Contact System: A Vascular Biology Modulator With Anticoagulant, Profibrinolytic, Antiadhesive, and Proinflammatory Attributes

Mapping of the discontinuous H-kininogen binding site of plasma prekallikrein. Evidence for a critical role of apple domain-2

Plasma prekallikrein, a zymogen of the contact phase system, circulates in plasma as heterodimeric complex with H-kininogen. The binding is mediated by the prekallikrein heavy chain consisting of four apple domains, A1 to A4, to which H-kininogen binds with high specificity and affinity (K(D) = 1.2 x 10(-8) M). Previous work had demonstrated that a discontinuous kininogen-binding site is formed by a proximal part located in A1, a distal part exposed by A4, and other yet unidentified portion(s) of the kallikrein heavy chain. To detect relevant binding segment(s) we recombinantly expressed single apple domains and found a rank order of binding affinity for kininogen of A2 > A4 approximately A1 > A3. Removal of single apple domains in prekallikrein deletion mutants reduced kininogen binding by 21 (A1), 64 (A2), and 24% (A4), respectively, whereas deletion of A3 was without effect. Transposition of homologous A2 domain from prekallikrein to factor XI conferred high-affinity kininogen binding from the former to the latter. The principal role of A2 for H-kininogen docking to the prekallikrein heavy chain was further substantiated by the finding that cleavage of a single peptide bond in A2 drastically diminished the H-kininogen binding affinity. Furthermore, the epitope of monoclonal antibody PKH6 which blocks kallikrein-kininogen complex formation with an IC(50) of 8 nM mapped to the center portion of domain A2. Our data indicate that domain A2 and two flanking sequence segments of A1 and A4 form a discontinuous binding platform for H-kininogen on the prekallikrein heavy chain. Domain-specific antibodies directed to these critical sites efficiently interfered with contact phase-induced bradykinin release from H-kininogen.

Plasma and tissue kallikrein in arthritis and inflammatory bowel disease

To ascertain the participation of the plasma kallikrein-kinin system (KKS) in arthritis and inflammatory bowel disease, we used two rat models resembling rheumatoid arthritis and Crohn's disease. Proteoglycan-polysaccharide from group A streptococcus (PG-APS) produced chronic destructive inflammation and systemic response in the genetically susceptible Lewis rat, in the joints when injected intraperitoneally and in the bowel when injected into the gut wall. In both models, the KKS is activated, as evidenced by decreased prekallikrein, factor XI and high molecular weight kininogen. A specific plasma kallikrein inhibitor, Bz-Pro-Phe-boroarginine, reverses the plasma changes as well as the clinical gross and microscopic pathology of both the experimental arthritis and the inflammatory bowel disease in the genetically susceptible rats. We have also shown that the tissue kallikrein system is involved in the intestinal inflammatory changes. Intestinal tissue kalikrein (ITK) is localized in goblet cells in both normal and inflamed tissue. In chronic granulomatous inflammation, ITK is localized in macrophages. ITK decreases in chronic inflammation, probably due to secretion, since the mRNA is unchanged. Kallikrein binding protein, the ITK inhibitor, decreases due to enzyme-inhibitor complexes. Both plasma and tissue kallikrein are appealing targets for drug therapy of rheumatoid arthritis and Crohn's disease.

Cytokeratin 1 and gC1qR mediate high molecular weight kininogen binding to endothelial cells

High molecular weight kininogen (HK) attaches to endothelial cells by separate sites on the heavy and light chains and requires 15-50 microM zinc. Previously identified binding proteins include gC1qR, cytokeratin 1, and the urokinase plasminogen activator receptor; however, their relative contributions to binding are not yet clarified. We have prepared affinity columns to which were coupled either cleaved HK or peptide LDCNAEVYVVPWEKKIYPTVNCQPLGM derived from heavy-chain domain 3. Endothelial cell membranes were solubilized and chromatographed in the presence or absence of zinc ion, the bound proteins were eluted, and active fractions were identified by dot blot using biotinylated HK, SDS/PAGE, and Western blot analysis. The peptide containing column eluate revealed but one band at 68 kDa if zinc ion was present which was identified as cytokeratin 1 by amino acid sequencing of an internal peptide. The HK affinity column revealed bands at 68 kDa (cytokeratin 1), 33 kDa (gC1qR), and 66 kDa (unidentified). HK or domain 3-derived peptide bound to the 68 kDa band; prekallikrein and Factor XII did not. HK or Factor XII bound to the 33-kDa band if zinc was present while no binding to the 66 kDa band was observed. Antibody to cytokeratin 1 inhibited HK binding to endothelial cells by 30%, antibody to gC1qR inhibited HK binding to endothelial cells by 72%, and a mixture of both inhibited binding by 86%. Our data suggest HK binding by interaction of the heavy-chain domain 3 with cytokeratin 1 and the light chain with gC1qR.

