Key messages

The focus of this study is to identify the genes modulated during the acute attack of HAE that might have a pathogenetic relevance in this rare disease.

Capsule summary

During an acute attack of HAE the up-regulation of two genes, ADM and uPAR, occurs and the products of these genes might be involved in the regulation of vascular tone and edema formation.

Key words:

Hereditary angioedema, C1 inhibitor deficiency, peripheral blood mononuclear cells, genes, acute attacks, vascular permeability; plasmin (3,4). Other forms of HAE with normal C1-INH have been classified (5)(6)(7). Some of these patients carry mutation in Factor XII gene (FXII-HAE), in others the genetic defect is still unknown and inheritance is derived from the segregation of symptoms (8). In one family a mutation in angiopoietin 1 gene has been recently shown to segregate with angioedema symptoms (9). Severe episodes of edema localized at the sub-cutaneous and mucosal layers characterize the acute attacks of HAE that can be extremely debilitating, particularly if the airways are affected (10). The symptoms are caused by the local leakage of fluids from the capillaries as a result of uncontrolled activation of the plasma contact system and generation of vasoactive mediators (11)(12). C1-INH is a serine protease inhibitor, which acts as a modulator of different pathways such as coagulation and fibrinolytic cascades, complement and contact-kinins systems (13)(14)(15). During the attacks of HAE, the uncontrolled activation of these pathways is enhanced as a result of endothelial cell activation (16) and generates factors that increase vascular permeability, provoking edema and inflammation (17). It is now well established that bradykinin is the principal mediator of symptoms of acute attacks of angioedema and that local trauma or emotional stress can represent triggering factors for an acute attack (18)(19)(20)(21). However, the molecular events that culminate with the release of bradykinin and generation of tissue edema are not completely understood (22)(23).

Different evidences (24)(25) report that biomarkers for coagulation system activation (FXII) and for fibrinolytic system activation (plasmin) are functionally linked to bradykinin production (26) associated with endothelial cell activation (27).

Finally, data from the literature suggest that leukocytes might be involved in the pathogenesis of edema formation (28) and cytokines are actively released in HAE (29). Circulating and tissue infiltrating leukocytes can migrate at cutaneous and mucosal levels where they can produce C1-IHN and other Complement factors and might potentially contribute to edema formation (30).

In this study we analyzed the transcriptomic profile of patients during the acute attack of HAE in comparison with the same patients in the remission phase and also comparing HAE patients with healthy subjects in order to identify the genes that may have a pathogenetic relevance in HAE.

Methods

Patient recruitment

We studied 8 patients with C1-INH-HAE (2 men and 6 women, aged 46.8 ± 22.2 years [mean ± standard deviation (SD), range 78-16]) during the remission phase (HAE-R) (free of attacks for at least 1 month) and during the acute attack (HAE-A) involving oral submucosal tissues in 1 patient, subcutaneous tissue in 4 patients and the abdomen in 3 (Table 1A). The diagnosis of abdominal acute attack in the patients with known C1-INH deficiency was based on the presence of acute abdominal pain with or without vomiting and diarrhea, the absence of any other cause of abdominal pain and prompt reversal upon the infusion of C1-INH. The control group consisted of 8 healthy subjects (HS, 2 men and 6 women) with a mean age of 50.8 ± 11.7 years (range 60-24). Blood samples were collected in PAX-gene Blood RNA tubes (Becton Dickinson, Italia) and sodium citrate as anti-coagulant. Blood samples were collected from these patients within 5 hours from the beginning of symptoms. All patients provided written informed consent in accordance with the Declaration of Helsinki. The protocol was approved by the Institutional Review Board of Milan, "Luigi Sacco" Hospital.

Laboratory Measurements

The following measurements were performed in non-activated plasma: C1-INH activity was measured using a chromogenic assay (Technochrom C1-inhibitor; Techno-clone GmbH, Vienna, Austria), and C1-INH antigen by means of radial immunodiffusion (RID) (NOR-Partigen, Behring, Marburg, Germany) (Table 1B).

