Hereditary angioedema (HAE) is an inherited disorder clinically characterized by recurrent and self-limited episodes of localized edema in various organs. Clinical symptoms include skin swellings, abdominal pain attacks, tongue swellings and potentially life-threatening laryngeal attacks. Types of HAE include the classical HAE due to the deficiency of functional C1 inhibitor (HAE-C1-INH), and various new types of HAE with normal activity of C1-INH (HAE with normal C1-INH, HAEnCI). Currently, 6 types of HAEnCI are recognized, based on underlying mutations of factor XII (HAE-FXII) (Dewald and Bork, 2006; Bork et al., 2011), angiopoietin-1 (HAE-ANGPT1) (Bafunno et al., 2018), plasminogen (HAE-PLG) (Bork et al., 2018), kininogen 1 (HAE-KNG1) (Bork et al., 2019), myoferlin (HAE-MYOF) (Ariano et al., 2020), and heparan sulfate-glucosamine 3-O-sulfotransferase 6 (HAE-HS3ST6) (Bork et al., 2021).

Bradykinin, a 9-amino acid peptide chain generated by proteolytic cleavage of its precursor high molecular weight kininogen (HMWK), is considered to be the main mediator of HAE (Kaplan and Joseph, 2010). It is well described how the kallikrein-kinin system (KKS) leads to bradykinin release: binding of negatively charged macromolecules to factor XII induces a conformational change causing activation. As activated factor XII (factor XIIa), it converts prekallikrein (PK) to kallikrein which cleaves HMWK to release bradykinin (Kaplan and Joseph, 2014). The presence of C1 inhibitor prevents this reaction and its deficiency in HAE-C1-INH results in an overproduction of bradykinin (De Maat et al., 2018). Mutations in HAE-FXII can introduce new sites in the mutant FXII protein that are sensitive to enzymatic cleavage by plasmin (de Maat et al., 2016). These FXII mutants rapidly activate after cleavage by plasmin, escape from inhibition through C1-INH, and elicit excessive bradykinin formation. Another study showed that the FXII mutants Thr328Lys and Thr328Arg are cleaved after residue 328 by coagulation proteases like thrombin and FXIa, producing a truncated form of FXII (delta-FXII) which converts PK to kallikrein more efficiently (Ivanov et al., 2019).

In HAE-PLG (Bork et al., 2018), there was a clinical response to a bradykinin B2 receptor 2 antagonist, an indirect sign for bradykinin involvement (Bork et al., 2020). It was unknown whether the KKS cascade pathway or another mechanism is linking the mutant protein to bradykinin overproduction. A recent study did show that plasminogen can directly release bradykinin from kininogens, and more efficiently if carrying the HAE mutation (Dickeson et al., 2022). Our findings in a large family and the following experiments, both presented here, are now shedding some more light on the pathomechanism and prove clinically that the KKS is bypassed for bradykinin production in HAE-PLG.

Here, we investigated a four-generation HAE-PLG family, in which affected family members carry the PLG mutation c.988A>G; p.K330E. Ten out of 23 individuals in this family were affected by HAE, DNA of seven of these affected individuals was sequenced. In these seven HAE patients the PLG mutation c.988A>G; p.K330E was identified. In addition to the HAE-causing PLG mutation, we identified an additional pathogenic mutation in the F12 gene in 6 family members (Figure 1). This additional F12 mutation (c.1681-1G>A) is not resulting in HAE, but has been identified in FXII deficiency (Schloesser et al., 1995). The F12 mutation was identified in five HAE patients (in combination with the PLG mutation) and one unaffected family member (wild type PLG) (Figure 1).

Individuals carrying the F12 mutation c.1681-1G>A in a heterozygous state do have, depending on the presence of the -4C/T polymorphism affecting the Kozak region of the F12 gene on the second allele (Bach et al., 2008), either about 25 or about 50% clotting activity. Individuals homozygous for the F12 c.1681-1A mutation do have no residual clotting activity (Figure 2).

Considering the role of FXII in the pathomechanism of HAE-C1-INH and HAE-FXII, we concluded that there would be an impact of the loss of one F12 allele if FXII is involved in the pathomechanism of HAE-PLG. Thus, we performed a clinical investigation based on the presence of the PLG and F12 mutations. We could not identify any differences in the clinical presentation of the five family members carrying the PLG mutation c.988A>G; p.K330E and the F12 mutation c.1681-1G>A compared to the two family members carrying only the PLG mutation c.988A>G; p.K330E (Table 1).

Based on these results, we concluded that it is possible that mutant plasminogen is not acting via the KKS to produce increased amounts of bradykinin. To investigate how the PLG mutation is causing HAE we first looked into plasminogen structure. The p.K330E mutation was modeled into the crystal structure of the native plasminogen (Xue et al., 2012). It disrupts the bonding with D328 that is one of the ligand binding residues in Kringle domain 3. The modified side chain and the disrupted D328 bonding cause a widening within the ligand binding groove formed by the hydrophobic surface of this domain (Figure 3). This likely allows for an altered proteolytic activity of the mutant plasminogen.