Activation of the plasma kallikrein/kinin system on endothelial cell membranes

For more than three decades, it has been known that the plasma kallikrein/kinin system becomes activated when exposed to artificial, negatively charged surfaces. The existence of an encompassing in vivo, negatively charged surface capable of activation of the plasma kallikrein/kinin system has, however, never been convincingly demonstrated. In this report, we describe current knowledge on how the proteins of the plasma kallikrein/kinin system assemble to become activated on cell membranes. On endothelial cells, the activation of the plasma kallikrein/kinin system is not initiated by factor XII autoactivation as seen on artificial surfaces. On endothelial cells, prekallikrein is activated by an antipain sensitive protease. Prekallikrein activation is dependent on the presence of high molecular weight kininogen and an optimal free Zn2+ concentration. Kallikrein generated on the surface of endothelial cell is capable of activating factor XII. Further, kallikrein formed on endothelial cell membranes is capable of cleaving its receptor and native substrate, high molecular weight kininogen, liberating bradykinin and the HK PK complex from the endothelial cell surface. Endothelial cell-associated kallikrein also is capable of kinetically favorable pro-urokinase and, subsequent, plasminogen activation.

Evaluation of markers of deep vein thrombosis in patients undergoing surgery for maxillofacial malignancies

During and following significant surgical intervention, deep venous thrombosis prophylaxis by application of anticoagulants is routinely used. However, patients with malignant disorders are subject to an especially high risk of deep venous thrombosis progressing in severe cases to subsequent pulmonary embolism. The present study focuses on appraising modern markers of deep vein thrombosis in 34 patients undergoing major maxillofacial surgery, with some malignant disorders. No significant differences between the two patient groups were noted using the markers of the kallikrein-kinin-system. From the first postoperative day plasma levels of the coagulation indicator thrombin-antithrombin-III complexes were significantly higher in the group of tumour patients. Markers of fibrinolysis indicated corresponding results: on the first postoperative day tissue-plasminogen activator values rose to 18.9 +/- 3.2 micrograms/l in the group of malignant patients, but only to 7.4 +/- 1.1 micrograms/l (P < 0.05) in the control group. Also postoperative D-dimer concentrations in the malignancy group were significantly above those of the control group. In the present study it could be demonstrated that patients with malignant neoplasia undergoing major maxillofacial surgery are exposed postoperatively to a particularly high risk of developing thromboembolic complications. All in all, the status of anti-thrombotic therapy requires reappraisal with respect to the current treatment approach adopted in tumour patients.

Fibrinolytic properties of activated FXII

Activated factor XII (FXIIa), the initiator of the contact activation system, has been shown to activate plasminogen in a purified system. However, the quantitative role of FXIIa as a plasminogen activator in contact activation-dependent fibrinolysis in plasma is still unclear. In this study, the plasminogen activator (PA) activity of FXIIa was examined both in a purified system and in a dextran sulfate euglobulin fraction of plasma by measuring fibrinolysis in a fibrin microtiter plate assay. FXIIa was found to have low PA activity in a purified system. Dextran sulfate potentiated the PA activity of FXIIa about sixfold, but had no effect on the PA activity of smaller fragments of FXIIa, missing the binding domain for negatively charged surfaces. The addition of small amounts of factor XII (FXII) to FXII-deficient plasma induced a large increase in contact activation-dependent PA activity, as measured in a dextran sulfate euglobulin fraction, which may be ascribed to FXII-dependent activation of plasminogen activators like prekallikrein. When more FXII was added, PA activity continued to increase but to a lesser extent. In normal plasma, the addition of FXII also resulted in an increase of contact activation-dependent PA activity. These findings suggested a significant contribution of FXIIa as a direct plasminogen activator. Indeed, at least 20% of contact activation-dependent PA activity could be extracted from a dextran sulfate euglobulin fraction prepared from normal plasma by immunodepletion of FXIIa and therefore be ascribed to direct PA activity of FXIIa. PA activity of endogenous FXIIa immunoadsorped from plasma could only be detected in the presence of dextran sulfate. From these results it is concluded that FXIIa can contribute significantly to fibrinolysis as a plasminogen activator in the presence of a potentiating surface.

