Scientific Report 2007

Molecular and Experimental Medicine

Translational Vascular Medicine Research

The Antithrombotic, Anti-inflammatory and Antiapoptotic Protein C Pathway

J.H. Griffin, B.N. Bouma, H. Deguchi, D.J. Elias, S. Eichinger,* J.A. Fernández, E. Kerschen,** P. Kyrle,* S. Li, L. Mosnier, S. Navarro, N. Pecheniuk, H. Weiler,** X. Xu, X. Yang, S. Yegneswaran, B.V. Zlokovic***

* Medical University of Vienna, Vienna, Austria
** Milwaukee Blood Center, Milwaukee, Wisconsin
*** University of Rochester, Rochester New York

Various host defense systems act in concert in normal physiology, and their dysfunctioning contributes markedly to pathophysiologic changes. Coagulation pathways, fibrinolysis pathways, and anticoagulant mechanisms prevent bleeding while avoiding harmful blood clots. The protein C pathways provide antithrombotic, anti-inflammatory, and cytoprotective activities and are a major focus of our research.

Antiapoptotic and Cytoprotective Effects of Activated Protein C

The antiapoptotic activity of activated protein C (APC), first described in 2001, may provide cytoprotective activity that reduces cell death after a variety of cellular injuries. Recombinant APC, a well-defined anticoagulant enzyme, reduced mortality in patients with severe sepsis and multiorgan failure in a large phase 3 clinical trial; the only adverse side effect was serious bleeding. The contribution of distinct APC activities to the overall therapeutic efficacy of the enzyme in patients with sepsis is unknown.

To generate recombinant APC variants that have reduced anticoagulant activity and thus are less likely to cause bleeding, we dissected APC's anticoagulant activity from its cytoprotective activity by using site-directed mutagenesis. Using staurosporine-induced endothelial cell apoptosis assays, we showed that several specific mutations in 2 APC surface loops that severely reduced anticoagulant activity resulted in APC variants that retain normal antiapoptotic activity. Like wild-type APC, these mutants require protease-activated receptor-1 and endothelial cell protein C receptor for cytoprotective activity. In contrast, 2 APC light-chain mutants had about 350% anticoagulant activity but only about 5% cytoprotective activity.

In collaboration with H. Weiler and colleagues, Milwaukee Blood Center, and using a lethal mouse endotoxemia model of sepsis, we showed that the survival benefit conferred by APC was abolished in mice genetically deficient in endothelial protein C receptor (<10% of normal) or lacking protease-activated receptor-1. An APC variant with normal cytoprotective activities but weak anticoagulant activity (<8% of normal) was as effective as wild-type APC in reducing mortality. Our data suggest that the survival-promoting efficacy of APC in this lethal sepsis model requires APC's 2 cellular receptors and its cytoprotective actions. Moreover, these results suggest that minimally anticoagulant APC variants may reduce serious bleeding risks while providing the life-saving beneficial effects of APC in sepsis and other serious injuries.

Neuroprotective Activities of APC

Stroke is a major cause of morbidity and mortality, and few therapeutic options are available for ischemic stroke. Thrombolytic therapy with tissue plasminogen activator (tPA) is one option, but use of tPA is problematic because of its neurotoxic effects, including induction of bleeding. Previously, in collaboration with B. Zlokovic, University of Rochester, we showed multiple neuroprotective activities of APC. More recently, we found that APC reduces the neurotoxic effects of tPA and blunts tPA-induced apoptosis of both endothelial cells and neurons.

Remarkably, studies in mouse and rat models of ischemic stroke indicate that recombinant murine APC reduces bleeding induced by tPA. APC stabilizes the blood-brain barrier against bleeding because it blunts the tPA-induced increases in mRNA and protein levels of matrix metalloprotease-9, which causes breakdown of the barrier.

We think that APC merits clinical trials as a neuroprotective agent in patients with ischemic stroke. Furthermore, we speculate that APC may add substantially to the effectiveness of tPA therapy for stroke in humans.