RNA extraction

For RNA isolation, 2.5 ml blood was drawn in PAX-gene Blood RNA tubes, incubated for two hours at room temperature and stored at -20°C. Tubes were thawed overnight at room temperature prior to RNA isolation. Total RNA was isolated using PAX-gene Blood RNA Kit (PreAnalytiX, Hombrechtikon, Switzerland) according to the manufacturer's instructions. Total RNA concentration was measured using the Nanodrop spectrophotometer (Nanodrop Technologies, Wilmington, DE, USA) and RNA purity and integrity was verified using lab-on-chip technology (Agilent 2100 Bioanalyzer, Palo Alto, CA, USA).

Gene expression profiling

Transcriptome data were generated using the HumanHT-12 v3 Expression BeadChip (Release 38, Illumina, San Diego, CA. USA). In this process, 500 ng total RNA was used to synthesize biotinlabeled cRNA using the Illumina®TotalPrep™ RNA amplification kit (Applied biosystems/Ambion, USA). Quality of labelled cRNA was measured using the NanoDrop® ND-100 spectrophotometer and the Agilent 2100 Bioanalyzer. 750 ng biotinylated cRNA was used for hybridization to gene-specific probes on the Illumina microarrays. The Illumina arrays were then scanned with the HiScanSQ.

Microarray statistical analyses were performed by Genespring GX 11.0 software (Agilent Tech Inc., Santa Clara, CA, USA). Identification of genes differentially expressed between HAE-A and HAE-R patients was carried out with false discovery rate (FDR) method of Benjamini-Hochberg (31) and gene probe sets were filtered on the basis of the,FDR, (adjusted-P value with multiple testing on 1000 permutations) and fold-change (FC) (32). Fold change filter was set to 1.5-fold in each comparison. Only genes that were significantly (adjusted-P value <0.05 and FC >1.5) modulated were considered for further analysis.

Moreover, gene set enrichment analysis (GSEA) was performed in pairwise comparisons of HAE-A versus HAE-R patients. GSEA analyzes gene expression data and determines whether a particular set of genes is over-or under-represented in the compared samples (31). C2 curated gene sets from the Broad Institute (http://www.broad.mit.edu/index.html), based on prior biological knowledge, were used for the analysis. Significance of differential expression, as determined by the enrichment analysis, was recalculated 1000 times. A corrected P-value was obtained from the analysis using the FDR q-value correction. On the basis of this correction, the cutoff for significance was established at a P-value <0.05.

To assess biological relationships among differently regulated genes, we used the Ingenuity Pathway Analysis software (IPA, Ingenuity System, Redwood City, CA, USA; http://www.ingenuity. com). The reference gene selection was performed by own software written in Java program language. IPA generates networks based on the connectivity of the genes and computes a score for each network according to the fit of the set of supplied focus genes. These scores indicate the likelihood of focus genes to belong to a network versus those obtained by chance. The canonical pathways generated by IPA are the most significant for the uploaded data set.

Fischer's exact test with FDR option was used to calculate the significance of the canonical pathway.

The Illumina microarray data are MIAME (Minimum Information About a Microarray Experiment) compliant and the raw data have been deposited in the database The European Bioinformatics Institute (EMBL-EBI) and are accessible through Experiment ArrayExpress accession: E-MTAB-5897 .

cDNA synthesis and quantitative real-time PCR analysis

Quantitative real-time PCR (RT-PCR) was performed on total RNA samples from the patients in the validation group (n=20) using an ABI Prism 7900HT Sequence detection system (Applied Biosystems, Foster City, CA, USA). RNA (500 µg) was reverse transcribed into cDNA using a cDNA synthesis kit (QuantiTect Reverse Transcription Kit, Qiagen) according to the manufacturer's instructions. RT-PCR amplification reactions were performed using the QuantiTect Primer Assay and the Quantifast SYBR Green PCRmix (Qiagen). Gene expression levels of one housekeeping gene, glyceraldehyde-3-phosphate dehydrogenase (GAPDH), ADM and uPAR were determined. Expression levels of target genes were expressed relative to GAPDH.

Cell culture

The Jurkat T-cell lines, purchased from ATCC, were cultured in RPMI 1640 medium supplemented with 10% fetal bovine serum and 2 mM glutamine, at 37°C and 5% CO2 and 100 µg/ml penicillin and streptomycin (Sigma Aldrich, USA).

