Adenosine A_2A receptor deficiency in leukocytes increases arterial neointima formation in apolipoprotein E-deficient mice

Abstract

Background

The A_2A receptor (A_2AR) plays a complex role in inflammation and tissue injury. In this study, we used the mice deficient in both A_2AR and apolipoprotein E (A_2AR^−/−/apoE^−/−) to investigate the role of A_2AR in mediating the interactions of leukocytes with injured arterial walls and the formation of arterial neointima induced by a guide wire.

Methods and Results

In apoE^−/− mice, A_2AR deficiency increased the size of arterial neointima in injured carotid arteries by 83%. Arterial neointima formation was also enhanced in bone marrow transplanted chimeric mice that lacked A_2AR in their bone marrow-derived cells. Epifluorescence intravital microscopy showed that neutrophil rolling and adherence to the injured arterial area was enhanced by 80% and 110% in A_2AR^−/−/apoE^−/− mice, respectively. This phenomenon occurred even though the protein levels of homing molecules on A_2AR-deficient neutrophils were unchanged from those of wild type neutrophils. A_2AR-deficient neutrophils exhibited an increase in the phosphorylation of p38 mitogen-activated protein kinase (MAPK), PSGL-1 clustering, and the affinity of b₂ integrins. Inhibition of p38 phosphorylation abrogated the increased PSGL-1 clustering and b₂ integrin affinity, thereby reversing the increased homing ability of A_2AR-deficient leukocytes.

Conclusion

The deficiency of A_2AR enhances the homing ability of leukocytes and increases the formation of arterial neointima after injury. A_2AR antagonists are being tested for the treatment of neurodegenerative diseases and other chronic diseases. Our results suggest that an evaluation of the effect of A_2AR antagonists on arterial restenosis following arterial angioplasty should be conducted.

Introduction

Adenosine receptor A₂ (A_2AR) is one of the four G-protein coupled receptors for adenosine. It is present on many inflammatory cells, including neutrophils, monocytes, platelets, and all vascular cells ^{1, 2}. A_2AR plays different roles in inflammation and tissue injury under different conditions. In many acute inflammatory or injury models of peripheral organs, A_2AR acts as an anti-inflammatory molecule. For example, loss of A_2AR increases inflammatory responses and causes tissue damage in the liver, lung, and spleen in vivo ^{3, 4}, while the activation of A_2AR with agonists reduces inflammation and protects tissues from injury ². In contrast to the anti-inflammatory effect of A_2AR in the acute injury or inflammatory models, the absence or blocking of A_2AR appears to offer mice protection from chronic diseases, such as atherosclerosis and liver cirrhosis ^{5, 6}, as well as neurodegenerative diseases ⁷. Accordingly, many A_2AR antagonists are being developed to treat neurological disorders including Parkinson’s disease, and some of them are being evaluated in clinical trials ⁸.

Arterial restenosis is a serious complication of angioplasty, including percutaneous transluminal coronary intervention ⁹. In human, VSMCs predominate neointimal hyperplasia. However, it has been demonstrated that in human neointimal hyperplasia, the number of leukocytes in neointima correlates with the severity of restenosis ¹⁰, indicating the causal role of infiltrated leukocytes in the formation of arterial restenosis. To study the effect of infiltrated leukocytes on arterial neointima in patients with arterial restenosis, a model of wire-induced neointima formation in the mouse carotid artery has been described and widely used in the field of research ¹¹. The inflammatory response, including the platelet and leukocyte accumulation on the injured arterial areas, as well as the smooth muscle cell migration, is requisite for arterial neointima formation ^12–14. Immediately after arterial injury, platelets interact with the injured area via many factors including glycoprotein Ib and glycoprotein IIb/IIIa ^{15, 16}. Upon adherence, platelets become activated and orchestrate the leukocyte recruitment and endothelial regeneration on the injured site ^17–19. Studies from our and other groups have demonstrated that the formation of arterial neointima is significantly suppressed following the inhibition of platelet accumulation, leukocyte adhesion, and the improvement of endothelial regeneration on the injured area ^20–22.

