C1-Inhibitor Decreases the Release of Vasculitis-Like Chemotactic Endothelial Microvesicles

Abstract

The kinin system is activated during vasculitis and may contribute to chronic inflammation. C1-inhibitor is the main inhibitor of the kinin system. In this study, we investigated the presence of the kinin B1 receptor on endothelial microvesicles and its contribution to the inflammatory process. Compared with controls (n=15), patients with acute vasculitis (n=12) had markedly higher levels of circulating endothelial microvesicles, identified by flow cytometry analysis, and significantly more microvesicles that were positive for the kinin B1 receptor (P<0.001). Compared with microvesicles from wild-type cells, B1 receptor-positive microvesicles derived from transfected human embryonic kidney cells induced a significant neutrophil chemotactic effect, and a B1 receptor antagonist blocked this effect. Likewise, patient plasma induced neutrophil chemotaxis, an effect decreased by reduction of microvesicle levels and by blocking the B1 receptor. We used a perfusion system to study the effect of patient plasma (n=6) and control plasma (n=6) on the release of microvesicles from glomerular endothelial cells. Patient samples induced the release of significantly more B1 receptor-positive endothelial microvesicles than control samples, an effect abrogated by reduction of the microvesicles in the perfused samples. Perfusion of C1-inhibitor–depleted plasma over glomerular endothelial cells promoted excessive release of B1 receptor-positive endothelial microvesicles compared with normal plasma, an effect significantly decreased by addition of C1-inhibitor or B1 receptor-antagonist. Thus, B1 receptor-positive endothelial microvesicles may contribute to chronic inflammation by inducing neutrophil chemotaxis, and the reduction of these microvesicles by C1-inhibitor should be explored as a potential treatment for neutrophil-induced inflammation.

Activation of the kinin/contact system is proinflammatory, inducing pain and capillary leakage symptomatic of severe vascular inflammation.¹ The kinin system consists of a cascade of proteins ultimately leading to the release of bradykinin from high molecular weight kininogen.¹ In addition to bradykinin, other kinins are released during inflammation, including a 13-amino acid peptide termed proteinase 3-kinin (PR3-kinin) cleaved by neutrophil-derived proteinase 3 (PR3) and thus presumably active during conditions associated with neutrophil influx, such as the chronic vascular inflammatory condition vasculitis.² Bradykinin binds to the constitutively expressed B2-receptor, whereas des-arg⁹-bradykinin and PR3-kinin bind to the B1-receptor (B1R), expressed primarily during chronic inflammation.^2,3 The major inhibitor of the kinin system is C1-inhibitor.¹

The kinin system may be assembled and activated on endothelial cells and neutrophils,⁴ as well as on blood cell-derived microvesicles (MVs)⁵ with a proinflammatory effect. MVs are vesicular membrane blebs (0.1–1.0 μm in diameter) shed by cells during activation, injury and/or apoptosis that contain proteins, RNA, histones, and other cellular components from the parent cell.^6–8 MVs can be found in the plasma of healthy individuals but are increased during inflammatory conditions.⁹ MVs can promote inflammation¹⁰ and thrombosis¹¹ and thus prolong the inflammatory process. This may be achieved by membrane exposure of, or release of, proinflammatory cytokines, chemokines, selectins, adhesion molecules, and complement factors^7,12 as well as phosphatidylserine, which may activate coagulation factors, and tissue factor (predominantly on monocytic MVs),¹³ thus promoting thrombosis.¹¹ MVs may also transfer the kinin B1R.¹⁴ In the circulation, MVs may interact with cells by receptor binding, membrane fusion or endocytosis/engulfment of the entire MV, as reviewed by Meckes and Raab-Traub.¹⁵ Thus, MVs may facilitate intercellular communication.

