is adsorbed to negatively charged materials together with prokallikrein or coagulation factor XI (FXI) in complex with high molecular weight kininogen [21]. The obvious question is whether FXII is subject to similar activating and inhibitory mechanisms as those found for purified FXH. The interaction between plasma and the surface was studied not only with the standard heparin surface, but also with the surface prepared from fractionated heparin molecules lacking the specificantithrombin binding sequence (low affinity, LA, heparin surface) [5]. During exposure of the standard heparin surface to plasma, the uptake of AT increased with time from 1.2picomol/cm2after 1 second to 7.5 picomol/ cmzafter 240 seconds, in agreement with previous findings [13].No measurable AT uptake occurred on the LA heparin surface. FXII adsorbed to the standard heparin surface during an exposure time of 1 second occurred largely in its spontaneously activated form (Figure 3, left panel, A, B). At prolonged exposure times, the active form of FXII decreased and was no longer above background at 240 seconds (Figure 3B). Evidently, the inhibitory mechanism on the heparin surface was not so rapid after contact with plasma as it was for the purified components. Despite the appearance of surfaceassociated activity on short exposure to plasma, no enzymatic activity was detected in the plasma samples nor was there any measurable formation of FXIIa-AT/Cl INH or FXIa-AT/Cl INH complexes, except after the longest exposure time when a slight increase above background levels was

46

Abs

J. Sanchez et al./Thrombosis Research 89 (1998) 41-50

M

Standard

405 nm

2.50

I

2.25 2.00 1.75

Fig. 3. Uptake and expontaneous activation of FXII on the standard heparin surface and the LA heparin surface during exposure to plasma. (A) Total FXII taken up by the surfaces. (B) Surfaceassociated, spontaneously activated FXII. Box plot. Each box corresponds to six determinations.

Y

1.50

El ISA

1.25

61 0

1.00 0.75

0

0.50

&‘4$ B

0.25 0.00

Uu

Uuuu 1

2

6

12

24

240

u 1

u 2

Juuu 6

12

24

240

Exposuretime ( see)

noted (Figure 4, left panel). Resorption of surfaceadsorbed material from the plasma-exposed heparin surface, by incubating the tubing segments as closed rotating loops with 400 I.Llof 150 IU of heparin/ml for 30 minutes, resulted in release of 1.3–1.4 pmol/cm2 FXIIa-AT complexes, but no FXIIa-Cl INH complex. Although this method for resorption of enzyme inhibitor complexes may not be quantitative, the findingsdefinitely indicate that AT plays a role for the inhibition of FXIIa on the immobilized heparin, as was found for the purified components. FXII adsorbed from plasma on the LA heparin surface seemed to undergo complete spontaneous activation at all contact times (Figure 3, right panel, A, B), and subsequent exposure to popcorn inhibitor had no effect on this surface-bound enzymatic activity. The surface-exposed plasma aliquots already exhibited at 2.5 seconds a high kallikrein generating activity when assayed with FXIIa substrate plasma (absorbance values between 1.0 and 1.10).No futher increase in enzymatic activitythen occurred at longer exposure times. Incubation of the surface-exposed plasma with popcorn inhibitor (4 ~mol/1)before the reaction with FXIIa substrate

plasma completely abolished the generation of kallikrein activity. Hence, the enzymatic activity susceptible to popcorn inhibitor that was released from the surface into the plasma phase would be ~-FXIIa. The LA heparin surface-exposed plasma contained no FXIIa-AT complexes. Instead there was a steep rise in the formation of FXIIa-Cl INH complexes after 15 seconds exposure (Figure 4, right panel). Addition of heparin (10 anti-FXa units/ml) to the plasma resulted in the formation of FXIIa-AT complex, in addition to FXIIa-Cl INH complex, likewise starting after 15 seconds exposure (Figure 4, right panel). Activation of FXI in plasma during exposure of plasma to the LA-heparin surface was detected by the formation of FXIa-Cl INH complex, which increased with the exposure time. When heparin was present in plasma, FXIa-AT complexes were also formed (Figure 4, right panel). These experiments demonstrate that the standard heparin surface catalyzes rapid inhibition of FXIIa by AT, which is taken up from plasma along with FXII. The neutralization of adsorbed and activated FXII is effective enough to prevent progression of the surface-induced activation to the kalli-

J. Sanchez et al./Thrombosis Research 89 (1998) 41-50

47

Standard

LA

T /1

-50 -40

-30

Fig. 4. Measurement of FXIIa and FXIa in complex with AT or Cl INH in plasma exposed to the standard heparin surface and the LA heparin surface. Circles represent complex with AT, and triangles complex with Cl INH. Open symbols represent plasma samples without heparin and filled symbols plasma samples containing 10 IU/ml systemic heparin. Each point corresponds to five determinations (M>SD).

