Myeloid cell-derived interleukin-6 induces vascular dysfunction and vascular and systemic inflammation

Abstract Aims The cytokine interleukin-6 (IL-6) plays a central role in the inflammation cascade as well as cardiovascular disease progression. Since myeloid cells are a primary source of IL-6 formation, we aimed to generate a mouse model to study the role of myeloid cell-derived IL-6 in vascular disease. Methods and results Interleukin-6-overexpressing (IL-6OE) mice were generated and crossed with LysM-Cre mice, to generate mice (LysM-IL-6OE mice) overexpressing the cytokine in myeloid cells. Eight- to 12-week-old LysM-IL-6OE mice spontaneously developed inflammatory colitis and significantly impaired endothelium-dependent aortic relaxation, increased aortic reactive oxygen species (ROS) formation, and vascular dysfunction in resistance vessels. The latter phenotype was associated with decreased survival. Vascular dysfunction was accompanied by a significant accumulation of neutrophils, monocytes, and macrophages in the aorta, increased myeloid cell reactivity (elevated ROS production), and vascular fibrosis associated with phenotypic changes in vascular smooth muscle cells. In addition to elevated Mcp1 and Cxcl1 mRNA levels, aortae from LysM-IL-6OE mice expressed higher levels of inducible NO synthase and endothelin-1, thus partially accounting for vascular dysfunction, whereas systemic blood pressure alterations were not observed. Bone marrow (BM) transplantation experiments revealed that vascular dysfunction and ROS formation were driven by BM cell-derived IL-6 in a dose-dependent manner. Conclusion Mice with conditional overexpression of IL-6 in myeloid cells show systemic and vascular inflammation as well as endothelial dysfunction. A decrease in circulating IL-6 levels by replacing IL-6-producing myeloid cells in the BM improved vascular dysfunction in this model, underpinning the relevant role of IL-6 in vascular disease.


Introduction
Cardiovascular disease (CVD) remains the leading cause of death worldwide.Vascular inflammation plays a crucial role in the development of CVD.Among the different mediators of the immune system, the cytokine interleukin-6 (IL-6) is of central relevance and has been repeatedly implicated in the pathogenesis of CVD.Interleukin-6 is a predictor of long-term cardiovascular mortality in patients with acute coronary syndrome, 1,2 involved in heart failure, 3 and has been suggested to predict the onset of coronary artery disease (CAD). 4It was described to be independently associated with the risk of major adverse cardiovascular events, cardiovascular and all-cause mortality, myocardial infarction, and heart failure in patients with stable coronary heart disease. 5Within the inflammation cascade, IL-6 has pivotal functions at the crossroads of the innate and adaptive immune responses.][18][19] Thus, IL-6 represents a potential mediator driving increased cardiovascular comorbidities in patients with these chronic inflammatory disease states. 20Interleukin-6 deficiency was described to attenuate angiotensin II (AngII)-induced vascular dysfunction in mice. 21Nevertheless, the mechanisms and interactions of IL-6 in CVD are incompletely understood. 22As myeloid cells are one of the relevant physiological sources of IL-6, 23 we generated mice with chronic myeloid IL-6 overexpression and analysed their vascular phenotype in order to determine the role of myeloid cell-derived IL-6 for systemic and vascular inflammation.

Methods Mice
Interleukin-6-overexpressing mice were generated by gene targeting of embryonic stem (ES) cells.In brief, we have constructed a targeting vector by inserting the cDNA coding for IL-6 followed by an internal ribosome entry site and the cDNA coding for enhanced green fluorescent protein (eGFP) (as shown in Figure 1A).The conditional 'knock-in' approach targeted the endogenous gt(ROSA)26Sor locus, introducing a lox-P-flanked transcriptional STOP cassette (Figure 1A).Following homologous recombination into C57BL/6-V6.5 ES cells, we generated the IL-6 OE mice by injection of ES cells into blastocysts.This new mouse strain established on the C57BL/6J background allows cell-specific overexpression of IL-6 combined with eGFP expression (IL-6 OE mouse strain) upon Cre-mediated control.By crossing homozygous IL-6 OE/OE female mice with LysMCre + male mice, 24 we obtained LysM-IL-6 OE mice with heterozygous expression of LysMCre and heterozygous overexpression of IL-6 in LysM + cells.Control mice were LysMCre negative and heterozygous for the IL-6 OE allele (IL-6 OE mice).At the age of 8-12 weeks, mice were used for experiments.Both female and male mice were examined, and the results were pooled as no sexspecific effect was observed.
Mice were euthanized by exsanguination under deep isoflurane anaesthesia (5% inhalant in room air) combined with tamasic (buprenorphine) as a single subcutaneous injection (0.075 mg/kg body weight) 30 min before the procedure to avoid any pain.Then, blood was collected by right ventricular puncture.Afterwards, organs, namely aorta, spleen, bone marrow (BM), heart, brain, colon, lungs, and kidneys, were harvested.
Housing, treatment, and euthanasia of animals were performed in accordance with the relevant laws and institutional guidelines of the Central Animal Facility of the UMC Mainz, Germany, and state authorities (G15-1-051, G15-1-101, G21-1-025, and G21-1-030).Experimental procedures including anaesthetic and analgesic treatments followed the guidelines from Directive 2010/63/EU of the European Parliament on the protection of animals used for scientific purposes.Experiments were conducted according to the ARRIVE guidelines.

