Blockade of endothelial adenosine receptor 2A suppresses atherosclerosis in vivo through inhibiting CREB-ALK5-mediated endothelial to mesenchymal transition

Abstract

Cardiovascular diseases (CVDs) are the leading cause of death worldwide, and morbidity and mortality rates continue to rise. Atherosclerosis constitutes the principal etiology of CVDs. Endothelial injury, inflammation, and dysfunction are the initiating factors of atherosclerosis. Recently, we reported that endothelial adenosine receptor 2A (ADORA2A), a G protein-coupled receptor (GPCR), plays critical roles in neovascularization disease and cerebrovascular disease. However, the precise role of endothelial ADORA2A in atherosclerosis is still not fully understood. Here, we showed that ADORA2A expression was markedly increased in the aortic endothelium of humans with atherosclerosis or Apoe^−/− mice fed a high-cholesterol diet. In vivo studies unraveled that endothelial-specific Adora2a deficiency alleviated endothelial-to-mesenchymal transition (EndMT) and prevented the formation and instability of atherosclerotic plaque in Apoe^−/− mice. Moreover, pharmacologic inhibition of ADORA2A with KW6002 recapitulated the anti-atherogenic phenotypes observed in genetically Adora2a-deficient mice. In cultured human aortic endothelial cells (HAECs), siRNA knockdown of ADORA2A or KW6002 inhibition of ADORA2A decreased EndMT, whereas adenoviral overexpression of ADORA2A induced EndMT. Mechanistically, ADORA2A upregulated ALK5 expression via a cAMP/PKA/CREB axis, leading to TGFβ-Smad2/3 signaling activation, thereby promoting EndMT. In conclusion, these findings, for the first time, demonstrate that blockade of ADORA2A attenuated atherosclerosis via inhibition of EndMT induced by the CREB1-ALK5 axis. This study discloses a new link between endothelial ADORA2A and EndMT and indicates that inhibiting endothelial ADORA2A could be an effective novel strategy for the prevention and treatment of atherosclerotic CVDs.

Introduction

Cardio-cerebral vascular diseases are the most frequent cause of global morbidity and mortality^{1, 2}. Atherosclerosis, a progressive chronic disease of the large and medium arteries characterized by the production of atherosclerotic plaques, is a primary cause of cardiovascular and cerebrovascular disorders. The lesions mostly affect the large and medium elastic arteries, which can result in ischemia of the heart, brain, limbs, or stroke. The scientific hypothesis of atherosclerosis has evolved from simple hyperlipidemia (lipid hypothesis) to a complex pathological process that includes not only lipid abnormalities but also endothelial cell (EC) activation and injury, as well as inflammatory processes that regulate the occurrence and progression of lesions^{3, 4}. The initiation of atherosclerosis is commonly attributed to endothelial damage, inflammation, and dysfunction^5–7. Therefore, gaining insight into the molecular mechanisms that govern endothelial injury and dysfunction is crucial for comprehending and advancing novel therapeutic strategies aimed at preventing or treating atherosclerotic cardiovascular and cerebrovascular diseases.

ECs constitute a contiguous monolayer that envelops the luminal surface of blood vessels. As a vital structure of the vascular system, ECs play a crucial role in regulating vascular permeability and maintaining vascular homeostasis⁸. Hyperglycemia, hyperlipidemia, reactive oxygen species (ROS), blood flow shear stress, cytokines (e.g., TGFβ), and endotoxin biochemical stimulus are the main factors leading to endothelial inflammation and dysfunction^{6, 7}. Endothelial-to-mesenchymal-transition (EndMT) is the type of endothelial inflammation and dysfunction in which endothelial cells lose their markers and functions and acquire mesenchymal cell markers and functions⁹. Accumulating evidence indicates that EndMT plays a substantial role in the pathogenesis of atherosclerosis^{9, 10}.

Adenosine, an endogenous purine nucleoside, functions to modulate a wide variety of physiological and pathological processes via interaction with its G protein-coupled adenosine receptors located on the cell membrane^{11, 12}. There are four known adenosine receptor subtypes: A1, A2A, A2B and A3, which are encoded by the ADORA1, ADORA2A, ADORA2B and ADORA3 genes¹³. Among them, our previous studies have shown that ADORA2A is highly expressed in ECs and significantly contributes to pathological retinal angiogenesis¹⁴ and cerebral ischemia-induced brain injury¹⁵. However, the role and regulatory mechanisms of endothelial ADORA2A in atherosclerosis remain unclear. In this study, we investigated that the effect of ADORA2A blockade on CREB1-ALK5-induced EndMT and the subsequent atherosclerotic plaque formation.

