Factor XII blockade inhibits aortic dilatation in angiotensin II-infused apolipoprotein E-deficient mice

Abdominal aortic aneurysm (AAA) is an important cause of death in older adults due to aortic rupture or complications of surgical repair [1]. Currently the only treatment for AAA is surgical repair but previous clinical trials suggest this is not beneficial for small aneurysms or people that are deemed unfit [2,3]. AAA is frequently diagnosed when the aneurysm is small and at low risk of rupture [4]. The natural history of most small AAAs is expansion to a larger size when the risk of rupture is much greater. Drug therapies that are effective at halting AAA growth could be very beneficial in the management of people with early-stage AAA [5].

Kinins are potent bioactive peptides important in regulating vascular permeability and inflammation following tissue injury [6]. Administration of a kinin-receptor agonist has been reported to amplify the response to angiotensin II (AngII) within a mouse model of AAA and promote the formation of large AAAs that commonly rupture [7]. A kinin-receptor antagonist limited aneurysm growth and reduced aortic rupture in the same experimental model [7]. These findings suggested that blocking kinin signalling could be a potential treatment strategy to limit AAA progression, although how to achieve this most effectively and safely remained unclear.

The contact system, initiated by Factor XII (FXII), can stimulate both prothrombotic and proinflammatory pathways, which are both implicated in AAA pathogenesis [1,7–11]. The inflammatory arm of the contact system is initiated when FXII activation leads to release of the the most well-known kinin, bradykinin, through activation of the plasma kallikrein–kinin system [12,13]. Plasma kallikrein released from prekallikrein by activated FXII (FXIIa) generates bradykinin from high molecular weight kininogen at the luminal surface of blood vessels [14]. It was hypothesised that FXII blockade would limit kinin activation and prevent progression of AAA pathological processes.

The current study investigated the effect of FXII deficiency and FXIIa neutralisation on the severity of aortic dilatation in apolipoprotein E-deficient (ApoE^−/−) mice in response to AngII.

The FXIIa neutralising antibody 3F7 (ch3F7-mG1L-aFXII) and control peptide were kindly provided by Dr. Con Panousis of CSL Limited, Australia.

The mouse studies were performed in accordance with institutional and ethical guidelines of James Cook University, Australia (AEC approvals A1883 and A2252), and conformed to the Guide for the Care and Use of Laboratory Animals (National Institutes of Health, U.S.A.) and the ARRIVE criteria set by The National Centre for the Replacement, Refinement and Reduction of Animals in Research (London, U.K.).

The Animal Resources Centre, Western Australia, supplied male ApoE^−/− mice. ApoE^−/− mice deficient in FXII (ApoE^−/−FXII^−/−) were generated through crossbreeding with FXII^−/− mice kindly provided by Professor Christoph Kleinschnitz, University Hospital Essen, Germany. All mice were housed under a 12:12-h light–dark cycle (relative humidity: 55–60%; temperature: 22 ± 1°C) and were given standard chow and water ad libitum.

Aortic dilatation was induced in 6-month-old male ApoE^−/− mice by subcutaneous infusion of AngII (1.0 µg/kg/min) for 28 days as previously described [7,15]. Briefly, an osmotic micro-pump (ALZET Model 1004, Durect Corporation) containing AngII (#A9525, Sigma–Aldrich) dissolved in sterile water was inserted into the subcutaneous space left of the dorsal midline under general anaesthesia (4% isoflurane) to administer AngII at a rate of 1.0 μg/kg/min over 28 days. Intraperitoneal administration of 3F7 commenced 24 h prior to placement of the AngII pumps, then every 2 days over the 28-day AngII infusion period. Sudden mouse fatality required necropsy within 24 h to confirm aortic rupture as cause of death. Mice completing the study protocol were isoflurane-sedated and CO₂-killed. Harvested aortas were phosphate buffered saline-perfused in situ and digitally photographed (Coolpix 4500, Nikon).

Two studies investigated the effect of FXIIa down-regulation on AngII-induced aortic dilatation:

• FXII deficiency: ApoE^−/−FXII^−/− (n=26) and ApoE^−/− (control, n=20) mice were infused with AngII over 28 days;

• FXIIa neutralisation: ApoE^−/− mice were allocated to receive the FXIIa-blocking antibody 3F7 (10 mg/kg; n=15) or isotype control (n=15) and infused with AngII over 28 days.