Vascular relaxation studies
Murine aortas were isolated and cut into 4 mm long segments, and the thoracic part of the aortas was utilized for aortic constriction and dilatation assays.They were mounted in organ chambers to measure endothelium or smooth muscle-dependent relaxation or contraction in response to acetylcholine (ACh), glyceryl trinitrate, potassium chloride (KCl), and prostaglandin F2α (PGF2α) as described previously. 25he ROS scavengers catalase (2000U/mL), superoxide dismutase (SOD) (2000U/mL), or the endothelin-1 (ET-1) antagonist bosentan (10 µM) were suspended in full medium (MV2 kit, PromoCell), and the isolated aorta segments were incubated in it for 90 min at 37°C and 5% CO 2 .Afterwards, relaxation studies were performed.
Besides the aortic analysis, the first and second branches of mesenteric arteries were isolated, cut into 2 mm long segments, and mounted in a myograph to measure relaxation in response to ACh or sodium nitroprusside or contraction in response to phenylephrine or KCl as described previously. 26,27

Echocardiography
Mouse echocardiography was performed using a VEVO-3100 ultrasound machine (FUJIFILM VisualSonics Inc., Toronto, Canada) equipped with a 38 MHz (MX400) linear array transducer.Standardized images in parasternal long axis (PLAX) and parasternal short axis as well as apical four-chamber view were recorded under isoflurane sedation with 1.5-2 vol%.In parallel, electrocardiogram and respiration rate were monitored and body temperature was kept stable using a heating system within the handling platform.Post-acquisition analysis was performed with the Vevo LAB software (VisualSonics, FujiFilm, Toronto, Canada) for left ventricular (LV) ejection fraction (PLAX B-mode), LV mass, stroke volume, and cardiac output (PLAX M-mode) and right heart function with right ventricular end-diastolic (mid-ventricular diameter) (4C and SAX) and pulmonary artery velocity time integral (SAX).

Blood pressure measurement
Systolic blood pressure was measured using a tail-cuff non-invasive blood pressure system as previously described. 28,29

Analysis of reactive oxygen and nitrogen species in the blood and aortic tissue
Reactive oxygen species (ROS)/reactive nitrogen species (RNS) were measured in whole blood by L012-enhanced chemiluminescence after incubation with phorbol 12,13-dibutyrate (PDBu) (10 μmol/L) for 20 min: [30][31][32] 200 IU of heparine was injected into the beating heart of the mouse and venous blood drawn from the right ventricle subsequently.Chemiluminescence was measured directly in the heparinized blood which was kept at room temperature (RT).L012-enhanced chemiluminescence signals were counted in 10 μL of samples with 100 μmol/L of L012 in the presence of PDBu using a Lumat LB 9507 from Berthold Technologies (Bad Wildbad, Germany).Chemiluminescence was expressed as counts per minute after incubation for 10 min.
Visualization of ROS in aortic cryosections was performed by staining with the superoxide-sensitive dye dihydroethidium (DHE, 1 µM) as previously described. 32Mouse aortas were rinsed, cleaned, and then cut into 4 mm sections which were then incubated in Krebs-Henseleit solution (containing 0.1 mg/mL of aprotinin, 0.2 mg/mL of pepstatin, and 0.5 mg/ mL of leupeptin) for 10 min at 37°C.Then, the aortic sections were embedded in Tissue-Tek and frozen in liquid nitrogen.8 µm aortic cryosections were cut and stained with DHE (incubation over 30 min at 37°C).Green autofluorescence from aortic lamina and red ethidium fluorescence inside the ROS-producing cells were detected by fluorescence light microscopy using a Zeiss Axiovert 40 CFL microscope, Zeiss lenses and Axiocam MRm camera (Zeiss, Oberkochen, Germany) and analysed with the AxioVision data acquisition software (Zeiss).