-20 T -10

J. /

-o

-1J.

‘u%=R=& 2.5

5

15

30

60

1 600

o [ 2.5

I 5

I 16

1

30

I 60

I 600

Exposure Time (see)

krein and coagulation systems in the plasma phase. It is also evident that heparin in solution cannot suppress these activation phenomena efficiently enough to prevent activation of FXI, the initial enzyme in the intrinsic coagulation system. 4. Discussion

The sequence of enzymatic reaction steps in the autocatalytic process leading to complete activation of the plasma contact activation system has previously been studied almost exclusively in plasma by triggering the process with negatively charged substancesin solution [20,22–26].An initial confirmational change in surface adsorbed FXII leads to conversion of enzymaticallyactivecx-FXIIa which, in turn, activates prokallikrein. Kallikrein splits off &FXIIa from a-FXIIa, and both &FXIIa and kallikrein induce further activation of FXII,

leading to autocatalytic amplification of the contact activation mechanism. Furthermore, a-FXHa has been shown to be the main activator of FXI and the intrinsic coagulation pathway [20]. The activation of the contact system is controlled by several inhibitors. Cl INH is thought to be a major inhibitor of ~-FXIIa, others being a2-macroglobulin and cx2-antiplasmin[11,24].The influence of AT on control of activation has been considered to be minor, although the rates at which FXIIa, FXIa, and kallikrein are inhibited by AT have been shown to be potentiated by heparin [27,28].Possible mechanisms for inhibition of surface-bound a-FXIIa have not been reported previously, apart from earlier investigations from this laboratory, demonstrating AT-mediated inhibition of activated FXII on surfaces modified with end-point immobilized heparin [5,15]. It has been shown that AT in solution or in plasma is rapidly taken up by surface-immobilized

48

J. Sanchez et al./Thrombosis Research 89 (1998) 41-50

heparin molecules, provided that their functional domains are not compromised by the coupling chemistry [13]. In the present experiments with purified reagents, the process of adsorption, binding, and activation and subsequent inhibition of FXIIa on the heparin surface by AT occurred within about 1 second. Since, in the system of purified components, no kallikrein is present to catalyze the formation of ~-FXII, the enzyme inhibited on the surface should be a-FXII, in keeping with the fact that cx-FXIIabut not (3-FXIIacontains the structural elements required for binding to negatively charged surfaces [11,20].However, the finding that the surface-exposed FXII solution activated prokallikrein in substrate plasma and that enzyme inhibitor complex was generated in the presence of Cl INH shows that some release of a-FXIIa in the liquid phase must have occurred. Despite the presence of Cl INH, the surfaceexposed reagent solution activated prokallikrein in FXIIa substrate plasma, confirming that Cl INH is a poor inhibitor of a-FXIIa [11]. Although the AT-mediated inhibition of surface-bound cx-FXIIaevidently is a rapid reaction it is obviously not instantaneous. If it had been instantaneous, the demonstration of surface-bound FXII by the activation with Cephotest solution would not have been possiblein the presence of AT. The exposure of plasma to the LA-heparin surface that lacks the specific AT binding sequences in the immobilized heparin molecules resulted in surface-bound FXIIa after an exposure time of 1 second; (3-FXIIa was found in the liquid phase within 2.5 seconds. Since (3-FXIIa is cleaved off from a-FXIIa and, unlike cx-FXIIa,does not bind to negatively charged surfaces [11,20],the remaining surface-bound activity is probably cx-FXIIa. The observation that FXIIa-Cl INH complex did not form in the liquid phase until more than 10seconds after the occurrence of enzymatic activity supports the hypothesis that Cl INH is a comparatively slow inhibitor of (3-FXIIa.Since the addition of heparin to plasma induced formation of FXIIa-AT complexes, but had no effect on the overall rate of inhibition, it seems that AT even in the presence of heparin, is a poor inhibitor of ~-FXIIa. Hence, the activation of FXI by surface-associated a-FXII released into the liquid phase, is not likely to be significantly suppressed by systemic heparin. The result, showing that Cl INH was a principal