Many senior patients with neurodegenerative diseases also suffer from atherosclerotic coronary diseases. Therefore, the patients, who might take A_2AR antagonists for the treatment of their neurological disease, could possibly need percutaneous transluminal coronary intervention for their coronary artery disease. Given this clinical scenario, it is relevant to study whether the blocking or inactivation of A_2AR affects the arterial repair following injury. To our best knowledge, no reports have been published on the effects of the blocking or inactivation of A_2AR on the formation of arterial neointima. In this study, we evaluated whether A_2AR deficiency affects the injury-induced arterial neointima by using the mice deficient in both A_2AR and apolipoprotein E (A_2AR^−/−/apoE^−/−).

Materials and Methods

A_2AR^−/− mice in a C57BL/6J background ²³ were bred with apoE^−/− (C57BL/6J background) mice to generate A_2AR^−/−/apoE^−/− mice and their littermate controls. Chimeric mice, with or without A_2AR in their bone marrow-derived cells, were produced by bone marrow transplantation ²⁴. The 8-week-old mice were fed a Western diet containing 21% fat, 0.15% cholesterol, and 19.5% casein without sodium cholate for 2 weeks prior to the wire injury of arteries and maintained on the same diet after wire injury until euthanization. Arterial injury was induced by a guide wire ²¹. Leukocyte interactions with the injured arteries were examined in vivo by intravital epifluorescence microscopy, or ex vivo on a coated surface by a technique described previously ²⁵. All animal experiments and care were approved by the University of Minnesota Animal Care and Use Committee in accordance with AAALAC guidelines. Detailed methods are available inthe supplemental material.

Results

A_2AR deficiency leads to the formation of large neointima in the injured arteries of apoE^−/− mice

To determine the role of A_2AR in the formation of arterial neointima, the carotid arteries of A_2AR^−/−/apoE^−/− mice and the littermate A_2AR^+/+/apoE^−/− mice were injured with a guide wire. Four weeks after the injury, their arteries were excised and processed for analysis. The neointima lesions were 83% larger in A_2AR^−/−/apoE^−/− mice than those in A_2AR^+/+/apoE^−/− mice (Fig. 1a). Additionally, the numbers of macrophages and α-actin-positive smooth muscle cells in the neointima of A_2AR^−/−/apoE^−/− mice were 57% and 68% greater than the numbers in A_2AR^+/+/apoE^−/− mice, respectively (Fig. 1b and 1c). A_2AR^−/−/apoE^−/− mice had slightly higher body weights and blood cholesterol levels than those of A_2AR^+/+/apoE^−/−mice after 6 weeks on a Western diet (Supplementary table I and II); however, these differences were not statistically significant.

Figure 1. A_2ARdeficiency increases injury-induced arterial neointima formation.

a, Movat staining and size quantification of neointima (I), media (M) and ratio of intima to media (I/M). b, Anti-F4/80 staining of infiltrated macrophages. c, Anti-α-actin staining of smooth muscle cells (SMCs). Data from 12 sections of 12 mice.

A_2AR deficiency in bone marrow–derived cells increases injury-induced arterial neointima formation

We next determined whether, and to what extent, the presence of A_2AR in bone marrow–derived cells (BMDCs) influences the formation of arterial neointima. We used bone marrow transplantation to generate the apoE^−/− chimeric mice that lacked A_2AR in their BMDCs and the apoE^−/− chimeric control mice that retained A_2AR in their BMDCs. Their carotid arteries were injured as described above. At the time of euthanization, both groups of mice were identical in blood cholesterol level, peripheral leukocyte count, and body weight (data not shown). However, the neointima lesions in chimeric mice lacking A_2AR in their BMDCs were 60% larger than the lesions in chimeric control mice (Fig. 2 and supplementary Fig. Ia). The arterial neointima was stained heavily for the marker of macrophages but not that of vascular smooth muscle cells (supplementary Figs. Ib and Ic). Additionally, we performed wire injury in chimeric mice that were A_2AR^−/−/apoE^−/− mice receiving bone marrow from A_2AR^+/+/apoE^−/− mice. The size of their arterial neointima was similar to that in the A_2AR^+/+/apoE^−/− control mice, which were A_2AR^+/+/apoE^−/− mice receiving bone marrow from A_2AR^+/+/apoE^−/− mice (data not shown). This result suggests that the enhanced neointima formation observed in A_2AR^−/−/apoE^−/− mice was mainly due to the deficiency of A_2AR in BMDCs

Figure 2. A_2AR deficiency in bone marrow derived cells increases injury-induced arterial neointima formation.