Ultrasound measurements of the suprarenal aorta (SRA) were obtained prior to AngII infusion (day 0) then at days 14 and 28 of the study period. Scans were performed in sedated mice (4% isoflurane) using yLab™ 70 VETXV platform (Esaote, Italy) with a 40-mm linear transducer at an operating frequency of 10 MHz (LA435; Esaote, Italy) to provide a sagittal image of the SRA. Maximum outer-wall SRA diameter (SRA_Max) was measured at peak systole using the calliper measurement feature. Good interobserver reproducibility of these measurements in our laboratory has been previously reported [15].

Mouse serum samples were assayed using commercial ELISAs for plasma kallikrein (MBS705761, MyBioSource) and bradykinin (EK-009-01, Phoenix Pharmaceutical Inc). For aortic samples, protein was obtained by RIPA buffer (Sigma–Aldrich) homogenisation of tissue harvested at the end of the study and sample protein concentration determined using the Bradford reagent (Bio-Rad Laboratories) according to the manufacturer’s recommendations. Protein samples were assayed using commercial kits for plasma kallikrein (MBS705761, MyBioSource), bradykinin (ADI-900-206, Enzo Life Sciences), phosphorylated epidermal growth factor receptor (pEGFR; MBS751770, MyBioSource), myeloperoxidase (MPO; HK210-02, Hycult Biotech), active matrix metalloproteinase (Mmp)-2 (QZBmmp2M, QuickZyme Biosciences) and Mmp-9 (QZBmmp9M, QuickZyme Biosciences), total and phosphorylated protein kinase B (Akt1, ab176657, SimpleStep ELISA™, Abcam), and total and phosphorylated extracellular signal-regulated kinase (Erk1/2, ab176660, SimpleStep ELISA™, Abcam). Tumor necrosis factor-α-converting enzyme (TACE/a disintegrin and metalloproteinase-17 (aka tumor necrosis factor-α-converting enzyme) (ADAM-17)) activity in SRA protein samples was measured using the SensoLyte® 520 Activity Assay (AS-72085, AnaSpec).

QuantiTect® Primer Assays were used to determine gene expression for kinin receptor B1 (Bdkrb1; QT00326886) and B2 (Bdkrb2; QT00111027) in mouse aortic tissue using quantitative real time (qPCR) as previously described [16]. The relative expression of these genes in experimental and control samples were calculated by using the concentration-C_t-standard curve method and normalised using the average expression of glyceraldehyde-3-phosphate dehydrogenase (Gapdh; QT01658692) for each sample using the Rotor-Gene Q operating software version 2.0.24. The QuantiTect SYBR® Green one-step RT-PCR Kit (Qiagen) was used according to the manufacturer’s instructions with 40 ng of total RNA as template. All reactions were independently repeated in duplicate to ensure the reproducibility of the results and the mean of the two values for each sample was used for analyses.

Data were analysed using GraphPad Prism (version 7) and TIBCO Spotfire S+ (version 8.2). D’Agostino and Pearson’s test was used to assess data normality. Parametric or non-parametric tests were applied appropriately to data distribution. Data are presented as median and interquartile range with maximum and minimum points unless otherwise stated. Student’s t or Mann–Whitney U tests were used for two-group comparisons of parametric and non-parametric data, respectively. Repeated measures one-way ANOVA followed by Tukey’s multiple comparisons test (parametric), or Kruskal–Wallis test followed by Dunn’s multiple comparisons test (non-parametric), were used to compare data within each group obtained as a function of time, with statistical significance between groups (control vs intervention) determined by mixed-effects linear regression. A contingency analysis applied Fisher’s exact test. A P-value less than 0.05 was considered statistically significant in all analyses.

Circulating and aortic levels of plasma kallikrein and bradykinin were determined in serum and SRA samples harvested from ApoE^−/−FXII^−/− mice, ApoE^−/− mice administered the FXIIa-neutralising antibody 3F7 (10 mg/kg/48 h, i.p.) and ApoE^−/− control mice. ApoE^−/−FXII^−/− mice and ApoE^−/− mice receiving 3F7, had significantly lower serum concentration of plasma kallikrein and bradykinin compared with ApoE^−/− controls (Figure 1). Concentrations of plasma kallikrein and bradykinin within the SRA of ApoE^−/− mice were also significantly reduced by FXII deficiency in comparison with controls (Figure 2). FXIIa neutralisation significantly reduced SRA concentration of plasma kallikrein, but not bradykinin, compared with controls (Figure 2).