Materials and methods

Human samples

Frozen sections of human peripheral artery with the atherosclerotic lesion (female, 66 years old, #CS627113, OriGene) or normal peripheral artery (male, 73 years old, #CS646954) were purchased from OriGene Technologies, Inc (Rockville, MD, US).

Animals

Adora2a^Flox/Flox (Adora2a^F/F) mice were kindly provided by Dr Joel Linden (La Jolla Institute for Allergy and Immunology, CA). Cdh5^Cre mice (The Jackson Laboratory, Stk#006137) and Apoe^−/− (The Jackson Laboratory, Stk#002052) mice were purchased from The Jackson Laboratory (Bar Harbor, ME, USA). Adora2a^F/FCdh5^Cre mice were obtained by crossing Adora2a^F/F mice with Cdh5^Cre mice. Apoe^−/−Adora2a^F/FCdh5^Cre mice were obtained by crossing Adora2a^F/FCdh5^Cre mice and Apoe^−/− mice. PCR genotyping analysis was performed using the following primers: Adora2a^F/F (5’-GGGCAAGATGGGAGTCATT-3’ and 5’-ATTCTGCATCTCCCGAAACC-3’), Cdh5^Cre (5’-GCGGTCTGGCAGTAAAAACTATC-3’ and 5’-GTGAAACAGCATTGCTGTCACTT-3’) and Apoe^−/− (5’-GCCTAGCCGAGGGAGAGCCG-3’, 5’-TGTGACTTGGGAGCTCTGCAGC-3’ and 5’-GCCGCCCCGACTGCATCT-3’).

Mouse atherosclerotic lesion analysis

Apoe^−/−Adora2a^F/FCdh5^Cre male mice (n = 13) and Apoe^−/−Adora2a^F/F male mice (n = 14) at 7 weeks were fed a Western diet (TD88137, ENVIGO, IN, USA) for 16 weeks. Mice were anesthesia followed by exsanguination and perfusion with phosphate-buffered saline (PBS) (BP665–1, Fisher Scientific, PA, USA) and 4% PFA (sc281692, Santa Cruz, TX, USA) through the left ventricle. Aortas were collected and stained with 2% Oil Red O (#O0625, Sigma-Aldrich, MO, USA) for 30 minutes at room temperature and then opened longitudinally for photography. The upper one-third of the heart was dissected and embedded in optimum cutting temperature compound (OCT, BDH Laboratory Supplier), then cryosectioned into serial sections (7 μm) and stained with 2% Oil Red O to evaluate the lipid deposition in the aortic sinus. Images were quantified with Image-Pro Plus software (Medical Cybernetics, MD, USA).

KW6002 treatment

To test the effect of Adora2a antagonism in atherosclerosis, istradefylline (KW6002, SML0422, Sigma-Aldrich), the potent Adora2a antagonist, was used in Apoe^−/− male mice. Mice were randomly divided into two groups and injected intraperitoneally with KW6002 or vehicle (n = 13) at 3.3 mg/kg/day for 8 weeks. KW6002 was dissolved in 5% DMSO, 5% castor oil and 90% saline.

Plasma triglyceride, cholesterol, and glucose measurement

Blood was collected from Apoe^−/− male mice, Apoe^−/−Adora2a^F/FCdh5^Cre male mice, and Apoe^−/−Adora2a^F/F male mice after 16 weeks of Western diet. The concentrations of triglyceride, cholesterol, and glucose in plasma were determined by enzyme-linked immunosorbent assay (TR15421, TR13421, TR22421, Thermo Scientific, MA, USA) according to the manufacturer’s protocol.

Hematoxylin and eosin (HE) staining

HE staining was performed on frozen sections of the aortic sinus with hematoxylin (#22050111, Thermo Scientific, MA, USA) and esion (#22050110, Thermo Scientific, MA, USA) according to the manufacturer’s protocol. Briefly, slides were washed with PBS and stained with Hematoxylin for 3 minutes. Then, sections were washed with water and dehydrated in 70% and 90% alcohol solution. Sections were stained with Eosin for 1 minute and dehydrated with alcohol solution and xylene. Then, coverslips were mounted onto the sections with neutral resin. Images were captured with an Olympus BX51 inverted microscope and analyzed with Image-Pro Plus software.

Immunohistochemistry (IHC) staining

Frozen sections from the aortic sinus were washed with PBS, and then incubated with H₂O₂ for 30 minutes at room temperature to destroy the endogenous peroxidase activity. Sections were heated at 98°C for 10 minutes in citric acid buffer (10 mM, pH 6.0) for antigen retrieval and blocked with avidin blocking solution with 10% rat serum for 1 hour at room temperature. Then, sections were incubated with Mac-2 (3 μg/ml, #ACL8942F, Accurate Chemical & Scientific Corporation, Westbury, NY) at 4 °C overnight. After washing in FSGP/PBS, sections were incubated with biotinylated secondary antibodies and ABC solution (PK-6100, Vector Labs, CA, USA) according to the manufacturer’s instructions, followed with DAB solution for 1 minute. Images were captured with an Olympus BX51 inverted microscope and analyzed with Image-Pro Plus software.

