P2X4 deficiency reduces atherosclerosis and plaque inflammation in mice

Extracellular adenosine-5′-triphosphate (ATP) acts as an import signaling molecule mediating inflammation via purinergic P2 receptors. ATP binds to the purinergic receptor P2X4 and promotes inflammation via increased expression of pro-inflammatory cytokines. Because of the central role of inflammation, we assumed a functional contribution of the ATP-P2X4-axis in atherosclerosis. Expression of P2X4 was increased in atherosclerotic aortic arches from low-density lipoprotein receptor-deficient mice being fed a high cholesterol diet as assessed by real-time polymerase chain reaction and immunohistochemistry. To investigate the functional role of P2X4 in atherosclerosis, P2X4-deficient mice were crossed with low-density lipoprotein receptor-deficient mice and fed high cholesterol diet. After 16 weeks, P2X4-deficient mice developed smaller atherosclerotic lesions compared to P2X4-competent mice. Furthermore, intravital microscopy showed reduced ATP-induced leukocyte rolling at the vessel wall in P2X4-deficient mice. Mechanistically, we found a reduced RNA expression of CC chemokine ligand 2 (CCL-2), C-X-C motif chemokine-1 (CXCL-1), C-X-C motif chemokine-2 (CXCL-2), Interleukin-6 (IL-6) and tumor necrosis factor α (TNFα) as well as a decreased nucleotide-binding oligomerization domain-like receptor protein 3 (NLRP3)-inflammasome priming in atherosclerotic plaques from P2X4-deficient mice. Moreover, bone marrow derived macrophages isolated from P2X4-deficient mice revealed a reduced ATP-mediated release of CCL-2, CC chemokine ligand 5 (CCL-5), Interleukin-1β (IL-1β) and IL-6. Additionally, P2X4-deficient mice shared a lower proportion of pro-inflammatory Ly6Chigh monocytes and a higher proportion of anti-inflammatory Ly6Clow monocytes, and expressend less endothelial VCAM-1. Finally, increased P2X4 expression in human atherosclerotic lesions from carotid endarterectomy was found, indicating the importance of potential implementations of this study’s findings for human atherosclerosis. Collectively, P2X4 deficiency reduced experimental atherosclerosis, plaque inflammation and inflammasome priming, pointing to P2X4 as a potential therapeutic target in the fight against atherosclerosis.

www.nature.com/scientificreports/ Nucleotides such as adenosine-5′-triphosphate (ATP) play a crucial role as cellular energy carrier. During inflammatory activation, invoked by ischemia, cellular stress or cell death, nucleotides are released into the extracellular space 8 . Extracellular nucleotides act as important signaling molecules mediating inflammation via the purinergic P2 receptors, which in turn are divided into G-protein-coupled P2Y receptors and ligand-gated ion channel P2X receptors 9 . Purinergic P2 receptors are involved in the pathogenesis of a multitude of chronic inflammatory diseases like asthma, chronic obstructive pulmonary disease, inflammatory bowel disease or graft versus host disease [10][11][12] . Simultaneously, purinergic receptors perform an important function in the pathophysiology of atherosclerosis. Recent studies showed that the metabotropic receptors P2Y 1 , P2Y 2 , P2Y 6, and the ligandgated receptor P2X 7 lead to experimental atherosclerosis by mediating leukocyte recruitment to the vessel wall and caused vascular inflammation by promoting adhesion molecule expression, cytokine expression as well as nucleotide-binding oligomerization domain-like receptor protein 3 (NLRP3)-inflammasome activation [13][14][15][16][17] .