SRA_Max was measured by ultrasound prior to AngII pump insertion (baseline) and at days 14 and 28 of the experimental period. Baseline SRA_Max was comparable in ApoE^−/− and ApoE^−/−FXII^−/− mice and in ApoE^−/− mice allocated to receive either 3F7 or as controls (Supplementary Figure S1). SRA_Max in all mice increased in response to AngII infusion over 28 days (Figure 3A,B). The degree of SRA dilatation was significantly less in ApoE^−/−FXII^−/− mice and ApoE^−/− mice administered 3F7 compared with controls (Figure 3C,D and Supplementary Figures S2 and S3). Seven out of twenty-six ApoE^−/−FXII^−/− mice had aortic rupture compared with two out of twenty ApoE^−/− control mice (P=0.262; Fisher’s exact test). The incidence of aortic rupture in mice administered 3F7 was 4 out of 15 compared with 3 out of 15 in control mice (P>0.999; Fisher’s exact test).

SRA protein samples obtained from ApoE^−/−FXII^−/− mice, ApoE^−/− mice administered 3F7 and ApoE^−/− controls at completion of the study (day 28) were assayed. ADAM-17 activity and levels of phosphorylated EGFR were significantly lower in ApoE^−/−FXII^−/− mice and ApoE^−/− mice administered 3F7, compared with ApoE^−/− controls (Figure 4).

Gene expression was assessed in SRA samples from ApoE^−/−FXII^−/− mice, ApoE^−/− mice administered 3F7 and ApoE^−/− controls at completion of the study (day 28). FXII deficiency resulted in significant down-regulation in gene expression for both the kinin B1 and the B2 receptors (Figure 5A and Supplementary Figure S4). Gene expression for the kinin B2 receptor, but not the kinin B1 receptor, was also significantly lower within the SRA of ApoE^−/− mice receiving the FXIIa neutralising antibody by comparison with controls (Figure 5A and Supplementary Figure S4). SRA protein samples obtained from ApoE^−/−FXII^−/− mice, ApoE^−/− mice administered 3F7 and ApoE^−/− controls at completion of the study (day 28) were assayed for MPO, as a marker of neutrophil activation. The SRA concentration of MPO was significantly lower in ApoE^−/−FXII^−/− mice than ApoE^−/− control mice (Figure 5B). Administration of FXIIa neutralising antibody did not significantly effect SRA MPO concentration (Figure 5C).

The concentrations of total Akt1 and Erk1/2 kinases within the SRA tissue of ApoE^−/−FXII^−/− mice and mice receiving 3F7 were similar to concentrations in control mice (Supplementary Figure S5). In contrast, phosphorylated Akt1 and Erk1/2 (and the ratio of phosphorylated-to-total kinase) were present at significantly lower concentrations in the SRA of mice with deficiency or neutralisation of FXII compared with controls (Figure 6 and Supplementary Figure S6). SRA protein concentrations for the active forms of Mmp-2 (Figure 6C,D) and Mmp-9 (Figure 6E,F) in ApoE^−/−FXII^−/− mice and ApoE^−/− mice administered 3F7 were significantly lower than in ApoE^−/− controls.

The current study demonstrated that blockade of the FXII/kallikrein pathway reduced AngII-induced aortic expansion. The contact system stimulates prothrombotic and proinflammatory pathways implicated in AAA pathogenesis [1,7–11]. The inflammatory arm of the contact system is initiated when FXII activation leads to generation of bradykinin through activation of the plasma kallikrein–kinin system [12,13]. Two separate approaches were used to investigate the role of the FXII/kallikrein pathway in aortic dilatation induced by AngII. There was whole body genetic deficiency of FXII and antibody-directed FXIIa neutralisation in the ApoE^−/− mouse model.

AngII-induced SRA dilatation was inhibited in both ApoE^−/−FXII^−/− mice and ApoE^−/− mice administered FXIIa-neutralising antibody in the present study. This effect was associated with significant reductions in circulating concentrations of plasma kallikrein and bradykinin, consistent with a previous study demonstrating significantly lower plasma levels of bradykinin in FXII-deficient compared with wild-type mice [17]. In mice with whole body FXII deficiency there were reduced SRA concentrations of plasma kallikrein and bradykinin, down-regulation in expression of kinin B1 and B2 receptors, evidence of decreased Akt1/Erk1/2 signalling and lower concentrations of active Mmp2 and Mmp9 and MPO, compared with controls. These findings were similar, but less marked, in mice in which FXIIa was neutralised via 3F7, possibly due to less severe FXII blockade. Importantly, FXIIa neutralisation in contrast with FXII-deficiency had no significant effect on SRA bradykinin concentration despite reduced levels of aortic plasma kallikrein. Furthermore, the similarity in SRA levels of MPO and kinin B1 receptor expression between mice receiving 3F7 and controls suggested a more limited effect of FXIIa neutralisation on the inflammatory response to AngII than whole body FXII deficiency.

