Angiotensin-converting enzyme inhibition down-regulates the pro-atherogenic chemokine receptor 9 (CCR9)-chemokine ligand 25 (CCL25) axis

by

Many experimental and clinical studies suggest a relationship between enhanced angiotensin II release by the angiotensinconverting enzyme (ACE) and the pathophysiology of atherosclerosis. The atherosclerosis-enhancing effects of angiotensin II are complex and incompletely understood. To identify antiatherogenic target genes, we performed microarray gene expression profiling of the aorta during atherosclerosis prevention with the ACE inhibitor, captopril. Atherosclerosis-prone apolipoprotein E (APOE)-deficient mice were used as a model to decipher susceptible genes regulated during atherosclerosis prevention with captopril. Microarray gene expression profiling and immunohistology revealed that captopril treatment for 7 months strongly decreased the recruitment of proatherogenic immune cells into the aorta. Captopril-mediated inhibition of plaqueinfiltrating immune cells involved down regulation of the C-C chemokine receptor 9 (CCR9). Reduced cell migration correlated with decreased numbers of aorta-resident cells expressing the CCR9-specific chemoattractant factor, chemokine ligand 25 (CCL25). The CCL25-CCR9 axis was pro-atherogenic, because inhibition of CCR9 by RNA interference in hematopoietic progenitors of APOEdeficient mice significantly retarded the development of atherosclerosis. Analysis of coronary artery biopsy specimens of patients with coronary artery atherosclerosis undergoing bypass surgery also showed strong infiltrates of CCR9positive cells in atherosclerotic lesions. Thus, the C-C chemokine receptor, CCR9, exerts a significant role in atherosclerosis.
Several clinical studies show that pharmacological inhibition of the reninangiotensin-aldosterone system (RAAS) could mediate atheroprotection (1)(2)(3)(4). In agreement with those clinical data, experimental studies demonstrated a causal relationship between angiotensin II release by the angiotensinconverting enzyme (ACE), the angiotensin II AT 1 receptor and the development of atherosclerosis because genetic inactivation of either ACE or the AT 1 receptor prevents atherosclerosis development in various animal models of atherosclerosis (5,6). Likewise, inhibition of angiotensin II activity by ACE inhibitors or AT 1 receptor antagonists is curative in such animal models of atherosclerosis (7,8).
In view of the major antiatherosclerotic effects of ACE inhibitors and AT 1 antagonists, many studies investigated mechanisms that could contribute to atheroprotection upon angiotensin II inhibition. The effects of angiotensin II in atherosclerosis are complex involving actions on circulating blood cells and arterial smooth muscle cells as major targets (9)(10)(11). Established pro-atherogenic effects of angiotensin II on smooth muscle cells are related to a phenotype transformation from contractile to synthetic (11). The effects of angiotensin II on circulating immune cells are less clear, although, gene inactivation studies clearly demonstrated the involvement of circulating blood and progenitor cells in the atherosclerosis-enhancing activity of ACE and the AT 1 receptor (9).
Several studies revealed that AT 1 receptor-stimulated signalling enhances the pro-atherogenic potential of monocytes and macrophages by sustaining monocyte/macrophage migration into the subendothelial layer of the arterial wall, where transformation into foam cells occurs (9,10). In addition to macrophages, the AT 1 receptor may also exert effects on T lymphocytes by supporting T lymphocyte migration into the perivascular tissue (12). Infiltrating T cells also play a role in the pathogenesis of atherosclerosis as demonstrated by studies with animal models and patients (13,14). However, the full impact of angiotensin II on peripheral immune cells in the course of atherosclerosis development is still not understood.
To further analyze the role of immune cells in the atherogenic process, and the antiatherogenic potential of angiotensin II inhibition, we performed microarray gene expression profiling of the aorta as the target organ of atherosclerosis. As a prototypic ACE inhibitor we chose captopril with established anti-atherosclerotic actions in vivo, in various animal models (7,10,15), and patients (16). We report here that inhibition of angiotensin II generation and atherosclerosis development by captopril in hypercholesterolemic APOEdeficient mice led to a profound change of the aortic immune cell infiltrate, and down regulated the chemoattractant C-C chemokine receptor type 9 (CCR9), with its specific ligand, C-C motif chemokine 25 (CCL25).