The purinergic receptor P2X 4 is a ligand-gated ion channel responsive to ATP and permeable to Na + , K + as well as Ca 2+ ions. It is the most sensitive purinergic receptor at nanomolar ATP concentrations and at the same time the most expressed P2X receptor subtype in the heart 18,19 . The distribution of P2X 4 extends over a variety of stromal cell types including endothelial cells, smooth muscle cells, cardiomyocytes, epithelial cells, fibroblasts, peripheral and central neurons as well as leukocytes like dendritic cells, eosinophils and macrophages. While P2X 4 is primarily stored intracellularly in lysosomes, the expression of P2X 4 on the cell surface is extremely sensitive and particularly upregulated by stimuli such as ischemia or inflammation 20,21 . Hence, P2X 4 is known to mediate inflammation in several disease models like kidney injury, chronic lung disease, rheumatoid arthritis and graft-versus-host disease. Activation of P2X 4 in response to high extracellular ATP levels evokes the release of cytokines like Interleukin 6 (IL-6) and tumor necrosis factor α (TNFα), promoting NLRP3-inflammasome signaling as well as activation of endothelial cells [22][23][24] .
Based on this background information, the aim of this study is to investigate the role of P2X 4 in experimental atherosclerosis and vascular inflammation.

Materials and methods
Mice. All methods are reported according to the ARRIVE guidelines. P2X 4 -deficient mice on homozygous C57Bl/6 background (P2rx4 tm1Rass ) were kindly provided by Prof. Dr. Idzko (Medical University of Vienna). LDL-receptor knockout mice on C57Bl/6 background (Ldlr tm1Her ) were purchased from Jackson Laboratories (Bar Harbor, Maine, USA). By crossing P2X 4 -deficient mice with LDLR-deficient animals, P2X 4 −/− /LDLR −/− double knockout animals were generated. Animals were kept under specifically pathogen-free (SPF) conditions at the animal facility of the University of Freiburg. All procedures were approved by the governmental animal care committee (Regierungspraesidium Freiburg, Germany, 35-9185.81/G-17/172) and conformed to the guidelines from the EU directive 2010/63 EU of the European Parliament.
In vivo atherosclerosis study. 6 weeks old, male, P2X 4 −/− /LDLR −/− mice (n = 24) and P2X 4 +/+ /LDLR −/− mice (n = 15) were fed high cholesterol diet (sniff EF R/M acc. D12108 mod., ssniff Spezialdiäten GmbH, Soest, Germany) for 16 weeks. After 16 weeks of diet, mice were harvested as previously described 15 . Histology. From the aortic samples of the atherosclerosis study, sections were prepared for histological and immunohistochemical preparations according to the protocol described before 15 . Stainings with Oil-red-O, anti-Mac-3, Sirius red and anti-α-actin were performed as described previously 15 . Anti-CD4 staining was realized using a CD4 immunofluorescence kit (Sigma-Aldrich, St. Luis, MO, USA) according to the manufacturer's protocol. Plaque size was determined in sections of the aortic root, arch and abdominal aorta by analyzing total wall area, intimal lesion area, and medial area. Necrotic core size was analyzed by measuring acellular intima lesion areas in sections of the aortic root as previously described 25 . For the analysis of plaque composition, the percentage of positively stained area was quantified for lipids (Oil-Red-O), macrophages (anti-mouse Mac-3), collagen (Sirius red), smooth muscle cells (anti-α-actin), and CD4+ cells (anti-CD4) within the intimal lesion area.
Staining of active caspase-1 in aortic roots was made as described previously using a FLICA-fmk kit (Immunohistchemistry Technologies) 17 . The percentage of FLICA positive area was quantified.
To determine the expression of P2X 4 in aortic roots, we performed a P2X 4 staining using a P2X 4 antibody (United States Biological, Salem, MA, USA).
For analysis of endothelial expression of VCAM-1 and ICAM-1, 3-colour-immunofluorescence with additional stainings for cell nuclei (DAPI) and endothelial cells (anti-CD31) were prepared as described previously 14,15 . The area of the CD31/VCAM-1 and ICAM-1 positive signals was measured.
Quantitative reverse transcript polymerase chain reaction. In order to determine mRNA expression levels in aortic lesions of the atherosclerosis study, RNA expression in aortic arches was analyzed. RNA extraction from aortic arches as well as purification and reverse transcription of isolated RNA were carried out as specified before 15,17 . To conduct real-time polymerase chain reaction, fluorochrome-tagged TaqMan primers were used (Thermo Fisher Scientific, Waltham, MA, USA). Detection and amplification of samples were carried out using a CFX96 TaqMan system (Biorad Laboratories, Hercules, CA, USA). Analysis of semiquantitative mRNA expression was performed by ddCt method, results were referred to β-Actin as the housekeeping gene.
