Prevention of CaCl2-induced aortic inflammation and subsequent aneurysm formation by the CCL3–CCR5 axis

Inflammatory mediators such as cytokines and chemokines are crucially involved in the development of abdominal aortic aneurysm (AAA). Here we report that CaCl2 application into abdominal aorta induces AAA with intra-aortic infiltration of macrophages as well as enhanced expression of chemokine (C-C motif) ligand 3 (CCL3) and MMP-9. Moreover, infiltrating macrophages express C-C chemokine receptor 5 (CCR5, a specific receptor for CCL3) and MMP-9. Both Ccl3−/− mice and Ccr5−/− but not Ccr1−/− mice exhibit exaggerated CaCl2-inducced AAA with augmented macrophage infiltration and MMP-9 expression. Similar observations are also obtained on an angiotensin II-induced AAA model. Immunoneutralization of CCL3 mimics the phenotypes observed in CaCl2-treated Ccl3−/− mice. On the contrary, CCL3 treatment attenuates CaCl2-induced AAA in both wild-type and Ccl3−/− mice. Consistently, we find that the CCL3–CCR5 axis suppresses PMA-induced enhancement of MMP-9 expression in macrophages. Thus, CCL3 can be effective to prevent the development of CaCl2-induced AAA by suppressing MMP-9 expression.

A n abdominal aortic aneurysm (AAA) is characterized by enlarged aorta with the structural destruction composed of the fragmentation of the extracellular matrix. Several conditions including aging, hypertension, smoking, hypercholesterolemia, diabetes, infection, genetic conditions, and trauma are denoted as risk factors for AAA formation 1,2 . Moreover, males are reported to have a much higher risk of developing AAA than females 3 . AAA usually causes no severe symptoms, at most with faint abdominal, back, or leg pains, and as a consequence, it frequently ruptures without preceding symptoms and progresses rapidly, resulting in high acute mortality up to about 80% 2 . Thus, AAA is one of the major causes of sudden unexpected natural death.
Multiple factors such as atherosclerosis, hemodynamics, genetic deficiency, microbiological infection, trauma, environmental conditions, and immunological states are presumed to be involved in the pathogenesis of AAA. Inflammation is temporally and spatially associated with disruption of the orderly lamellar structure of the aortic media and eventually leads to AAA development and progression. In support of this assumption, human AAA lesions exhibit dense infiltration of a myriad of innate and adaptive immune cells, including macrophages, lymphocytes, neutrophils, mast cells, and dendritic cells 4 . Among these infiltrates, macrophages and T lymphocytes are the predominant cell types 5,6 . Consistently, the intense inflammatory response can cause gradual aortic dilatation and subsequent aneurysm formation in the aorta and carotid artery in animal models 7,8 . Moreover, blocking aortic wall inflammation can limit the development and progression of aortic aneurysm 9 , and accentuation of inflammation can increase the frequency and size of the aneurysms 10,11 .
Inflammation in the aorta can frequently induce the breakdown of normally long-lived matrix macromolecules present in aortic walls, such as collagen and elastin, and eventually dilate the aortic wall, resulting in aneurysm formation. Collagen and elastin are mainly degraded by a family of endopeptidases, matrix metalloproteinases (MMPs). Moreover, several lines of evidence indicate that MMPs with a high capacity to degrade collagen, particularly MMP-9, was expressed by infiltrating macrophages as well as resident vascular smooth muscle cells in aneurysm lesions [12][13][14][15] . Based on these observations, the efficacy of MMP inhibitors was examined in preclinical animal models [16][17][18] and in early clinical trials for patients with aneurysm 19,20 . However, MMP inhibitors could only delay the speed of aneurysm formation without preventing its progression. Thus, it is necessary to establish an effective therapeutic strategy for aneurysm based on the understanding of the molecular and cellular mechanisms underlying aneurysm formation 21 .
The difficulties in obtaining human aneurysm tissues at the early phase have hampered the understanding of the molecular and cellular changes at the early step of aneurysm formation. As a consequence, animal models are still indispensable for clarifying the early step of aneurysm formation, aortic dilatation. Currently, three methods have been commonly used to induce aortic dilatation and subsequent aneurysm formation; periaortic application of CaCl 2 , transient intraluminal elastase perfusion, and angiotensin II (Ang II)-infusion 22 . In this study, we use mainly periaortic application of CaCl 2 but also Ang II-infusion to induce aortic dilatation and aneurysm formation, because these models exhibit abundant inflammatory infiltrates and MMP expression, similarly as observed on human AAA 13 .