Influence of Lipids on Blood Coagulation

Venous thrombosis is clinically distinct from arterial thrombosis; the two differ in thrombus appearance, pathogenic mechanisms, and therapeutic approaches. High-density lipoprotein (HDL) protects against arterial atherothrombosis, but it is not known if HDL protects against recurrent venous thromboembolism. We hypothesized that HDL protects against recurrent venous thrombosis because the lipoprotein has multiple antithrombotic and anti-inflammatory actions. These protective activities include downregulating thrombin generation by acting as an anticoagulant cofactor for APC/protein S, enhancing the protective activity of endothelial nitric oxide synthase, reducing leukocyte adhesion to endothelium, and exerting antiapoptotic effects on endothelium.

To test this hypothesis, in collaboration with S. Eichinger and P. Kyrle, Medical University of Vienna, we studied prospectively 772 patients who had had a first episode of spontaneous venous thromboembolism. During a mean follow-up of 48 months, 100 of the patients had recurrent venous thromboembolism. Compared with patients who had no recurrence, those with recurrence had significantly lower mean levels of apolipoprotein AI but similar levels of apolipoprotein B. For patients with levels of apolipoprotein AI greater than the 67th percentile of the values in the study population, the relative risk of recurrence was 0.51. Comparisons of plasma levels of lipoprotein particles in patients with recurrence vs those without recurrence gave similar results; HDL parameters were lower in patients with recurrence. Patients with high levels of apolipoprotein AI, HDL cholesterol, and large HDL particles had decreased risk of recurrent venous thromboembolism. Thus, HDL appears to protect against recurrent venous thrombosis, leading to speculations that HDL parameters might be predictive of the risk for venous thrombosis and that lipid-altering drugs that increase HDL levels might reduce the risk for first or recurrent venous thrombotic events.

Antithrombotic Mechanisms

M.J. Heeb, B.N. Bouma, A. Gruber,* S.R. Hanson,* U. Marzec*

* Oregon Health and Sciences University, Portland, Oregon

Natural anticoagulant proteins regulate blood coagulation and can prevent thrombosis. Recently, we focused on protein S, but we also investigated protein Z and protein Z–dependent protease inhibitor, all inhibitors of procoagulant factor Xa and other targets. Deficiency of anticoagulant protein S or protein Z is associated with increased risk of stroke and venous thrombosis, illustrating the importance of these proteins.

Protein S in a Thrombosis Model

Protein S is a vitamin K–dependent plasma protein known as a cofactor for the anticoagulant activated protein C. We showed that protein S also has direct anticoagulant (PS-direct) activity independent of activated protein C via inhibition of coagulation factors Xa, Va, and VIIIa. Purified protein S varies in PS-direct activity, depending on the methods used to isolate the protein. We examined whether protein S purified by different methods has PS-direct activity in vivo.

In the Hanson thrombosis model, platelets and fibrinogen are radiolabeled and infused into a baboon. Then saline or a test agent is infused for 1 hour along the wall of a femoral arteriovenous shunt, upstream from a thrombogenic segment of the shunt. The diameters of the segments vary to simulate arterial or venous flow. Deposition of platelets is measured during a 2-hour period by using a gamma camera. Deposition of fibrin is measured after the platelet label has decayed.

Immunoaffinity-purified protein S inhibited platelet deposition 39%–97% in "arterial" segments and 65%–100% in "venous" segments; protein S purified by anion exchange inhibited deposition 5%–34% in arterial segments and 0%–73% in venous segments. Protein S was also infused in the absence or presence of antibody (HPC4) that blocked protein C activation. Protein S suppressed platelet deposition in the presence of HPC4 nearly as well as in its absence (79%–97% vs >95%). Fibrin deposition after 2 hours was suppressed 34%–83% by protein S in the presence of HPC4 and 67%–90% in the absence of the antibody. Protein S significantly depressed formation of thrombin-antithrombin complexes in the presence and absence of HPC4.

Thus, immunoaffinity-purified protein S had PS-direct activity in vivo, and the activity did not depend on activated protein C. These studies suggest potential for protein S as an antithrombotic agent.