To obtain BMDMs, 10 6 progenitor cells per well were differentiated by the addition of 3 µg/mL M-CSF for 7 days. Fully differentiated bone marrow-derived macrophages (BMDMs) were first stimulated with 100 ng/mL LPS (E. Human atherosclerosis study. Human carotid arteries of patients with atherosclerotic carotid plaques were prepared after endarterectomy for histological sections. The carotid sections were stained against P2X 4 (Thermo Fisher Scientific, Waltham, MA, USA). Expression of P2X 4 was determined in non-diseased and atherosclerotic areas of the same carotid sections. Additionally, 3-colour-immunofluorescence stainings for P2X 4 , cell nuclei (DAPI) and endothelial cells (anti-CD31) were performed. Analysis was performed by blinded investigators using Image Pro software (Media Cybernetics, Rockville, MD, USA). Data analysis. Data of > 3 experiments were pooled and results are presented as mean ± SEM. Statistical analyses were performed using unpaired t-test for parametric data or Mann-Whitney-U-test for non-parametric data. Normal distribution of results prior to further analysis was verified by Shapiro-Wilk Test. A probability value ≤ 0.05 was considered as statistically significant.
Ethics approval. This study was approved by the governmental Animal Care committee. All experiments conformed to the guidelines from the EU directive 2010/63 EU of the European Parliament.

P2X 4 is expressed in human atherosclerotic lesions.
Since our data indicate an important role of P2X 4 in experimental atherosclerosis, we further evaluated whether P2X 4 is expressed in human atherosclerotic plaques. Human atherosclerotic lesions from carotid endarterectomy (n = 10) were stained against P2X 4 for immunohistochemical analysis of P2X 4 expression. P2X 4 expression was significantly increased in atherosclerotic-diseased vessel areas (n = 10), compared to non-atherosclerotic parts of the vessel wall (n = 7, p < 0.05, Figure Supplemental 5A and 5B). As shown by the 3-colour immunofluorescence staining, P2X 4 was particularly colocalized with CD31 positive endothelial cells in human atherosclerotic plaques (Supplemental Fig. 5C). These data suggest that P2X 4 is involved in human atherosclerotic disease.

Discussion
This study demonstrated for the first time that P2X 4 -deficiency reduces atherosclerosis and diminishes vascular inflammation. Various studies identified extracellular ATP to mediate inflammation by inducing the release of various pro-inflammatory chemokines through activation of P2X 4 23,29,30 . Since P2X 4 expression on cell surfaces is particularly sensitive being upregulated by stress stimuli, ischemia or inflammation, P2X 4 expression has been described in numerous chronic inflammatory diseases on stromal cell types including endothelial and epithelial cells 18,19 . Indeed, we found an increased expression of P2X 4 in atherosclerotic lesions histologically as well as on the RNA level. In line with the literature, 3-colour immunofluorescence indicated that P2X 4 is particularly colocalized with endothelial cells, which in turn have an important function in atherogenesis. −/− LDLR −/− mice (n = 12) and P2X 4 +/+ LDLR −/− mice (n = 13). Two-step multiplex TaqMan RT-PCR was performed to determine expression of cytokines and adhesion molecules. Expression fold change was calculated by ddCt method, results were referred to β-Actin as the housekeeping gene (C). BMDMs were isolated from the bone marrow of 8-week-old P2X4-competent (n = 5) and P2X4-deficient mice (n = 5). Fully differentiated BMDMs were first stimulated with 100 ng/mL LPS for 4 h, followed by stimulation with either 100 µM or 5 mM ATP for 1 h. Multiplex fluorescence-encoded beads assay was performed to determine concentrations of inflammatory cytokines (D). Results are presented as mean ± SEM. Statistical significance was calculated using Shapiro-Wilk Test followed by an unpaired t-test for parametric or Mann-Whitney-U-test for non-parametric data. *p < 0.05; **p < 0.01; ***p < 0.001. www.nature.com/scientificreports/ Further, P2X 4 −/− LDLR −/− mice developed significantly smaller atherosclerotic lesions, indicating a functional role of P2X 4 in atherogenesis. Beyond decreased lesion size, atherosclerotic lesions from P2X 4 −/− LDLR −/− mice revealed substantially less plaque inflammation. In