In vivo high-resolution endoscopic analysis of the colon
A high-resolution video endoscopic system (Karl Storz SE & Co AG, Tuttlingen, Germany) was used for the analysis of the colitis-like phenotype.Colitis scores were determined on mice anaesthetized with a mixture of ketamine (Ketavest 100 mg/mL; Pfizer, New York, NY) and xylazine (Rompun 2%; Bayer Healthcare, Leverkusen, Germany) intraperitoneally.Endoscopic scoring of five parameters, namely translucency, granularity, fibrin, vascularity, and stool, was performed as previously described. 33

Flow cytometric reactive oxygen species analysis
For flow cytometric ROS analysis, whole blood was drawn by cardiac puncture and anticoagulated with ethylenediaminetetraacetic acid.Cells were treated with and without PDBu for 15 min at RT followed by incubation with CellROX® reagent (a fluorogenic indicator for the detection of ROS in cells, which is non-fluorescent but shows a strong fluorescence when oxidized; 5 μM, Thermo Fisher Scientific) for 30 min at 37°C, as previously described. 34

Cytokine and chemokine quantification
Plasma IL-6 levels were measured with BD OptEIA™ mouse ELISA kit.Absorption was measured using the infinite M200 PRO NanoQuant plate reader (Tecan, Austria) or with the Bioplex Multiplex Immunoassay System ( BioRad, Germany), according to the manufacturer's protocol.

Isolation and cultivation of vascular smooth muscle cells
Vascular smooth muscle cells were isolated from the aortas of LysM-IL-6 OE mice and cultured as previously described 36 In brief, aortas were carefully dissected and perfused with 1 × Phosphate buffered Saline (Gibco) to remove any blood.The aorta was kept in Dulbecco's modified Eagle medium (DMEM; GlutaMAX-I; Gibco) with 10% fetal bovine serum (FBS; Gibco), 100 U/mL penicillin, and 100 μg/mL streptomycin (Gibco).After removing the perivascular fat under the microscope, aortas were cut into small pieces and digested in medium containing collagenase II (10 mg/mL) for 3 h in a thermoshaker.Then, the medium was removed by centrifugation (400 g for 10 min), and cells were resuspended in DMEM containing 10% FBS, 100 U/mL penicillin, and 100 μg/mL streptomycin on 0.1% gelatin-coated six-well plates.Cells were analysed at passage 0.

Histology
At the time of sacrifice, aorta, spleen, BM, heart, brain, colon, lungs, and kidneys were isolated from the experimental mice, fixed in 4% paraformaldehyde, paraffin-embedded, cut, and stained with haematoxylin and eosin according to standard protocols. 37Aortic sections were stained for collagen fibres using Sirius Red and analysed by polarized light microscopy.

Immunohistochemistry
Aortic segments were fixed in 4% paraformaldehyde, paraffin-embedded, and stained with the primary antibody (anti-ET-1, 1:500, Meridian Life Science, USA) followed by incubation with the biotinylated secondary antibody after dilution.For immunochemical detection, ABC reagent (Vector) and then DAB reagent (peroxidase substrate Kit, Vector) as substrates were applied. 38

RNA isolation and quantitative real-time PCR
Aortas and cells were lysed and homogenized in TRIzol® (Thermo Fisher) and processed following the manufacturer's instructions.RNA concentration was measured using a NanoDrop spectrophotometer (Thermo Fisher).The TaqMan® Gene Expression Assay (Applied Biosystems™) was used with 0.5 μg of total RNA for quantitative realtime PCR for Tatabox-binding protein (Tbp and MMP9 (Mm00442991_m1).For human material the following Primers were used: Tbp (Hs00427620_m1), Il6 (Hs00174131_m1), Et-1 (Hs00174961_m1).Besides, the following primers were used: glyceraldehyde-3-phosphate dehydrogenase (Gapdh) (forward: TAC CCC CAA TGT GTC CTG CTG G; reverse: CCT TCA GTG GGC CCT CAG ATG C); Col3α1 (forward: GAG GGC CAT AGC TGA ACT GA; reverse: TGA CTG TCC CAC GTA AGC AC), Vimentin (forward: CGG AAA GTG GAA TCC TTG CA; reverse: CAC ATC GAT CTG GAC ATG CTG T), and αSma (forward: GGACGTACA ACTGGTATTGTGC; reverse: CGGCAGTAGTCACGAAGGAAT).Quantification of mRNA expression as relative expression levels of the respective samples to Tbp or Gapdh as endogenous control (housekeeping genes) was calculated with the delta-delta threshold cycle method.