Masson’s trichrome staining

Frozen sections from the aortic sinus were washed with PBS and fixed with 4% PFA for 20 minutes and Bouin’s solution overnight. After that, Masson trichrome staining was performed on sections with Masson’s trichrome stain kit (#25088, Polysciences) according to the manufacturer’s instructions.

En face staining of aortic endothelium

Mice were euthanized with terminal general anesthesia followed by exsanguination and perfusion with PBS and 4% PFA through the left ventricle. The harvested aortas were fixed with 4% PFA for 30 minutes and permeabilized with 0.5% TritonX-100 for 20 minutes. The aortas were blocked with 10% goat serum (#50062Z, Thermo Fisher, Waltham, MA) for 1 hour at room temperature and incubated with primary antibodies against ADORA2A (1 μg/ml, Santa Cruz, sc-32261), CD31 (1 μg/ml, Dianova, DIA310), ACTA2 (1 μg/ml, Santa Cruz, sc-56499) or ALK5 (1 μg/ml, Santa Cruz, sc-518086) overnight at 4°C. The aortas were incubated with Alexa Fluor 488-conjugated or Alexa Fluor 594-conjugated secondary antibody for 1 hour at room temperature. Nuclei were stained with DAPI (1 μg/ml, Thermo Scientific, MA, USA) for 10 minutes. Images are obtained by using an inverted confocal microscope (Zeiss 780, Carl Zeiss) and were analyzed using Image J software.

Immunofluorescent staining

Cells were cultured in 8-chambered cell culture slides (#08–774-26, Fisher Scientific) with different treatments. Cell slides or slides of frozen sections were washed with PBS, fixed with 4% PFA for 20 minutes at room temperature, and permeabilized with 0.5% TritonX-100 in PBS for 20 minutes at room temperature. Then slides were blocked with 10% normal goat serum (#50062Z, Thermo Fisher, Waltham, MA) for 1 hour at room temperature and incubated with primary antibodies against ADORA2A (1 μg/ml, Millipore, 05–717), CD31 (1 μg/ml, Dianova, DIA310) for mouse samples, CD31 (1 μg/ml, Santa Cruz Biotechnology, sc-376764) for human samples and ACTA2 (1 μg/ml, Santa Cruz, sc-32251) overnight at 4°C. The slides were incubated with Alexa Fluor 488-conjugated or Alexa Fluor 594-conjugated secondary antibody for 1 hour at room temperature. Nuclei were stained with DAPI (1 μg/ml, Thermo Scientific, MA, USA) for 10 minutes. Images were obtained using an inverted confocal microscope (Zeiss 780, Carl Zeiss) and analyzed using Image J software.

Human aortic endothelial cells (HAECs) culture and treatment

Primary human aortic endothelial cells (HAECs) were purchased from ATCC (PCS-100–011) and cultured with endothelial growth medium (EGM-2, CC-3162, Lonza) in a humidified incubator with 5% CO₂ at 37°C. HAECs were used in experiments from passage 4 to 6. HAECs were induced EndMT by hTGFβ1 (10 ng/mL, 100–21C, Pepro Tech) and IL-1β (10 ng/mL, 200–01B, Pepro Tech) treatment in 25% EGM-2. And cell medium with indicated treatment was refreshed every other day. Cells were pretreated with KW6002 (100 nM) for 1h or transfected with siRNA for 1 day before EndMT induction.

siRNA transfection and constructed reporter plasmid transfection of HAECs

Cells were transfected with siRNAs targeting human ADORA2A (sc-39850, Santa Cruz Biotechnology) or negative control siRNA (sc-37007, Santa Cruz Biotechnology) using Lipofectamine RNAiMax reagent (#13778–150, Invitrogen, CA, USA) according to the manufacturer’s protocol. Cells were harvested at the indicated time points for assays after different treatments.

Cells were seed in 96-well plates and transfected with 100 ng constructed reporter plasmids using Lipofectamine 2000 (#11668019, Thermo Fisher Scientific, USA) according to the manufacturer’s instructions. To construct the ALK5 promoter reporter plasmid, a 2143 bp fragment of the ALK5 gene promoter (−1960/+183 relative to the transcription start site) was amplified using PCR and then inserted into the pGL3-basic plasmid (Sangon Biotech, Shanghai, China) at specific sites (KpnI and XhoI), resulting in the generation of pGL3–1960. Modification were made to the pGL3–1960 plasmid to generate the pGL3–1113, pGL3–136 and pGL3-mut. For pGL3-mut plasmid, the predicted CREB-1 binding site (TGACG) located in the ALK5 promoter was converted to ‘CTATA’ by PCR. The primers used were shown in Table S4. Cells with different treatments were used for dual-luciferase reporter assay.