Chemokine system is essential for leukocyte trafficking and eventual inflammatory responses and tumor development 23 . Moreover, several chemokines could have diverse roles in the aneurysm formation in a context-dependent manner [24][25][26] . Hoh et al. 25 demonstrated that CCL2 could promote elastase-induced aneurysm formation of the carotid artery by using the CCL3mediated pathway. MacTaggart and colleagues 24 revealed the involvement of CCR2 but not CCR5 in CaCl 2 -induced AAA development in SV129 mice. CCR5 and CCR1 are specific receptors for CCL3, a macrophage-tropic inflammatory chemokine 23 . Our immunohistochemical analysis of human AAA samples detects the expression of CCL3 and CCR5 as well as MMP-9 ( Supplementary Fig. 1). Moreover, intra-aortic MMP9 mRNA expression is significantly enhanced in human AAA samples, compared with human normal aorta ones (Supplementary Fig. 1). However, it still remains elusive on the pathophysiological roles of CCL3 in AAA formation.
CaCl 2 treatment enhances the intra-aortic CCL3 expression in mice, preceding AAA formation. In contrary to our expectation that the lack of CCL3 may prevent the development of an abdominal aneurysm, AAA formation is exaggerated in Ccl3 −/− mice, compared with WT mice, with enhanced macrophage infiltration and Mmp9 expression in the aorta. Moreover, the absence of CCR5 but not CCR1 also exacerbates CaCl 2 -induced AAA formation like Ccl3 −/− mice. Similar observations are obtained on the Ang II-infusion AAA model. Moreover, CCL3 treatment can prevent CaCl 2 -induced AAA formation in both wild-type and Ccl3 −/− mice with dampened MMP-9 expression. Thus, CCL3, hitherto considered as a typical inflammatory chemokine, can exhibit anti-inflammatory activities by acting its specific receptor, CCR5, in these AAA models.

Results
Exaggerated AAA formation in the absence of CCL3. We initially examined immunohistochemically the expression of CCL3 in human aortic aneurysmal tissues. CCL3 was mainly detected in infiltrating leukocytes present in aortic tissues (Fig. 1a). Double-color immunofluorescence analysis demonstrated that CD68 + macrophages were the main cellular source of CCL3 (Fig. 1a). These observations would indicate the potential involvement of CCL3 in aortic aneurysm formation, probably by regulating macrophages, which express abundantly CCL3. We next examined Ccl3 mRNA expression in the aorta of WT mice after CaCl 2 treatment. Ccl3 mRNA could be faintly detected in the untreated aortic tissue. CaCl 2 treatment significantly enhanced Ccl3 mRNA expression in the aorta at 1 and 2 weeks, decreasing thereafter (Fig. 1b). Likewise, intra-aortic CCL3 protein contents were transiently increased after CaCl 2 treatment (Fig. 1c). Doublecolor immunofluorescence analysis further detected CCL3 protein mainly in F4/80 + macrophages recruited into the adventitia of the aorta (Fig. 1d). Moreover, apoptotic macrophage numbers were increased simultaneously with a decline in CCL3 expression ( Supplementary Fig. 2). Thus, macrophages could produce CCL3 after migrating into CaCl 2 -induced aneurysm lesions but might lose a capacity to express CCL3 when they became apoptotic. To determine the roles of CCL3 in aneurysm formation, we treated simultaneously WT and Ccl3 −/− mice with CaCl 2 application into the aorta. Despite no significant differences in the aortic diameter between untreated WT and Ccl3 −/− mice, Ccl3 −/− mice displayed increased aortic diameter to a larger extent than WT mice 6 weeks after CaCl 2 application (Fig. 1e, f and Supplementary Table 1). Moreover, 7 out of 8 Ccl3 −/− mice exhibited a more than 35% increase in the aortic diameter. There was no difference in the histological structure of the aorta between untreated WT and Ccl3 −/− mice (Fig. 1g, HE). On the contrary, at 6 weeks after CaCl 2 application, Ccl3 −/− mice displayed disruption and fragmentation of medial elastic fibers, and lamellar of the aortic wall, to a larger extent, than WT mice (Fig. 1g, h). These observations implied that macrophage-derived CCL3 would be protective for CaCl 2 -induced AAA formation.