EXPERIMENTAL PROCEDURES
Animal experiments-For the study, 4-6 week-old APOE-deficient mice on a C57BL/6J background, and non-transgenic C57BL/6J controls were used. Mice were kept on a 12 h light/12 h dark regime, had free access to food and water, and were fed a standard rodent chow containing 7 % fat and 0.15 % cholesterol (AIN-93-based diet). As indicated, APOE-deficient mice received captopril in drinking water (20 mg/kg; prepared fresh every day) for 7 months, or the angiotensin II AT 1 receptor-specific antagonist, losartan (30 mg/kg). At an age of 32-34 weeks, mice were anesthetized with ketamine/xylazine (100/10 mg/kg), and perfused intracardially with sterile PBS. The aorta was isolated, rapidly dissected free of adherent adipose tissue, and immediately frozen in liquid nitrogen, or processed for further use. Total plasma cholesterol was determined using a standard enzymatic kit, and systolic blood pressure was measured by the tail-cuff method (10). Atherosclerotic lesion area of the aortic root was quantified by computerized image analysis as described (10). Mouse peripheral blood mononuclear cells and monocytes were isolated from heparin-anticoagulated blood diluted with PBS by density gradient centrifugation through Ficoll Paque or Optiprep followed by plastic adherence or by Nycoprep density gradients (10). Monocyte purity was > 80% as determined by immunofluorescence staining of cell-specific antigens (10). Monocytes were cultivated as described (10). All solutions and chemicals used in monocyte isolation and activation were endotoxin-free (endotoxin <0.008 ng/ml).
All animal experiments were performed in accordance with the NIH guidelines, and were reviewed and approved by the local committees on animal care and use (University of Zurich and Hamburg).
Patients-The study was performed with coronary artery biopsy specimens obtained form 15 patients with coronary artery disease undergoing coronary artery bypass graft surgery (diseased vessels 2.3±0.5; age 51±6 years; 9 males). All the patients had a recent history of an acute coronary syndrome (3 weeks to 3 months prior to bypass surgery).
Coronary artery biopsy specimens without atherosclerotic lesions as confirmed by the absence of oil-red O-stained atherosclerotic lesions served as controls. Informed consent was obtained from all the participants. The study was approved by the local ethics committee (Ain Shams University, Cairo).
Bone marrow transplantation-For lentiviral-mediated RNA interference (RNAi) inhibition of CCR9 in hematopoietic progenitors, transduced bone marrow cells isolated from APOE-deficient donors (1 x 10 6 ) were injected into the tail vein of irradiated [11][12] week-old, syngeneic APOE-deficient recipients as described previously (10). For RNA interference, lentiviral constructs with RNA polymerase II promoter-driven expression of micro-RNAs targeting CCR9 (nucleotides 359-379), or β-galactosidase as a control (nucleotides 1298-1318) were used. Generation of high-titer lentiviruses and murine hematopoietic cell transduction were performed essentially as described (10,17). Four months after bone marrow transplantation, mice were sacrificed, blood was collected by cardiac puncture, and mononuclear cells were isolated for assessment of CCR9 protein levels (10). Persistent reduction of CCR9 in peripheral blood mononuclear cells (18) was maintained throughout the observation period of 4 months after transplantation.
