Comprehensive study into the activation of the plasma enzyme systems during attacks of hereditary angioedema due to C1-inhibitor deficiency

C1-inhibitor (C1-INH) – the deficiency of which causes hereditary angioedema (C1-INH-HAE) types I and II – is a regulator of the complement, contact, coagulation, and fibrinolytic systems, as it inhibits rapidly activated factor XII (FXIIa), activated factor XI (FXIa), and kallikrein [1‐5]. The contact system, also known as the plasma kallikrein-kinin system, consists of factor XII, prekallikrein, high-molecular-weight kininogen (HK), and factor XI. Factor XII can be activated by contact with negatively charged surfaces, extracellular RNA [6], misfolded proteins [7], inorganic polyphosphates released from bacteria [8, 9], and bacterial surfaces [10].

When proteins of the contact system bind to endothelial cells, the activation of factor XII leads to the conversion of prekallikrein to kallikrein [11], which in turn reciprocally activates factor XII. Furthermore, physical trauma and surgery (known trigger factors of C1-INH-HAE) [12] can activate the contact system via the stress-induced release of HSP-90, or the direct endothelial cell activation of circulating factor XII, during which large areas of endothelium are exposed [13]. Moreover, these stimuli are important activators of the coagulation cascade through the expression of tissue factor and the activation of factor VII [14]. Activated factor XII can initiate blood coagulation via factor XI that links the extrinsic and intrinsic arms of the coagulation system. Factor XI circulates bound to high-molecular-weight kininogen in the plasma, and it is activated by FXIIa. FXIa in turn can activate factor XI, thereby contributing to the formation of a fibrin clot, and also to thrombin generation via the intrinsic pathway of coagulation. Remarkably, thrombin may increase vascular permeability by the cleavage and activation of a thrombin-susceptible receptor on the endothelial cell surface. Binding induces intercellular gap formation [15], which leads to edema [14, 16, 17].

Once activated by factor XII, kallikrein cleaves its enzyme cofactor, high-molecular-weight kininogen, leading to the release of the vasoactive substance, bradykinin – the key mediator of edematous attacks. The binding of bradykinin to the bradykinin B2 receptor on the surface of endothelial cells activates several intracellular signaling pathways that induce vasodilatation, as well as increase vascular permeability and fluid efflux [18, 19]. Furthermore, bradykinin can stimulate the release of tissue plasminogen activator in the human vasculature [20]; it also suppresses PAI-1 gene expression [21].

Additionally, kallikrein activates the fibrinolytic system either directly – by converting plasminogen to plasmin [22‐24], or indirectly – by activating tissue-type (tPA) and urokinase-type (uPA) plasmin activators, which convert plasminogen to plasmin. Interestingly, plasmin can activate factor XII to factor XIIa as an alternative to the kallikrein feedback [25]. In vitro studies showed that plasmin enhances the production of bradykinin through the action of kallikrein on high-molecular-weight kininogen [26].

Previous studies have analyzed the changes that occur during C1-INH-HAE attacks in each plasma enzyme system separately. According to their findings activation of the plasma enzyme systems occurs during attacks as suggested by the elevated levels of cleaved HK, FXIIa, activated factor VII [14, 27, 28], prothrombin fragments 1 + 2 (a marker of thrombin generation), and D-dimer (a marker of fibrin degradation) [14, 29]. Despite all these facts, thrombotic complications during the attacks, or increased thrombotic risk in patients with C1-INH-HAE are not characteristic, even though the activation of the contact system is poorly regulated. It is hypothesized that in patients with C1-INH-HAE, the activation of factor XII preferentially triggers prekallikrein activation, rather than the production of FXIa by FXIIa.

Although the studies mentioned in the foregoing were milestones in the exploration of the pathomechanism of C1-INH-HAE, they had their limitations. They enrolled only a small number of C1-INH-HAE patients, and usually did not compare the symptom-free and during-attack states of the same patients. Furthermore, measuring the components of the plasma enzyme systems separately has the disadvantage that the results of the assays do neither reflect the true balance between the activators and inhibitors in the original blood sample, nor the interaction of the four plasma enzyme systems. In our study, we tried to overcome these limitations by the simultaneous evaluation of the changes occurring in the coagulation, fibrinolytic, and kinin-kalllikrein systems during C1-INH-HAE episodes. Furthermore, we compared this large set of parameters in the same patients during symptom-free periods and during attacks – that is, each patient served as his/her own control.