Endothelial cells also react to stress by shedding MVs (termed endothelial cell-derived microvesicles; EMVs).¹⁶ EMVs are increased in various conditions associated with vascular damage, such as diabetes and hypertension.¹⁷ Circulating EMVs as well as platelets and leukocytes have been detected systemically in patients with vasculitis.^9,14,18–21 Vasculitis is a disease characterized by inflammation in and around vessel walls affecting multiple organs, including the kidneys, skin, joints, intestines, and lungs. The most common vasculitis in childhood is Henoch–Schönlein purpura (HSP), whereas adults are more frequently affected by ANCA-associated vasculitis including granulomatosis with polyangiitis, microscopic polyangiitis, and eosinophilic granulomatosis with polyangiitis.²² The kinin system has been shown to be activated during vasculitis.^2,23 Levels of MVs have correlated to disease activity as assessed by the Birmingham vasculitis activity score (BVAS),^18,24 thus serving as disease biomarkers.

The presence of B1R on EMVs would reflect a phenomenon occurring on the endothelial cell during the inflammatory process and may also exacerbate the condition. The aim of this study was to investigate if B1R is present on EMVs shed during vasculitis and to study the contribution of B1R on EMVs to the inflammatory process, specifically neutrophil chemotaxis. Furthermore, the study aimed to explore if C1-inhibitor could regulate the release of B1R-positive EMVs from glomerular endothelial cells.

Results

EMVs in Vasculitis Plasma Are B1R-Positive

Plasma from patients with vasculitis (n=12) and healthy controls (n=15) were analyzed by flow cytometry for the presence of EMVs, defined as positive for CD105 and/or CD144. Patient plasma exhibited significantly more EMVs than controls. Furthermore, a significantly higher amount of EMVs were B1R-positive compared with controls (Figure 1). Of the total number of patient EMVs, a median of 41% (range, 8%–69%) were B1R-positive compared with 13% (range, 0%–49%) in controls. No correlation was found between B1R-positive EMV levels and variables such as specific diagnosis, presence/absence of immunotherapy, creatinine levels, or BVAS (Table 1). Patient EMVs were also found to be IL-8–positive, with a median of 86% (range, 51%–97%; n=5). Most patient EMVs that were B1R-positive were also IL-8–positive, with a median of 92% (range, 86%–96%). Fewer control EMVs were IL-8–positive, with a median of 46% (range, 6%–65%; n=5, data not shown).

Figure 1.

EMVs in vasculitis plasma express B1R. Plasma samples from patients with vasculitis (n=12, patients 1–7 and 9–13 in Table 1) and healthy controls (n=15) exhibited significantly higher levels of EMVs, positive for CD105 and/or CD144. Circulating EMVs were B1R-positive and samples from patients with vasculitis had significantly higher levels of B1R compared with controls. Samples were run using a BD FACSCanto Cytometer. ***P<0.001.

Table 1.

Description of patients included in this study

Patient No.	Sex	Age at Sampling	Diagnosis	ANCA	Clinical Findings	Creatinine,a μmol/L	BP at Admission, mmHg	BVAS Score	Blood Sample Taken Before/After Start of Immunosuppressive Therapy
1	M	43	HSP	—	Purpura, nephritis, arthritis	89	170/100	15	B
2	M	16	HSP	—	Purpura, arthritis	69	122/70	3	B
3b	M	80	MPA	MPO	Rash, nephritis	259	168/72	23	B
4	M	62	GPA	PR3	URT symptoms, pulmonary granuloma, temporal arteritis	61	130/90	6	B
5b	F	19	SLE	—	Rash, myalgia, scleritis, nephritis	62	110/60	17	A
6b	F	76	GPA	PR3	Gingivitis, rhinitis, hematuria	74	130/70	14	B
7b	F	12	GPA	PR3 MPO	Fever, arthralgia, nephritis	312	120/70	15	A
8b	F	61	GPA	MPO	Polychondritis, cough, rhinitis, nephritis	54	150/80	13	B
9	M	10	HSP	—	Purpura, arthritis, nephritis	47	102/60	7	B
10	F	51	GPA	PR3	Hemoptysis, cough, nephritis	132	164/96	8	A
11	M	73	MPA	MPO	Hearing loss, nephritis	173	155/93	15	A
12	F	68	GPA	PR3	Myalgia, arthralgia, nephritis	86	136/79	13	A
13	F	67	MPA	MPO	Myalgia, confusion, nephritis	148	130/60	22	B
14	F	9	HSP	—	Purpura, myalgia, nephritis, arthritis	51	110/62	18	B
15	M	13	GPA	PR3	Dyspnea, pulmonary infiltrates, arthralgia, nephritis	84	120/70	24	B
16	F	17	GPA	MPO	Dyspnea, myalgia, arthralgia, gastrointestinal symptoms, nephritis	86	115/75	26	B
17	F	13	GPA	PR3	Myalgia, pansinusitis, exophthalmos, pulmonary infiltrates	48	110/75	17	B