fact, mechanistically, we found a significantly lower expression of the pro-inflammatory cytokines CCL-2/MCP-1, CXCL-1, CXCL-2, IL-6 and TNFα at mRNA level in atherosclerotic lesions of P2X 4 −/− LDLR −/− mice. It is well established that these cytokines exhibit highly pro-atherogenic properties 31,32 . Supporting our findings, P2X 4 -depression was shown to decrease serum levels of TNF-α and IL-6 in a mouse model of collagen-induced arthritis 24 . Furthermore, it could be shown that ATP-mediated activation of P2X 4 induced the release of C-X-C motif chemokine-5 (CXCL-5) in primary monocyte-derived human macrophages 33 . Besides the expression of the pro-inflammatory cytokines in plaque, we observed reduced IL-12 levels in the plasma of P2X 4 −/− LDLR −/− mice in our in vivo study. Consistent with our results, previous studies demonstrated that functional blockade of endogenous IL-12 resulted in a significant decrease in atherogenesis in LDLR −/− mice 34 . In addition to our in vivo data, we were able to demonstrate a reduced ATP-mediated release of the pro-inflammatory cytokines CCL-2, CCL-5, IL-6 and IL-1ß by BMDMs from P2X 4 -deficient mice in vitro. Although P2X 4 -deficient bone marrow-derived dendritic cells revealed decreased IL-1ß secretion as demonstrated an in vitro model of chronic respiratory inflammation and our in vitro model indicated a reduced release of IL-1ß in BMDMs, in our in vivo study, we could not observe a reduction of IL-1ß expression in atherosclerotic lesions from P2X 4 −/− LDLR −/− mice 35 . Considering these data, the pro-atherogenic effects of P2X 4 are most likely to be explained by the modulation of lesional inflammatory cytokine expression. Although no differential plaque proportions of inflammatory cells were detected in vivo, the cytokine phenotype observed among P2X4-deficient mice suggests that at least parts of the anti-inflammatory effects may be related to reduced macrophage-mediated cytokine and chemokine release. Finally, the P2X 4 -mediated, limited inflammatory activity in the plaque microenvironment may explain an overall decreased atherogenesis with unaltered plaque composition.
NOD-like receptor 3 (NLRP3) inflammasome mediated inflammation is crucially involved in modulating atherosclerotic plaque progression 36 . ATP-P2X 4 signaling was described to stimulate the activation of the NLRP3 inflammasome in podocytes, increasing the expression and release of IL-1β and Interleukin-18 thereby causing tubular necrosis and aggravating renal fibrosis 22,28 . In our study, NLRP3 inflammasome priming was reduced on the RNA level in atherosclerotic lesions of P2X 4 −/− LDLR −/− mice. Overall, in atherosclerosis P2X 4 may be involved in the initiation of inflammasome priming, but does not appear to exert a pronounced influence on inflammasome assembly since we did not find difference in IL-1β RNA expression and caspase 1 activity between P2X 4 -deficient and P2X 4 -competent mice as assessed by FLICA staining. Compared to the P2X 7 receptor, an ATP-responsive ligand gated ion channel playing an important role in atherogenesis via mediating NLRP3 inflammasome activation, P2X 4 has similar characteristics, but a lower Ca 2+ flow due to a smaller ion channel size 33,37 . Interestingly, the literature describes the formation of a heterotrimer from P2X 4 and P2X 7 38 . Corresponding to this, in our in vitro study, no difference in the release of IL-1β was observed after stimulation of BMDMs from P2X 4 -deficient and P2X 4 -competent mice with a dose of 100 µM ATP. After stimulation with a higher dose of 5 mM ATP, we found a significantly lower release of IL-1β by BMDMs from P2X 4 -deficient mice. In this context, a stimulation dose of 100 µM ATP dose is merely able to activate the more sensitive P2X 4 receptor alone, whereas with the stimulation dose of 5 mM ATP the receptor P2X 7 is also activated. In the light of these data, it can be assumed that the P2X 4 receptor alone may not be able to influence the secretion of IL-1β. Due to the significantly higher sensitivity of P2X 4 concerning their common ligand ATP, a possible initiation of the inflammatory response including inflammasome priming through P2X 4 and subsequent pronounced mediation of inflammasome activation with the secretion of IL-1β by P2X 7 is conceivable [39][40][41] .