Bradykinin ELISA

The measurement of Bradykinin in cell lysate and cell culture supernatant was assessed using a human Bradykinin enzyme-linked immunosorbent assay (ELISA) kit (Cloud-Clone Corp.,USA), according to the manufacturer's instructions. The cells, seeded in 6-well plates at a density of 2 × 10 6 cells/well, were pre-incubated with 10µg/ml and 20µg/ml anti-uPAR antibody (Immunological Sciences, Italy), to neutralize the uPA-uPAR interaction, for 1 hour and then treated with 100 nM urokinase-type Plasminogen Activator (uPA) (Urokinasi Hospira, Italy), for 15 minutes. Cell lysates and their supernatants were subsequently collected. Bradykinin concentration was measured at 450 nm and calculated from a standard curve.

Statistical analysis and bioinformatics

For the remaining experimental analysis, we analyzed data with statistical software (GraphPad Prism; GraphPad, San Diego, CA, USA). All data were expressed as the mean ± SD of data obtained from at least three independent experiments. Statistical analysis was carried out using twotailed paired and unpaired Student's t test and one-way ANOVA and Tukey post-hoc tests. P-values lower than 0.05 were considered statistically significant.

Results

Differences in gene expression during the acute attack of hereditary angioedema

In the first step of the study, in order to identify genes/pathways specifically modulated during acute attacks, we compared the whole-genome gene expression profiles of HAE patients' PBMC collected during the remission phase (HAE-R) with samples from the same patients during the acute attack (HAE-A). We identified 23 genes significantly modulated during the acute attack with a FDR <0.05 and a FC >1.5 (Table 2). These genes were all upregulated during the acute attacks and the Principal Component Analysis (PCA, Figure 1A) showed that these 23 genes well discriminated the acute attack and the remission phase. Then, to determine whether in HAE-A patients there was a coordinated expression or "enrichment" in a set of functionally related genes, we performed a gene set enrichment analysis (GSEA). This computational method allowed us to determine whether an a priori defined set of genes (involved in a specific pathway or biological process) shows statistically significant, concordant differences between PBMCs from HAE-A and HAE-R patients. We identified 10 gene sets, each identifying a specific biological process, significantly modulated (q<0.05) during acute attacks (Table 3). Specifically, the sets of genes up-regulated during the acute attack were mostly associated with "response to external stimulus" and "protein processing" processes (q<0.05).Among genes contributing to "response to external stimulus" process, we found uPAR, called also PLAUR, and ADM (supplemental Table 1). These two genes are factors involved in the regulation of vascular tone and could be involved during the acute HAE attack.

Pathway analysis

We performed a pathway analysis to get a global characterization of the biological functions overrepresented by the sets of genes modulated during the acute attack. The analysis revealed that the genes identified during the acute attack were primarily involved in "natural killer cells signaling" and "leukocyte extravasation signaling" canonical pathways, as indicated in Figure 1B. In particular, the top-ranked network included several genes encoding for regulators of these pathways (score 24.8 associated genes, p<0.01, Figure 1C). Analysis of this network revealed a central role of genes important for the regulation of the vascular tone. Adrenomedullin (ADM) gene is coding for a pre-pro-hormone, which is cleaved to form a potent hypotensive and vasodilatator agent. Dysferlin (DYSF) gene is coding for a key calcium ion sensor involved in the Ca(2+)-triggered synaptic vesicle-plasma membrane fusions and in muscle contraction. Finally the Fc Fragment of IgG Receptor III (FCGR3) gene is coding for a receptor of the Fc region of immunoglobulin gamma that binds IgG complexes and mediates antibody-dependent cellular cytotoxicity and other antibodydependent responses, such as phagocytosis.

Differences in gene expression between HAE patients and healthy subjects

The second step of the study was to perform a whole-genome gene expression analysis using microarray technology in order to identify genes specifically expressed or modulated in the disease.

For this purpose, we compared the genomic profile of PBMCs from HAE-R, HAE-A and healthy subjects (HS).

