The Role of Factor XII in Contact System Activation

To the Editor:

Although the excellent review on the contact system by Colman and Schmaier1 that appeared recently in Blood examines the subject in great detail, there are some points that deserve further discussion.

A major item in that review is the novel observation, described recently in Blood,² that prekallikrein (PK) becomes activated when it is bound to high molecular weight kininogen (HK) on endothelial cells, independently of factor XII. This activation is mediated by an ill-defined cell-associated thiol-protease and is said to be regulated by HK, because increasing concentrations of HK upregulate this unknown protease. The investigators speculate that this novel mechanism may be the predominant activation pathway for the contact system in vivo, because it does not require an artificial surface. However, some considerations need to be made. First, the stimulus for this novel activation mechanism of PK is unclear. The investigators mention that increasing HK concentrations trigger the activation of PK by upregulating the thiol-protease. In the original study describing this novel pathway, activation of PK was observed at concentrations of HK of 20 nmol/L,2 which is more than 30-fold lower than the plasma concentration of HK.1 So, this factor XII-independent activation of PK should be turned on under physiological conditions and may explain why normal persons have higher levels of kallikrein-C1-inhibitor than of factor XIIa-C1-inhibitor.3 However, as it is now, this pathway does not provide an explanation for enhanced contact activation under pathophysiological conditions, because under these conditions HK levels do not increase markedly. For example, we have observed activation of PK (and factor XII) in patients with insect sting-induced angioedema and shock, which was accompanied by decreasing rather then increasing concentrations of HK.4 Second, there is in vivo evidence for a factor XII-dependent activation of PK. We have observed activation of PK in healthy persons upon intravenous injection of desamino D-arginine vasopressin (DDAVP).5 This activation was not observed in factor XII-deficient persons,5indicating the existence of a factor XII-dependent activation of PK in vivo (and, notably, not triggered by an artificial surface). In our experience, activation of PK in clinical situations is often, if not always, accompanied by activation of factor XII, which is not expected in case of a factor XII-independent activation of PK. Therefore, factor XII-dependent activation of PK likely predominates contact activation under pathophysiological conditions.

A second comment concerns the fibrinolytic activities of the contact system. In their review, Colman and Schmaier1 propose that factor XII-independent activation of prekallikrein on endothelial cells is involved in two pathways for fibrinolysis, one involving the release of tissue-type plasminogen activator induced by bradykinin and another involving the activation of pro-urokinase by kallikrein. They do not mention the existence of another fibrinolytic pathway involving the contact system, ie, factor XII-dependent activation of plasminogen. There is in vivo evidence for the existence of this pathway, which is independent of kallikrein and urokinase.5 Direct activation of plasminogen by factor XIIa is often considered to be an unlikely scenario for this pathway, because plasminogen is a too poor substrate for this contact enzyme. For example, β-factor XIIa is about 300,000 times less active then urokinase in activating plasminogen. However, we have recently observed that the interaction of factor XII with plasminogen can be considerably potentiated by the presence of dextran sulphate, yielding conditions that urokinase is 10,000 and 1,500 times more active then factor XII in activating Glu- and Lys-plasminogen, respectively (Ravon et al, unpublished observations). Thus, considering the relative plasma concentrations of factor XII and urokinase (that of factor XII is 4 orders of magnitude higher), factor XII may be equally potent as urokinase in activating plasminogen, as indeed has been found in vivo.5 Notably, the theme of a plasminogen-like molecule being activated by a factor XII-like molecule has been used more often by nature, as is illustrated by hepatocyte growth factor (scatter factor), which is homologous to plasminogen, and its activating protease, which is homologous to factor XII.6 The observation that genetic deficiencies of factor XII, rather than those of PK or HK, are associated with thromboembolic disorders underscores the importance of factor XII-dependent activation of plasminogen for in vivo fibrinolysis.

We appreciate the opportunity to respond to Dr Hack's comments because it provides a forum to discuss the implications of our work.1-1,1-2 Dr Hack's questions focus on the novel mechanism for prekallikrein (PK) activation presented in recent publications.1-2,1-3 We agree that we do not know yet what regulates the membrane-associated PK activating cysteine protease. However, we propose that this PK activation mechanism is the first physiologic pathway by which the kallikrein/kinin system can be activated in vivo. We do not believe that this PK activation system is constitutively active in vivo. The amount of high molecular weight kininogen (HK) necessary for maximal PK activation is limited by the number of kininogen receptors on cells and not the ambient plasma concentration. How HK modulates the activity of this enzyme needs to be examined further. Furthermore, HK is not the single regulator of activation of this pathway. In a study now in press, we show that activation of this pathway also is dependent on an optimal free zinc ion concentration.1-4 In the absence of an optimal free zinc ion concentration, the system is quiescent. These data suggest that the local liberation of zinc ion may be the immediate regulator of activation of this system. Furthermore, we show that on endothelial cells, factor XII (FXII) does not autoactivate in any reasonable period of time (<1 hour) and that FXII activation is dependent on PK activation and not vice versa.1-4 Activated FXII is then able to reciprocally activate more plasma PK.

We agree that our proposed pathway for contact system activation is not a substitute for the role of anionic surfaces (dirt, cardiopulmonary bypass tubing, bacteria, etc) in FXII activation in nonphysiologic states. Furthermore, we agree that Dr Hack's DDAVP infusion experiments suggest that, in response to endothelial cell agonists, FXII activation can initiate PK activation.1-5 It was an unintended oversight not to refer to the FXII-dependent pathway of plasminogen activation described by Levi et al.1-5 As described by Dr Hack, the presence of an artificial surface potentiates FXIIa's activation of plasminogen. This pathway, which may become operative in nonphysiologic states, can be conjoined with the proposed physiologic contact system pathways for fibrinolysis mediated by bradykinin-induced tPA liberation and kallikrein activation of single-chain urokinase.1-2,1-3 1-6 Last, it is not rigorously proven that FXII-deficient patients are at increased risk for thrombosis. Furthermore, because PK and HK deficiencies are so rare, there have not been sufficient number of individuals described with these defects to determine if they have less of a risk for thrombosis than FXII-deficient patients.