Size quantification of neointima (I), media (M) and ratio of intima to media (I/M). Data from 12 sections of 12 mice.

A_2AR deficiency increases leukocyte interactions with the injured arteries

Using epifluorescence intravital microscopy, we examined the interactions of leukocytes with injured mouse carotid arteries in vivo. Immediately after the arterial injury, circulating leukocytes (fluorescently labeled with rhodamine 6G) were able to roll onto, and adhere to, the injured arterial vessel wall. Within 30 minutes after arterial injury, the number of leukocytes rolling on and adhering to the arterial wall were 1.8 and 1.7 fold greater, respectively, in A_2AR^−/−/apoE^−/− mice than in A_2AR^+/+/apoE^−/− mice (Fig. 3a).

Figure 3. A_2AR deficiency increases leukocyte interactions with injured arteries.

a, Images and quantification of leukocyte rolling (←) and adhesion (▲) in injured carotid arteries. b to f, Quantification of carotid arteries stained for (b) platelets 1 hour after wire injury (WI), (c) neutrophils 1 hour after WI, (d) macrophages 1 hour after WI, (e) neutrophils 7 days after WI and (f) macrophages 7 days after WI (n=5).

To determine which types of cells were adhering to the injured arteries, we immunostained arterial cross-sections with markers specific for platelets, neutrophils, and monocytes. At 1 hour after the arterial injury, platelets and leukocytes covered the denuded luminal surface (Fig. 3b and supplementary Fig. IIa). Many more leukocytes bound to the injured area of carotid arteries in A_2AR^−/−/apoE^−/− mice than in A_2AR^+/+/apoE^−/− mice. Nearly all of the adherent leukocytes were neutrophils (Fig. 3c and supplementary Fig. IIb); the presence of monocytes (macrophages) was rare (Fig. 3d and supplementary Fig. IIc). At 7 days after wire injury, the adhesion and infiltration of both neutrophils and macrophages were observed in the injured carotid arteries, and more cells were observed in the injured vessel walls of A_2AR^−/−/apoE^−/− mice than in A_2AR^+/+/apoE^−/− mice (Fig. 3e, 3f and supplementary Fig. IId, IIe).

A_2AR deficiency increases the interactions of leukocytes with the P-selectin expressing surface

The homing ability of wild type (wt) and A_2AR-deficient neutrophils was further studied by intravital microscopy in an in vivo trauma model of mouse cremaster muscle. Mild trauma caused by the exteriorization of the cremaster muscle is an inflammatory stimulus. This stimulus induces the presentation of P-selectin and fibrinogen on the endothelium of postcapillary venules, and it mediates both P-selectin-dependent neutrophil rolling and b2 integrin-dependent neutrophil adhesion ²⁶. To evaluate the role of A_2AR in leukocyte-endothelial interactions, we measured the neutrophil rolling flux fraction, defined as the number of rolling neutrophils divided by the total numberof neutrophils passing through the same vessel. Compared to wt mice, A_2AR-deficient mice exhibited a 2-fold increase in the neutrophil rolling flux fraction (Fig. 4a). Additionally, many more neutrophils adhered to the endothelium of postcapillary venules in A_2AR^−/− mice than in wt mice.

Figure 4. A_2AR deficiency increases neutrophil homing ability.

a and b, Images and quantification of neutrophil rolling (←) and adhesion (▲) on the endothelium of postcapillary venules of mouse cremaster muscle (n=6). c and d, Neutrophil rolling and adhesion on activated platelets through micro-flow chambers (n=6).

To further determine which cellular A_2AR is responsible for these increased neutrophil-endothelial interactions in A_2AR^−/− mice, we performed an intravital microscopy study in chimeric mice with (control) or without A_2AR in their BMDCs. Compared to control mice, chimeric mice with A_2AR-deficient BMDCs had significantly increased neutrophil rolling and adhesion, and chimeric mice with A_2AR-deficient vessel walls had a similar level of neutrophil rolling and adhesion (Fig. 4b).

