Evidence for bradykinin release in chronic spontaneous urticaria

Chronic spontaneous urticaria (CSU) is characterized by recurrent itchy weals and/or angioedema and is believed to be driven by mast cell activation. It was shown that excessive mast cell activation during anaphylaxis initiates contact activation, resulting in bradykinin release. Evidence for bradykinin release was never demonstrated in CSU.


| INTRODUC TI ON
Chronic spontaneous urticaria (CSU) is hallmarked by recurrent itchy weals and/or angioedema. The pathophysiology of CSU is not fully understood. 1 Mast cell and histamine involvement are strongly implicated, and patients demonstrate clinical response to the anti-IgE monoclonal antibody omalizumab and antihistamine therapy. 2,3 Angioedema is a common symptom of CSU. When patients experience angioedema without weals, this CSU subtype can be referred to as idiopathic angioedema. 4 Angioedema results from prekallikrein (PK) into PKa that will cleave bradykinin from high molecular weight kininogen (HK) 6 ; this enzyme system is called the contact system. Therefore, C1-INH deficiency results in uncontrolled bradykinin production.
The various subtypes of angioedema are classified as bradykinin induced or mast cell induced. Clinically, bradykinin-and histamine-mediated swellings are indistinguishable. Especially when weals and angioedema coincide, swellings are considered mast cell and not bradykinin induced. 4,7 However, research in anaphylaxis suggests that mast cell activation and bradykinin release co-exist. A study performed in wasp-venom allergic individuals demonstrated increased markers of bradykinin generation. After a wasp sting, markers increased in those with severe anaphylaxis, for example angioedema and/or shock, while markers were low prior to the sting and in individuals with a mild response. 8 In addition, activation of the fibrinolytic system was also observed during anaphylaxis as plasmin-antiplasmin (PAP) complexes increased. 9 Later studies found increased levels of cleaved high molecular weight kininogen (cHK), a marker for bradykinin release, in patients visiting the emergency medicine department with a (food or drug) allergic response 10 and explained these observations demonstrating bradykinin involvement during anaphylaxis in mouse models. 10,11 Mast cell-released heparin was found to activate FXII 11 connecting mast cell activation to bradykinin release as FXIIa activity will lead to PKa activity and HK cleavage. While histamine is primarily released from mast cells or basophils, the contact system is always present in circulation.
Contact system activation is not only limited to excretion of mast cell-released heparin but can also be triggered by, among others, platelet polyphosphate, 12 activation of the endothelium 13,14 and neutrophil extracellular traps. 15 Little is known about bradykinin generation in CSU, as it is generally considered a mast cell-and histamine-mediated disease. We recently developed a sensitive ELISA-based method to assess cHK generation. 16 Using this method, we revisited cHK generation in CSU. In addition, C1-INH in complex with FXIIa or PKa were determined, as well as markers of the fibrinolytic system as these were previously associated with mast cell activation in anaphylaxis. 9 2 | ME THODS

| Study participants and blood collection
All patients with a diagnosis of CSU defined as having spontaneous, recurrent weals and/or angioedema for at least 6 weeks, visiting the dermatology and allergology outpatient department, were asked to provide blood samples at first and selected follow-up visits. This study was approved by the local ethical committee (proto-col#13-272). Of these, all patients with written informed consent and at least one collected sample at first visit between January 2014 and February 2016 were included and data were analysed for this study. Patient characteristics (age, sex, CSU subtype, disease severity, treatment) were retrieved from medical records by the physician of the research team. In CSU patients that presented with angioedema without weals, HAE-1 and HAE-2 were excluded based upon medical history, family history and C4 levels as part of routine clinical care. This CSU subtype is referred to as idiopathic angioedema throughout this article. 4 Disease severity was categorized into 1) symptom-free: complete disease control for at least one month, 2) mild: recent symptoms but no need for additional treatment and 3) moderate to severe: recent symptoms, need for additional treatment and/or frequent use of rescue medication.
When available, the urticaria control test (UCT) score was collected. 17 During the timeframe of this study, the UCT was not yet collected routinely at every visit. Antihistamine resistance was defined as persistent symptoms despite using a four times daily dose Clinical relevance: If elevated bradykinin generation has clinical implications in the pathology of CSU is open to debate.