Using ANOVA with Tukey HSD post hoc test we found 603 genes significantly modulated in HAE patients compared to HS (Supplemental Table 2). PCA showed that these genes could well distinguish the three analyzed sample classes (Figure 2A). Most of identified genes had more different levels of expression than HS whether during the attack or the remission phase (Figure 2B).

Then, we performed a GSEA comparing HAE-R and HS to identify the biological processes that could be directly involved in the HAE pathogenesis, since HAE-R patients are in a physiological condition similar to that of HS. The analysis revealed that "cell matrix adhesion", "cell substrate adhesion" and "g protein signaling coupled to camp nucleotide second messenger" gene sets were the most represented processes differentially regulated in the remission phase of HAE (Table 4).

A further pathway analysis of genes modulated in HAE-R compared to HS revealed among most significant biological processes primarily involved in "inflammatory response" and "immunological disease" canonical pathways, as indicated in Supplementary Figure 1.

In particular, 53 genes belonged to the category of "inflammatory response" (p-value = 2.73E-07), and 51 genes belonged to the category of "immunological disease" (p-value = 8.26E-09) (Table 5).

Surprisingly, among genes involved in these biological functions we found again ADM and uPAR, that resulted significantly down-regulated in HAE-R patients compared to HS, while Granzyme and Perforin1 were up-regulated.

Moreover, also "Granzyme B signaling" and "VEGF signaling" pathways, have been found significantly modulated in HAE-R compared to HS (Supplemental Figure 2).

Differential expression of ADM and uPAR genes at transcriptional level

Two of the genes identified by microarray analysis were of great importance for the pathogenesis of the disease because they are involved not only in the regulation of vascular tone and in the inflammatory response but also in bradykinin-forming mechanisms controlled by C1-IHN. Therefore, in order to confirm the gene expression modulation of these two genes, we performed a quantitative RT-PCR for ADM and uPAR in an independent set of 20 HAE patients with the same clinical and demographic characteristics of those used for microarray experiments. We found that ADM and uPAR gene expression was significantly higher during the acute attack episode compared to the same subject in the remission phase (**p<0.02, *p<0.05 respectively, Figure 3 A-B). These results were in line with those obtained by the gene expression array.

Blocking the uPA-uPAR interaction alters bradykinin pathway

The mechanism of action of uPAR in the production of bradykinin was examined.

We neutralized the uPA-uPAR interaction with a blocking anti-uPAR antibody both at 10µg/ml and at 20µg/ml, for 1 hour and then we activated the plasminogen with 100 nM uPA for 15 minutes. Then, we measured the T cell bradykinin production.

The treatment with anti-uPAR antibody attenuated the production of bradykinin, induced by urokinase plasminogen activator (uPA), in a dose-dependent manner, both in the cell lysate (p=0.0002 at 10 µg/ml and p<0.0001 at 20 µg/ml, Figure 3C) and in the supernatant (p=0.01at 10 µg/ml and p=0.001 at 20 µg/ml, Figure 3D). Anti uPAR-treated cells showed a significantly reduction of bradykinin production also in the control medium without uPA stimulation (supplemental figure 3).

Discussion

The first finding emerging from our PBMCs microarray gene expression analysis on RNA isolated from HAE patients is that the genes identified during the acute attacks were primarily involved in "natural killer cells signaling" and "leukocyte extravasation signaling" pathways. Chiefly, from our bioinformatics analysis, we found, for the first time, that the expression of two genes, ADM and uPAR, was up-regulated during acute attacks when comparing HAE-A vs HAE-R and was downregulated in HAE-R vs HS. These genes codify for molecules with vasodilator effects and are involved not only in the regulation of vascular tone and in the inflammatory response, but also in the bradykinin-forming mechanisms, the principal mediator of HAE attacks.

Our results shed more light on the role for adaptive immunity in the development of HAE that remained controversial over time. Before us, only Lopez-Lera and colleagues analyzed the wholegenome RNA expression of PBMCs in three HAE type-I families according to the presence of mutation and the clinical symptoms (28). They could not find differentially regulated genes in PBMCs of HAE patients. However, the design of their study was very different and based on comparisons between patients with mutation versus patients without mutation within the analyzed families.

Moreover, recent lines of evidence demonstrate that neutrophil and mast cell activation is functionally linked to bradykinin production through the contact system activation on endothelial cells (24).