To compare the binding affinity of wt and A_2AR-deficient neutrophils, we employed an ex vivo micro-flow chamber^{27, 25} whose surface was coated with thrombin-activated wt platelets. Whole blood from A_2AR^+/+ or A_2AR^−/− mice was then perfused through the chamber. Immediately following the perfusion, the numbers of rolling and adherent neutrophils were 60% and 50% higher, respectively, in the blood of A_2AR^−/− mice than that of A_2AR^+/+ mice. After 5 minutes of perfusion, these numbers further increased to 80% and 90%, respectively (Fig. 4c). In addition, A_2AR -deficient neutrophils rolled much more slowly than wt neutrophils did (velocity data not shown).

Mechanisms for the increased homing ability of A_2AR-deficient neutrophils

To determine the influence of A_2AR deficiency in neutrophil homing ability, we first measured the expression of homing molecules on wt and A_2AR-deficient neutrophils by flow cytometry. The levels of PSGL-1, L-selectin, LFA-1, CD11b, and CXCR2 were identical between both types of neutrophils (supplementary Fig. IIIa).

PSGL-1 is the key molecule responsible for P-selectin-mediated neutrophil rolling on the injured arteries or postcapillary venules ²⁶. To explore the mechanisms responsible for the increased rolling of A_2AR-deficient neutrophils, we first used a flow cytometry assay to examine the binding of neutrophil PSGL-1 to P-selectin. In accordance with previously measured levels of PSGL-1 expression (above), the levels of P-selectin binding to wt and A_2AR-deficient neutrophils were identical (supplementary Fig. IIIb). Thus, under static conditions, PSGL-1 has the same level of P-selectin binding in A_2AR-deficient and wt neutrophils.

Using confocal microscopy, we investigated the distribution of PSGL-1 following the immunofluorescent labeling of neutrophils with PSGL-1 antibody. PSGL-1 was much more clustered on the surface of A_2AR-deficient neutrophils than that of wt neutrophils. This result was observed on neutrophils isolated from blood and on those adhering to injured arteries (Fig. 5a and 5b). Also, we explored several intracellular signaling mechanisms – i.e., the activity of phosphoinositide 3-kinase (PI3K)/Akt and mitogen-activated protein kinases (MAPKs) – that might be responsible for the increased homing ability of A_2AR-deficient neutrophils. In a western blot assay of neutrophils isolated from mouse spleen, the level of phosphorylated p38 MAPK in A_2AR-deficient neutrophils was 2.5 times higher than that in wt neutrophils (Fig. 5c). This phenomenon was also observed in injured arteries (Fig. 5d). In contrast, the levels of phosphorylated Akt and other MAPKs were not significantly different between wt and A_2AR-deficient neutrophils (supplementary Fig. IV).

Figure 5. A_2AR deficiency increases neutrophil p38 activity, and PSGL-1 clustering.

a, PSGL-1 distribution on neutrophil membranes. b, P-selectin and PSGL-1 on injured arteries. c and d, Western blot of P38 and phosphorylated p38 (pP38) of (c) spleen neutrophils and (d) carotid arteries 1 hour after injury (n = 3). e and f, Leukocyte rolling and adhesion through (e) P-selectin coated and (f) P-selectin and ICAM-1 coated micro-flow chambers (n=3).

Consistent with the PSGL-1 distribution under flow conditions, A_2AR-deficient neutrophils exhibited greater PSGL-1 binding than wt neutrophils did. This result was demonstrated by an ex vivo micro-flow chamber experiment ^{27, 25}, in which mouse whole blood was perfused through a P-selectin coated surface. After 5 minutes of perfusion, the whole blood from A_2AR^−/− mice presented an average of 1200 ± 80 rolling neutrophils per mm,² whereas the whole blood from wt mice presented 800 ± 70 rolling neutrophils per mm² (Fig. 5e). In a similar setup, 5-minute perfusion of mouse whole blood through the flow chamber coated with P-selectin and ICAM-1 led to a 2-fold higher number of A_2AR-deficient neutrophils adhering to the surface than the number of adhered wt neutrophils (Fig. 5f), indicating an increase in the affinity of b₂ integrins on A_2AR-deficient neutrophils.