Next, we wanted to investigate if plasminogen is capable of cleaving HMWK in vivo and if the mutant is having a different activity in doing so. We incubated either purified HMWK alone, HMWK plus wt plasminogen, or HMWK plus mutant plasminogen at 37°C for different durations. To test if any degradation depends on proteinase activity, we did this incubation in parallel with and without proteinase inhibitor. Using SDS-Page and Western blot analysis we investigated the cleavage efficacy. We observed that all three samples showed no degradation when incubated in the presence of proteinase inhibitor (Figure 4, left side). Without proteinase inhibitor HMWK alone showed a weaker band at ∼125 kDa band, indicating self-cleavage activity. The signal of both HMWK samples incubated with plasminogen decreased drastically compared to HMWK alone, indicating a cleavage activity of both, wt and mutant plasminogen. There was an additional band at ∼70 kDa in these samples, which was stronger if incubated with wt plasminogen compared to mutant plasminogen (Figure 4, right side), indicating an increased HMWK cleavage activity in mutant plasminogen compared to wt plasminogen.

To investigate if this increased cleavage of HMWK also results in the generation of bradykinin, we analyzed these samples with MALDI-TOF-MS (Matrix-Assisted Laser-Desorption/Ionization Time-Of-Flight Mass-Spectrometry) to monitor the formation of the bradykinin molecule. Samples of HMWK incubated without proteinase inhibitor revealed distinct mass signatures at the calculated weight of bradykinin (1060.56). Control measurements with synthesized bradykinin showed a similar pattern. The presence of proteinase inhibitor resulted in the total absence of any bradykinin signal (Figure 5).

Since the identification of PLG mutations in HAE (HAE-PLG) (Bork et al., 2018), the pathomechanism how mutant plasminogen results in HAE has not been solved completely. There are indications for bradykinin involvement (Bork et al., 2020), and a recent study did not only demonstrate the possibility of plasminogen to directly cleave kininogens to release bradykinin, but also the increased efficiency of mutant plasminogen to do so (Dickeson et al., 2022). Yet, it was still not solved completely whether the KKS cascade is involved in bradykinin overproduction in HAE-PLG.

Here, we identified in a big HAE-PLG family individuals carrying an additional F12 mutation, effectively resulting in the loss of FXII expression from one allele, and thus is reducing the presence of FXII protein at least by half. FXII plays an important role in the KKS: activated FXII (FXIIa) converts PK to kallikrein, which cleaves HMWK to release bradykinin. We hypothesized that, if mutant plasminogen acts via the KKS, HAE patients with reduced amounts of FXII would have a milder phenotype, as the loss of FXII would counteract the effects of the plasminogen mutation. Clinical comparison of patients carrying only the PLG mutation with patients carrying the additional F12 mutation on top of the PLG mutation revealed that the presence of the additional F12 did have no impact on the clinical presentation of the HAE-PLG patients. This gave a clear indication that the pathology in HAE-PLG might not be achieved via the KKS.

The next question was how the PLG mutation could result in increased bradykinin production, if it is not via the KKS. About 3 decades ago it has been shown that plasminogen is in principle capable of cleaving HMWK (Kleniewski and Donaldson, 1987; Kleniewski et al., 1992). Interestingly, 8 years before the PLG mutation p.K330E was identified to be causative for HAE (Bork et al., 2018), the plasminogen amino acid K330 had been investigated (Christen et al., 2010). This position within the Kringle domain 3 of plasminogen is broadly conserved amongst vertebrates, however as glutamic acid (E, which is the mutation found in HAE-PLG). The authors concluded that at some point during human evolution glutamic acid (E) was mutated to lysine (K). They could show that reverting it to glutamic acid increased docking of lysine-like zwitterionic ligands (Christen et al., 2010). This suggests that plasminogen p.K330E could be a gain-of-function mutation. This was confirmed by a recent in vitro study showing that mutant plasminogen releases bradykinin more effectively than wild type plasminogen (Dickeson et al., 2022). Modeling of p.K330E mutation indicates a widening within the ligand binding groove of the Kringle domain 3, which is in accordance with earlier modeling attempts of this domain (Christen et al., 2010). This likely allows for an altered proteolytic activity and is in line with the gain-of-function theory. We could confirm the gain of function by the p.K330E mutation in vitro using purified HMWK which we incubated with purified wild type respectively mutant plasminogen. Mutant plasminogen did not only cleave HMWK more efficiently, but also resulted in an increased bradykinin production compared to wild type plasminogen. We observed more intense signals at the bradykinin mass in the mutant plasminogen group in all experiments.

In summary, the clinical, structural, and biochemical data lead us to the conclusion that the HAE-causing plasminogen mutation p.K330E results in a gain of function, which enables mutant plasminogen to cleave HMWK more efficiently than the wt protein and produce a critical amount of bradykinin by bypassing FXII/kallikrein. Thus, treatment approaches for HAE-PLG should be aimed directly at plasminogen.