Western blot analysis
Isolated aortic tissue was homogenized, and protein was isolated using Radio-Immunoprecipitation assay buffer (Sigma-Aldrich) containing a protease/phosphatase inhibitor cocktail (Thermo Scientific).Aortic lysates were electrophorized on a 7-12% gradient gel.Proteins were transferred on a polyvinylidene difluoride membrane (pore size 0.45 μm).Immunoblotting was performed with the following primary antibodies: inducible NO synthase (iNOS) (BD Bioscience), ET-1 (Abcam), and alpha-/ beta-actinin (Cell Signaling Technologies).For detection with automated imaging equipment (FUSION CCD Imager), either anti-mouse or antirabbit (Cell Signaling Technologies) secondary antibodies coupled to horseradish peroxidase in combination with ECL Western Blotting Substrate (Thermo Scientific) were used.Final quantification was performed by densitometric analysis (GelPro-analyzer).

Statistical analysis
Data are displayed as mean ± SEM.Statistical calculation was done with GraphPad Prism software (version 9; GraphPad Software Inc.).Kolmogorov-Smirnov test was used for analysis of normal distribution.In the case of normal distribution, we performed the unpaired Student's t-test to compare two experimental groups and the one-way analysis of variance (ANOVA) test with Tukey's post hoc test for comparison of more than two groups.If non-normal distribution was given, Mann-Whitney test was applied for comparison of two groups and Kruskal-Wallis with Dunn's multiple comparison test to compare more than two groups as indicated in the figure legends.Comparison of aortic relaxation curves and survival curves was made using a two-way ANOVA with Bonferroni's post hoc test and log-rank (Mantel-Cox) test, respectively.P values of <0.05 were considered significant and marked by asterisks (*P < 0.05; **P < 0.01; ***P < 0.001).

Myeloid cell-derived interleukin-6 evokes a chronic systemic inflammatory state with an expansion of the myeloid cell compartment paralleled by vascular dysfunction in mice
To understand the function of IL-6 in vascular disease, we generated mice with conditional IL-6 overexpression (IL-6 OE/OE mouse strain).Crossing the IL-6 OE/OE mouse strain with LysM-Cre + mice 24 resulted in mice in which myeloid cells overexpress IL-6 and eGFP (LysM-IL-6 OE mice, Figure 1A).We confirmed this by flow cytometry of splenocytes (Figure 1B) and IL-6 ELISA of the supernatants of isolated and in vitro cultured CD11b + cells (Figure 1C, left).Myeloid IL-6 overexpression resulted in significantly elevated circulating IL-6 plasma levels compared with non-detectable concentrations in littermate controls (Figure 1C, right).Of note, the IL-6 levels in these mice were considerably lower than those reported during acute sepsis. 40he increase in systemic IL-6 was accompanied by an increase in splenic total cell numbers shown by flow cytometric analysis in the LysM-IL-6 OE mice (Figure 1D).We found an expansion of splenic CD11b + myeloid cells, CD11b + Ly6G + Ly6C + neutrophil granulocytes, and CD11b + Ly6G − Ly6C + monocytes/macrophages (shown as total cell counts and the percentage of viable splenic cells, Figure 1E and F; gating strategy is shown in Supplementary material online, Figure S1A).
The CD11b + myeloid cells and the CD11b + Ly6G + Ly6C + neutrophil granulocytes were also increased in the blood, whereas the number of systemic CD11b + Ly6G − Ly6C + monocytes/macrophages was not altered (see Supplementary material online, Figure S1B and C).There was no change in blood CD90.2 + T cells detectable in the LysM-IL-6 OE mice compared with controls (see Supplementary material online, Figure S1D).The total number of splenic T cells was not altered either, but a reduction in the percentage of splenic T cells of total living cells was apparent (see Supplementary material online, Figure S1E).
Chronic elevations of circulating IL-6 levels were associated with an increased mortality in mice.Starting premature death at 9 weeks, only 12.5% of LysM-IL-6 OE mice reached 15 weeks of age (Figure 1G).To better understand the reduced lifespan in the LysM-IL-6 OE mice, we cautiously monitored the mice and performed detailed (histological) analysis of 11-12-week-old LysM-IL-6 OE mice compared with control mice (Figure 1H and Supplementary material online, Figure S2).It was apparent that the LysM-IL-6 OE mice were prone to the development of intestinal prolapses.This was accompanied by a colitis-like phenotype shown by histological analysis and endoscopy (Figure 1H).Additionally, LysM-IL-6 OE mice revealed inflammatory changes in the lungs, liver, spleen, and kidney (see Supplementary material online, Figure S2A-D), all of which-in particular the pulmonary inflammationcould contribute to premature death.The brain was apparently histologically not affected (see Supplementary material online, Figure S2E).There were no macro-or microscopic signs of thrombosis or bleeding in all analysed organs (see Supplementary material online, Figure S2).In summary, LysM-IL-6 OE mice developed signs of systemic inflammation with an IL-6-driven colitis-like phenotype.
We then focused on the analysis of the vascular system.Maximal constriction of aortic rings from LysM-IL-6 OE mice to PGF2α was greater than that of control mice, whereas responses to KCl did not differ between genotypes (Figure 2A).Vessels from LysM-IL-6 OE mice also demonstrated endothelial dysfunction evidenced by an impaired ACh-induced relaxation (Figure 2A).On the other hand, endothelium-independent vascular relaxation was comparable in segments from wild-type (WT) and LysM-IL-6 OE mice, as was blood pressure (see Supplementary material online, Figure S3A and B).In mesenteric arteries, contractile responses to KCl were also unaffected by the overexpression of IL-6 in myeloid cells (Figure 2B).Additionally, ACh-induced relaxation was abrogated in mesenteric arteries from LysM-IL-6 OE mice (Figure 2B).This indicated a clear link between a myeloid cell-derived cytokine and severe endothelial dysfunction that was independent of any change in VSMC reactivity or blood pressure (see Supplementary material online, Figure S3C).Endothelial dysfunction is linked to increased ROS production, 41 and ROS/RNS levels were significantly increased in blood from LysM-IL-6 OE mice (Figure 2C).Furthermore, the CD11b + myeloid cells demonstrated an increased ROS production following PDBu stimulation by trend, underpinning their increased reactivity (Figure 2D).We detected no differences in the heart size or function in the LysM-IL-6 OE mice compared with controls.There were no structural or functional differences in the heart, as assessed by histology and echocardiography (see Supplementary material online, Figure S4).
Taken together, these data demonstrate that the chronic systemic inflammatory state evoked by chronic myeloid IL-6 overexpression is associated with vascular dysfunction without significant cardiac alterations.