While Ly6C high monocytes promote inflammation through the secretion of pro-inflammatory cytokines, Ly6C low monocytes are associated with anti-inflammatory properties 42,43 . Both at baseline and after a high cholesterol diet, P2X 4 -deficient mice were found to share a significantly lower proportion of pro-inflammatory Ly6C high monocytes and at the same time a higher amount of anti-inflammatory Ly6C low monocytes. At the same time, in the bone marrow, no differences in common myeloid progenitor cells were detected between P2X 4 -deficient and P2X 4 -competent mice. Overall, these results are most likely to be associated with systemic anti-inflammatory properties of P2X 4 -deficiency with lower expression of pro-inflammatory cytokines in atherosclerotic plaques and limited secretion of pro-inflammatory cytokines by BMDMs from P2X 4 -deficient mice 44 .
Because leukocyte recruitment plays a central role in atherogenesis, the influence of P2X 4 was characterized by intravital microscopy. In this study, P2X 4 -deficient mice were found to show a tendency towards limited leukocyte rolling. This is in line with the reduced lesional expression of the chemoattractants CCL-2, CXCL-1, CXCL-2 as well as the decreased concentrations of the chemotractants CCL-2 and CCL-5 in BMDM supernatant and an overall lower inflammatory activation and thereby potentially lower leukocyte reactivity 45 . In contrast, no effect of P2X 4 -deficiency on leukocyte adhesion to the vessel wall was observed. Mechanistically, we found a reduced expression of the endothelial adhesion factor VCAM-1 in P2X 4 -deficient mice. VCAM-1 promotes the adherence of leukocytes to endothelial cells, whereas stimulation with ATP is known to increase the expression of endothelial VCAM-1 46,47 . Although there were no differences in lesional expression of adhesion molecules within plaques between P2X 4 deficient mice and controls, the main effect of VCAM-1 on vascular adhesion of leukocytes is mostly related to endothelial cells 47 . Therefore, the decreased endothelial expression of VCAM-1 may explain, at least in part, the tendency toward limited leukocyte rolling. Nethertheless, because of the statistically significant but rather mild effect of leukocyte rolling and the unaffected plaque composition, reduced leukocyte rolling may only contribute to a part of the observed anti-atherogenic effects in P2X4 deficiency.
Additionally, P2X 4 -deficient mice, both before and after diet, presented a lower proportion of total T cells, CD4+ cells and a higher proportion of CD8+ cells. Despite the differences in blood count, immunohistochemistry showed no P2X 4 -mediated effects on T-cells in atherosclerotic plaques. Consequently, a P2X 4 -specific blood phenotype is most likely to be assumed in this context. www.nature.com/scientificreports/ Previous studies described the endothelial expression of P2X 4 and P2X 4 -mediated pro-inflammatory responses in human umbilical vein endothelial cells in vitro 23 . In line with this, we were able to demonstrate the involvement of P2X 4 in human atherosclerotic plaques for the first time. Consistent with the literature and our histochemical data from the murine atherosclerosis study, P2X 4 is also found in human atherosclerotic plaques to be predominantly colocalized with endothelial cells 23 . Since P2X 4 is also present in human atherosclerotic plaques, the results of this study could potentially be applicable to human atherosclerotic diseases and could make an important contribution to further research of new treatment options.
In conclusion, P2X 4 -deficiency was shown to improve atherosclerosis, reduced the expression and the release of pro-inflammatory cytokines and decreased inflammasome priming. Consequently, P2X 4 appears to be a promising target structure for the treatment of atherosclerosis and cardiovascular inflammation.