Reduction in plasma kallikrein within the SRA of mice administered 3F7 may have been due to the inhibition of FXIIa activity on prekallikrein within the artery wall. Following endothelial cell injury, which can be induced by AngII infusion, vascular smooth muscle cells become exposed to plasma constituents. Previous studies suggest that plasma-derived FXII binds to vascular smooth muscle cells and activates prekallikrein to plasma kallikrein at the cell surface [18]. Despite the observed reduction in plasma kallikrein within the SRA of mice receiving FXIIa neutralising antibody, bradykinin levels were not affected. It is known that infiltrating leucocytes are a significant source of kinins within the artery wall during inflammation. The proteases derived from these cells are important non-kallikrein mediators of bradykinin formation and their presence with the SRA may have accounted for the maintained levels of bradykinin [19–23].

Neutrophil kinin B2 receptors are believed to facilitate extravascular migration of inflammatory cells at the sites of inflammation where they participate in local release of bradykinin into inflamed tissue [24]. Bradykinin is the endogenous agonist for the kinin B2 receptor and it has been previously shown that stimulation of the kinin B2 receptor augments AngII-induced aortic dilatation and rupture in ApoE^−/− mice [7]. This effect appeared to occur via neutrophil activation, stimulation of an inflammatory phenotype in vascular smooth muscle cells and production of proinflammatory proteins [7]. A recent study by Stavrou et al. [25] showed that human and mouse neutrophils express FXII following activation and that neutrophils from FXII^−/− mice exhibit reduced adherence and migration compared with wild-type controls. Migration and chemotaxis in both wild-type and FXII^−/− neutrophils were stimulated by FXII, an effect that did not require its catalytic activity (FXIIa) but rather was a zymogen FXII function [25]. It is plausible that the persistent neutrophil activation within the SRA of mice administered 3F7 was contributed to by activity of the FXII zymogen in promoting SRA leucocyte infiltration and activation independent of FXIIa neutralisation.

These findings suggested that the effect of FXII-deficiency to limit SRA dilatation in response to Ang II infusion was due to two main mechanisms. Firstly, the inhibition of the plasma kallikrein–kinin system and subsequent down-regulation of bradykinin generation within the SRA. Secondly, the inhibition of neutrophil infiltration and activation within the SRA in response to AngII. The cumulative effect of these two processes appeared to be responsible for limiting matrix remodelling and SRA dilatation in ApoE^−/−FXII^−/− mice. SRA concentrations of bradykinin and MPO were not affected by FXIIa neutralisation. This finding suggested the need to investigate other potential mechanisms to explain how FXII/kallikrein blockade limited SRA dilatation.

Tumor necrosis factor (TNF)-α has been implicated in AAA pathogenesis [26,27]. Plasma and aortic concentrations of TNF-α have been reported to be elevated in AAA patients [26]. Inhibition of TNF-α limits experimental AAA by limiting aortic inflammation and matrix remodelling [27]. TNF-α is generated via the action of TACE, also known as ADAM-17. Expression of ADAM-17 is elevated in human AAA tissue and its deletion inhibits experimental AAA [28]. Pharmacological inhibition of ADAM-17 or vascular smooth muscle cell deficiency of ADAM-17 attenuates AAA formation and rupture induced by AngII [29]. In the current study, both FXII deficiency and FXIIa neutralisation significantly reduced SRA ADAM-17 activity. This may have contributed to the inhibition of SRA dilatation found.

Previous in vitro and experimental in vivo studies have suggested that ADAM-17 plays a role in AngII-induced activation of the EGFR in vascular smooth muscle cells [30]. Evidence of EGFR activation has been reported in human AAA samples [31]. EGFR activation has been implicated in vascular remodelling in mice infused with AngII [32,33]. Deletion of EGFR in vascular smooth muscle cells prevents AngII-induced vascular remodelling [34], while pharmacological inhibition of EGFR limits development of experimental AAA [31]. It has been previously reported that plasma kallikrein stimulates ADAM-17 activity in vascular smooth muscle cells via a bradykinin-independent mechanism involving EGFR activation and EGFR-dependent Erk1 signalling [35]. In line with this, in the current study reduced plasma kallikrein concentration within the SRA of FXII-deficient and 3F7-administered ApoE^−/− mice were accompanied by down-regulation in ADAM-17 and EGFR activity. These effects may have contributed to the ability of FXII deficiency and neutralisation to limited SRA expansion.