Microarray gene expression profiling-Microarray gene expression analysis of the aorta was performed essentially as described previously (19). Briefly, total RNA of the dissected aortas was isolated with the RNeasy Midi kit (Qiagen). Procedures for cDNA synthesis, labeling and hybridization were carried out according to the protocol of the manufacturer (Affymetrix). Antibodies-The following antibodies were used for immunohistochemistry, immunofluorescence, and immunoblotting: monoclonal rat anti-CD8a antibodies (BD Pharmingen); rabbit anti-CD22 antibodies (BD Pharmingen); monoclonal rat anti-MOMA-2 antibodies (Serotec); affinity-purified rabbit anti-AT 1 receptor antibodies raised against an antigen corresponding to positions 306-359 of the mouse AT 1A receptor sequence (10), affinity-purified rabbit/rat anti-CCR9 antibodies raised against recombinant CCR9 protein, or an antigen encompassing amino acids 321-351 of the human or mouse CCR9 sequence, respectively; rabbit/rat anti-mCCL25 antibodies raised against an antigen encompassing amino acids 1-27 of the mouse CCL25 sequence, or recombinant mouse CCL25. Recombinant CCL25 was produced in Escherichia coli and purified similarly as described (20), and the CCR9 protein was expressed in and purified from Spodoptera frugiperda cells infected with a recombinant baculovirus encoding CCR9 as detailed previously (21). Immunization of antigens (1 mg/ml) in phosphate-buffered saline mixed 1:1 with complete Freund`s andjuvant was done according to standard protocols followed by booster injections in incomplete Freund`s adjuvant 2 weeks after the first injection and monthly thereafter (17). Immunoblotting and immunohistochemistry were routinely used to determine and confirm cross-reactivity of the antibodies with the respective proteins.
Immunohistology and immunofluorescence-For immunohistology, cryosections or paraffin sections of the aortic sinus region (10 µm, taken at 50 µm intervals for analyses, 10-15 sections per set) were used (10,17). Immunohistological detection of CCR9 and CCL25 was performed with affinity-purified, polyclonal antibodies pre-absorbed to mouse or human tissue similarly as described (10,17). All sections were imaged with a Leica DMI6000 microscope equipped with a DFC420 camera.
Immunofluorescence localization studies were performed with cryosections (10 µm) of post-fixed and frozen aortas obtained from 8 month-old APOE-deficient mice. For co-localization of CCR9 and CCL25, affinitypurified rabbit anti-CCL25 antibodies and rat anti-CCR9 antibodies were applied (dilution 1:4000), followed by secondary antibodies or F(ab) 2 fragments of the antibodies labelled with Alexa Fluor 488 and Alexa Fluor 546, respectively (dilution 1:5000), and counterstaining with DAPI. Sections were imaged with a Leica DMI6000 microscope and a Leica (TCS) confocal laser microscope.
Immunoblot detection of proteins-Immunoblot detection of proteins was performed as described previously with affinity-purified antibodies pre-absorbed to mouse or human proteins, respectively (10,17). Bound antibody was visualized with F(ab) 2 fragments of enzyme-coupled secondary antibodies, or by enzyme-coupled protein A followed by enhanced chemiluminescence detection.
Statistical Analysis-Unless otherwise stated, data are expressed as mean ± S.D. To determine significance between two groups, we made comparisons using the unpaired twotailed Student`s t test. p values of < 0.05 were considered significant unless otherwise indicated.

Captopril inhibits atherosclerosis development of APOE-deficient mice.
To assess the effect of ACE inhibition on atherosclerosis development, atherosclerosisprone, 4-6 week-old, APOE-deficient mice were treated for 7 months with the ACE inhibitor, captopril, in drinking water. After 7 months, captopril-treated mice had a modest decrease in body weight and blood pressure relative to age-matched, non-treated APOEdeficient mice (Fig. 1A). Those observations are in agreement with established effects of ACE and ACE inhibitors, respectively (10,22). In contrast, captopril, did not affect total plasma cholesterol levels of APOE-deficient mice (Fig. 1A).
Concomitantly, captopril treatment largely prevented the development of atherosclerotic plaques in the aortic root of APOE-deficient mice as evidenced by histological analysis (Fig. 1B; middle vs. left panel). As a control, the aortic root of nontransgenic C57BL/6J mice was lesion-free ( Fig. 1B; right panel). Quantification of the plaque area demonstrated that captopril treatment had reduced the mean aortic root lesion area by 79 ± 9 % relative to untreated APOE-deficient mice (Fig. 1C). Altogether, these findings confirm that captopril efficiently inhibits atherosclerosis development of APOE-deficient mice (7,10).