Enhanced inflammation and MMP-9 expression in Ccl3 −/− mice. Both in humans and rodents, the development of aneurysm is closely associated with the local inflammatory responses including inflammatory cell infiltration into the adventitia and medium 13 . Exaggerated AAA formation in Ccl3 −/− mice prompted us to investigate the inflammatory responses in AAA lesions because CCL3 is a typical inflammatory chemokine acting mainly on inflammatory cells, particularly monocytes/macrophages 27,28 . After the CaCl 2 application, the total number of circulating immune cells was increased, to a similar extent, in WT and Ccl3 −/− mice (Supplementary Fig. 3a, b), with no significant difference in each immune cell population between WT and Ccl3 −/− mice (Supplementary Fig. 3a, c-g). Treatment with CaCl 2 enhanced the recruitment of F4/80 + macrophages into the adventitia of the aorta in Ccl3 −/− mice, more markedly than in WT ones (Fig. 2a, b), but without any significant differences in CD3 + T cell, CD4 + T cell, CD8 + T cell, and B220 + B cell numbers in the adventitia of the aorta between WT and Ccl3 −/− mice (Fig. 2c, d, Supplementary  Fig. 4). We next examined intra-aortic gene expression of MMPs and TIMPs, which are presumed to be involved in aortic aneurysm formation 12,13,15,29,30 and observed that Ccl3 −/− mice exhibited significantly enhanced intra-aortic Mmp9 mRNA expression compared with WT mice (Fig. 2e). However, there were no significant differences in intra-aortic gene expression of other MMPs and TIMPs between both strains ( Supplementary Fig. 5). Consistently, CaCl 2 application increased intra-aortic MMP-9 activities to a larger extent in CCL3 −/− than WT mice (Fig. 2f). Furthermore, F4/80 + macrophages infiltrating into aortas expressed MMP-9 ( Fig. 2g). In addition, the administration of anti-CCL3 antibody significantly aggravated CaCl 2 -induced aneurysmal formation in WT mice, together with exaggerated pathological changes, macrophage infiltration, and MMP-9 expression and activity ( Fig. 3 and Supplementary Table 2). These observations would indicate that the lack of CCL3 could aggravate CaCl 2 -induced AAA formation by augmenting macrophage recruitment and macrophage-derived MMP-9 expression.

Involvement of CCL3 produced by bone marrow cells in AAA.
In order to exclude the contribution of non-hematopoietic cellderived CCL3 to CaCl 2 -induced aneurysm formation, we next applied CaCl 2 in bone marrow (BM) chimeric mice generated between WT and Ccl3 −/− mice. The mice transplanted with Ccl3 −/− mouse-derived BM cells exhibited exaggerated CaCl 2induced aneurysm formation together with enhanced macrophage infiltration and Mmp9 expression compared with those transplanted with WT mouse-derived BM cells in both WT and Ccl3 −/− mice ( Fig. 4 and Supplementary Table 3). Thus, BM cell-derived CCL3, probably macrophage-derived one, could be protective for CaCl 2 -induced AAA formation by suppressing macrophagederived MMP-9 expression. Representative results from 4 independent experiments are shown here. b-d Intra-aortic CCL3 expression in WT mice after CaCl 2 treatment. b Intra-aortic Ccl3 expression (n = 6 in each time point). **P < 0.01, vs. pretreatment. c Intra-aortic CCL3 protein contents (n = 6 in each time point). **P < 0.01, vs. pretreatment. d Representative images of CCL3 expression by F4/80 + macrophages in the aortas from WT mice (upper panels, scale bar = 50 µm; lower panels, scale bar = 20 µm). Representative results from four independent experiments are shown here. Blue coloration, nuclear staining by DAPI. e Representative macroscopic appearances of aortas in WT and Ccl3 −/− mice after CaCl 2 treatment (from six independent experiments). f Aortic diameters were measured in WT and Ccl3 −/− mice (n = 6 independent experiments). **P < 0.01, vs. pretreatment in each strain; ## P < 0.01, vs. WT-posttreatment.  