Several limitations of the current study should be acknowledged. First, only one rodent model of AAA was used. AngII-infused ApoE^−/− mice were used since our prior findings implicating kinins in AAA were made in this model [7]. Also this model has a similar aortic gene profile to human AAA [36]. Second, histological analysis of the SRA was not performed as suitable samples were not available. The number of protein and mRNA-based assays performed in the present study required use of the entire SRA sample. We opted for quantitative measurement of inflammatory cell activation and inflammation markers over semi-quantitative histological assessments. Third, although ADAM-17 has a large substrate profile, its activation is typically only in response to stimuli that drive disease states [37]. The EGFR-ADAM-17 pathway was focussed on since this has previously been implicated in AAA in other experimental models. Other ADAM-17 substrates, such as Notch1 and angiotensin-converting enzyme 2, may also be important but these were not investigated [10,16,38]. Finally, conclusions regarding the effect of intervention using samples from the study end-point have limitations and analyses of serum and aortic tissue taken from earlier time points may have provided additional mechanistic insight.

In summary, the current study found that FXII/FXIIa blockade inhibited AngII-induced aortic dilatation in ApoE^−/− mice. ADAM-17/EGFR and Akt1/Erk1/2 signalling and Mmp-2 and Mmp-9 activity were down-regulated within the SRA. The findings suggest a novel axis to target in the design of an aortic aneurysm medical therapy.

• Currently there is no drug therapy for aortic aneurysm but past research has suggested that inflammation and the kallikrein–kinin system may be potential targets for a drug therapy.

• FXII is an important activator of the kallikrein–kinin system. Deficiency or neutralisation of FXII inhibited aortic expansion in a mouse model of aortic aneurysm through down-regulating the activity of a novel pathway generating the matrix remodelling enzymes Mmp-2 and -9.

• The findings suggest that FXII may be a novel target for a drug therapy for aortic aneurysm.

CP is an employee of CSL Limited. The other authors declare there are no competing interests associated with the manuscript.

This work was supported by the National Health and Medical Research Council [grant numbers 1098717, 1079369]; the Queensland Government [grant number SCRF]; in part by a Cardiac Health Institute Fellowship (to S.W.S.); a Practitioner Fellowship from the National Health and Medical Research Council, Australia [grant number 1117061 (to J.G.)]; a Senior Clinical Research Fellowship from the Queensland Government, Australia [grant number SCRF (to J.G.)]; the Deutsche Forschnungsgemeinschaft (German Research Foundation) [grant number SFB 688 (to C.K.)]; and employee of CSL Limited (Australia) (to C.P.). The funding bodies played no role in the production of this publication.

Open access for this article was enabled by the participation of James Cook University in an all-inclusive Read & Publish pilot with Portland Press and the Biochemical Society under a transformative agreement with CAUL.

C.S.M. designed and performed the mouse work, analysed the data, and drafted the manuscript. S.-W.S. and S.M.K. assisted with the animal model and ultrasound analysis. E.B. performed molecular analyses. S.K.M. performed protein assays. C.K. provided the FXII knockout mice to generate ApoE^−/−FXII^−/− mice used in the study. C.P. provided the FXIIa neutralising antibody 3F7 (ch3F7-mG1L-aFXII) and control peptide. J.G. obtained funding and contributed to experimental design, data analysis and interpretation, and co-wrote the paper. All authors contributed to critical revision of the manuscript.

AAA

abdominal aortic aneurysm

ADAM-17

a disintegrin and metalloproteinase-17 (aka tumor necrosis factor-α-converting enzyme)

AngII

angiotensin II

ApoE^−/−

apolipoprotein E-deficient mouse

ApoE^−/−FXII^−/−

whole-body FXII-deficient ApoE^−/− mouse

ARRIVE

Aanimal Research: Reporting of In Vivo Experiments

EGFR

epidermal growth factor receptor

FXII

factor XII

FXIIa

factor XIIa

IP

intra-peritoneal

Mmp

matrix metalloproteinase

MPO

myeloperoxidase

SRA

suprarenal aorta

TACE

tumor necrosis factor-α-converting enzyme