In this regard, the results reported by the last evidences and by our study, support the hypothesis that the adaptive immunity function may affect the pathogenesis of HAE. Adrenomedullin (ADM) is a 52-amino-acid vasoactive peptide with vasodilator activity mediated by the cyclic adenosine monophosphate (cAMP), NO and renal prostaglandin systems (33).

ADM acts via a G protein-coupled 7-transmembrane domain receptor, which associates the calcitonin receptor-like receptor (CRLR). It modulates proteins, namely receptor activity-modifying proteins (RAMP) 2 and RAMP 3 (34). ADM could act as an endogenous immunomodulatory factor, with predominant anti-inflammatory effects (35). It is produced in many types of tissue, but the endothelium (36) and the monocytes (37) are the principal source, in particular in response to proinflammatory cytokines (38). ADM reduces endothelial hyper-permeability and is an apoptosis survival factor (39).

Our microarray data confirmed the study recently published by Kajadacsi and colleagues (40). They demonstrated that during attacks of hereditary angioedema due to C1-IHN, the excess of bradykinin that enhanced the endothelial cells permeability and the activation of endothelial cells, was associated with elevated levels of several vasoactive peptides, such as ADM, AVP and ET-1. In particular, they suggested that the cooperation of these vasoactive peptides might be necessary to counterbalance the vasodilatory activity of bradykinin and to terminate the attacks. Recently, Xie and collegaues (37), confirmed these data in acute Systemic Capillary Leak Syndrome (SCLS), characterized by abrupt and transient episodes of hypotensive shock and edema due to plasma leakage into peripheral tissues.

Finally, it is important to note that Complement Factor H is a serum-binding protein for ADM. (41) The interaction between Complement Factor H (also known as adrenomedullin binding protein 1, AMBP-1) and ADM, has been shown to have possible therapeutic application in various inflammatory diseases (42) and emerged also by our transcriptomic data (Figure 1C).There is plenty of evidence on the reciprocal effects of binding on Complement factor H and ADM activity. (43).

Initially, the complement activation was proposed in the literature as the principal pathway involved in the pathogenesis of C1-IHN-HAE, but it was later discredited (23,(44)(45)(46). As known, the coagulation, fibrinolysis, and plasma kallikrein cascades are interrelated as shown in Figure 3E and these cascades are controlled by C1-IHN (47). In the last years several studies focused on the role of fibrinolysis as contributor to angioedema (26,48) and our data are in accordance with these evidences.

Cellular Receptor for Urokinase Plasminogen Activator (PLAUR or uPAR) is a glycosylphosphatydilinositol (GPI)-anchored protein (49) and is able to bind both the pro-and active forms of Urokinase Plasminogen Activator (uPA) (50). Once activated, the primary function of uPA is the conversion of plasminogen to plasmin, a broad-spectrum enzyme capable of widespread extracellular matrix degradation and activation of several pro-MMPs. Several lines of evidence indicate that, in addition to fibrinolysis, the uPA-uPAR system also modulate numerous steps of the inflammatory cascade and influence the development of immune responses (51).

The inflammation is an adaptive response to damage of vascularized tissues (52). There is a lot of evidence to suggest that the plasminogen activation system is implicated in the regulation of all phases of the inflammatory response, but it plays a crucial role in leucocyte recruitment to the site of the inflammation (53)(54).

uPAR is expressed constitutively in various immunologically active cells, included monocytes, macrophages and activated T-cells and also in endothelial cells (55).

Is well known that uPAR interacts with components of the bradykinin-forming cascade, but uPAR actually binds HK best when it has been cleaved (56) and is an important binding protein for Factor XII (57).

We demonstrated that when we neutralized the uPAR expressed on T cells, the production of bradykinin is reduced in a dose-dependent manner. These new data demonstrate the involvement of uPAR in the production of bradykinin, that is the responsible agent for edema, and show the mechanism of action: the uPAR overexpression in HAE-A patients, promotes the plasminogen activation by uPA with the consequent formation of plasmin and, finally, leads to increased levels of bradykinin. This mechanism is in addition to other pathways, such as that of the contact system in which kallikrein leads to bradykinin production (Figure 3E). Intriguingly, in endothelium when uPAR is expressed, plasminogen is activated also in the absence of fibrin (58). Moreover, recently, Joseph et al demonstrated that pro-inflammatory mediators (IL-1 and TNF-α), and estrogen stimulate release of urokinase from endothelial cells, which can convert plasminogen to plasmin and represents a possible source for plasmin generation (59).