Lack of CCR5 exaggerates AAA formation. CCL3 utilizes two distinct chemokine receptors, CCR1 and CCR5, as its specific receptors 23 . In order to evaluate the contribution of each receptor, the mice deficient in Ccr1 or Ccr5 were treated with CaCl 2 along with WT mice. No differences in the diameter of the aorta were observed among these mice when untreated ( Fig. 5a and Supplementary Table 4). CaCl 2 increased significantly the aortic diameters in each strain at 6 weeks after the treatment, and the increase, pathological changes, macrophage infiltration, and MMP-9 expression were augmented in Ccr5 −/− but not Ccr1 −/− mice, compared with WT ones (Fig. 5a-f and Supplementary  Table 4). Moreover, we examined in vivo effects of a specific CCR5 antagonist, maraviroc, on aneurysm formation in WT and Ccl3 −/− mice. Pharmacological blockage of CCR5 with maraviroc exaggerated CaCl 2 -induced aneurysm formation in WT mice similarly as observed in Ccr5 −/− mice (Fig. 5g). However, maraviroc treatment failed to have any effects on CaCl 2 -induced aneurysm formation in Ccl3 −/− mice (Fig. 5g). Moreover, CCR5 was expressed by macrophages in the aorta of CaCl 2 -treated WT mice (Fig. 5h). Thus, these observations implied that the interaction of CCL3 with CCR5 but not CCR1 can prevent CaCl 2induced AAA formation.
Protective effects of an MMP-9 inhibitor on AAA. In line with the previous studies [16][17][18] , we observed that the enhanced expression and activation of MMP-9 were closely associated with augmented aneurysm formation, indicating the crucial involvement of MMP-9 in aneurysm formation. Hence, we examined the in vivo effects of an MMP-9 inhibitor on aneurysm formation in WT, CCl3 −/− and Ccr5 −/− mice after CaCl 2 treatment. The administration of an MMP-9 inhibitor abrogated CaCl 2 -treatment-mediated enhancement in aortic diameters in CCl3 −/− and Ccr5 −/− mice as well as WT mice (Fig. 5i). These observations implied that MMP-9 would be a key molecule for the development of CaCl 2 -induced AAA formation.
Protective effects of CCL3 on AAA formation. In order to examine the preventive effects of CCL3, CCL3 was continuously administered to WT, Ccl3 −/− , or Ccr5 −/− mice, beginning immediately after the CaCl 2 application. CCL3 treatment abrogated AAA formation in WT and Supplementary Table 5). Moreover, attenuated aneurysm formation was accompanied by depressed macrophage infiltration and MMP-9 expression ( Fig. 7d-f, j-l and p-r). These observations would imply that CCL3 can dampen CaCl 2 -induced AAA formation by suppressing MMP-9 expression by macrophages. from Ccr5 −/− but not Ccr1 −/− mice (Fig. 8b). The potential capacity of p38 MAPK to inhibit LPS-induced Mmp9 expression 32 prompted us to evaluate MAPK signaling pathways. Indeed, CCL3 enhanced PMA-induced p38 MAPK phosphorylation but decreased ERK phosphorylation in macrophages ( Fig. 8c-f). Moreover, a specific p38 MAPK inhibitor, SB203580, but not an ERK inhibitor, PD98059, abrogated CCL3-mediated suppression of PMA-induced Mmp9 expression (Fig. 8g). Thus, the CCL3-CCR5 axis could depress Mmp9 gene expression by activating p38 MAPK in macrophages. When CaCl 2 -treated WT mice were administered with SB239063, aneurysm formation was significantly exaggerated, compared with vehicle treatment (Fig. 8h). Thus, the activation of p38 MAPK through the CCL3-CCR5 axis can be protective for CaCl 2 -induced aneurysm formation.    Roles of CCL3-CCR5 axis in angiotensin II-induced AAA. In order to validate the roles of the CCL3-CCR5 axis in aneurysm formation in general, we utilized another mouse aortic aneurysm model, where angiotensin II (Ang II) was continuously administered 22,33 . The continuous Ang II infusion elevated systolic blood pressure to similar extents in WT, Ccl3 −/− , Ccr1 −/− , and Ccr5 −/− mice (Fig. 9a). At 4 weeks after Ang II infusion, we measured the maximal external aortic diameter ex vivo. The aortic diameter was increased to a similar extent in WT and Ccr1 −/− mice (Fig. 9b, c), while Ccl3 −/− and Ccr5 −/− mice exhibited more exaggerated increases in the aortic diameter, compared with WT and Ccr1 −/− ones (Fig. 9b, c, Supplementary