Adenoviral transduction of HAECs

Cells at 50% to 70% confluence were transduced with human ADORA2A adenovirus (Vector Biolabs. Malvern, PA, USA) or negative control adenovirus (Vector Biolabs. Malvern, PA, USA) in basal medium for 4 hours and then followed with a culture with regular growth cell medium. Cells were harvested at the indicated time points for assays after different treatments.

Quantitative real-time PCR analysis

Total RNA was extracted from mouse aortas or HAECs with Trizol reagent (#155960269, Invitrogen, NY, USA) following the manufacturer’s instructions. 1 μg sample of RNA was used as a template for reverse transcription using the iScriptTM cDNA synthesis kit (#270–8891, Bio‐Rad, CA, USA). Real‐time PCR was performed via a Bio‐Rad CFX96 instrument (Bio‐Rad). The primers used for PCR were shown in Table S3. The following PCR protocol was used: 10 min at 95°C, followed by 50 cycles of denaturing at 95°C for 15 s, annealing at 60°C for 30 s, and extension at 72°C for 60 s. Quantification of relative gene expression was calculated with the efficiency-corrected 2^-△△CT method, and 18S RNA was used as an internal standard. All real‐time PCRs for each sample were performed in triplicate.

Western blotting

HAECs were harvested and lysed with RIPA supplemented with 1% protease inhibitor (#05892970001, Sigma-Aldrich) and 1% phosphatase inhibitors (#4906845001, Roche). The total amount of protein was quantified using the BCA protein assay kit (#23225, Thermo Fisher Scientific) according to the manufacturer’s guidelines. Equal amounts of protein per lane (10–20 μg) were subjected to 8–12% SDS‐PAGE. The first antibodies used in this study were as follows: ADORA2A (0.5 μg/ml, Millipore, #05–717), VCAM1 (0.2 μg/ml, Santa Cruz Biotechnology, sc-13160), ICAM1 (0.5 μg/ml, Boster Biological Technology, BA0541), ACTA2 (0.2 μg/ml, Santa Cruz Biotechnology, sc-56499), TAGLN (0.5 μg/ml, Abcam, ab14106), CDH5 (0.2 μg/ml, Santa Cruz Biotechnology, sc-9989), CD31 (0.2 μg/ml, Santa Cruz Biotechnology, sc-376764), Fibronectin (0.5 μg/ml, Abcam, ab2413), Collagen I (0.5 μg/ml, Novus, NB600–408), Smad2/3 (0.5 μg/ml, Cell Signaling Technology, #8685), p-Smad2/3 (0.5 μg/ml, Cell Signaling Technology, #8828), TGFBR1/ALK5 (0.2 μg/ml, Santa Cruz Biotechnology, sc-518086), ALK1 (0.5 μg/ml, Abcam, ab108207), CREB (0.5 μg/ml, Cell Signaling Technology, #9197), p-CREB-Ser133 (0.5 μg/ml, Cell Signaling Technology, #9198), GAPDH antibody (0.5 μg/ml, Cell Signaling Technology, #2118), and β-actin (0.2 μg/ml, Santa Cruz Biotechnology, sc-47778). The membranes were then incubated with anti-mouse IgG-HRP (1:2000, Cell Signaling Technology, #7076S) or anti-rabbit IgG-HRP (1:2000, Cell Signaling Technology, #7074S) for 1 hour at room temperature and followed washing three times with TBST. For detection, chemiluminescence assay (Millipore) and the ChemiDoc MP system (Bio-Rad) were used to acquire images. Quantification was done with ImageJ free software (Version 1.47), and each lane was normalized to GAPDH or β-actin.

THP-1 cells adhesion assay

THP‐1 cells (ATCC, TIB-202) were suspended in RPMI 1640 (1 × 10⁶ cells/ml). HAECs were cultured in 12-well plates for 48–72 hours to form a monolayer and washed twice with a serum-free medium. THP-1 cells were stained with Calcein-AM (C1430, Thermo Fisher Scientific, USA). 1 ml of the stained-cell suspension was added. After incubating in the cell incubator for 30 min, non-adherent THP-1 cells were removed by gently washing with serum-free medium three times. Cells were washed with PBS and fixed with 4% PFA. The number of adherent THP-1 cells was counted in four random fields per well for quantification.