K E Y W O R D S
angioedema, bradykinin, chronic spontaneous urticaria, cleaved high molecular weight kininogen, idiopathic angioedema, plasmin-antiplasmin, soluble urokinase receptor of antihistamines; need for add-on therapy such as cyclosporine and omalizumab; or frequent need for oral steroids in addition to antihistamine prophylaxis.
Venipuncture was performed at inclusion during a regular control visit at the outpatient clinic. One additional blood draw was performed in 20 patients during random control visits, two additional blood draws in two patients and three additional blood draws in one patient.
Blood was collected in standard blood collection tubes containing 1.8 mg EDTA per ml blood and specific sample collection/ anticoagulant tubes (SCAT) containing 25 µM PPACK (Phe-Pro-Arg-chloromethylketone, Sigma-Aldrich), 11 mM sodium citrate and 0.1% mannitol (w/v) to prevent post-blood draw enzymatic activity in samples. Blood was centrifuged at 2000 g for 10 minutes shortly after blood draw and stored at −80 until use. We previously reported cHK measurements in HAE-1 and HAE-2. Samples were collected in serine protease inhibitor-containing tubes. 16 As we collected our samples from CSU patients in the same way, we used these HAE analyses for comparison.
In addition, plasma from healthy donors was collected (with written informed consent; approval by the local ethical committee of the University Medical Center Utrecht; protocol#07-125).
Blood was collected in standard sodium citrate tubes (10% sodium citrate, 3.2% wt/vol). Blood was centrifuged twice at 2000 g for 10 minutes shortly after blood draw and stored at −80 until use.
Plasma from ~30 healthy donors was pooled for control pooled plasma, and plasma from 28 patients was stored individually for control plasma.

| Biomarker assays
cHK, 16 FXIIa-C1-INH complexes, 18 PKa-C1-INH complexes, 18  One hundred % cHK reflects total HK cleavage in control pooled plasma generated by incubating plasma with 1 µg/ml bèta-FXIIa (Hematologic Technology) at 37ºC for 10 minutes after which reaction was stopped by diluting plasma 64× in a phosphate-buffered saline buffer (mPBSt: 127.9 mmol/L NaCl, 6.2 mmol/L Na 2 HPO 4 , 3.7 mmol/L NaH 2 PO 4 , pH 7.0 supplemented with 0.1% Tween-20 wt/vol and 1% skimmed milk powder wt/vol, containing 50 µmol/L PPACK). One hundred % C1-INH complex indicates control pooled plasma incubated with dextran sulphate Mr ~ 500 000 (Sigma-Aldrich) at 37ºC for 30 minutes after which reaction was stopped by diluting 32× in mPBSt containing 50 µmol/L PPACK. Calibration curves were created by mixing 100% activated plasma with unactivated control pooled plasma both diluted in mPBSt containing PPACK. Plasma dilution in calibration curves was equal to plasma dilution of patient samples. 16,18 For 25 patients, tryptase levels were determined as a routine diagnostic via ImmunoCAP using the Phadia250 (Thermo Fisher Scientific). These levels were retrieved from medical records. For 14/25 patients, blood was collected for tryptase determination at the same day as collection of study samples that were used for cHK determination.