Characterization of Cell-Associated Plasminogen Activation Catalyzed by Urokinase-Type Plasminogen Activator, but Independent of Urokinase Receptor (uPAR, CD87)

The 55-kD urokinase (uPA) receptor (uPAR, CD87) is capable of binding uPA and may be involved in regulating cell-associated plasminogen activation and pericellular proteolysis. While investigating the relationship between uPAR levels and plasmin generation, we found that uPA-catalyzed plasminogen activation is stimulated by cells which do not express uPAR. This uPAR-independent mechanism appears to be at least as effective in vitro as uPAR-dependent stimulation, such that stimulation on the order of 30-fold was observed, resulting from improvements in both apparent kcat and apparent Km. The mechanism depends on simultaneous binding of both uPA and plasminogen to the cell and requires the presence of the amino-terminal fragment (ATF), available in single chain and two chain high-molecular-weight uPA, but not low-molecular-weight uPA. Stimulation was observed in all leukemic cell lines investigated at similar optimum concentrations of 106to 107 cells/mL and may be more general. A mechanism is proposed whereby uPA can associate with binding sites on the cell surface of lower affinity, but higher capacity than uPAR, but these are sufficient to stimulate plasmin generation even at subphysiologic uPA concentrations. This mechanism is likely to operate under conditions commonly used for in vitro studies and may have some significance in vivo.

Characterization of Cell-Associated Plasminogen Activation Catalyzed by Urokinase-Type Plasminogen Activator, but Independent of Urokinase Receptor (uPAR, CD87)

The 55-kD urokinase (uPA) receptor (uPAR, CD87) is capable of binding uPA and may be involved in regulating cell-associated plasminogen activation and pericellular proteolysis. While investigating the relationship between uPAR levels and plasmin generation, we found that uPA-catalyzed plasminogen activation is stimulated by cells which do not express uPAR. This uPAR-independent mechanism appears to be at least as effective in vitro as uPAR-dependent stimulation, such that stimulation on the order of 30-fold was observed, resulting from improvements in both apparent kcat and apparent Km. The mechanism depends on simultaneous binding of both uPA and plasminogen to the cell and requires the presence of the amino-terminal fragment (ATF), available in single chain and two chain high-molecular-weight uPA, but not low-molecular-weight uPA. Stimulation was observed in all leukemic cell lines investigated at similar optimum concentrations of 106to 107 cells/mL and may be more general. A mechanism is proposed whereby uPA can associate with binding sites on the cell surface of lower affinity, but higher capacity than uPAR, but these are sufficient to stimulate plasmin generation even at subphysiologic uPA concentrations. This mechanism is likely to operate under conditions commonly used for in vitro studies and may have some significance in vivo.

Activation of the plasma kallikrein/kinin system on endothelial cells

For more than two decades, it has been known that activation of the plasma kallikrein/kinin system only occurs when it is exposed to artificial, negatively charged surfaces. The existence of physiological, negatively charged surfaces has, however, never been demonstrated in vivo. In this report, we describe current knowledge about how the proteins of the plasma kallikrein/kinin system interact with and become activated on cell membranes. In this model, activation of the plasma kallikrein/kinin system on endothelial cells is not initiated by factor XII autoactivation, as seen on artificial surfaces. On endothelial cells, plasma prekallikrein is activated by a membrane-associated cysteine protease. This activation is dependent on the presence of high molecular weight kininogen and an optimal zinc (Zn2+) concentration. Although the initiation of activation of plasma prekallikrein is independent of factor XII, kallikrein-mediated factor XIIa generation, in turn, accelerates the activation of the system. Further kallikrein formed on endothelial cell membranes is capable of cleaving its receptor and native substrate, high molecular weight kininogen, liberating bradykinin and terminating activation. In addition, the kallikrein formed on the surface of endothelial cells results in kinetically favorable activation of prourokinase and, subsequently, plasminogen. Activation of the plasma kallikrein/kinin system on endothelial cells proceeds by a physiological mechanism to initiate cellular fibrinolysis independent of plasmin, fibrin, and tissue-type plasminogen activator.