Dual-luciferase reporter assay

HAECs were seeded in 96-well plates and transfected with Ad-ADORA2A or Ad-CTRL for 48h, then transfected with 100 ng of indicated constructed reporter plasmids (pGL3–1960, pGL3–1113, pGL3–136, pGL3-mut and pGL3-basic) using Lipofectamine 2000 for 48 hours. The luciferase activity was measured using the Dual-luciferase reporter assay kit (E1910, Promega) according to the manufacturer’s instructions.

Data and statistical analysis

Animal grouping was performed in a randomized manner. Data were presented as the mean ± SEM in cellular studies and mean ± SD in animal experiments. Data distribution was assessed by the Shapiro-Wilk test for normality, and T test or Brown-Forsythe test was performed to test the equality of variance. For comparisons of two groups, when the variances were equal, an unpaired two-tailed Student’s t-test was performed. When variances were unequal, an unpaired two-tailed Student’s t-test with Welch’s correction was performed. Otherwise, non-parametric Mann–Whitney tests were performed when the data distribution was not satisfied. For multiple groups, the Brown–Forsythe test was used to test the homogeneity of variance. Differences among multiple groups were assessed by one-way ANOVA followed by a Bonferroni post hoc analysis for the data sets was satisfied by the Brown–Forsythe test, otherwise by one-way ANOVA and Welch’s ANOVA test with Dunnett’s T3 multiple comparisons. Statistical analyses were performed using GraphPad Prism 8.0 software (La Jolla, CA, USA). *P < 0.05 was considered to be statistically significant.

Results

ADORA2A is upregulated in the aortic endothelium of humans and mice with atherosclerosis

To investigate the role of ADORA2A in the formation and development of atherosclerotic lesions, we first analyzed ADORA2A expression in the arteries of humans with atherosclerosis and Apoe^−/− mice challenged with a high-cholesterol diet. The HE staining of human peripheral arteries with or without atherosclerosis was performed (Fig. S1). Immunofluorescence stainings with specific monoclonal antibodies against ADORA2A or the vascular EC marker CD31 showed that Adora2a expression was increased in the aortic ECs of human atherosclerotic artery (Fig. 1a) and Apoe^−/− mice (Fig. 1b), compared to controls. Consistently, both mRNA and protein levels of ADORA2A were significantly increased in the aorta of Apoe^−/− mice compared to control aortas (Fig. 1c-d). Similarly, en face staining of the aortic endothelium showed that ADORA2A was markedly increased in the aortic endothelium of Apoe^−/− mice compared with that of control mice (Fig. 1e), suggesting that ADORA2A upregulation is associated with atherosclerosis and endothelial ADORA2A may contribute to the development of atherosclerosis.

Figure 1.

A2AR was upregulated in ECs under atherosclerosis circumstance.

Endothelial-specific Adora2a deficiency inhibits atherogenesis and EndMT in Apoe^−/− mice

To study the endothelial-specific effects of ADORA2A on the formation of atherosclerotic lesions, we generated endothelial-specific Adora2a deficient mice (Adora2a^ΔVEC) mice by breeding Adora2a^F/F with Cdh5^cre mice (Fig. S2). Adora2a^ΔVEC mice were further crossed to an Apoe^−/− background to develop atherosclerotic susceptibility, with Apoe^−/−Adora2a^F/F mice serving as the control group. 6-week-old Apoe^−/−Adora2a^ΔVEC mice and control Apoe^−/−Adora2a^F/F mice were challenged with a high-cholesterol diet for 12 weeks. There was no difference in body weight or blood glucose levels between groups (Table S1). Oil Red O staining revealed a reduction in the lesion area of both the atherosclerotic surface (Fig. 2a) and the aortic sinus (Fig. 2b) in Apoe^−/−Adora2a^ΔVEC mice, compared to those of control mice. To evaluate the effect of EC-specific Adora2a deficiency on the morphological composition of atherosclerotic plaques, immunohistochemical, Masson Trichrome, and H&E staining were performed on the cross-sections of the aortic sinus. Remarkably, the necrotic core areas and the number of macrophages in Apoe^−/−Adora2a^ΔVEC mice were reduced, while the collagen content was increased compared with the control mice (Fig. 2b), indicating that loss of Adora2a in ECs stabilized atherosclerotic plaques. Adora2a depletion in the endothelium also decreased the levels of serum triglyceride and cholesterol (Table S1). The process known as endothelial-to-mesenchymal transition (EndMT) is widely recognized as a significant contributor to the development of atherosclerosis. To determine the role of ADORA2A in regulating EndMT, we performed en-face staining of the aortic endothelium with specific antibodies against EndMT marker ACTA2 and EC surface marker CD31. The results showed increased ACTA2 and decreased CD31 expression in the aortic endothelium of Apoe^−/−Adora2a^F/F, compared to those of Adora2a^F/F control mice, while endothelial-specific Adora2a deletion reversed EndMT in Apoe^−/− mice (Fig. 2c). These data indicate that suppressed atherosclerosis in Apoe^−/− mice with loss of EC Adora2a might be related to inhibition of EndMT.