Interleukin-6-induced vascular dysfunction is associated with vascular inflammation, increased reactive oxygen species formation and endothelin-1 expression, vascular smooth muscle cell phenotype alterations, and fibrosis in the vasculature
Flow cytometric analysis of CD11b + cells in aorta showed that vascular dysfunction in LysM-IL-6 OE mice was accompanied by an increase in Ly6G + Ly6C + neutrophils and by Ly6G − Lys6C + monocytes/macrophages, but neither CX3CR1 + nor the CD115 + CD11b + monocytes/ macrophages, in the aortic vessel wall (Figure 3A and Supplementary material online, Figure S5A (gating strategy) and B).Total aortic CD11b + cell numbers were not increased, but we noted a shift towards the Ly6G + Ly6C + neutrophils within this population (Figure 3A).The quantity of CD90.2 + T cells was increased in LysM-IL6 OE aortas, whereas no significant change in aortic CD4 + or CD8 + T cells was observed (see Supplementary material online, Figure S5C and D).In parallel to the described inflammation in the aortic vessel wall, increased ROS formation in the aortic vasculature was apparent (Figure 3B).In line, Mcp1, VCAM1, Cxcl1, Tnf, Nox2, Il6, STAT3, Rorc, iNOS, and VEGFa mRNA expression in the aortic vessel wall of LysM-IL-6OE mice was significantly increased (Figure 3C).Next to the increased protein expression of the inducible NO synthase (iNOS) in the LysM-IL-6 OE aortas contributing to increased vascular ROS formation (Figure 3D), ET-1, a potent vasoconstrictor and mediator of vascular dysfunction, 42 was elevated in protein expression (Figure 3D) and histological analysis (Figure 3E).Stimulation of HPAECs with IL-6 and sIL-6R resulted in a significantly increased ET-1 expression (Figure 3F), suggesting that IL-6 directly triggers ET-1 expression in endothelial cells.This observation, in combination with the reported capacity of ET-1 to drive IL-6 expression in endothelial cells, 43 may trigger a vascular dysfunction vicious circle.Of note, IL-6 synthesis in pulmonary endothelial cells isolated from LysM-IL-6 OE mice was increased compared with controls (see Supplementary material online, Figure S6A).Expression of the IL-6 receptor IL-6R was increased in the aortic tissue of LysM-IL6 OE mice, further suggesting that IL-6 directly acts on the vessel wall in our experimental mice (see Supplementary material online, Figure S6B).Focusing on the role of ROS in this context, we applied the ROS scavengers SOD and catalase, respectively, over 90 min in the organ bath (see Supplementary material online, Figure S7) and found no relevant improving effect.Additionally, application of the ET-1 antagonist bosentan displayed no effect (see Supplementary material online, Figure S7).This indicated that both ROS and ET-1 do not function as acute drivers of vascular dysfunction in this mouse model, but rather contribute to the chronic effects of myeloid cell-derived IL-6 on the vasculature.
In regard to the complex long-term vascular effects evoked by myeloid cell-derived IL-6 overexpression on the vasculature, we found that the amount of interstitial collagen thickness in the vascular wall of LysM-IL-6 OE mice was moderately increased, whilst the aortic wall thickness remained unchanged (Figure 4A).This is in line with the concept that vascular inflammation promotes vascular fibrosis. 44To investigate the mechanistic underpinnings, we examined cultured VSMCs isolated from LysM-IL-6 OE and control mice by immunocytochemistry (Figure 4B): here, no significant differences in proliferation (measured via Ki67) were apparent, although the presence of 10% serum in the medium may have masked minor changes.Expression of the mechanical stress marker F-actin did not differ between the VSMCs isolated from LysM-IL-6 OE and control mice.However, there was a trend in SMA to be less expressed in LysM-IL-6 OE VSMCs compared with controls (Figure 4B).This was corroborated in aortic tissue, where we found significantly less SMA mRNA expression in the LysM-IL-6 OE compared with control mice (Figure 4C).We found increased matrix metallopeptidase-9 (Mmp9) mRNA levels in the LysM-IL-6 OE mice (see Supplementary material online, Figure S8).The mRNA levels of Mmp2, collagen (Col1a1, Col1a2, Col3a1), or vimentin did not differ significantly (see Supplementary material online, Figure S8).