| Data analysis
GraphPad Prism 8 was used for data analysis. cHK and C1-INH complex levels were interpolated from a calibration curve using a sigmoidal 4PL fit model. The first collected sample per patient was included for analysis unless indicated otherwise. Groups were compared using Mann-Whitney t test or Wilcoxon test for paired samples and multiple groups with Kruskal-Wallis test using Dunn's correction for multiple testing. Correlation was tested with Spearman's rank test.
Eighteen patients (20%) did not receive any prophylaxis but all had on-demand treatment available also mainly consisting of antihistamines (Table 1). Most patients (n = 59, 67%) described moderate to severe symptoms at first visit, needing additional therapy. Four (5%) patients were symptom-free at first visit. When analysing all visits including follow-up visits, 16 (14%) patients reported to be symptom-free (Table 1).
When classifying by disease severity, cHK levels in healthy controls were significantly lower than in CSU patients with mild symptoms (median cHK 10.6% range 1.4%-17.2%, P < .0001, Figure 1B) and moderate to severe symptoms (median cHK 9.0% range 2.7%-18.71%, P = .0043) but not compared to symptom-free patients (median cHK 6.5% range 1.5%-20.8%, P > .9999, Figure 1B). cHK levels in symptom-free patients were significantly lower than in patients with mild symptoms (P = .0296) and lower compared to patients with moderate to severe symptoms. However, the difference between symptom-free patients and those with moderate to severe symptoms was not statistically significant (P = .3018, Figure 1B). An urticaria control test (UCT) score was reported at 31 visits. The average score was 11 (SD ± 5.6) and trended towards a negative correlation with cHK (P = .06, r = −.28, Supplemental Figure 1). There were no differences in cHK levels among patients with weals, weals and angioedema or idiopathic angioedema (P = .91, Figure 1C).
In three patients with idiopathic angioedema, one or more samples were collected during an angioedema attack. All of these patients experienced a high attack frequency with 2 to 10 attacks monthly while using antihistamine prophylaxis of 1-4 times the daily recommended dose. In addition, one of these patients received tranexamic acid and one tranexamic acid and omalizumab which only led to partial improvement of symptoms. No increase in cHK could be observed in samples collected during an angioedema attack compared to samples in between attacks (P = .42, Figure 1D).
Furthermore, no differences could be observed among CSU patients with good response to prophylactic antihistamine therapy and patients with insufficient response to antihistamine therapy (P = .9269, Figure 1E). Tryptase levels were routinely determined in 25 patients as part of their diagnostic work-up to rule out systemic mastocytosis. In 24 patients, tryptase was within the normal range (defined as levels < 11.4 µg/L), and in one patient, tryptase levels were elevated but further diagnostic work-up did not change the CSU diagnosis. For 14 patients, tryptase was determined at the moment of sample collection for cHK measurement. When analysing these 14 samples, we found a moderate negative correlation (r = −.65, P = .0137, n = 14, Figure 1F Distribution of cHK levels among age and sex was analysed but no relation with age or sex was found ( Figure S3).
We analysed changes in cHK levels detected in samples collected from the same patients over time. In only 11 of the 23 cases, disease severity corresponded to cHK levels ( Figure S4: A Figure S4 (red lines indicate disease severity, and blue lines cHK).

| Plasmin-antiplasmin levels are increased in CSU patients
Activation of the fibrinolytic system was analysed by measurement of PAP complexes and suPAR levels. We observed increased PAP levels in CSU patients compared to healthy controls (P =< .0001). This difference was present for symptom-free patients (P = .0021), patients with mild symptoms (P = .0002) and patients with moderate to severe symptoms (P < .0001, Figure 3A). There was no association with disease phenotype (P = .28, Figure 3B) or response to antihistamine therapy (P = .21, Figure 3C). Soluble uPAR levels were only determined in patient samples as reference ranges from healthy control populations were available. Measurements were done in 116 samples. In only two patients, levels were slightly increased and there was no difference in suPAR levels among the disease severity groups (P = .24, Figure 3D) or disease phenotypes (P = .42, Figure 3E). Furthermore, PAP and suPAR did not correlate to cHK levels or each other ( Figure 4A-C).