Microarray analysis of atherosclerosis prevention by captopril. To assess the mechanisms underlying atherosclerosis prevention mediated by captopril, we performed microarray gene expression profiling of the aorta of 8 months-old APOEdeficient mice with overt atherosclerosis relative age-matched APOE-deficient mice treated for 7 months with the ACE inhibitor, captopril ( Fig. 1D; middle vs. left panel). As a control, we used non-transgenic C57BL/6J mice without atherosclerosis ( Fig. 1D; right panel). Oil-red O staining of representative aortas of the three study groups confirmed that captopril treatment for 7 months prevented the development of atherosclerotic plaques in the aorta of APOE-deficient mice whereas agematched APOE-deficient mice showed high plaque load of the aortic arch ( Fig. 1D; middle vs. left panel).
For microarray analysis, total aortic RNA isolated of the three study groups were used, i.e. untreated APOE-deficient, captopriltreated APOE-deficient, and non-transgenic C57BL/6J control mice. Data obtained after chip scanning revealed uniform quality of the microarrays as assessed by the comparable number of probe sets present, and the 3´/5`ratio of probe sets detecting GAPDH and beta actin (GEO database no. GSE19286).

Microarray analysis reveals infiltrating immune cells in the aorta of APOE-deficient mice.
After data normalization, microarray data were evaluated. Because ACE inhibitors target inflammatory immune cells (9,10), we determined the impact of captopril on immune cell recruitment. To identify microarray probe sets specific of aorta-infiltrating immune cells, we applied a stringent filtering strategy (19). Aortic probe sets of captopril-treated APOE-deficient mice were selected displaying, (i) a significantly different signal compared to untreated APOEdeficient mice (*, p < 0.05), (ii) a more than 50 % reduction of signal intensity relative to untreated APOE-deficient mice, (iii) membrane localization according to gene ontology (GO) analysis, and (iv) immune cell specificity. The very stringent approach identified only 9 differentially expressed probe sets, which showed effects of captopril on the infiltration of the aorta with immune cells (Fig.  2A). Notably, captopril normalized the increased aortic expression of T cell-and macrophage-specific marker proteins of APOE-deficient mice, i.e. CD8a, CD8b, CD4, CD28, CCR9, and Mmd (monocyte to macrophage differentiation-associated) ( Fig.  2A). In contrast, B lymphocyte-specific markers, such as CD79a and CD22, were not significantly affected by captopril ( Fig. 2A). As a control, atherosclerosis development of APOE-deficient mice was accompanied by a strong up-regulation of all mononuclear cellspecific probe sets relative to non-transgenic C57BL/6J controls, i.e. T cell-, macrophageand B cell-specific marker proteins ( Fig. 2A). Thus, the microarray approach was capable to detect aorta-infiltrating immune cells in the course of atherosclerosis development.
Captopril suppresses aortic T cell infiltration of APOE-deficient mice. To validate the microarray data regarding the observed effect of captopril on aortic T cell recruitment, we performed immunohistology analysis of the aortic sinus region. Immunostaining with antibodies specifically recognizing CD8a revealed that captopril significantly suppressed the infiltration of the aortic sinus region with CD8a-positive cells (Fig. 2B, middle vs. left panel). As a control, non-transgenic C57BL/6J control mice did not show substantial infiltration of the aorta with CD8a-positive cells compared to APOEdeficient mice (Fig. 2B, right vs. left panel).
The CD8a positive cells did not stain positive for the dendritic cell-specific marker protein, CD11c (data not shown). In addition, the microarray data did not show a significant difference in the expression of CD11c between untreated and captopril-treated APOEdeficient mice (cf. GSE19286).
Altogether, the data are compatible with the notion that captopril suppresses aortic T cell infiltration of APOE-deficient mice. Since T cells are considered pro-atherogenic (13), suppression of T cell migration may contribute to atherosclerosis prevention exerted by captopril.