M, male; —, no ANCA detected; B, before; MPA, microscopic polyangiitis; MPO, myeloperoxidase; GPA, granulomatosis with polyangiitis; URT, upper respiratory tract; F, female; A, after.

^aNormal creatinine values of 45–90 μmol/L for women and 60–105 μmol/L for men.

^bThese patients were previously described, including their kinin levels.^2,23

MVs with B1R Induce Neutrophil Migration

Experiments were designed to investigate if the presence of B1R on MVs induced neutrophil chemotaxis. Neutrophils migrated toward MVs derived from human embryonic kidney (HEK) cells transfected with B1R (HEK^B1R cells) or MVs derived from wild-type cells (HEK^wt cells; 70 μg/ml MVs). Enhanced migration was demonstrated when MVs derived from HEK^B1R cells were placed in the lower compartment compared with MVs derived from HEK^wt cells (Figure 2A). Neutrophil migration toward MVs derived from HEK^B1R cells was significantly reduced by preincubation with the B1R-antagonist R715. Neutrophil migration toward MVs derived from HEK^wt cells was comparable with MVs derived from HEK^B1R cells preincubated with the R715 B1R-antagonist.

Figure 2.

Neutrophils migrate toward B1R-expressing MVs. Isolated neutrophils from healthy donors migrated through a membrane filter in a Boyden chamber assay. The y-axis represents the distance of neutrophil migration. (A) HEK^wt and HEK^B1R cell-derived MVs (70 µg/ml) and DMEM (negative control) were placed in the lower compartment. More neutrophil migration was induced by MVs from HEK^B1R compared with HEK^wt and HEK^B1R preincubated with the B1R-antagonist R715. (B) Purified neutrophils migrated toward plasma from patients with vasculitis as well as toward the same samples with reduced MV content (median 7.5% of the original EMVs). Neutrophil migration was decreased in MV-reduced vasculitis plasma. (C) Purified neutrophils migrated toward plasma from patients with vasculitis preincubated with or without the B1R-antagonist R715. Neutrophil migration was decreased in samples preincubated with the B1R-antagonist. Experiments in (B and C) were carried out at separate occasions. DMEM and PBS were used as the negative controls. N-formyl-met-leu-phe (fMLP) was used as the positive control. ***P<0.001; **P<0.01; *P<0.05.

Patient-Derived B1R-Positive MVs Are Chemotactic for Neutrophils

Plasma from patients with vasculitis induced neutrophil migration (Figure 2B). The chemotactic effect was lowered significantly by reduction of the amount of MVs (Figure 2B). Likewise, addition of the B1R-antagonist decreased neutrophil migration (Figure 2C). Addition of anti-human CXCL8/IL-8 further reduced chemotaxis (35.7 μm; range, 31–42.5, data not shown). This implies that plasma from patients with vasculitis contains MVs with chemotactic properties, indicated by the presence of B1Rs, as well as other chemotactic substances, such as IL-8.