Our aim was to perform a comprehensive study in a large population of C1-INH-HAE patients, in order to investigate the changes occurring in the plasma enzyme systems during edematous attacks. In the current study, we simultaneously evaluated such changes in the coagulation, the fibrinolytic, and the contact systems. Furthermore, we compared this large set of parameters recorded from the same patients in symptom-free periods and during attacks – that is, each patient served as his/her own control. The design of our study enabled us to conduct an investigation of unprecedented comprehensiveness and complexity into the changes accompanying C1-INH-HAE attacks. Figure 6 provides a comprehensive and schematic representation of the changes that occur in the plasma enzyme systems during edematous attacks.

In recent years, many studies have been published on the activation of the plasma enzyme systems during hereditary angioedema attacks. However, these included only a small number of patients and in most cases, the “symptom-free” and the “during attack” samples were obtained from different subjects.

Based on our results, the activity of factor XII is higher in the patients during the symptom-free period than in the healthy subjects and it increases further significantly during attacks; the latter is in accord with previous findings [14, 32].

Although a previous study did not detect any difference between healthy controls and C1-INH-HAE patients as regards the level of factor XI [33], we ascertained significant elevation of the latter in symptom-free C1-INH-HAE patients. However, we could not detect any further increase during attacks, compared with the symptom-free period.

In agreement with previous studies [29, 34], the level of prothrombin fragments 1 + 2 was higher in symptom-free C1-INH-HAE patients than in healthy controls, and it increased further in the same patients during attacks. Intriguingly, patients with C1-INH deficiency and with elevated levels of prothrombin fragments 1 + 2 (which indicate a state of hypercoagulability) do not have an increased risk for thrombotic events, as would be expected in such a situation. We had not measured the levels of coagulation inhibitors and thus, could not ascertain the magnitude and direction of their changes – or the possible development of a steady state of some kind. It is conceivable that the hyperfibrinolytic state, which we confirmed in our current study, and that was demonstrated in previous studies as the elevation of D-dimer levels – is sufficient to counteract the observed hypercoagulability [14, 33, 35]. D-dimer is produced during the breakdown of the fibrin mesh by plasmin. Therefore, in the state of hypercoagulation, any clot that forms is cleared by the fibrinolytic activity and thus, D-dimer is generated. In the context of our study, this may mean secondary hyperfibrinolysis, because the plasminogen levels measured during symptom-free periods did not differ from those determined in the healthy controls, or from those recorded during edematous attacks. If a major increase of hyperfibrinolysis would have occurred, we should have detected decreased plasminogen levels. Hypercoagulation, on the other hand, appeared prominent – because its presence was indicated even by the reduction of prothrombin time and of aPTT. It should be noted that the reduction of clotting times can be shown only occasionally, although this succeeds sometimes in disseminated intravascular coagulation [36]. In addition to the decreased clotting times, the elevated levels of the markers of the activation of coagulation were also characteristic in our study during the attacks. Of these, prothrombin fragments 1 + 2 indicates the formation of thrombin, whereas the elevation of TAT level reflects the neutralization of the formed thrombin by antithrombin. As thromboembolic events did not occur in our patients, it is possible that the thrombin thus formed is kept under control by the inhibitors.

The decrease of thrombin time indicates activation of the extrinsic pathway in the first place. However, judged by the activation of factors XII and XI, activation of the intrinsic pathway of blood coagulation is more typical. This is the possible reason why the reduction of aPTT is more conspicuous in symptom-free periods – and much more during attacks – even in comparison with healthy controls. To our best knowledge, activated partial thromboplastin time was not studied previously. In our study, we could detect a decreased level during the symptom-free period, compared to the healthy subjects, and its level decreased further during attacks.

One previous study did not find any difference between the TAT levels of healthy controls and of symptom-free patients [33]. Notwithstanding this, we detected a stepwise elevation in its level, both in the symptom-free period, and during the attacks. Thrombin generation appears to increase both in symptom-free/symptomatic C1-INH-HAE patients (as shown by the elevated level of the relevant plasma marker, prothrombin fragments 1 + 2). However, the effect of thrombin on vascular permeability may be weaker in vivo, owing to its rapid inactivation by anti-thrombin – as shown by the elevated levels of the TAT complex in our study. This inactivation may also explain why C1-INH-HAE patients do not develop thrombosis during edematous attacks.

To our best knowledge, the levels of fibrinogen and plasminogen have not yet been studied before in C1-INH-HAE patients. In our study, we could not find any difference in these parameters between healthy subjects and C1-INH-HAE patients. However, fibrinogen level was slightly elevated during attacks, compared with the symptom-free period – this may be explained by the fact that fibrinogen is an acute phase protein [37, 38]. Remarkably, PTT decreases during an acute phase reaction, just as we have seen in the C1-INH-HAE patients during attacks. Thus, the latter may be accompanied by an acute phase reaction.