inhibitor of FXIa in the absence of heparin, is in accordance with a previous report [29]. More surprisingly, however, this was the case even in the presence of heparin, despite the fact that heparin has been shown to accelerate the inhibition of FXIa by AT substantially [28,30]. Plasma exposed to the standard heparin surface exhibited no kallikrein-generating activityand contained no measurable FXIIa or FXIa inhibitor complexes. Apparently, the efficient inhibition of contact activation on immobilized heparin at an early stage of the process leads to prevention of the autocatalytic amplification of this activation mechanism. The results showing formation of FXIIa-AT—but not of FXIIa-Cl INH—complexes on the surface, indicate a mechanism involving rapid neutralization of surface-associated cx-FXIIa by AT. The fact that no @-FXIIaactivity was detected in the plasma phase, even at short exposure times when surface-bound activity was present, could indicate that additional inhibitory mechanisms are active. For example, it has been demonstrated that high molecular weight kininogen, in complex with kallikrein, enhances the inhibitory effect of AT and heparin on this enzyme [31]. Clearly, AT taken up by the functionally active, immobilized heparin molecules mediates rapid inhibition of the plasma contact activation system on the surface. Under the shear stress conditions in the present experimental setting, AT binds almost exclusively to the specific AT-binding sequence [13]. Hence, AT bound to the specific sequences exerts an inhibitory control of the initial enzyme(s) of the plasma contact activation system, resulting in complete suppression of the activation of the intrinsic plasma coagulation pathway. It was demonstrated that a certain surface density of the specific AT binding sequence is required to achieve this functional property of the heparin surface [15]. The suppressionof FXI activation on the present heparin surface is in contrast to recently published findings on a surface with multipoint attached heparin. It was instead concluded that the thromboresistant properties of immobilized heparin is based on thrombin inhibition resulting in prevention of feedback activation of factor V and VIII [32].The density, however, of the specificAT binding sequences on that particular surface was not defined. Heparin in solution does not catalyze ATmediated inhibition of surface-adsorbed FXIIa and

J. Sanchez et al./Thrombosis Research 89 (1998) 41-50

FXIa and thus cannot effectively control contact activation. Our findings strongly indicate that the thrombo-resistant properties of the present heparin surface with end-point immobilized heparin are mainly due to its capacity to suppress FXII activation, thereby preventing subsequent activation of the intrinsic plasma coagulation system. The possible compatibility of the end-point immobilized heparin surface with the fibrinolytic and the complement systems may be related to the prevention of &FXIIa generation, since both of these systems are said to be triggered by ~-FXIIa [21,33,34].

8.

9.

This study was supported by grants from the Swedish Board for Industrial and Technical Development and the Karolinska Institute, Stockholm, Sweden.

10. References 1. Larm O, Larsson R, Olsson P. A new nonthrombogenic surface prepared by selectivecovalent binding of heparin via a modified reducing terminal residue. Biomat Med Devic Artif Organs 1983;11:161-3. 2. Mottaghy K, Oedekoven B, Poppel K, Bruchmuller K, Kovacs B, Spahn A, Geisen C. Heparin free long-term extracorporeal circulation using bioactive surfaces. Trans Am Soc Artif Intern Organs 1989;35:635-9. 3. Koul B, Vesterqvist O, Egberg N, Steen S. Twenty-four-hour heparin-free veno-right ventricular ECMO: An experimental study. Ann Thorac Surg 1992;53:1046-51. 4. Elgue G, Blomback M, Olsson P, Riesenfeld J. On the mechanism of coagulation inhibition on surfaces with end point immobilized heparin. Thromb Haemostas 1993;70:289–93. 5. Sanchez J, Elgue G, Riesenfeld J, Olsson P. Control of contact activation on end-point immobilized heparin: The role of antithrombin and the specificantithrombin-bindingsequence. J Biomed Materials Res 1995;29:655-61. 6. Thelin S, Bagge L, Hultman J, Borowiec J, Nilsson L, Thorelius J. Heparin-coated cardiopulmonary bypass circuits reduce blood cell trauma. Experiments in the pig. Eur J Cardiol Thorac Surg 1991;5:486-91. 7. Videm V, Mollnes TE, Garred P, Aasen OA, Svenning JL. Biocompatibility of extracorpo-

11.

12.

13.

14.

15.

16.

17.

49

real circulation: In vitro comparison of heparin-coated and uncoated oxygenator circuits. J Thorac Cardiovasc Surg 1991;101:654-60. Borowiec J, Thelin S, Bagge L, Nilsson L, Venge P, Hansson H-E. Heparin-coated circuits reduce activation of granulocytes during cardiopulmonary bypass. A clinical study. J Thorac Cardiovasc Surg 1992;3:642-7. Mollnes TE, Riesenfeld J, Garred P, Nordstrom E, Hogasen K, Fosse E, Gotze O, Harboe M. A new model for evaluation of biocompatibility: Combined determination of neoepitopes in blood and on artificial surfaces demonstrates reduced complement activation by immobilization of heparin. Artif Organs 1995;19:909-17. Garred P, Mollnes TE. Immobilized heparin inhibits the increase in leukocyte surface expression of adhesion molecules. Artif Organs 1997;21:293-9. Pixley RA, Schmaier A, Colman RW. Effect of negatively charged activating compounds on inactivation of factor XIIa by Cl inhibitor. Arch Biochem Biophys 1987;256:49&8. Elgue G, Sanchez J, Egberg N, Olsson P, Riesenfeld J. Effect of surface-immobilized heparin on the activation of adsorbed factor XII. Artif Organs 1993;17:721-6. Pasche B, Elgue G, Olsson P, Riesenfeld J, Rasmuson A. Binding of antithrombin to immobilized heparin under varying flow conditions. Artif Organs 1991;15:481-91. Storm K, Artursson G, Nilsson U. Purification and characterization of porcine C3: Studies of the biologically active protein and its spilt products. Vet Immunol Immunopathol 1992; 34:47-61. Sanchez J, Elgue G, Riesenfeld P, Olsson P. Inhibition of plasma contact activation system of immobilized heparin: Relation to surface density of functional antithrombin binding sites. J Biomed Materials Res 1997;37:37-42. Elgue, G, Pasche, B, Blomback, M, Olsson, P. The use of a commercial ELISA for assay of thrombin-antithrombin complexes in purified systems. Thromb Haemostas 1990;63:435–8. Swartz, M J, Mitchell, H L, Cox, D J, Reeck, G R. Isolation and characterization of trypsin inhibitor from opaque-2 corn seeds. J Biol Chem. 1997;252:8105-7.