Release of B1R-Positive EMVs from Primary Glomerular Endothelial Cells

The release of B1R-positive EMVs was analyzed after perfusion of plasma from patients with vasculitis (n=6; ANCA-positive vasculitis n=4, patients 3, 4, 7, and 15; HSP n=2, patients 2 and 14; Table 1) and controls (n=6) over primary glomerular endothelial cells (PGECs) using a microfluidic perfusion system. Patient preperfusion samples exhibited higher amounts of circulating EMVs positive for B1R compared with controls. Perfusion of plasma over PGECs led to release of B1R-positive EMVs. Significantly more EMVs were released from PGECs when perfused with patient plasma compared with controls, and these EMVs were B1R-positive (Figure 3). No marked difference in PGEC response was noted when using samples from patients with ANCA-positive vasculitis or patients with HSP. Samples from two patients with ANCA-positive vasculitis (patients 7 and 15) were compared with the same samples after IgG depletion by protein-G to remove ANCA. Removal of IgG did affect the release of EMVs in the perfusion system (data not shown), suggesting that ANCA did not contribute to EMV release.

Figure 3.

B1R is found on EMVs released after perfusion of PGECs. (A) Plasma samples from patients with vasculitis (n=6) and healthy controls (n=6) perfused over PGECs at a shear stress of 2 dynes/cm² for 15 minutes induced the release of EMVs (CD144-positive and CD105-positive MVs). Patient plasma samples contained more EMVs than control samples before perfusion (presample). Perfused plasma samples from patients with vasculitis exhibited significantly higher levels of EMVs than controls. (B) Patient samples contained more B1R-positive EMVs before perfusion (presample). Perfused samples from patients with vasculitis had significantly higher levels of B1R-positive EMVs compared with controls (the lowest value of B1R-positive EMVs in control plasma was 0.03×10⁶/ml). Statistical comparisons were carried out between patients and controls, but not between presamples and perfused samples. Samples were analyzed using a BD FACSCanto Cytometer. **P<0.01; *P<0.05.

Plasma from patients with vasculitis in which MV levels were reduced (7.5% of the original quantity of EMVs) did not exhibit an increase in the number of EMVs after perfusion over PGECs (Supplemental Figure 1A). Likewise, an increase of B1R-positive EMVs was not registered when the vasculitis plasma with reduced MVs was perfused over the cells (Supplemental Figure 1B).

C1-Inhibitor and B1R-Antagonist Reduced Release of B1R-Positive EMVs from PGECs

C1-inhibitor–depleted plasma exhibited lower total EMVs (CD144-positive MVs) and B1R-positive EMVs before perfusion (median EMVs, 90; range, 38–250×10⁶/ml; B1R-positive EMVs median, 8; range, 1–24×10⁶/ml) in comparison to normal plasma (median EMVs, 330; range, 80–536×10⁶/ml; B1R-positive EMVs median, 53; range, 8–76×10⁶/ml). After perfusion, EMV release was higher in the C1-inhibitor–depleted plasma compared with normal plasma (Figure 4A). C1-inhibitor–depleted plasma with added C1-inhibitor (Berinert) exhibited lower EMV release that did not reach statistical significance when compared with C1-inhibitor–depleted plasma and control plasma (Figure 4A). Perfused C1-inhibitor–depleted plasma also exhibited significantly elevated B1R-positive EMVs that were reduced in the presence of C1-inhibitor (Figure 4B), suggesting that the lack of C1-inhibitor triggered kinin system activation on PGECs and EMVs. Kinin system activation was confirmed by measuring kinin levels in C1-inhibitor–depleted plasma perfused over PGECs showing a higher level in the absence of C1-inhibitor (227.2 pg/ml) compared with a sample with added C1-inhibitor (20.6 pg/ml). Likewise, addition of the B1R-antagonist (R715) to C1-inhibitor–depleted plasma reduced total EMV release and B1R-positive EMVs (Figure 4, C and D, respectively). Taken together, these results suggest that C1-inhibitor and B1R-antagonist reduced the total number of EMVs as well as levels of B1R-positive EMVs.