Figure 2.

Loss of endothelial Adora2a suppressed the formation of atherosclerotic lesions.

Blockade of ADORA2A with KW6002 alleviates atherogenesis and EndMT in Apoe^−/− mice

To explore the therapeutic potential of targeting ADORA2A for the treatment of atherosclerosis, we treated Apoe^−/− mice with KW6002 (3.3 mg/kg, i.p.) or vehicle daily for 8 weeks (Fig. 3a). The ADORA2A-specific antagonist of KW6002 has been extensively studied and received approval from the US FDA in 2019 for the treatment of early Parkinson’s disease. Oil Red O staining revealed significantly decreased lesion sizes in the whole aorta (Fig. 3b) and aortic sinus (Fig. 3c) of KW6002-treated Apoe^−/− mice compared to control mice. Furthermore, the areas of Mac-2 staining of macrophages and hematoxylin and eosin (H&E) staining of the necrotic core were much smaller in sections of lesions from KW6002-treated Apoe^−/− mice than in Apoe^−/− mice treated with vehicle (Fig. 3c). In contrast, Masson Trichrome staining of collagen in the aortic sinus was much stronger in lesions from KW6002-treated Apoe^−/− mice than control mice (Fig. 3c). En face staining of the aortic endothelium showed decreased ACTA2 and increased CD31 expression in the aortic endothelium of KW6002-treated Apoe^−/− mice, compared to those of vehicle-treated mice (Fig. 3d). Collectively, these observations demonstrate that the ADORA2A antagonist KW6002 effectively reduces the formation and instability of atherosclerotic plaque by inhibiting EndMT.

Figure 3.

Adenosine 2A receptor inhibition suppressed the formation of atherosclerotic lesions.

ADORA2A knockdown or inhibition reduces EndMT

To investigate the regulatory role of ADORA2A in EndMT induction in vitro, we cultured HAECs and incubated them with human TGFβ1 (hTGFβ1) and IL-1β, a well-recognized potent inducer of EndMT¹⁶. Treatment with hTGFβ1 and IL-1β markedly increased both the mRNA (Fig. 4a) and protein (Fig. 4b) levels of ADORA2A as well as EndMT markers (e.g., ACTA2, TAGLN and COL1), and reduced endothelial cell markers (e.g., CDH5 and CD31) (Fig. 4c-d), indicating that ADORA2A may be associated with EndMT induction. Using a loss-of-function approach, ADORA2A expression on HAECs was knocked down with siRNA targeting ADORA2A (siADORA2A) (Fig. 4c). Remarkably, both RT-PCR and Western blot analysis showed that ADORA2A knockdown downregulated the expression of ICAM-1, ACTA2, TAGLN and COL1A1, and upregulated the expression of CDH5 and CD31 in HAECs treated with hTGFβ1 and IL-1β (Fig. 4c-d). Consistently, immunofluorescent staining showed that ADORA2A knockdown decreased the protein level of TAGLN and increased the expression of CD31 in hTGFβ1 and IL-1β-treated HAECs (Fig. S3). Because monocyte adhesion to the vascular endothelium is an essential inflammatory event in EndMT-mediated atherosclerosis, we conducted monocyte adhesion experiments with HAECs and THP-1 co-culture. The results showed that the number of firmly adhering human monocytes to HAECs treated with hTGFβ1 and IL-1β increased substantially, compared with those to vehicle-treated HAECs, while knockdown of ADORA2A with siRNA markedly decreased the effect of hTGFβ1 and IL-1β treatment on monocyte adhesion (Fig. S4). Similar results were obtained when we treated hTGFβ1 and IL-1β-stimulated HAECs with the ADORA2A pharmacologic antagonist KW6002. Both the mRNA and protein levels of ACTA2, TAGLN, COL1, and FN were inhibited with KW6002 treatment compared to vehicle treatment, while endothelial cell markers CD31 and CDH5 were increased (Fig. 5a-c). The number of monocytes adhered to HAECs treated with hTGFβ1 and IL-1β was significantly decreased with KW6002 treatment compared to vehicle treatment (Fig. S5). Overall, these findings suggest that both gene knockdown of ADORA2A with siRNA and pharmacologic inhibition of ADORA2A with KW6002 are capable of reducing EndMT in HAECs.

Figure 4.

Adenosine 2A receptor deficiency significantly reduced EndMT.

Figure 5.

Adenosine 2A receptor inhibition significantly reduced EndMT.