50

J. Sanchez et al./Thrombosis Research 89 (1998) 41-50

18. Hojima, Y, Pierce, JV, Pisano, JJ. Hageman

factor fragment inhibitor in corn seeds: Purification and characterization. Thromb Res 1980;20:149-62. 19. Pedersen LC, Yee VC, von Dassow G, Hazeghazam M, Reeck GR, Stenkamp RE, Teller DC. The corn inhibitor of blood coagulation factor XIIa. Crystallization and preliminary crystallographic analysis. J Mol Biol 1994; 236:385-7. 20. Revak SD, Cochrane CG, Bouma BN, Griffin, J H. Surface and fluid phase activities of the two forms of activated Hageman factor produced during contact activation in plasma. J Exp Med 1978;147:719-29. 21. Saito H. The contact phase of blood coagulation. In: Bloom AL., Forbes CD, Thomas DP, Tuddenham EGD, editors. Haemostasis and Thrombosis. Edinburgh: Churchill Livingston; 1994,p. 289–307. 22. Cochrane CG, Revak SD, Wuepper KD. Activation of Hageman factor in solid and fluid phases. A critical role of kallikrein. J Exp Med 1973;138:1564-83. 23. Griffin JH. Role of surface in surface-dependent activation of Hageman factor (blood coagulation factor XII). Proc Natl Acad Sci USA 1978;75:1998-2002. 24. de Agostini AI, Lijnen HR, Pixley RA, Colman RW, Schapira M. Inactivation of factor XII active fragment in normal plasma. Predominant role of Cl-inhibitor. J Clin Invest 1984;73:1452-89. 25. Pixley RA, Cassello A, de la Cadena RA, Kaufman N, Colman RW. Effect of heparin on the activation of factor XII and the contact system in plasma. Thromb Haemostas 1991; 66:540-7. 26. Silverberg M, Diehl SV. The autoactivation of Factor XII (Hageman factor) induced by

1ow-MWheparin and dextran sulphate. Biochem J 1987;248:715–20. 27. Pixley RA, Schapira M, Colman RW. Effect of heparin on the inactivation rate of human activated factor XII by antithrombin III. Blood 1985;66:198-203. 28. Holmer E, Kurachi K, Soderstrom G. The molecular-weight dependence of the rate-enhancing effect of heparin on the inhibition of thrombin, factor Xa, factor XIa, factor XIIa and kallikrein by antithrombin. Biochem J 1981; 193:395-400. 29. Wuillemin WA, Minnema M, Meijers JC, et al. Inactivation of factor XIa in human plasma assessed by measuring factor XIa-protease inhibitor complexes:Major role for Cl-inhibitor. Blood 1995;85:1517-26. 30. Wuillemin WA, Eldering E, Citarella F, de Ruig CP, ten Cate H, Hack CE. Modulation of contact system proteases by glycosaminoglycans. Selective enhancement of the inhibition of factor XIa. J Biol Chem 1996;271:12913–8. 31. Olson ST, Sheffer R, Francis AM. High molecular weight kininogen potentates the heparinaccelerated inhibition of plasma kallikrein by antithrombin: Role for antithrombin in the regulation of kallikrein. Biochemistry 1993;32: 12136-47. 32. Blezer R, Fouache B, Willems G, Lindhout T. Activation of blood coagulation at heparincoated surfaces. J Biomed Materials Res 1997;37:108-13. 33. Ghebrehiwet B, Randazzo BP, Dunn JT, Silverberg, M, Kaplan, AP. Mechanisms of activation of the classical pathway of complement by Hageman factor fragment. J Clin Invest 1983;71:1450-6. 34. DiScipio RG. The activation of the alternative pathway C3 convertase by human plasma kallikrein. Immunology 1982;45:587-95.