Figure 4.

C1-inhibitor and B1R-antagonist reduce release of B1R-positive EMVs after perfusion of PGECs. (A) Perfusion of normal plasma, C1-inhibitor–depleted plasma, and C1-inhibitor–depleted plasma complemented with C1-inhibitor, over PGECs at a shear stress of 5 dynes/cm² for 5 minutes induced the release of EMVs (CD144-positive). C1-inhibitor–depleted plasma showed higher levels of total EMVs compared with control plasma. Complementing the C1-inhibitor–depleted plasma with C1-inhibitor reduced the total number of EMVs. (B) Cells perfused with C1-inhibitor–depleted plasma exhibited significantly higher levels of B1R-positive EMVs than control plasma. Addition of C1-inhibitor significantly reduced the number of B1R-positive EMVs. (C) The amounts of total EMVs released from PGECs perfused with C1-inhibitor–depleted plasma to which the B1R-antagonist R715 was added were significantly reduced compared with cells perfused with C1-inhibitor–depleted plasma. (D) PGECs exposed to C1-inhibitor–depleted plasma in the presence of the B1R-antagonist exhibited significantly lower levels of B1R-positive EMVs. Samples were run using a CyFlow Cube 8 flow cytometer. **P<0.01; *P<0.05. C1INH, C1-inhibitor; C1INH-dpl plasma, C1-inhibitor–depleted plasma.

B1R-Positive EMVs Are Chemotactic for Neutrophils

C1-inhibitor–depleted plasma perfused over PGECs and containing B1R-positive EMVs (Figure 4) exhibited chemotactic activity toward neutrophils. This activity was specific for B1R because it was significantly reduced in the presence of the B1R-antagonist R715 (Figure 5).

Figure 5.

Neutrophils migrate toward B1R-positive endothelial MVs. Purified neutrophils, from healthy donors, migrated through a filter in a Boyden chamber toward a chemoattractant. The y-axis represents the distance of neutrophil migration. PBS was used as the negative control. C1-inhibitor–depleted plasma (preincubated with or without the B1R-antagonist R715) derived from perfusion experiments over PGECs was used as the chemoattractant. Migration of neutrophils was decreased in plasma incubated with the B1R-antagonist R715. *P<0.05. C1INH-dpl plasma, C1-inhibitor–depleted plasma.

Discussion

MVs have been shown to be important mediators of cell-to-cell communication and they also reflect the state of activation of the parent cell from which they were shed. Here, we show that EMVs carry kinin receptors during chronic vascular inflammation, and that the presence of B1R on MVs induced neutrophil chemotaxis. Thus, MVs actively promote and sustain the inflammatory state. B1R-positive EMVs were observed in patients with vasculitis and reproduced in an in vitro perfusion system using patient plasma as well as C1-inhibitor–depleted plasma. Importantly, both C1-inhibitor and a B1R-antagonist inhibited release of B1R-positive EMVs by blocking activation of the kinin system on the endothelium. These results suggest that C1-inhibitor can block the release of chemotactic EMVs during vascular inflammation.

C1-inhibitor inhibits plasma kallikrein, thereby preventing release of kinins.^25,26 Patients with vasculitis have elevated levels of kinins, both systemically and locally in the kidneys.^2,23 High levels of bradykinin are also found in patients with hereditary angioedema because of mutations in C1-inhibitor or neutralizing autoantibodies.^27,28 We have previously shown that patients with vasculitis have normal C1-inhibitor levels.²³ Moreover, genetic sequencing of the C1-inhibitor reactive site did not show alterations.²³ Thus patients with vasculitis most probably have circulating C1-inhibitor capable of reducing the release of B1R-positive EMVs. Kinin system activation may, however, overwhelm the inhibitor and thus levels of C1-inhibitor, although normal, may not suffice to inhibit excessive release of chemotactic B1R-positive EMVs. The results presented here suggest that treatment with C1-inhibitor may, to a certain extent, prevent neutrophil influx at inflammatory sites.