Captopril maintains the presence of B lymphocytes in the aorta of atherosclerosisprone APOE-deficient mice. We also used immunohistology to validate the microarray data on B cell migration. In contrast to T cells, B lymphocytes are considered atheroprotective (23). In agreement with the microarray data, captopril treatment did not significantly reduce the presence of B cells in the aorta of APOEdeficient mice according to immunohistology analysis applying CD22-specific antibodies (Fig. 2C, middle vs. left panel). As a control, B cells were largely absent in non-transgenic C57BL/6J control mice (Fig. 2C, right vs. left  panel). Altogether, captopril treatment may sustain the migration of atheroprotective B cells into the aorta of atherosclerosis-prone APOE-deficient mice while suppressing the aortic infiltration with atherosclerosispromoting T cells.
Captopril suppresses the recruitment of CCR9-positive immune cells into the aorta of APOE-deficient mice. In addition to the lymphocyte markers for T cells and B cells, the microarray data revealed the strong upregulation of three different probe sets detecting CCR9 in the atherosclerotic aorta of APOE-deficient mice relative to nontransgenic controls ( Figs. 2A and 3A). All three probe sets were substantially down regulated by captopril treatment (Fig. 3A). Immunoblotting with CCR9-specific antibodies (Fig. 3B, lane 1 vs. 2) confirmed the microarray data revealing a strong reduction of aortic CCR9 protein levels upon captopril treatment (Fig. 3B; lane 4 vs. 3; and cf. Fig.  8D). As a control, non-transgenic control mice did not show significant CCR9 protein levels in the aorta (Fig. 3B, lane 5).
CCR9 is expressed on various immune cells including T cells and macrophages (20,24). Immunohistology was performed to identify the aorta-infiltrating cells, which were CCR9-positive. Immunohistology of aortic sinus sections with CCR9-specific antibodies showed localization of CCR9 on plaqueresident cells of the atherosclerotic aorta (Fig.  3C, left panel). CCR9-positive cells were strongly reduced by captopril treatment and virtually absent in non-transgenic C57BL/6J controls (Fig. 3C, middle and right panel).
Some CCR9-positive cells were docking to the atherosclerotic intima from the side of the aortic lumen suggesting that CCR9positive cells could be derived from circulating immune cells (Fig. 3D; cf. Fig. 4). As a control, significant staining of endothelial cells by CCR9-specific antibodies was not observed (data not shown).
Immunofluorescence was applied to determine the cellular localization of CCR9. Membrane-localized CCR9 was present on cells, which displayed characteristics of macrophages differentiating into foam cells with multiple cell nuclei (Fig. 3E). CCR9positive cells in the atherosclerotic aortic intima of APOE-deficient mice were localized in close proximity to cells releasing the CCR9specific chemoattractant, CCL25 (Fig. 3E). Thus, atherosclerotic plaques of APOEdeficient mice show CCR9-positive cells, and captopril treatment strongly reduced the recruitment of aorta-infiltrating CCR9-positive cells.
CCR9 localization on plaque-resident macrophages and circulating monocytes. We next asked whether the plaque-resident CCR9positive cells showed characteristics of macrophages as a major source of foam cells. Immunofluorescence studies revealed the colocalization of CCR9 with the monocyte/macrophage-specific marker, MOMA-2, on plaque-resident macrophages/foam cells of the aorta (Fig. 4A). In agreement with the migration of bloodderived monocytes into the inflamed aortic intima, circulating MOMA-2-positive monocytes of APOE-deficient mice also stained positive for the CCR9 protein (Fig.  4B). In contrast, circulating mononuclear cells of captopril-treated and non-transgenic control mice did not show significant CCR9 protein levels (cf. Figs. 6A, 7C). In addition to monocytes and macrophages, plaqueinfiltrating and circulating CD8a + T cells were also CCR9-positive (Fig. 4C, D).
Captopril reduces plaque-resident CCL25-positive cells. The C-C chemokine, CCL25, attracts CCR9-positive cells (20,24). Immunohistology revealed the CCR9-specific chemoattractant, CCL25, in the atherosclerotic aortic root with advanced plaques adjacent to necrotic core areas (Fig. 5A). In agreement with the immunohistology data, there was a significantly increased CCL25 expression in the atherosclerotic aorta relative to nontransgenic C57BL/6J control mice (Fig. 5B). Interestingly, atherosclerosis prevention by captopril treatment largely suppressed the increase in aortic CCL25 expression of APOEdeficient mice (Fig. 5B). The microarray data on CCL25 expression in the atherosclerotic aorta were confirmed by immunoblotting with CCL25-specific antibodies (Fig. 5C).