| D ISCUSS I ON
We evaluated if there is evidence for bradykinin release in CSU and found that cHK, a marker for bradykinin production, was elevated in CSU patients compared to healthy controls. cHK levels normalized in patients that were in disease remission. Moreover, the observed increases in cHK levels were comparable to previously determined levels in HAE-1 and HAE-2. Increased cHK was not limited to CSU patients with angioedema and not associated with antihistamine responsiveness. However, cHK was inversely correlated to tryptase levels. In addition, PAP-complex levels were also increased reflecting fibrinolytic activity in CSU.
The increase in cHK levels observed in this CSU cohort was within the healthy control range, roughly not surpassing 20% cHK.
This raises the question if the cHK increase observed is of possible clinical relevance. To some extent, bradykinin release appears to be a physiological process, as cHK is detectable in healthy individuals. Cleaved HK only moderately discriminates between healthy individuals and those with bradykinin driven disease (HAE-1 and HAE-2). However, the cHK increase detected in CSU was comparable to levels detected in HAE-1 and HAE-2 patients, suggesting possible clinical relevance. The here-reported cHK levels in HAE-1 and HAE-2 are lower than described in previous studies, 16,21,22 and this can be explained by the use of protease inhibitor blood collection tubes in this study preventing post-blood draw ex vivo cHK generation. Moreover, we previously found that a slight rise of cHK within the range of healthy controls could indicate an angioedema attack in individual HAE patients. 16,21 It is therefore possible that the overall slightly increased cHK levels in CSU do reflect pathological bradykinin release.
Previous studies to CSU only investigated bradykinin levels or HK degradation in very small numbers of patients and were limited to cases with angioedema. One study reported no decrease in HK antigen on Western blot in 7 cases with idiopathic angioedema. 10 This different outcome can be explained by the sensitivity of cHK detection versus loss of HK antigen on Western blot. Another study observed decreased HK levels in 10 patients with angioedema provoked by oral contraceptive intake, notably 8 of them also reported weals, symptoms improved and HK normalized after cessation of oral contraceptives. 23 Evidence is surfacing that bradykinin contributes to idiopathic angioedema that is unresponsive to antihistamine therapy (idiopathic non-histaminergic angioedema). Increased bradykinin levels were detected in four cases of idiopathic non-histaminergic angioedema during angioedema attacks. 24 Moreover, ex vivo plasma stimulation with a FXII activator in patients with idiopathic non-histaminergic angioedema resulted in increased PKa activity, 25 suggesting increased sensitivity to bradykinin production. Both studies did not found increased bradykinin levels or sensitivity to bradykinin production in idiopathic angioedema patients that did responded to antihistamine therapy (idiopathic histaminergic angioedema). 25,26 We determined cHK in three patients with idiopathic angioedema during an acute attack but did not observe a further increase in cHK. All three showed no clear benefit from antihistamine therapy but two continued to have a partial and complete response to omalizumab suggesting mast cell involvement in their disease pathology. Our findings appear to be in line with previous observations in idiopathic histaminergic angioedema as the cHK increase detected is subtle and within the range of healthy controls. However, using our sensitive method of cHK detection we observed a possible clinically relevant difference and are the first to report elevated cHK levels in CSU, unrelated to the occurrence of angioedema.
We hypothesized that mast cell activity is the cause of increased cHK levels in CSU. A causal relation between mast cell activity and contact activation in allergy was repeatedly demonstrated before. [9][10][11]27 Tryptase is a marker for mast cell degranulation that is used in the clinic as a biomarker in anaphylaxis and mastocytosis. In this study, we found that cHK negatively correlated to tryptase levels. At the moment, we do not know the mechanism behind this inverse correlation. However, it is noteworthy that human mast cell tryptase has the ability to cleave HK at multiple sites. 28 It is possible that this destroys the immunoreactive epitope for the antibodies that recognize cHK in our ELISA.
Furthermore, it remains uncertain if mast cell degranulation is the source of cHK production in CSU. Previous work showed a positive correlation between tryptase and cHK in anaphylaxis. 10 To confirm our finding in CSU, replication in a larger cohort is necessary, as only limited data were available.
Involvement of the fibrinolytic system in CSU is topic of debate. The fibrin degradation product D-dimer is associated with CSU, and it correlates with disease severity and antihistamine-resistant urticaria. 29,30 However, urokinase-type plasminogen activator (uPA), suPAR and plasminogen activator inhibitor-1 were not elevated in one study in CSU patients. 31 We add to these previous observations our finding of increased PAP-complex levels in CSU compared to healthy controls. There is a large bout of evidence demonstrating that activation of the fibrinolytic system and contact system go hand in hand. 9 Mast cell-released heparin is a likely source of contact activation in CSU, but we cannot exclude involvement of other triggers for bradykinin production such as platelets 12 or endothelial cells. 13,14 We here report on increased cHK levels in CSU reflecting increased bradykinin release. CSU is considered a multifactorial, mast cell-mediated disease, and we cautiously introduce the idea of including bradykinin in this multifactorial model. If and to what extent bradykinin indeed contributes to CSU remains to be answered.

ACK N OWLED G EM ENTS
We thank Ans Lebens, Jos Beutler, Stans den Hartog Jager, Stefan Nierkens and Edward Knol for their assistance in setting up the biobank, patient inclusion and sample collection. We thank Arjan Barendrecht for advice on the ELISA set-up. We thank Bram Nuiten from ELITechGroup for suPAR determinations. All authors read and approved the final manuscript.

DATA AVA I L A B L E S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.

R E FE R E N C E S F I G U R E 4
Cleaved HK levels and biomarkers of the fibrinolytic system do not correlate. Cleaved HK (cHK), plasmin-antiplasmin (PAP) and soluble urokinase-type plasminogen activator receptor (suPAR) were determined by ELISA. cHK plotted against A, PAP levels and B, suPAR levels. C, suPAR levels plotted against PAP levels. Spearman's rank test was used