These findings are in agreement with previous observations that, in patients deficient in C1-INH the role of plasmin, the key enzyme of the fibrinolytic cascade, appears more prominent (15,23,25).

We are aware of the limitations of our study. Although we have performed a detailed gene expression analysis both comparing HAE-A versus HAE-R patients and HAE patients versus HS, we have not investigated ADM and uPAR serum levels because our approved clinical protocol did not consider the serum collection.

In conclusion, our study suggests, for the first time, that during the acute attack of HAE two genes involved in the regulation of vascular tone, ADM and uPAR, are up-regulated and the products of these genes could provide a further mechanism involved in the bradykinin production.

.. complex: Implications for hereditary angioedema. J Allergy Clin Immunol. 2016. pii: S0091-6749( 16)31278-7.

Figure legends 200 words

Figure 1. Differences in gene expression during the acute attack of C1-IHN-HAE and functional analysis of the top selected genes identified by microarray during acute attack. A.

The PCA showed that gene expression profile was different between HAE-A and HAE-R patients.

B.

The most representative canonical pathways dysregulated during the acute attack were mainly involved in "natural killer cells signaling" and "leukocyte extravasation signaling". Graph shows category scores; "threshold" indicates the minimum significance level [scored as -log(p-value) from Fisher's exact test, set here to 1.25]. "Ratio" (differential yellow line and markers) refers to the number of molecules from the dataset that map to the pathway listed divided by the total number of molecules that map to the canonical pathway from within the IPA knowledgebase. C.

The network was algorithmically constructed using IPA-software based on the functional and biological connectivity of genes. Genes were graphically symbolized as nodes and the biological relationships between genes as edges. Red nodes represented genes containing identified variants; empty nodes are biologically linked to the studied genes based on evidence in the literature. The analysis of this top-ranked network (score 24,8, p<0.01) revealed the presence of a gene important for the regulation of the vascular tone (ADM). The red line evidences the interaction between Complement Factor H and ADM.

Figure 2. Differences in gene expression between HAE patients and healthy subjects

The PCA (A) showed that gene expression profile was different among the three analyzed sample classes. The profile plot (B) represented the gene expression patterns at FC ≥1.5 relative at 90 of 603 genes up-and down-regulated between HAE patients and healthy subjects. Up-regulated genes are shown in red, and down-regulated ones in blue.

Figure 3. Validation of gene expression levels and functional study. The amount of ADM (A)

and uPAR (B) in an independent set of 20 HAE patients was evaluated by real-time PCR. ADMand uPAR-normalized gene expression levels were significantly higher during the acute attack of the disease compared with HAE-R (**p<0.02, *p<0.05 respectively, p-values obtained by paired ttest). The histograms represent the mean ± SD. The treatment with anti-uPAR antibody reduced the production of bradykinin, in a dose-dependent manner, in the Jurkat cell lysate (C, p=0.0002 at 10 µg/ml and p<0.0001 at 20 µg/ml compared to uPA) and in the supernatant (D, p=0.01at 10 µg/ml and p=0.001 at 20 µg/ml compared to uPA). E.Schematic overview of interactions between the contact, complement and plasminogen activation system with a central role of uPAR.in the production of bradykinin. Anti uPAR-treated cells showed a significantly reduction of bradykinin production also in the control medium without uPA stimulation. *p<0.05, p-values obtained by paired t-test.

Supplemental

Table 3

The most represented processes (q<0.05) identified by the GSEA during the acute attack of HAE.

The "response to external stimulus" gene set built-in genes important for the structure or regulation of extracellular matrix (uPAR) and for the regulation of the vascular tone (ADM).

Table 5

The most representative canonical pathways dysregulated in HAE-R patients compared to HS, identified by Anova and Tukey tests. The genes ADM and uPAR were built-in both in the "inflammatory response" and "immunological disease" categories.