Immunohistology revealed the CCL25 protein on plaque-infiltrating cells of the atherosclerotic aortic intima of APOEdeficient mice (Fig. 5D, left panel). Infiltration of the aortic intima with CCL25 positive cells was strongly reduced by captopril treatment (Fig. 5D, middle panel). As a control, the aortic intima of non-transgenic C57BL/6J control mice did not show significant numbers of CCL25 positive cells (Fig. 5D, right panel). Thus, atherosclerosis development is accompanied by the appearance of CCL25 and its specific receptor, CCR9, on plaqueinfiltrating cells of APOE-deficient mice. On the other hand, prevention of atherosclerosis development by captopril down-regulated the CCL25-CCR9 axis. CCL25 co-localizes with the angiotensin II AT 1 receptor on plaque-resident cells. Enhanced activation of AT 1 receptors by ACE-dependent angiotensin II generation is considered to be causally involved in atherogenesis (5)(6)(7)(8)10). To assess whether there could be a relationship between the AT 1 receptor and plaque-infiltrating CCL25 positive cells, we performed immunofluorescence localization of AT 1 and CCL25. Immunofluorescence data with AT 1specific and CCL25-specific antibodies, respectively, showed localization of the AT 1 receptor on plaque-infiltrating CCL25-positve cells of APOE-deficient mice (Fig. 5E). This finding strongly suggests that AT 1 -stimulated signalling could directly modulate the expression of CCL25. Previous findings support such a conclusion demonstrating that AT 1 -mediated signalling triggers the expression of Egr-1 (early growth response protein 1) target genes such as CCL25 (25,26).
In agreement with that notion, treatment of APOE-deficient mice with the AT 1 -specific antagonist losartan reduced the aortic protein levels of CCL25 by 89±6 % similarly as did the ACE inhibitor captopril (n=3 mice/group, *, p<0.0004; Fig. 5F). Together these data suggest a causal relationship between enhanced AT 1 receptor activation and plaque-infiltrating CCL25positive cells.
Inhibition of ACE-dependent angiotensin II AT 1 receptor activation reduces CCR9 protein levels of circulating mononuclear cells. We next analyzed the relationship between the AT 1 receptor and the CCL25-specific receptor, CCR9. The CCR9 protein is expressed on plaque-infiltrating cells and on circulating mononuclear cells of APOE-deficient mice (cf. Fig. 4). Treatment with the ACE inhibitor captopril strongly decreased mononuclear cell CCR9 protein levels of APOE-deficient mice by 84±6 % (n=4 mice/group, *, p<0.001; Fig. 6A). In addition, treatment of APOE-deficient mice for 7 months with the angiotensin II AT 1 receptor-specific antagonist, losartan, similarly reduced the CCR9 protein by 81±6 % (n=4 mice/group; *, p<0.001; Fig. 6A). Together these observations indicate, that atherosclerosis development of APOEdeficient mice is accompanied by an AT 1 receptor-dependent up-regulation of the CCR9 protein on circulating mononuclear cells.
AT 1 receptor activation increases CCR9 protein levels of cultivated monocytes. The causal relationship between increased CCR9 protein levels and AT 1 receptor activation was analyzed with cultivated monocytes isolated from APOE-deficient mice. Incubation of cultivated monocytes for 48 h with the AT 1 receptor antagonist, losartan, significantly reduced the angiotensin II-stimulated increase in CCR9 protein levels by 61± 7 % (n=3, *, p<0.01) as determined by immunoblotting (Fig. 6B). As a control, losartan had no significant effect on cultivated monocytes in the absence of angiotensin II (Fig. 6B). Thus, AT 1 receptor activation increases CCR9 protein levels of cultivated monocytes isolated from APOE-deficient mice.