Fig. 7a). Moreover, the incidence of WT mice and that of Ccr1 −/− ones were 8.3% (1 out of 12 in WT mice) and 10% (1 out of 10 in Ccr1 −/− mice), respectively, and there was no significant difference between WT and Ccr1 −/− mice (Fig. 9d). On the contrary, the incidences of AAA development in both Ccl3 −/− and Ccr5 −/− mice were significantly higher, compared with WT ones (64.3%, 9 out of 14 in Ccl3 −/− mice; 66.7%, 8 out of 12 in Ccr5 −/− mice) (Fig. 9d). However, we never observed the aortic rupture in each mouse strain after Ang II infusion. Histopathological analyses revealed more extended destruction of aortic structure in Ccl3 −/− and Ccr5 −/− mice, than in WT and Ccr1 −/− ones (Fig. 9e). We further examined intra-aortic MMP-9 activities and observed that Ccl3 −/− and Ccr5 −/− mice, but not Ccr1 −/− mice, exhibited significantly enhanced intra-aortic MMP-9 activities, compared with WT mice (Fig. 9f). Furthermore, Ang II application increased the numbers of MMP-9 + macrophages which infiltrated into aortas, in Ccl3 −/− and Ccr5 −/− mice to a larger extent, compared with WT ones (Supplementary Fig. 7b, d). Altogether, the CCL3-CCR5 axis can have a protective role also in another AAA model, the Ang II-infused aneurysm formation model.

Discussion
Accumulating evidence implied the pathophysiological roles of the chemokine system in the development of AAA [24][25][26]  Simultaneously with MMP overexpression, abundant inflammatory cell infiltration was frequently observed in human AAA lesion 5,6 . These observations further suggest the association of matrix destruction with inflammatory cell infiltration but the results on animal studies are inconsistent in terms of a causal role of local inflammation in AAA pathogenesis 7,13 . Here, we revealed that CaCl 2 application induced abundant infiltration of macrophages, which expressed MMP-9, an MMP with a high capacity to degrade collagen. These observations would imply that macrophages could contribute to AAA development by producing MMP-9 37 .
It is still controversial the cellular source of MMP-9 in the course of AAA development. Macrophages could enhance MMP-9 production by human vascular smooth muscle cells 37,38 , by providing various pro-inflammatory cytokines, which were detected within AAA tissue 5,39 . This assumption may be supported by the observation that TNF-α and IL-1β expression was detected in the aorta of CaCl 2 -treated mice ( Supplementary Fig. 8a-d). The analysis using BM chimeric mice revealed that aneurysm formation and MMP-9 expression were enhanced by CCL3 deficiency restricted to BM-derived cells. Moreover, among various inflammatory cells, only intra-aortic macrophage numbers changed in parallel with MMP-9 expression in the aorta in WT and Ccl3 −/− mice treated with CaCl 2 . Furthermore, we revealed that infiltrating macrophages were the main cellular source of MMPs in the aortic wall of CaCl 2treated mice. These observations would imply an important contribution of macrophage-derived MMP-9 to CaCl 2 -induced AAA formation. This assumption may be supported by the previous study that MMP-9 was predominantly derived from macrophages present in AAA 40 .
It remains elusive how MMP-9 can contribute to AAA formation. Pyo et al. 12 revealed that MMP-9, particularly macrophage-derived one, destructed local matrix proteins and eventually induced aneurysm development in the elastase AAA infusion model. Moreover, Mmp9 −/− mice were resistant to CaCl 2 -induced aneurysm development, together with preserved lamellar morphology of aortic walls, compared with WT mice. Longo et al. 13 further proved that macrophage-derived MMP-9 cooperatively induced the development of experimental AAA in mice. We observed that CaCl 2 application enhanced intra-aortic MMP-9 expression. Thus, enhanced activities of MMP-9 can promote AAA formation upon the CaCl 2 application. Supporting this notion, an MMP-9 inhibitor significantly suppressed CaCl 2 -induced AAA formation in WT, Ccl3 −/− , and Ccr5 −/− mice.