In our study, D-dimer (the plasma marker of fibrin degradation) showed a significant elevation in the symptom-free period compared with the controls; this is in agreement with findings from previous studies [29, 34]. Interestingly, its level increased further during attacks of the same patients – as shown by a previous study as well [35]. At variance with a previous study, the level of PAI-1 decreased continuously in the symptom-free period and even further during attacks. However, no difference was found previously between the symptom-free and the symptomatic period [34]. As discussed by van Geffen et al., decreased levels of PAI-1 may be a consequence of enhanced thrombin production in association with protein C activation, which leads to the consumption of PAI-1 [34].

Taken together, our study confirmed the results of previous research into the activation of the coagulation system (as shown previously by elevated levels of prothrombin fragments 1 + 2) [33], as well as into the activation of the fibrinolytic system (as shown previously by the increased concentrations of D-dimer, TAFI, thrombomodulin and plasmin-α2-antiplasmin complex) [29, 34] in the symptom-free period.

Unlike the earlier studies [29, 33, 34], we detected differences in multiple parameters (factor XI, D-dimer, PAI-1) between the symptom-free and the during-attack states. The likely explanation for these is the small number of patients evaluated in the previous studies, of which only a few analyzed blood samples obtained from the same patients during the symptom-free period as well as during attacks.

The design of our study – that is, simultaneous evaluation of the components of the various plasma enzyme systems in the same patients – enabled us to investigate the correlations within and between the individual enzyme systems as far as the study parameters are concerned. We described for the first time that in C1-INH-HAE patients, the interactions among the plasma enzyme systems change both during symptom-free periods and during attacks, as suggested by the diverse correlation patterns, compared with the healthy subjects. The correlation between aPTT and factor XI or factor XII was present in all three sets of data (healthy subjects, symptom-free C1-INH-HAE, during attack C1-INH-HAE). Factor XII and aPTT exhibited a significant, negative correlation (which was the weakest in healthy subjects). This means that the higher is the level of factor XII, the shorter (smaller) is aPTT, and coagulation time decreases accordingly. However, this mechanism is possibly counterbalanced by the activation of the fibrinolytic system, as suggested by our results and by previous studies [14, 33, 35]. In healthy subjects, the C1-inhibitor – the key regulator in C1-INH-HAE – correlated with aPTT, the activity of factor XI, and with fibrinogen level. Interestingly, C1-INH activity showed no correlation in the symptom-free period, but during attacks, its activity exhibited a significant negative correlation with prothrombin time, or with the level of prothrombin fragments 1 + 2. The latter observation may indicate the involvement of the activation of the coagulation system during C1-INH-HAE attacks, in agreement with the findings of a previous study [34]. During attacks, the relationship between prothrombin fragments 1 + 2 and TAT became stronger than that observed in the symptom-free period. Despite this, the relationship between the levels of the D-dimer and of the prothrombin fragments 1 + 2 became weaker during attacks, compared with that seen in the symptom-free period. We can conclude that the strongest correlations were found in HAE-C1-INH patients during the symptom-free period.

The novelties of our study include the comparison of several plasma enzyme system components between attacks occurring at single vs. multiple sites, and in subcutaneous vs. submucosal locations. During attacks with multiple sites, we observed enhanced activation of the plasma enzyme systems, compared with attacks restricted to a single site. This was reflected by the significant elevation of D-dimer levels (that was also shown by Reshef et al.) [35], and by the decreased prothrombin time and plasminogen concentration in multiple-site attacks. Interestingly, we could not find any difference in the study parameters between subcutaneous and submucosal attacks. To our best knowledge, two studies have analyzed so far the possible differences between subcutaneous vs. submucosal attacks [29, 35], and only one of them showed significant difference in the level of D-dimer between these two locations [35].

As several during-attack samples were available from the same C1-INH-HAE patients, we analyzed whether there is a fluctuation between the attacks occurring in different patients, or between the successive attacks of the same patient. Interestingly, we found differences between the patients as regards the patterns of coagulation and fibrinolysis markers, namely in the levels of fibrinogen, thrombin-antithrombin complex, D-dimer and C1-inhibitor showed large inter-individual variations. Furthermore, D-dimer level and the activity of the C1-inhibitor exhibited a large fluctuation between the successive attacks of the same patient.

The strengths of our study include the comparison of a large number of paired samples from symptom-free periods or from edematous episodes of the same patients, as well as the parallel monitoring of the parameters of the plasma enzyme systems. This approach allowed accurate appraisal of the changes occurring during C1-INH-HAE attacks. Moreover, our study pointed out that the individual edematous episodes may be characterized by different marker patterns among the studied individuals. We also showed that attacks with multiple sites are characterized by more enhanced activation of the plasma enzyme systems, compared with attacks restricted to a single site.