Interleukin-6 produced by bonemarrow-derived cells correlates with vascular dysfunction and systemic inflammation
To elucidate the contribution of myeloid cells derived from the BM, we generated chimeric mice transplanted with different ratios of BM isolated from LysM-IL-6 OE mice (100, 50, and 10%) mixed with BM of control mice up to 100% (Figure 5A).Interestingly, circulating IL-6 levels in plasma (Figure 5B), vascular endothelial dysfunction (Figure 5C), and oxidative stress levels in blood (Figure 5D) all correlated with the amount of LysM-IL-6 OE mice BM transferred and were most significantly pronounced in WT mice reconstituted with 100% LysM-IL-6 OE BM cells.Importantly, vascular function in WT mice transplanted with only 10% of LysM-IL-6 OE BM was unchanged to controls (Figure 5C), indicating a dose-dependent effect of IL-6 on systemic inflammation and vascular dysfunction.The amount of CD11b + myeloid cells invading in the aortic vessel and spleen as markers of the IL-6-driven chronic systemic inflammation also correlated with the transferred amount of LysM-IL-6 OE mice BM (Figure 5E and F).
These findings show that vascular dysfunction and inflammation, as well as systemic inflammation and ROS formation, are driven by BM cell-derived IL-6 in a dose-dependent manner.