High affinity binding of factor XIIa to an electronegative surface controls the rates of factor XII and prekallikrein activation in vitro

The incubation of normal human plasma in the presence of sulphatide vesicles results in the generation of amidolytic activity due to factor XIIa (FXIIa) and to kallikrein (KRN). The progress of the generation of the enzymes distinguished a high initial rate of enzyme generation, a decline of this rate to maximum amidolytic activity ([FXIIa]m and [KRN]m) and a negative pseudo-first-order rate attributed to enzyme inactivation by plasma C1-inhibitor (C1INH). [FXIIa]m and [KRN]m were determined after the treatment of various dilutions of plasma in the presence of 4, 15, or 40 microM sulphatide vesicles. At all levels of sulphatides, [FXIIa]m and [KRN]m initially increased with the concentration of plasma, to reach a plateau at higher concentration of plasma. The plateau activities of the generated enzymes and the optimal concentration of plasma both increased with the level of sulphatide vesicles. The pseudo-first-order inactivation rate for KRN increased progressively with the concentration of plasma but the respective rate for FXIIa was independent of the plasma concentration. The data suggest that contiguous binding of plasma FXIIa, factor XII (FXII), and the complexes of high molecular weight kininogen (HK) with prekallikrein (HK-PKRN) and factor XI (HK-FXI) to an electronegative surface induces a rapid generation of FXIIa and KRN. The concentration of the electronegative surface controls the levels of generated FXIIa and KRN and their release to the bulk phase. The released FXIIa and KRN are both inactivated by C1INH.

A chromogenic substrate assay kit for factor XII: evaluation and use for the measurement of factor XII levels in cardiopulmonary bypass patients

Factor XII levels were determined in plasma samples from 75 patients before undergoing aortocoronary bypass grafting and from 40 healthy age-matched donors by using a microtitre plate adaptation of a new chromogenic peptide substrate assay kit for factor XII. The chromogenic peptide substrate assay values for factor XII correlated well with those obtained in clotting (r=0.90; y= 15.811+0.8236x) and immunochemical (r=0.88; y=17.90+0.817x) assays in the normal donor samples. Factor XII levels in the patients were significantly lower than those in the normal donors (83.3+/-23.2% versus 103.4+/-23.1: p=0.004), and nine patients (12%) had factor XII values below 50% compared with only one of the normal donors (2.5%). Factor XII levels and kallikrein-like activities (a measure of contact system activation) were followed before, during, and one day after cardiopulmonary bypass in 20 patients. Factor XII levels were significantly reduced, and kallikrein-like activities significantly elevated after 5 and 30 minutes cardiopulmonary bypass. One day after cardiopulmonary bypass both factor XII levels and kallikrein-like activities were significantly lower than preoperation values.

Mapping binding domains of kininogens on endothelial cell cytokeratin 1

Human cytokeratin 1 (CK1) in human umbilical vein endothelial cells (HUVEC) is expressed on their membranes and is able to bind high molecular weight kininogen (HK) (Hasan, A. A. K., Zisman, T., and Schmaier, A. H. (1998) Proc. Natl. Acad. Sci. U. S. A. 95, 3615-3620). New investigations have been performed to demonstrate the HK binding domain on CK1. Four overlapping recombinant (r) CK1 proteins were produced in Escherichia coli by a glutathione S-transferase gene fusion system. Biotin-HK specifically bound to rCK128 and rCK131 in the presence of Zn2+ but not to Deleted1-6rCK131. Recombinant CK128 and rCK131 also inhibited biotin-HK binding to HUVEC with IC50 of 0.4 and 0.5 microM, respectively. Alternatively, rCK114 and Deleted1-6rCK131 did not inhibit binding at concentrations >/=1 microM. Seven sequential 20 amino acid peptides of CK1 were prepared to cover the protein coded by exons 1-3. Only the first peptide (GYG20) coded by exon 1 significantly inhibited HK binding to HUVEC with an IC50 of 35 microM. Fine mapping studies isolated two overlapping peptides also coded by exon 1 (GPV15 and PGG15) that inhibited binding to HUVEC with IC50 of 18 and 9 microM, respectively. A sequence scrambled peptide of PGG15 did not block binding to HUVEC and biotin-GPV20 specifically bound to HK. Peptides GPV15 and PGG15 also blocked prekallikrein activation on endothelial cells. However, inhibition of PK activation by peptide PGG15 occurred at 10-fold lower concentration (IC50 = 1 microM) than inhibition of biotin-HK binding to HUVEC (IC50 = 10 microM). These studies indicate that HK binds to a region of 20 amino acids coded by exon 1 on CK1 which is carboxyl-terminal to its glycine-rich amino-terminal globular domain. Furthermore, HK binding to CK1 modulates PK activation on HUVEC.