To determine whether the enhanced p38 MAPK activity is responsible for the increased homing ability of A_2AR-deficient neutrophils, we treated A_2AR^−/− mice with the p38 inhibitor SB203580 (1mg/kg/day) intraperitoneally for 3 days before the perfusion. The treatment of SB203580 reduced the number of rolling cells on both the P-selectin coated surface and the surface coated with P-selectin and ICAM-1 to the levels in wt mice (Figs. 5e and 5f).

Discussion

Our results demonstrate that leukocyte A_2AR is critical for the protection of atherosclerotic mice from the injury-induced arterial neointima formation. A_2AR deficiency in neutrophils increases p38 activation, resulting in an increase in the clustering of PSGL-1 and the affinity of b₂ integrins. Ultimately, A_2AR deficiency enhances neutrophil recruitment to the injured arteries and augments the formation of the injury-induced arterial neointima.

Adenosine is known to inhibit a variety of neutrophil functions, including the production of TNF-a and superoxide anions ²⁸. Several studies have demonstrated that adenosine or adenosine receptor agonists inhibit the neutrophil adhesion to endothelial cells and fibrinogen by suppressing the upregulation of neutrophil b₂ integrins, as stimulated by N-formyl-methionyl-leucyl-phenylalanine (fMLP) ^29–31. When neutrophils are treated with an inhibitor of adenosine kinase (a major intracellular adenosine removal enzyme), adenosine levels increase, and the adhesion of activated neutrophils to cultured endothelial cells decreases through the alteration of L-selectin and the neutrophil cytoskeleton ³². Recently, adenosine or A_2AR agonists have been shown to induce heterologous desensitization of chemokine receptors and suppress the expression of very late antigen-4 (VLA-4) on stimulated human neutrophils ^{33, 34}. Notably, these previous studies revealed the pharmacological effects of adenosine or adenosine receptors on leukocytes. Also, the neutrophils used in the studies were activated in vitro by fMLP, TNF-a or other stimuli. The present study shows a physiological role of A _2AR in the regulation of neutrophil homing ability. Under physiological conditions, A_2AR deficiency did not affect the expression of adhesion molecules on neutrophils. Rather, A_2AR deficiency altered the cell membrane distribution of PSGL-1and the affinity of b₂ integrins, thereafter inhibiting neutrophil recruitment to the injured arteries in vivo.

We explored several possible intracellular signaling mechanisms for the increased homing ability of A_2AR-deficient neutrophils, including the activity of PI3K/Akt and MAPKs. We found that A_2AR regulates the neutrophil homing ability through MAPK p38. Our finding that A_2AR-deficient neutrophils exhibited increased p38 MAPK activation could result from a change in the signaling associated with intracellular cAMP ¹. A_2AR occupancy elevates the intracellular cAMP, thereafter inhibiting the activation of p38 via the cAMP response element–binding protein–induced dynein light chain ^{35, 36}. Leukocyte p38 is usually activated by outside-in signals, and the activated p38 is able to signal many pathways related to different cellular functions ^{35, 36}. In the early phase of arterial injury, the binding of neutrophil PSGL-1 with platelet P-selectin predominates neutrophilic interactions with the vessel wall. Therefore, we focused on the role of p38 activation on the regulation of PSGL-1 binding function. A_2AR-deficient neutrophils exhibit increased PSGL-1 clustering. Consequently, these cells have elevated levels of P-selectin binding. In our ex vivo micro-flow chamber assay, the inhibition of p38 activation significantly reversed the increased neutrophil rolling and adhesion, demonstrating that p38 is a key modulator in regulating A_2AR-mediated leukocyte homing. The enhanced PSGL-1 clustering can be regulated directly by p38 or indirectly through p38 altered cytoskeletal protein organization ^{37, 38}.