Zinc as Essential for PS-Direct Activity

Immunoaffinity-purified protein S has good PS-direct activity that matches the PS-direct activity of protein S in plasma, but protein S purified by anion exchange has poor PS-direct activity. We discovered that immunoaffinity-purified protein S contained a mean of 1.4 moles of zinc per mole of protein S, whereas anion exchange–purified protein S contained 0–0.3 moles of zinc per mole of protein S. Zinc correlated with PS-direct activity and with the method used for purification of protein S (P < .002).

Active protein S lost PS-direct activity when treated with the zinc chelator phenanthroline and then regained most of the activity when incubated with zinc at pH 7.4. Inactive protein S gained zinc content associated with PS-direct activity when incubated with zinc at pH 2.7 or in 6 M urea but not when incubated with zinc at pH 7–8.

The previously unrecognized role of zinc and of deleterious purification methods has led to confusion about the validity of PS-direct activity. A postulated zinc-binding site was located at the interface of 2 laminin G-type domains in protein S. Zinc may thus be essential for interdomain interactions required for PS-direct activity.

Low Protein Z Levels and Stroke

Protein Z is a vitamin K–dependent protein that acts as a cofactor for inhibition of factor Xa by protein Z–dependent protease inhibitor. We previously showed an association between low protein Z levels and increased risk for ischemic stroke in men but not in women. We hypothesized that hormonal changes with age in women may play a role in this difference. After dividing the data on the patients in our study to data on those older or younger than the median age of 57 years, we found a similar association between low protein Z levels and stroke in younger women and in younger and older men but not in older women, compared with matched controls.

Structure and Function of Coagulation Cofactors

A.J. Gale, T. Cramer, J. Cruz, D. Rozenshteyn, X. Xu, J.-L. Pellequer*

* CEA Valrhô, DSV/iBEB/SBN/LIRM, Bagnols ser Cèze, Cedex, France

Coagulation factors Va and VIIIa are highly homologous cofactors of the serine proteases factor Xa and factor IXa, respectively. These cofactors are the primary targets of activated protein C (APC) in its downregulation of the procoagulant pathway. In collaboration with J.-L. Pellequer in France, we used homology modeling techniques to model the 3-dimensional structures of these multidomain proteins. We are using the models as guides to create mutants of these cofactors and APC to investigate mechanisms of cofactor function and regulation. For example, we engineered disulfide bonds between domains in both factor Va and factor VIIIa. In factor Va, the disulfide bond facilitated investigation of the mechanisms of inactivation of factor Va by APC cleavage.

Factor VIIIa, however, is inactivated by 2 mechanisms. Thrombin activation of factor VIII results in a heterotrimer that consists of the A1 subunit, the A2 subunit, and the light chain. Both spontaneous dissociation of the A2 subunit and proteolytic cleavage of factor VIIIa by APC inactivate factor VIIIa. Hemophilia A, a deficiency of factor VIII, is treated by infusions of purified recombinant factor VIII. But the usefulness of factor VIII is limited because it is unstable after activation by thrombin as a result of the spontaneous dissociation of the A2 subunit.

We generated 2 mutants of factor VIII in which 2 newly introduced cysteine residues form a de novo disulfide bridge that cross-links the A2 and A3 domains. These interdomain disulfides prevent the spontaneous dissociation of the A2 subunit. These variants may provide a new, improved therapy for hemophilia A. We are using both in in vivo assays in mice and ex vivo assays in whole blood and plasma to evaluate the therapeutic potential of these stabilized variants. One disulfide variant clearly has improved functional properties both ex vivo and in vivo.

We are also studying the mechanisms of inactivation of factor VIIIa and factor Va by APC. In particular, we are investigating the APC-cofactor activity of factor V during APC proteolysis of factors VIIIa and Va. We are using disulfide-stabilized variants as tools to investigate mechanisms of factor VIIIa inactivation alone and in combination with mutants of APC cleavage sites. In other studies, we are investigating modulation of function of factors VIII and VIIIa by the neutrophil proteases cathepsin G and elastase and of factors VIII and V by the intrinsic pathway of coagulation.

Publications

Bouma, B.N., Mosnier, L.O. Thrombin activatable fibrinolysis inhibitor (TAFI): how does thrombin regulate fibrinolysis? Ann. Med. 38:378, 2006.