The kinin system is activated in both children and adults with vasculitis.^2,23,29 The components of the kinin system play a role in the intrinsic coagulation pathway as well as in the induction of inflammation.¹ During this process there is a local reduction of BP, promotion of inflammation and pain, inhibition of platelet aggregation, induction of fibrinolysis, and capillary leakage.³ B1R, expressed during chronic inflammation, binds des-arg⁹-bradykinin and des-arg¹⁰-kallidin, as well as PR3-kinin^2,3 and the receptor remains on the plasma membrane after ligand binding. Thus, the presence of the B1R on the surface of cells suggests that the ligand is bound, as the receptor is internalized under resting conditions, in the absence of signaling.³ Ligand binding activates endothelial formation of nitric oxide³⁰ and partakes in the induction of pain.^31,32 EMVs most probably do not possess the cellular machinery required for internalization of the B1R, and would thus be expected to continuously express the receptor.

B1R stimulation leads to neutrophil migration, as shown in both in vivo and in vitro models.^32–34 Our results strengthen these findings, showing that even circulating B1R-positive MVs induce chemotaxis, and reduction of MV levels in plasma of patients with vasculitis or blockade of the B1R in the plasma reduced chemotaxis. Importantly, the B1R was demonstrated in kidneys from patients with GN, including vasculitis, and a B1R-antagonist ameliorated the inflammatory response in a mouse model of GN.³⁵ Thus, the presence of B1R on EMVs suggests that it is expressed on these cells during vasculitis. Perfusion experiments demonstrated that patient plasma induced release of B1R-positive EMVs from PGECs, which would be expected to be a major source of EMVs in vasculitis as glomerular injury is predominant. The results suggest that B1R-positive EMVs could induce neutrophil chemotaxis at sites of inflammation.

Endothelial cell injury is a major manifestation of vasculitis in general, and circulating endothelial cells have been detected, indicating vascular wall injury.^20,36 EMVs are elevated during vasculitis^18,19 and decrease upon treatment,²⁰ suggesting that they are adequate biomarkers of endothelial injury during vasculitis. Circulating EMVs in patients with vasculitis were shown to be E-selectin positive.⁹ Furthermore, levels of EMVs were higher in pediatric patients with vasculitis that developed thromboembolic events.²¹ In this study, high levels of EMVs could not be correlated to patient diagnosis, BVAS, renal function, or immunosuppressive treatment, probably because of the limited number of patients and their heterogeneity. Nevertheless, reducing MV levels in vasculitis plasma decreased EMV release from PGECs, and a novel function of EMVs contributing to the local inflammation was demonstrated. Both B1R and IL-8, demonstrated on EMVs, were found to be chemotactic. Blockade of B1R reduced the chemotactic potential of EMVs, an effect further enhanced, albeit minimally, by additional blockade of IL-8 on the EMVs, showing that EMVs have potent inflammatory properties. The chemotactic potential of B1R-positive MVs was not specific for EMVs, as even MVs derived from HEK^B1R cells induced neutrophil migration. Although not addressed here, we speculate that even other MVs bearing B1R, such as neutrophil-derived MVs,¹⁴ could induce a similar chemotactic and proinflammatory effect.

The factor/s inducing kinin system activation and EMV release during vasculitis are most probably numerous. EMVs have been associated with vascular injury and metabolic derangements^16,18–20 and may in themselves be injurious to endothelial cells.³⁷ MV release may be beneficial for the parent cell as a means to rid the cell of unwanted components,³⁸ and general abrogation of MV release may thus not be favorable. C1-inhibitor reduced the total release of EMVs and, more specifically, those that were B1R-positive. As B1R-positive MVs induced neutrophil chemotaxis, a major feature of inflammation, the effect of C1-inhibitor on B1R-positive EMV release could be beneficial. We therefore suggest that C1-inhibitor should be explored as a treatment for neutrophil-associated vascular inflammation.