ADORA2A overexpression induces EndMT

Having shown that ADORA2A knockdown or inhibition reduces EndMT, we next asked whether ADORA2A overexpression increases EndMT in HAECs. To this end, we overexpressed ADORA2A in HAECs using an adenoviral vector of ADORA2A (Ad-ADORA2A). As expected, Ad-ADORA2A transfection increased both the mRNA and protein levels of ADORA2A, ICAM-1, ACTA2, TAGLN, and COL1A1, FN and decreased CDH5 or CD31 expression (Fig. 6a-b), compared with transfection of control adenovirus (Ad-CTRL) in HAECs. In line with these results, immunofluorescent staining also showed that ADORA2A overexpression increased the protein level of TAGLN and reduced the expression of CD31 in HAECs (Fig. 6c). Taken together, both the loss-of-function and gain-of-function studies demonstrate the role of ADORA2A in regulating EndMT.

Figure 6.

Adenosine 2A receptor overexpression induced EndMT.

ADORA2A promotes EndMT via activation of the cAMP-CREB1-ALK5 axis

TGFβ signaling pathway via two type I TGFβ receptors, including activin receptor-like kinase 1 (ALK1, also known as ACVRL1) and TGFβ receptor I (TβRI, activin receptor-like kinase 5 (ALK5)) in vascular ECs, plays a crucial role in the regulation of EndMT. Therefore, we further examined the regulatory effect of ADORA2A on the expression of ALK1 and ALK5 in HAECs. As shown in Fig. 7a-6b, ADORA2A knockdown downregulated hTGFβ1 and IL-1β-induced protein expression of ALK5, but not ALK1, in HAECs. Similar results were obtained using the ADORA2A pharmacologic antagonist KW6002 (Fig. 7c-d). In line with the in vitro data, en face immunofluorescence staining showed increased TβRI expression in the aortic endothelium of Apoe^−/− mice, compared to that of control mice, while endothelial-specific Adora2a deficiency markedly decreased TβRI expression in that of Apoe^−/− mice (Fig. 7e). Furthermore, ADORA2A knockdown or inhibition suppressed hTGFβ1 and IL1β-induced phosphorylation of Smad2/3, the major effector of TGFβ signaling (Fig. 8a-b). These results suggest ADORA2A is required for the activation of the TGFβ1-ALK5-Smad2/3 axis.

Figure 7.

Adenosine 2A receptor deficiency significiantly supressed ALK5.

Figure 8.

Adenosine 2A receptor deficiency significiantly supressed TGFB1 signaling via CREB.

Because ADORA2A is a G protein-coupled receptor that controls a wide range of physiological and pathological processes via the cAMP/PKA/CREB signaling pathway, we posited that the cAMP/PKA/CREB pathway may be involved in ADORA2A-mediated ALK5 expression in HAECs. Indeed, ADORA2A knockdown or inhibition suppressed hTGFβ1 and IL1β-induced phosphorylation of CREB (Fig. 8c-d). We further analyzed the CREB binding sites in the ALK5 gene promoter. CREB regulates target gene expression by binding to the cAMP-response element (CRE) (TGACGTCA) of the target gene. By searching the NCBI database, we found that the ALK5 gene promoter region −407bp contains a putative CREB binding site 5’-TGACG-3’ (Fig. 8e). This putative CREB binding site in ALK5 promoter region was further validated in Genecards program (enhancement confidence score: 2.3). Since the information from two programs indicates that ALK5 functions as a target gene of CREB, we generated reporter plasmids containing TGACG or its mutant CTATA and transduced them to ADORA2A-overexpressing HAECs. Reporter gene assays showed that the luciferase activities of pGL3-plasmids without TGACG or with mutant CTATA were not significantly increased by ADORA2A overexpression compared with pGL3-plasmids containing TGACG, indicating that mutation of the CREB binding site completely abrogated the transcriptional activation of ALK5 reporter by CREB (Fig. 8f). Taken together, these findings imply that activation of the cAMP-CREB1-ALK5 axis is involved in ADORA2A-mediated EndMT.

Discussion

This study demonstrated that the formation of atherosclerotic plaques and EndMT were inhibited in endothelial Adora2a knockout mice or mice treated with antagonist KW6002. In the in vitro studies, EndMT induction of HAECs was inhibited by Adora2a knockdown and antagonist KW6002 but was promoted by adenovirus encoding ADORA2A. Mechanistically, ADORA2A promoted the expression of TβRI/ALK5 and the activity of its downstream P-Smad2/3 signaling pathway through the CREB signaling pathway.