We demonstrated that the CaCl 2 application augmented the expression of CCL3, hitherto considered as an inflammatory chemokine with a potent macrophage chemotactic activity, in the adventitia of aorta. However, to our surprise, Ccl3 −/− mice exhibited exaggerated CaCl 2 -induced AAA formation with augmented macrophage infiltration and MMP-9 expression in the aorta. Moreover, anti-CCL3 Abs induced similar phenotypes in WT mice, whereas CCL3 administration attenuated CaCl 2induced AAA formation. Furthermore, CCL3 reduced in vitro b The effects of CCL3 on PMA-induced Mmp9 expression in peritoneal macrophages obtained from WT, Ccr1 −/− , and Ccr5 −/− mice (n = 4 independent experiments). **P < 0.01, vs. no stimulation; ## P < 0.01, # P < 0.05, vs. PMA only. c Western blotting analyses on the expression of ERK, p-ERK, p38, p-p38, JNK, p-JNK, and GAPDH in WT macrophages stimulated by PMA (50 ng/ml) and/or CCL3 (10 3 ng/ml) for 24 h. Representative images from four independent experiments are shown. d-f Semi-quantitation was performed on the band intensities: d p-p38/p38 ratios, e p-JNK/JNK ratios, and f p-ERK/ ERK ratios (n = 4 independent experiments). **P < 0.01, *P < 0.05, vs. no stimulation. g The effects of a p38 inhibitor (SB203580) or an ERK inhibitor (PD98059) on Mmp9 expression in WT macrophages stimulated by PMA and/or CCL3 for 24 h (n = 4 independent experiments). **P < 0.01, vs. no stimulation. h The effects of a p38 inhibitor on CaCl 2 -induced AAA formation. Aortic diameters were measured in vehicle-treated and SB239063-treated mice before and 6 week after CaCl 2 application. (Pre: n = 5 each; 6 week-n = 5 each in-vehicle and). **P < 0.01, vs. pretreatment in each group; ## P < 0.01, SB239063 vs. vehicle. One-way ANOVA followed by Dunnett's post hoc test was used in (a), (b), and (d-g). Unpaired two-sided Student's t test was used in (h). Data are presented as mean values ± SEM.
PMA-induced MMP-9 expression in both human and mouse macrophages. These observations would indicate that CCL3 can dampen inflammatory responses in CaCl 2 -induced AAA formation, by suppressing MMP-9 expression in macrophages.
CCL3-induced depressed MMP-9 expression was observed only in the presence of CCR5 but not CCR1, among specific receptors for CCL3. Thus, CCL3-CCR5 interaction can have a negative impact on MMP-9 expression in mouse macrophages. Although it remains an open question on the CCL3-CCR5 axismediated signal pathways, we revealed that CCL3 could augment PMA-induced p38 MAPK but rather reduced PMA-induced ERK phosphorylation in macrophages. Given the fact that p38 MAPK activation could depress lipopolysaccharide-induced MMP-9 expression in astrocytes 32 , we further examined the effect of a p38 MAPK or an ERK inhibitor on CCL3-mediated attenuation of PMA-induced Mmp9 expression. Indeed, inhibition of p38 MAPK but not ERK pathway reversed CCL3-mediated inhibition of PMA-induced Mmp9 expression in macrophages. Since p38MAPK can counteract ERK1/2 activation 32 , CCL3 can dampen MMP-9 expression by activating p38 MAPK and eventually depressing the ERK pathway. This notion can be reinforced by the observation that the administration of a p38 inhibitor alleviated CaCl 2 -induced AAA.