Concise Methods

Patients and Controls

Blood samples were available from 17 patients (ten females and seven males, from children and adults; median age, 43 years) with vasculitis treated at the Department of Nephrology and the Department of Pediatrics, Section of Pediatric Nephrology, Skåne University Hospital, Lund and Malmö, Sweden. Details regarding patients’ age, sex, clinical and laboratory findings, diagnosis, and BP are given in Table 1. None of the patients were on hemodialysis at, or before, the time of sampling. Samples were obtained at onset or during the acute phase of disease and used for flow cytometry analysis and perfusion assays. Disease activity was assessed using the BVAS (Table 1).²⁴ Blood was also available from 27 healthy adult controls (15 women, 12 men; median age, 42 years) not using any medications at the time of sampling. These samples were used for flow cytometry analysis, perfusion and neutrophil chemotaxis assays. Samples from the patients and controls were obtained after written consent of the subjects or their parents. The study was carried out with the approval of the Regional Ethics Review Board of Lund University.

Blood Samples

Whole blood was drawn into citrated tubes and centrifuged to obtain platelet-poor plasma and platelet-free plasma. These samples were analyzed for the presence of EMVs and used in perfusion experiments. In certain plasma samples the EMV content was reduced by centrifugation to 7.5% (range, 5.7%–11.3%) of the original EMV content. For details see Supplemental Material.

Culture of HEK Cells and Isolation of MVs

HEK^wt and HEK cells stably expressing human B1R (HEK^B1R)³⁹ were cultured as detailed in the Supplemental Material. Medium from cells was collected and centrifuged at 2500×g for 5 minutes and further at 100,000×g for 3 hours to collect the pellets containing MVs. The concentration of MVs derived from HEK^wt and HEK^B1R was 70 µg/ml (corresponding to 36.7×10⁸ MVs/ml). Kinin levels were measured in the supernatants of HEK^wt and HEK^B1R cells using an ELISA kit described below, and were found to be below the detection limit.

Culture of PGECs

PGECs (Cell Systems, Kirkland, WA) were cultured as described in the Supplemental Material and used in perfusion experiments.

Perfusion Experiments over PGECs

A semiautomated microfluidic perfusion system (VenaFlux; Cellix, Dublin, Ireland) was used to mimic shear in the microvasculature. Microcapillary channels (Vena8 Endothelial+ biochips; Cellix) were coated with fibronectin (Sigma-Aldrich, St. Louis, MO) and the PGEC suspension as described.⁴⁰ PGECs were prestimulated with 200 µM histamine (Sigma-Aldrich) in Dulbecco phosphate buffered saline (DPBS; PAA Laboratories) for 10 minutes at a shear stress of 5 dynes/cm² obtained with a Mirus Evo nanopump (Cellix). Histamine prestimulated samples mimic the activated endothelium present during inflammation.⁴¹ Histamine-stimulated endothelial cells were found to release more EMVs, confirming previous results,⁴⁰ and were therefore used in all described experiments.

Plasma samples from patients and controls were perfused over the PGECs at 2–5 dynes/cm² for 5–15 minutes, as described in the Supplemental Material. In certain experiments, IgG was depleted from plasma by adsorption to a Protein G Sepharose column (Amersham Biosciences, Uppsala, Sweden). After perfusion, the presample (plasma before perfusion over PGECs) and the perfused sample were collected and centrifuged for 5 minutes at 10,000×g. Dilution of the perfused samples was adjusted for by protein concentrations. These samples were later assayed by flow cytometry for detection EMVs.