Co-localization of AT 1 and CCR9 on plaque-resident cells. In addition to circulating monocytes, CCR9 is present on plaqueresident macrophages (cf. Fig. 4A,B). Immunofluorescence co-localization studies with CCR9-specific and AT 1 -specific antibodies, respectively, revealed localization of AT 1 on CCR9-positive plaque-resident cells of APOE-deficient mice (Fig. 6C). Together these findings strongly suggest the involvement of enhanced AT 1 -stimulated signalling in the increased CCR9 protein levels of circulating monocytes and plaque-resident macrophages/foam cells of APOE-deficient mice.
Inhibition  (Fig. 7A). Four months after bone marrow transplantation, APOE-deficient mice with lentivirus-mediated RNAi inhibition of CCR9 showed a significant reduction of circulating mononuclear cell CCR9 protein levels relative to control APOE-deficient mice receiving bone marrow transduced with a control lentivirus (Fig. 7B). RNAi-mediated inhibition of CCR9 induced a decrease of CCR9 protein levels on circulating mononuclear cells by 89±6 % (n=5 mice/group, *, p<0.001) as determined by densitometric immunoblot scanning (Fig. 7B).
The RNAi-mediated down regulation of CCR9 in APOE-deficient mice was comparable to the captopril-induced inhibition of CCR9 protein levels (Fig. 7C). The induction of the CCR9 protein on circulating mononuclear cells was related to atherosclerosis development, because mononuclear cells of non-transgenic C57BL/6J control mice did not show significant CCR9 protein levels (Fig. 7C).
Immunofluorescence studies showed that plaque-resident cells were positive for CCR9 and the T cell-and monocyte/macrophage-specific immune cell markers, CD8a and MOMA-2 (cf. Fig. 4). Lentiviral-mediated RNAi inhibition of CCR9 in hematopoietic progenitors and circulating immune cells (cf. Fig. 7B) was accompanied by a significant reduction of aortic CCR9 protein levels of APOE-deficient mice by 67±8% (n=5, *, p<0.001; Fig. 7D). As a control, treatment with captopril and losartan led to a comparable decrease of the aortic CCR9 protein of APOE-deficient mice (Fig.  7D).
Next we assessed atherosclerosis progression upon inhibition of CCR9 expression.
Quantification of the atherosclerotic lesion area of the aorta revealed that inhibition of CCR9 by RNAi significantly reduced the development of the atherosclerotic lesion area by 55 ± 13 % (n=5 mice/group, *, p<0.001; Fig. 7E and F). As a control, down regulation of CCR9 did not significantly affect plasma cholesterol, body weight or blood pressure (Fig. 7G). Together these findings strongly indicate that the CCR9-CCL25 axis contributes to the development of atherosclerosis of APOE-deficient mice.
Infiltrates of CCR9-positive cells in atherosclerotic lesions of patients with coronary artery disease. To analyze the potential impact of CCR9 for atherosclerosis development in patients, we determined the CCR9 protein in atherosclerotic lesions of patients with coronary artery atherosclerosis undergoing coronary artery bypass surgery. Coronary artery biopsy specimens with atherosclerotic lesions as determined by positive oil-red O staining showed a strong infiltration of CCR9-positive cells (Fig. 8A, B; left panels). For comparison, coronary artery biopsy specimens without atherosclerotic lesions as verified by the absence of oil-red Opositive lesions, did not show significant CCR9 staining (Fig. 8A, B; right panels).
Immunoblot detection with CCR9specific antibodies confirmed the CCR9 protein on atherosclerotic lesions of patients (Fig. 8C). For comparison, vessel biopsy specimens without atherosclerotic lesions did not show significant CCR9 protein levels (Fig.  8C). As a control, the CCR9 antibodies crossreacted specifically with the human CCR9 protein of transfected HEK cells whereas the antibodies did not interact with mocktransfected control cells (Fig. 8C, lane 1 versus  2). Altogether, atherosclerotic lesions of patients and mice showed a strong infiltration with CCR9-positive cells.

