The hypotension and disseminated intravascular coagulation (DIC) in bacteremia is thought to be mediated by the combined actions of cytokines, prostaglandins, and complement. The contact system, via the release of bradykinin and the activation of Factor XI, has been postulated to be contributing to the observed hypotension and DIC. Using a mAb to Factor XII (C6B7), we blocked the activation of the contact system in an established experimental baboon model in which Escherichia coli was infused to produce lethal bacteremia with hypotension. The untreated group (n = 5) displayed contact activation, manifested by a significant decrease in high molecular weight kininogen (HK) and a significant increase in a2macroglobulin-kallikrein complexes (a2M-Kal). The C6B7-treated group (n = 5) showed an inactivation of Factor XII and the changes in HK and a2M-Kal complexes were prevented. Both groups developed DIC manifested by a decrease in platelet, fibrinogen, and Factor V levels. The untreated group developed irreversible hypotension. The treated group experienced an initial hypotension that was reversed and extended the life of the animals. This study suggests that irreversible hypotension correlates with prolonged activation of the contact system, and specific antibody therapy can modulate both the pathophysiological and biochemical changes. (J. Clin. Invest. 1993.91:61-68.) Key words: septic shock * bradykinin * prekallikrein * high molecular weight kininogen * a2-macroglobulin Introduction Activation of the kallikrein-kinin system concomitant with hypotension has been demonstrated to occur in humans ( 1, 2) and in a lethal baboon model (3) during gram-negative bacteremia. The consequences ofcontact activation include the generation of kallikrein, which releases bradykinin from high molecular weight kininogen (HK),' and the generation of Factor

An immunocytochemical study was performed to examine the cellular localization and the subcellular distribution of kininogens in human blood cells. Kininogens were visualised using the immunogold-silver staining method and confocal scanning laser microscopy. We confirmed the existence of high molecular weight kininogen in human neutrophils and describe for the first time the presence of low molecular weight kininogen on these cells. Both high and low molecular weight kininogens were restricted to the neutrophils where they localized as clusters of immunogold particles on the cell membrane. No labeling was observed intracellularly in organelles such as mitochondria, endoplasmic reticulum, and azurophilic or specific granules after permeabilization of the neutrophils with Triton X-100, a procedure that permitted the visualization of elastase in the azurophilic granules. Clusters of high molecular weight kininogen molecules attached to the neutrophil surface could serve as receptors for plasma kallikrein and/or be the source of substrate for a discrete and circumscribed formation of kinins that may in turn facilitate the local diapedesis of neutrophils and the transudation of plasma constituents during acute inflammation.

An immunocytochemical study was performed to examine the cellular localization and the subcellular distribution of kininogens in human blood cells. Kininogens were visualised using the immunogold-silver staining method and confocal scanning laser microscopy. We confirmed the existence of high molecular weight kininogen in human neutrophils and describe for the first time the presence of low molecular weight kininogen on these cells. Both high and low molecular weight kininogens were restricted to the neutrophils where they localized as clusters of immunogold particles on the cell membrane. No labeling was observed intracellularly in organelles such as mitochondria, endoplasmic reticulum, and azurophilic or specific granules after permeabilization of the neutrophils with Triton X-100, a procedure that permitted the visualization of elastase in the azurophilic granules. Clusters of high molecular weight kininogen molecules attached to the neutrophil surface could serve as receptors for plasma kallikrein and/or be the source of substrate for a discrete and circumscribed formation of kinins that may in turn facilitate the local diapedesis of neutrophils and the transudation of plasma constituents during acute inflammation.

We have isolated and probed the mechanism of action of two naturally occurring antibodies (Baltimore and Winston-Salem) against factor XI (FXI), that developed in patients congenitally deficient in FXI after replacement therapy. Purification on immobilized protein A and neutralization with monospecific antibodies against IgG heavy and light chain subtypes indicated that both antibodies were of restricted heterogeneity. Both Winston-Salem (IgG3 kappa) and Baltimore (IgG1 kappa) completely inhibited FXI coagulant activity at titers of 200 and 8 Bethesda units, respectively. Immunoaffinity columns prepared from each antibody were able to bind the heavy but not the light chain of reduced and alkylated activated FXI (FXIa). The activation of purified FXI by activated bovine factor XII (FXIIa), a reaction independent of high molecular weight kininogen (HK), was not inhibited by either antibody. The active site on the FXIa light chain was unaffected by either patient's IgG, as measured by its amidolytic activity. In contrast, one antibody (Baltimore) or its Fab' blocked the surface- mediated proteolytic activation of FXI by human FXIIa in a concentration-dependent fashion by preventing its binding to HK, but had no effect on the rate of activation of FIX by FXIa. In contrast, the other antibody (Winston-Salem) or its Fab' inhibited the activation of FIX by FXIa in a concentration-dependent fashion but did not inhibit binding of FXI to HK. We conclude that each of these two naturally occurring antibodies is directed against a specific, separate, and distinct epitope located in the heavy chain of FXIa, one near or at the domain essential for the activation of FIX by FXIa and the other close to the domain required for binding to HK.

We have isolated and probed the mechanism of action of two naturally occurring antibodies (Baltimore and Winston-Salem) against factor XI (FXI), that developed in patients congenitally deficient in FXI after replacement therapy. Purification on immobilized protein A and neutralization with monospecific antibodies against IgG heavy and light chain subtypes indicated that both antibodies were of restricted heterogeneity. Both Winston-Salem (IgG3 kappa) and Baltimore (IgG1 kappa) completely inhibited FXI coagulant activity at titers of 200 and 8 Bethesda units, respectively. Immunoaffinity columns prepared from each antibody were able to bind the heavy but not the light chain of reduced and alkylated activated FXI (FXIa). The activation of purified FXI by activated bovine factor XII (FXIIa), a reaction independent of high molecular weight kininogen (HK), was not inhibited by either antibody. The active site on the FXIa light chain was unaffected by either patient's IgG, as measured by its amidolytic activity. In contrast, one antibody (Baltimore) or its Fab' blocked the surface- mediated proteolytic activation of FXI by human FXIIa in a concentration-dependent fashion by preventing its binding to HK, but had no effect on the rate of activation of FIX by FXIa. In contrast, the other antibody (Winston-Salem) or its Fab' inhibited the activation of FIX by FXIa in a concentration-dependent fashion but did not inhibit binding of FXI to HK. We conclude that each of these two naturally occurring antibodies is directed against a specific, separate, and distinct epitope located in the heavy chain of FXIa, one near or at the domain essential for the activation of FIX by FXIa and the other close to the domain required for binding to HK.