Increased neutrophil recruitment to the injured arteries contributes to neointima formation through many mechanisms ^{39, 40, 25}. Additionally, monocyte recruitment and activation could also play a role in the increased neointima formation following the arterial injury in A_2AR^−/−/apoE^−/− mice. Similar to neutrophils, A_2AR-deficient monocytes express adhesion molecules at the same levels seen in wt cells ⁵. However, A_2AR-deficient monocytes showed increased adhesion on atherosclerotic endothelium in an ex vivo mouse carotid artery perfusion model (data not shown). Furthermore, NF-kB was highly activated in A_2AR-deficient macrophages in atherosclerotic lesions ⁵. In a short-term arterial neointima formation model, these factors might contribute to the aggravation of arterial neointima formation. However, in a long-term spontaneous atherosclerosis model, A_2AR-deficiency results in a decrease in atherosclerosis ⁵. This discrepancy may be due to the fact that many resolution aspects of inflammation take actions to reduce the size of lesions during a long-term chronic process of spontaneous lesion formation. These resolution aspects include macrophage apoptosis, macrophage emigration from lesions, and the reduction of the vasa vasorum density of atherosclerotic arteries ^{41, 42, 43, 44}. We have previously shown that A_2AR-deficiency increases the apoptosis of macrophages ⁵. In addition, CD8 T cells target the neovascularization of atherosclerotic arteries to reduce atherosclerosis ⁴⁵. Dr. Sitkovsky’s group has reported that A_2AR-deficient T cells are able to suppress or even regress the tumor growth ⁴⁶, which may be achieved through the elimination of neovascularization by A_2AR-deficient CD8 T cells (personal communication). Therefore, the reduction of atherosclerosis in A_2AR^−/−/apoE^−/− mice may also be due to the suppression of neovascularization by A_2AR-deficient CD8 T cells.

Increased neointima formation in A_2AR^−/− mice is not attributed to the A_2AR deficiency in endothelial cells. In a mouse carotid artery wire injury model, the regeneration of endothelial cells on the injured area occurred at a few days after wire injury. These regenerated endothelial cells were inflamed and expressing a variety of adhesion molecules, which is the key for leukocyte recruitment to the injured arteries and the formation of arterial neointima ⁴⁷. In a similar mouse carotid artery injury model, endogenous extracellular adenosine, generated through CD73, protected the regenerated endothelial cells from inflammatory activation ⁴⁸. Also, in a mouse carotid artery ligation model, an A_2AR agonist inhibited adhesion molecule expression and consequent neointima formation in injured arteries ⁴⁹. Thus, it has been speculated that adenosine reduces endothelial activation through A_2ARs ¹. In this study, we found that A_2AR deficiency in endothelial cells does not affect the inflammatory response of endothelial cells that are regenerated on the injured area. This finding is supported by our immunostaining and western blot results from an in vivo carotid artery injury model and our flow cytometry results of an in vitro cell culture system to determine the role of A_2AR deficiency in endothelial inflammatory responses (Supplementary Figs. Va to Vc). Thus, different from its pharmacological effect as reported in other studies ^{50, 49}, endothelial A_2AR does not play a physiological role in determining the inflammatory status of endothelial cells in the arterial neointima model.

Augmented arterial neointima formation in A_2AR^−/− mice may not be associated with theA_2AR deficiency in platelets. Immediately after wire injury, platelets accumulated on the injured area. These platelets were activated and presenting P-selectin and other integrins to serve as a platform for leukocyte recruitment to the injured arterial vessel wall ¹⁷. In a dog coronary hypoperfusion model, it has been demonstrated that platelet A_2AR is critically involved in the effect of adenosine on the inhibition of platelet-leukocyte interactions and platelet aggregation ⁵¹. In humans, upregulated A_2AR on platelets, as a result of chronic caffeine intake, is beneficial for the prevention of platelet aggregation ^{52, 53}. Different from above studies, we did not find increased platelet activation mediated by thrombin in A_2AR- deficient platelets (Supplementary Fig. VIa). In response to thrombin stimulation, A_2AR-deficient mouse platelets presented P-selectin and exhibited platelet-leukocyte interactions at the same levels as seen in wt mouse platelets (Supplementary Fig. VIb). It is likely that the role of A_2AR in platelet activation is strain dependent. Otherwise, adenosine and adenosine receptors are not critically involved in platelet activation either in response to thrombin or in the milieu of arterial injury.

Taken together, the deficiency of A_2AR enhances leukocyte recruitment to the injured arteries and aggravates the formation of injury-induced arterial neointima. A_2AR antagonists are being tested in clinical trials for the treatment of neurodegenerative diseases and other chronic conditions. The results from this study indicate that the evaluation on the effect of A_2AR antagonists on arterial restenosis following arterial angioplasty should be considered.