C1-inhibitor–depleted plasma (Milan Analytica AG, Rheinfelden, Switzerland) as well as vasculitis plasma were diluted 1:1 in filtered DPBS and similarly perfused over the cells at a shear stress of 5 dynes/cm² for 5 minutes. In certain experiments C1-inhibitor (Berinert, final concentration of 1 IU in suspension; CSL Behring, Marburg Germany) or B1R-antagonist R715 (1 μM; a kind gift from D. Regoli, Universite de Sherbrooke, Quebec, Canada) was added to the plasma just before perfusion.

Kinin ELISA

Kinin levels in HEK supernatants and perfused C1-inhibitor–depleted plasma samples were measured using the MARKIT-M bradykinin-ELISA kit (Dainippon Pharmaceutical, Osaka, Japan) as per the manufacturer’s instructions.

Detection of MVs Derived from Endothelial Cells and B1R on the EMVs

Plasma samples and samples collected from perfusion experiments were incubated with mouse anti-human CD144 (conjugated with phycoerythrin, 1:200; peridinin chlorophyll protein-cyanin5.5, 1:600; or FITC, 1:200) and mouse anti-human CD105:peridinin chlorophyll protein-cyanin5.5 at 1:800. MVs labeled with either or both antibodies were considered to derive from endothelial cells.

To detect B1R, rabbit anti-human B1R from two sources (in-house developed antibody⁴² at 10 μg/ml final concentration or BDKRB1 at 1:50; Abgent, San Diego, CA) was incubated with the plasma or perfusion samples. To detect IL-8, mouse anti-human IL-8:phycoerythrin at 1:800 was used. Flow cytometry was performed using a BD FACSCanto Cytometer and FACSDiva Software version 6.0 (Becton Dickinson Immunocytometry Systems, San Jose, CA) or a CyFlow Cube 8 flow cytometer (Sysmex Partec, Norderstedt, Germany) with FCS Express 4 Flow Research Edition software version 4.07.0003 (De Novo Software, Glendale, CA). The latter flow cytometer detects vesicles of submicron size and therefore identifies more events corresponding to MVs. Details are given in the Supplemental Material.

Neutrophil Migration Assay

Neutrophils were purified from citrated whole blood using Lymphoprep (Fresenius Kabi, Oslo, Norway), as previously described⁴³, and adjusted to a concentration of 2×10⁶ cells/ml. A neutrophil migration assay was performed with a Boyden chamber micropore filter assay (Neuro Probe, Gaithersburg, MD).⁴⁴ Purified neutrophils from healthy donors (n=10) were placed in the upper compartment of the chamber. MVs derived from HEK^wt cells, HEK^B1R cells, and PGECs in perfusion experiments were placed in the lower compartment with or without preincubation with 1 μM B1R-antagonist R715, for 30 minutes at room temperature. Alternatively, plasma (diluted at 1:4) was placed in the lower compartment, in the presence or absence of the B1R-antagonist R715. In certain experiments, anti–hCXCL8/IL-8 (1 μg/ml; R&D Systems, Minneapolis, MN) was added to the plasma 30 minutes before running the assay. Migration in the membrane (measured in micrometers) was assayed by light microscopy (Zeiss Axiostar Plus; Zeiss, Göttingen, Germany). For details, see the Supplemental Material.

Statistical Analyses

To compare EMV levels between patient and control samples, in plasma as well as in the perfusion experiments, the Mann–Whitney U test was used. Correlations between B1R-positive EMVs and variables such as diagnosis, treatment, creatinine levels, and the BVAS were carried out using Spearman correlation coefficient. Neutrophil migration and perfusion experiments were compared using Kruskal–Wallis and Dunn multiple comparisons test. The Wilcoxon test was used to compare neutrophil migration in paired samples. A P value ≤0.05 was considered significant. Statistical analysis was performed using GraphPad prism software (version 7; GraphPad Software, La Jolla, CA).

Disclosures

None.