The most puzzling observation in the present study is that intra-aortic macrophage infiltration was augmented in the absence or the inhibition of the CCL3-CCR5 axis, which can exhibit a potent in vitro monocyte/macrophage chemotactic activity [41][42][43] . Our previous study suggested that macrophages were recruited independently of CCR5-mediated signals 44 . However, the evidence is accumulating to indicate that leukocyte extravasation requires the degradation of matrix proteins present in the basement membrane in the vasculature by various proteinases, particularly MMPs. Indeed, the recruitment of leukocytes such as neutrophils and macrophages was impaired in Mmp9-deficient mice 45,46 . Thus, the absence of the CCL3-CCR5 pathway may enhance the expression of MMP-9, which can facilitate macrophage infiltration. Moreover, Ccl3 −/− mice exhibited enhanced intra-aortic gene expression of Ccl2, a potent macrophage-tropic chemokine, compared with WT mice (Supplementary Fig. 8e), but CCL3 failed to induce in vitro Ccl2 gene expression in macrophages (Supplementary Fig. 8f). Thus, the enhanced CCL2 expression in Ccl3 −/− mice can arise from an increase in recruited macrophages. Collectively, augmented MMP-9 activity in the absence of CCL3 increased macrophage recruitment, and CCL2 produced by the recruited macrophages further increased macrophage infiltration.
Several lines of evidence implicate a crucial involvement of several inflammatory mediators including TNF-α 47 and IL-1β 48 in the development of experimental AAA. Moreover, the expression of these mediators was augmented in human AAA walls 49 . Consistently, intra-aortic TNF-α and IL-1β expression was enhanced after CaCl 2 application and was detected mainly in infiltrating macrophages (Supplementary Fig. 8a-d). Moreover, the increments were further augmented in the absence of CCL3. Since the ERK pathway has an indispensable role in macrophage activation, particularly their expression of pro-inflammatory cytokines including TNF-α 50 , CCL3-mediated inhibition of the ERK pathway may account for exaggerated intra-aortic TNF-α expression in the absence of CCL3.
Our experimental study demonstrates that the CCL3-CCR5 pathway plays protective roles in AAA formation (Fig. 10). Moreover, a substantial proportion of Caucasians have a 32-base pair deletion allele, called Δ32, in human CCR5 gene 51 , and the deletion results in a frame-shift, with a premature stop codon. As a consequence, homozygous persons bearing CCR5 Δ32 mutation do not express functional CCR5 protein similarly as Ccr5 −/− mice do not. The association of CCR5 Δ32 mutation with some diseases was reported previously [52][53][54][55] . Moreover, CCR5 Δ32 mutation increased the incidence of AAA 56,57 , which was consistent with our observations in CaCl2-or Ang II-treated Ccr5 −/− mice. Thus, it is reasonable to examine the therapeutic activity of CCL3 for an aneurysm. Furthermore, given that CCL3 or its derivatives have been tried in humans to induce hematopoietic cell mobilization to peripheral blood, without severe adverse effects 58,59 , it may be a therapeutic strategy for an aneurysm.

Methods
Animals. Specific pathogen-free 8-week-old male C57BL/6 mice were obtained from Japan SLC and designated as WT mice in this study. Homozygous Ccl3 −/− mice were obtained from Jackson Laboratories (Bar Harbor, ME).   47 . Briefly, after deep anesthesia with an intraperitoneal injection of pentobarbital (50 mg/kg body weight), a 2-cm incision was made along the abdominal midline. The abdominal aorta between the renal arteries and bifurcation of the iliac arteries was exposed from the surrounding retroperitoneal structures, and 0.25 M CaCl 2 was applied to the external surface of the aorta. After 15 min, the aorta was rinsed with 0.9% sterile saline and the incision was closed, and mice were returned to their cages after recovery. Aortas were obtained to measure their diameters before and 6 weeks after CaCl 2 application and were processed to further analyses. The development of AAA was defined as a more than 35% increase relative to the mean aortic diameter of pretreatment mice in each group. In some experiments, WT mice were i.p. given goat anti-mouse CCL3 pAbs (100 μg/mouse, AB-450-NA, R&D) or normal goat IgG control (50 μg/mouse, AB-108-C, R&D) every week beginning at the day of aneurysm induction for 6 weeks until sacrifice. In another series of experiments, WT mice were i.p. given recombinant CCL3 (2 μg/200 μl/mouse, 450-MA-050, R&D) or PBS in an equal volume twice a week for 6 weeks beginning at the day of CaCl 2 application. At the indicated time intervals after the CaCl 2 application, mice were sacrificed to obtain aortas for subsequent analyses.