Cheng, T., Petraglia, A.L., Li, Z., Thiyagarajan, M., Zhong, Z., Wu, Z., Liu, D., Maggirwar, S.B., Deane, R., Fernández, J.A., LaRue, B., Griffin, J.H., Chopp, M., Zlokovic, B.V. Activated protein C inhibits tissue plasminogen activator-induced brain hemorrhage. Nat. Med. 22:1278, 2006.

Eichinger, S., Pecheniuk, N.M., Hron, G., Deguchi, H., Schemper, M., Kyrle, P.A., Griffin, J.H. High-density lipoprotein and the risk of recurrent venous thromboembolism. Circulation 115:1609, 2007.

Feistritzer, C., Schuepbach, R.A., Mosnier, L.O., Bush, L.A., Di Cera, E., Griffin, J.H., Riewald, M. Protective signaling by activated protein C is mechanistically linked to protein C activation on endothelial cells. J. Biol. Chem. 281:20077, 2006.

Fernández, J.A., Heeb, M.J. Role of protein C inhibitor and tissue factor in fertilization. Semin. Thromb. Hemost. 33:13, 2007.

Fernández, J.A., Lentz, S.R., Dwyre, D.M., Griffin, J.H. A novel ELISA for mouse activated protein C in plasma. J. Immunol. Methods 314:174, 2006.

Fernández, J.A., Pecheniuk, N.M., Deguchi, H., Elias, D.J., Griffin, J.H. Is adiponectin implicated in venous thromboembolism? J. Thromb. Haemost. 4:1151, 2006.

Fernández, J.A., Vento, A.E., Jormalainen, M., Griffin, J.H., Pesonen, E., Syrjälä, M., Repo, H., Jansson, S.E., Rämö, O.J., Petäjä, J. Activated protein C in the cardioplegic solution on a porcine model of coronary ischemia-reperfusion has deleterious hemodynamic effects. Cardiovasc. Drugs Ther. 20:113, 2006.

Gruber, A., Marzec, U.M., Bush, L., Di Cera, E., Fernández, J.A., Berny, M.A., Tucker, E.I., McCarty, O.J., Griffin, J.H., Hanson, S.R. Relative antithrombotic and antihemostatic effects of protein C activator versus low-molecular-weight heparin in primates. Blood 109:3733, 2007.

Heeb, M.J., Fisher, M., Paganini-Hill, A. Association of low protein Z levels with ischemic stroke in young women. Thromb. Haemost. 97:495, 2007.

Heeb, M.J., Radtke, K.-P., Fernández, J.A., Tonnu, L. Plasma contains protein S monomers and multimers with similar direct anticoagulant activity. J. Thromb. Haemost. 4:2215, 2006.

Mineo, C., Deguchi, H., Griffin, J.H., Shaul, P.W. Endothelial and antithrombotic actions of HDL. Circ. Res. 98:1352, 2006.

Mosnier, L.O., Bouma, B.N. Regulation of fibrinolysis by thrombin activatable fibrinolysis inhibitor, an unstable carboxypeptidase B that unites the pathways of coagulation and fibrinolysis. Arterioscler. Thromb. Vasc. Biol. 26:2445, 2006.

Mosnier, L.O., Zlokovic, B.V., Griffin, J.H. The cytoprotective protein C pathway. Blood 109:3161, 2007.

Pecheniuk, N.M., Deguchi, H., Elias, D.J., Xu, X., Griffin, J.H. Cholesteryl ester transfer protein genotypes associated with venous thrombosis and dyslipoproteinemia in males. J. Thromb. Haemost. 4:2080, 2006.

Radtke, K.P., Griffin, J.H., Riceberg, J., Gale, A.J. Disulfide bond-stabilized factor VIII has prolonged factor VIIIa activity and improved potency in whole blood clotting assays. J. Thromb. Haemost. 5:102, 2007.

Raivio, P., Fernández, J.A., Kuitunen, A., Griffin, J.H., Lassila, R., Petäjä, J. Activation of protein C and hemodynamic recovery after coronary artery bypass surgery. J. Thorac. Cardiovasc. Surg. 133:44, 2007.