Haemocompatibility of paediatric membrane oxygenators with heparin-coated surfaces

Extracorporeal circulation (ECC) in paediatric patients with heparin-coated oxygenation systems is rarely investigated. The objective of this study was to evaluate, preclinically, the haemocompatibility of paediatric membrane oxygenators with heparin-coated surfaces. We compared 16 paediatric membrane oxygenators (Minimax, Medtronic) in an in vitro heart-lung machine model with fresh human blood. Eight of these oxygenation systems had a covalent heparin coating (Carmeda bioactive surface). After 90 min simulated ECC, the heparin-coated systems showed significantly higher platelet count, lower platelet-factor 4 release, reduced contact activation (factor XIIa and kallikrein), and lower neutrophil elastase levels (p < 0.05), compared to the noncoated oxygenator group. More biocompatible materials for paediatric operations may ameliorate the various postperfusion syndromes arising from ECC procedures, particularly unspecific inflammation, hyperfibrinolysis and blood loss.

The pharmacokinetics of the kininogens

Recent evidence has shown that plasma high molecular weight kininogen and both kininogens have the ability to modulate prekallikrein activation and thrombin-induced platelet activation, respectively. However, nothing is known about the plasma clearance and tissue distribution of these proteins. We examined the in vivo pharmacokinetics of high (HK) and low (LK) molecular weight kininogens in rats. 125I-HK and -LK molecular weight kininogens' clearance in rats best-fitted a biexponential model. For HK, the t1/2alpha and t1/2beta were 0.6 and 9.5 h and for LK, 0.78 and 7.4 h, respectively. 125I-kinin-free HK (cleaved HK) was cleared with a t1/2alpha and t1/2beta of 0.45 and 9.9 h, respectively. 125I-Domain 3 of kininogens was cleared with a t1/2beta and t1/2c of 0.99 and 13.3 h, respectively. HK was mostly concentrated in lung; LK, domain 3, and cleaved HK were mostly concentrated in kidney. The kininogens were also concentrated in liver, spleen, and skin. These studies indicate that protein size rather than form is the major determinant of its clearance. Furthermore, the distribution of the kininogens is where bradykinin metabolism and activity are well described.

Structure and function of gC1q-R: a multiligand binding cellular protein

gC1q-R is a 33 kDa, single chain, highly acidic protein, which was first isolated from membrane preparation of Raji cells and now appears to be ubiquitously distributed. Although, gC1q-R was originally identified as a protein which binds to the globular "heads" of C1q, recent evidence suggests that the molecule is in fact a multiligand binding, multifunctional protein with affinity for diverse ligands which at best are functionally related. These molecules include: thrombin, vitronectin, and high molecular weight kininogen. The gC1q-R molecule, which is identical to the transcription factors SF2 and the Tat-associated protein, or TAP, is the product of a single gene localized on chromosome 17p13.3 in human, and chromosome 11 in mouse, and is encoded by an approximately 1.5-1.6 kb mRNA. The full length cDNA encodes a primary translation protein of 282 residues and the 'mature' or membrane form of the protein isolated from Raji cells corresponds to residues 74-282 and is presumed to be generated by a site-specific cleavage and removal of the highly basic, 73-residues long, N-terminal segment during post-translational processing. The translated amino acid sequence does not predict for the presence of a conventional sequence motif compatible with a transmembrane segment and does not have a consensus site for a GPI anchor. However, there is strong evidence which indicates that gC1q-R is expressed both inside the cell and on the membrane. First, certain mAbs raised against gC1q-R react moderately with intact Raji cells in suspension and this binding increases when the cells are first bound to poly-L-lysine coated surfaces and then fixed with glutaraldehyde. Second, surface labeling of cells using the membrane impermeable sulfo-NHS-LC-biotin shows that gC1q-R on the surface incorporates biotin whereas intracellular gC1q-R does not. In addition, the membrane expression of gC1q-R can be upregulated with inflammatory cytokines such as INF-gamma, TNF-alpha, or LPS. These results suggest, that gC1q-R, is localized both as an intracellular and as a cell surface protein and may have important biological functions in both compartments of the cell.