Discussion
Our results highlight the crucial long-term impact of myeloid cellderived IL-6 on the development of vascular inflammation and dysfunction and underline a clear dose-dependent effect.
For us as clinical scientists based in cardiology, this mouse model was generated with the main aim to decipher the role of myeloid cellderived IL-6 in CVD.This is of high clinical relevance, as elevated IL-6 levels in patients with CVD (as also seen for example in obesity 45 ) seem to predict a poorer outcome, 22,46 underlining the need for a better understanding of IL-6 in CVD.
The well-established central role of IL-6 in systemic inflammation 47 was confirmed in our mouse model.Additionally, our mouse model highlights the central role of IL-6 in inflammatory bowel disease (IBD). 15This is in line with the observation that patients with IBD have a higher risk of cardiovascular events compared with patients without IBD despite an unchanged incidence of conventional cardiovascular risk factors. 17Our mouse model also reflects the cardiovascular comorbidity of patients with IBD.One relevant factor of the cardiovascular comorbidity in IBD might be the chronic systemic inflammation. 17he microbiome, which is relevant in CVD development 48 and IBD, 49 might represent a further interconnecting factor.
To the best of our knowledge, this is the first report on the chronic effects of IL-6 on the vasculature in a mouse model.Elevated levels of IL-6, activated downstream of the cytokine IL-1β, are known to be associated with an increased rate of adverse cardiovascular events in patients with chronic atherosclerotic disease. 50,51The inflammation concept in CAD supports the idea to specifically target the IL-6 receptor pathways to prevent or attenuate CAD. 52Interleukin-6 was found in atherosclerotic plaques 53 and shown to promote atherosclerotic lesion progression in mice. 54Recently, it has been demonstrated that treatment approaches of antagonizing IL-6 could reduce various biomarkers of systemic inflammation associated with the atherothrombotic process in patients with high atherosclerotic risk. 55,56In contrast, a potentially atheroprotective role of IL-6 has also been reported. 57,58his underscores that experimental results on the role of IL-6 in the cardiovascular system must be interpreted carefully. 22n our study, we detected that IL-6 induces vascular dysfunction systemically-in both large and small vessels.This was paralleled by an increased ROS formation, recapitulating findings by Wassmann et al. 59 in mice treated with IL-6 over 18 days.In parallel, Mcp1, Vcam1, Cxcl1 and 2, Tnf, iNOS, Nox2, Stat3, and Rorc, key drivers and relevant components of vascular dysfunction and inflammation, were upregulated in the LysM-IL-6 OE aortas.The increased infiltration of myeloid cells into the aortic vessel wall and the increased vascular ROS formation detected in LysM-IL-6 OE mice are fully compatible with the activating effect of IL-6 on myeloid cells. 8,60In comparison with mice with AngII-induced vascular dysfunction (1 week of AngII treatment), the number of myeloid cells invading into the aortic vessel wall seemed higher in the AngII-treated mice. 61This suggests that it is necessary to differentiate between long-term inflammation, as seen in the LysM-IL-6 OE mouse model, and the short-term inflammation in the AngII model.However, since ROS scavengers as catalase could not attenuate vascular dysfunction when applied to LysM-IL-6 OE aortas in vitro, this indicates that not acute exposure to oxidative stress, but rather long-term impact of chronic IL-6-driven inflammation renders the vasculature dysfunctional.
We therefore focussed on these long-term effects: we found a moderately increased fibrosis in the aortic vessel walls of the LysM-IL-6 OE mice in line with the reported fibroblast stimulating effect of IL-6. 62Although this increase in the interstitial collagen thickness was significant, we were not faced with a major difference in size.This finding may pave the way for long-term IL-6-driven alterations of the vasculature and reflect an ongoing process.In any case, the increased mRNA expression of MMP-9 is compatible with an increased collagen turnover. 63Further, we found an altered phenotype of VSMCs in the form of loss of contractile markers, suggesting a phenotypic switch towards a more synthetic phenotype. 64,65We therefore hypothesize that the vessels in the LysM-IL-6 OE mice are most likely in the process of an altered collagen formation.Nevertheless, they still provide an increased capacity to constrict upon stimulus.Taken together, these findings conclude that the vascular dysfunction initiated by myeloid cell-derived IL-6 overexpression involves the activation of multiple vascular cell types and release of downstream mediators and finally impacts the vascular collagen and VSMC phenotype.

Interleukin-6 drives vascular inflammation
Vascular dysfunction in LysM-IL-6 OE mice was associated with an altered VSMC phenotype but did not evoke blood pressure elevation and systemic hypertension.The nitric oxide release of iNOS, which was upregulated in the LysM-IL-6 OE aortas, could potentially blunt a potential blood pressure increase.The increased aortic ET-1 expression in the LysM-IL-6 OE mice is strengthened by reports that chronic endothelial ET-1 overexpression is associated with normal blood pressure 66 (in contrast to short-term endothelial overexpression of ET-1 resulting in hypertension 67 ).Interestingly, our data demonstrate IL-6 to be capable of driving vascular ET-1 expression.In total, this suggests that chronic IL-6 expression has the potential to trigger a vicious cycle of vascular dysfunction and inflammation.High expression of the IL-6 receptor in the aortic tissue in LysM-IL-6 OE mice suggests that IL-6 can directly impact the vasculature.This assumption is in line with data reported for human Endothelial cells (ECs). 68Further analysis has to follow on the exact IL-6 signalling in ECs and VSMCs under chronic IL-6 exposure.Here, the dose-response curve of IL-6 in the cardiovascular system must be kept in mind to determine at which IL-6 quantity chronic IL-6 exposure evokes changes in the cardiovascular system.The BM transfer experiments we performed suggest that the number of IL-6-producing myeloid cells in the BM correlates directly with systemic and vascular inflammation.Further studies on the impact of anti-IL6 treatment in this mouse model-and in a mouse model of vascular dysfunction-are needed.
The reported mouse model, as all genetic mouse models, certainly is not fully physiologic.However, extrapolating murine experimental knowledge from bench to bedside can raise necessary awareness that patients with elevated IL-6 levels (e.g.due to chronic IL-6-driven inflammatory diseases) are at risk to develop vascular inflammation.Since IL-6 levels correlate not only with systemic inflammation but also with vascular dysfunction and inflammation, it must be one treatment goal to keep the underlying inflammatory situation limited.