ADORA2A is highly expression in vascular cells and immune cells and plays an important role in inflammation and immune reactions^{12, 13}. In a genome-wide association study (GWAS) analysis, the ADORA2A gene has been identified as one of the important genes affecting the risk of human coronary artery disease (CAD)¹⁷. Our previous study showed that inhibition of ADORA2A, especially in myeloid cells, decreases the formation of atherosclerotic lesions¹⁸. However, the role of endothelial ADORA2A in atherosclerosis and the underlying mechanism is unclear. Here, with endothelial-specific Adora2a deficient atherosclerotic mice, we demonstrated that endothelial ADORA2A participates in the development of atherosclerotic lesions by induction of EndMT.

EndMT is a manifestation of endothelial polarity and dysfunction under inflammatory conditions⁹. Endothelial cells adopt mesenchymal markers such as fibronectin, ACTA2, and TAGLN and lose the expression of VE-cadherin and CD31⁹. EndMT has been involved in human and experimental atherosclerosis¹⁶. EndMT-derived fibroblasts infiltrate atherosclerotic lesions and secrete matrix metalloproteinase^{6, 19}, leading to the progression and destabilization of atherosclerotic plaques. TGFβ1 was considered the main EndMT inducer^{16, 19, 20}. To induce EndMT, TGFβ first binds to the constitutively active TGFβ type II receptor (TβRII), then promotes the recruitment of TGFβ type I receptors which include ALK1 and ALK5 (TβRI)^21–23. TGFβ-ALK5 signaling increases the phosphorylation of Smad2/3 and induces the expression of mesenchymal genes. For example, Snail is a transcriptional factor of mesenchymal cells, including EndMT. Its transcription is highly regulated by Smad2/3²⁴. Knockout of Smad3 decreases EndMT and reduces EndMT-induced cardiac fibrosis²⁵. Therefore, ALK5 and its related signaling have been regarded to be critical for EndMT and the development of atherosclerosis²⁶. In this study, we found ALK5 was markedly downregulated in the absence or blocking of ADORA2A. Thus, ALK5 is a critical mediator in ADORA2A-induced EndMT and the subsequent formation of atherosclerotic lesions.

CREB is a nuclear transcription factor and is known to be activated by various extracellular stimuli²⁷. The CREB participates in the TGFβ pathway²⁸ and plays an important role in the development of cardiovascular disease²⁹. It has been reported that CREB phosphorylation was induced by vascular endothelial growth factor via p38MAPK pathway and PKA in endothelial cells³⁰ and was involved in the expression of genes important to angiogenesis. Also, CREB phosphorylation is critical for vascular smooth muscle cell proliferation and vascular fibrosis³¹. CREB can be activated by various types of GPCR, including receptors of adenosine. CREB pathway is the downstream pathway of adenosine in astrocytes³². In immune cells, adenosine increases CREB phosphorylation through ADORA2A and ADORA2B³³. In line with these studies, we found that CREB1 was a transcription factor activated by ADORA2A and could bind to the CRE site of the ALK5 promoter region to promote gene transcription, suggesting that ALK5 regulation of ADORA2A was closely related to the cAMP/PKA/CREB1 signaling pathway.

Currently, drugs for the treatment of atherosclerosis target hyperlipidemia and platelet activation. These medications are either lipid-lowering drugs to prevent plaque development or anti-platelet and vasodilator drugs to reduce complications³⁴. In this study, we have shown that blocking ADORA2A with KW6002 results in the inhibition of EndMT, which is a critical vascular pathological change important for the development of atherosclerosis. KW6002 has been approved for the treatment of Parkinson’s disease³⁵. Oral use of KW6002 showed a very consistent safety profile in clinical trials with advanced Parkinson’s Disease patients^{36, 37}. Importantly, potential side effects related to different biological roles of ADORA2A, such as the control of the immune inflammatory system³⁸, sleep³⁹, and vascular tone⁴⁰, are not evident in clinical trials. Therefore, antagonists of ADORA2A, including KW6002, hold promise in the treatment of atherosclerosis. Further investigation on the effect of ADORA2A antagonists on the development of atherosclerotic lesions in large animal models or in humans is warranted.

Nonstandard abbreviations and acronyms

ACTA2

Alpha-smooth muscle actin

ADORA2A

adenosine receptor 2A

ALK1/ACVRL1

activin receptor-like kinase 1

BMDMs

bone marrow-derived cells

CAD

coronary artery disease

CVDs

cardiovascular diseases

ECs

endothelial cells

EndMT

endothelial to mesenchymal transition

GPCR

G protein-coupled receptor

GWAS

genome-wide association study

HAECs

human aortic endothelial cells

ICAM-1

intercellular cell adhesion molecule-1

ROS

reactive oxygen species

TβRI/ALK5

TGFβ type I receptor/activin receptor-like kinase 5

TβRII

TGFβ type II receptors

TGFβ

transforming growth factor beta

Data availability

Data will be made available on request.