DISCUSSION
Inflammation plays an important role in the pathogenesis of atherosclerosis because atherosclerotic plaque development involves the migration of pro-inflammatory immune cells into the arterial wall. Several studies with animal models of atherosclerosis showed that C-C chemokine receptors are important for the initiation and progression of atherosclerotic plaque formation by recruiting monocytes/macrophages into the subendothelial layer (27). Evidence for the causal involvement of individual C-C chemokines and their receptors is provided by gene inactivation and over-expression studies in mice (28)(29)(30). In addition to the experimental data, clinical studies with patients confirmed the up-regulation of several C-C chemokines in patients with atherosclerosis, unstable angina pectoris and myocardial infarction (31,32). The important role of C-C chemokines in atherogenesis was further confirmed by broad-spectrum C-C chemokine blockers, which inhibited atherosclerotic plaque formation in mice (33,34). The latter studies point to the therapeutic potential of C-C chemokine receptor blockade for atherosclerosis.
The present study investigated the impact of pro-inflammatory immune cells for atherosclerosis development by microarray gene expression analysis of aortic genes from atherosclerotic APOE-deficient mice relative to ACE inhibitor-treated mice. Microarray analysis and immunohistology revealed substantial effects of ACE inhibitor therapy on immune cell recruitment into the aorta. The study identified the strong up-regulation of a C-C chemokine receptor axis in atherosclerotic lesions, i.e. atherosclerosis development was accompanied by a strong increase of plaqueresident immune cells, which were positive for the chemoattractant CCL25, and its receptor CCR9. Atherosclerosis treatment by captopril reduced the aortic infiltration with CCR9-and CCL25-positive immune cells. That observation is intriguing because there seems to be a causal relationship between atherosclerosis-induced angiotensin II-AT 1 receptor activation and enhanced CCR9-CCL25 expression: The expression of CCL25 as an Egr-1 target gene (25), could be directly triggered by angiotensin II-mediated AT 1 receptor signalling, which induces Egr-1 activation (26). In agreement with that notion, the AT 1 receptor protein was co-localized with CCR25 on plaque-resident cells of APOEdeficient mice. Moreover, treatment with the AT 1 -specific antagonist, losartan, reduced the number of plaque-resident CCL25-positive cells in vivo.
Analogously to CCL25, the upregulation of CCR9 in the course of atherosclerosis could be also mediated by angiotensin II-dependent AT 1 receptor activation because ACE inhibition and AT 1 antagonism blunted the increased mononuclear cell CCR9 protein levels of APOE-deficient mice, and the aortic recruitment of CCR9positive immune cells. Moreover, angiotensin II-dependent AT 1 receptor activation directly stimulated an increase in CCR9 protein levels of isolated monocytes. Moreover, the AT 1 receptor protein was expressed on CCR9positive cells of atherosclerotic plaques. A causal relationship between CCR9 expression and AT 1 receptor activation is also suggested by previous data, and could involve the AT 1dependent induction of tumour necrosis factor alpha (TNF-alpha) synthesis (35,36).
In addition to the relationship between the angiotensin II system and CCR9-CCL25 protein expression, several lines of evidence support the concept of a causal involvement of the CCR9-CCL25 axis in atherosclerosis development. (I) Immunohistology and immunofluorescence data revealed circulating CCR9-positive monocytes and plaque-resident CCR9-positive macrophages with characteristics of multinucleated, plaqueforming foam cells in APOE-deficient mice. In agreement with an atherosclerosisenhancing function, the CCR9/CCL25 axis is induced by several pro-inflammatory and pro-atherogenic stimuli such as interferon-gamma, macrophage-colony stimulating factor (M-CSF), and TNF-alpha (37)(38)(39). Those CCR9/CCL25-inducing factors are causally involved in atherosclerosis development (40,41). Therefore targeting of the CCL25-CCR9 axis may have potential significance in diagnosing and treating atherosclerotic plaque formation independently of the angiotensin II system.

FOOTNOTES
We thank A. Abd-elbaset for excellent assistance in animal experiments. This work was supported in part by the Swiss National Science Foundation.