Conclusion
Long-term myeloid cell-derived IL-6 expression evokes a chronic systemic inflammatory state with autoimmune colitis as a leading symptom.This is associated with significant vascular inflammation and dysfunction, ROS formation, elevated vascular ET-1 expression, alterations of the VSMCs, and increased collagen formation.Vascular dysfunction and ROS formation were driven by BM cell-derived IL-6 in a directly dose-dependent manner.Based on our findings, we conclude that long-term reduction in systemic IL-6 levels (e.g. by treating the existing autoimmune disease) is essential to reduce the associated vascular phenotype.
Further research focused on the complex impact of central cytokines as IL-6 on vascular regulation in (chronic) inflammatory diseases will lead to a better understanding of the vascular component in inflammatory diseases and new targets amenable for drug therapy.

Lead author biography
Tanja Knopp studied anthropology in Mainz, Germany, and focused on the impact of interleukin-6 on the vasculature and on the coagulation cascade during her thesis.As a post-doc in Bern, Switzerland, she is now working in the field of thrombosis and haemostasis.
Rebecca Jung studied biology in Mainz, Germany.As a PhD candidate, she concentrated on the role of the cytokines IL-6 and IL-17A in vascular dysfunction development and on the cardiovascular comorbidity in the IL-17A-driven autoimmune skin disease psoriasis.She meanwhile switched from university to industry and continued working in the field of autoimmune diseases.

Figure 1
Figure 1 Mice overexpressing interleukin-6 in myeloid cells have an increase in the splenic myeloid cell compartment and a reduced life expectancy.(A) Generation of the IL-6 OE allele involved homologous recombination in embryonic stem cells (V6.5).The conditional 'knock-in' approach targeted the endogenous gt(ROSA)26Sor locus, introducing a lox-P-flanked transcriptional STOP cassette.Upon Cre-mediated recombination, the cassette was excised, enabling dual expression of interleukin-6 and enhanced green fluorescent protein under the control of the chicken β-actin (CAG) promotor.(B) Flow cytometric analysis of splenocytes from LysM-IL-6 OE and control mice.Cells were stained for CD11b, B220, and CD90.2 with gating based on green fluorescent protein signal.Representative plots of n = 4 mice are shown.(C) Interleukin-6 levels in splenocytes and plasma.Left panel: interleukin-6 level in sorted and cultured CD11b + splenocytes of LysM-IL-6 OE mice and control mice.Splenocytes were cultured for 24 h.n = 3-4, Mann-Whitney test.Right panel: interleukin-6 plasma levels in 10-week-old LysM-IL-6 OE compared with control mice (n.d.= not detectable).n = 19-28, Mann-Whitney test.P < 0.0001.(D) Statistical analysis of the total living cells per spleen of LysM-IL-6 OE and control (continued)

Figure 4 Figure 5
Figure 4 Interleukin-6-induced vascular dysfunction is associated with an altered vascular smooth muscle cell phenotype and fibrosis formation.(A) Collagen deposition in aortic sections.Sirius Red staining of aortic sections.Left: representative image of aortic sections (images without polarized light are shown below), scale bar = 50 µm.Right: aortic wall thickness and collagen thickness measurement at 10 different points/section with ImageJ software, n = 5, unpaired Student's t-test.P = 0.0079 (collagen thickness).(B) Immunocytochemical analysis of vascular smooth muscle cell phenotype.Vascular smooth muscle cells were isolated from LysM-IL-6 OE mice and control mice and cultured.Cultured vascular smooth muscle cell were stained with smooth muscle alpha-actin, F-actin, Ki67, and DAPI (60×).Representative images are shown in the top row.Quantification is shown in the bottom row.n = 3 mice per group of two independent experiments.Single images of three biological replicates per mouse were analysed, and the mean values each mouse were compared, unpaired Student's t-test.(C ) Quantitative real-time PCR analysis for alpha SMA in LysM-IL-6 OE mice aortas (red) vs. control aortas.Housekeeping gene: Gapdh.n = 10-12 mice per group, unpaired Student's t-test.P = 0.03.Data are presented as mean ± SEM, and P values of <0.05 were considered significant and marked by asterisks (*P < 0.05; **P < 0.01; ***P < 0.001).