Ang II infusion-induced AAA model. Mice were deeply anesthetized and mini osmotic pumps (model 2004; Alzet, Cupertino, CA) were subcutaneously implanted in the neck region of anesthetized mice for 28 days in order to inject constantly Ang II (1 μg/kg/min, Sigma Aldrich) 33 . We employed the following condition as the criteria for Ang II-induced AAA formation: more than 30% increase in the external diameters of the suprarenal aorta, compared with the average of control mice 33 . After we observed the suprarenal abdominal aorta by ultrasound imaging, the aorta samples were obtained for the subsequent measurement of the external diameters of the suprarenal abdominal aortic diameters by an independent researcher ex vivo under a microscope. Since the average of the external diameters of the suprarenal aortic width was 765 µm in control mice, we judged that AAA developed when the external diameters were more than 995 µm.
Blood pressure measurement. Blood pressure was measured noninvasively on conscious mice using a CODA volume pressure recording tail-cuff system (Kent Scientific Corporation, Torrington, CT). The systolic blood pressure was measured at least five times, before and 4 weeks after Ang II pump implantation. The mean systolic blood pressure for each group was determined by averaging the systolic blood pressure of each mouse included in that group. The data from the 1-day measurement of each time point were used.
Transabdominal ultrasound imaging. For ultrasonic imaging, after removed abdominal hairs using hair removal cream, the animals were placed on a heated (41°C) imaging stage in the supine position while under anesthesia with ketamine-xylazine. Warmed ultrasound gel was applied to the abdominal surface and ultrasound prove applied to the gelled surface to collect images by the imaging system (APLIO 500 TUS-A500, Toshiba Medical System Corporation, Japan).
Generation of BM chimeric mice. The following BM chimeric mice were prepared 62 : male Ccl3 −/− BM→female WT mice, male WT BM→female Ccl3 −/− mice, male WT BM→female WT mice, and male Ccl3 −/− BM→female Ccl3 −/− mice. BM cells were collected from the femurs of donor mice by aspiration and flushing. Recipient mice were irradiated to 10 Gy using an RX-650 irradiator (Faxitron X-ray Inc., Wheeling, IL). Then, the animals received intravenously 5 × 10 6 donor-derived BM cells in a volume of 200 μl sterile PBS (−) under anesthesia. Thereafter, the mice were housed in sterilized micro isolator cages and were fed normal chow and autoclaved hyperchlorinated water for 60 days. To verify the successful engraftment and reconstitution of the BM in the transplanted mice, genomic DNA was extracted from peripheral blood and tail tissues of each chimeric mouse at 30 days after BM transplantation with a NucleoSpin tissue kit (Macherey-Nagel, Duren, Germany). Then, we performed polymerase chain reaction (PCR) to detect the Sry gene contained in the Y chromosome (forward primer, 5′-TTGCCTCAACAAAA-3′; reverse primer, 5′-AAACTGCTGCTTCTG CTGGT-3′). The amplified PCR products were fractionated on a 2% agarose gel and visualized by ethidium bromide staining. After durable BM engraftment was confirmed, mice were treated with CaCl 2 as described above.
Human macrophage assay. Human THP-1 cells were obtained from American Type Culture Collection (Rockville, MD). The cells were suspended in RPMI 1640 medium containing 10% FBS and were seeded in 35-mm culture dishes at a density of 0.5 × 10 6 cells/ml in the presence of 50 ng/ml PMA (P1585, Sigma-Aldrich) for 24 h at 37°C in a humidified atmosphere containing 5% CO to induce the differentiation into macrophages. The differentiated, adherent cells were washed with sterilized PBS (pH 7.4) and were cultured in fresh RPMI 1640 medium without PMA prior to further treatment. Macrophages were stimulated with 50 ng/ml of PMA (P1585, Sigma-Aldrich) in the presence or the absence of the indicated concentrations of human CCL3 (270-LD, R&D Systems) for the indicated time intervals. In some experiments, total RNAs were extracted for the determination of human MMP-9 mRNA expression by quantitative RT-PCR.
Enzyme-linked immunosorbent assay. Aortic tissues from WT mice were homogenized with PBS containing a complete protease inhibitor cocktail (Roche Diagnostics). Homogenates were centrifuged at 12,000×g for 15 min, 4°C. Supernatants were used to quantify CCL3 with a commercial ELISA kit (R&D Systems), according to the manufacturer's instructions. The detection limit was 1.5 pg/ml. The total protein in the supernatant was measured with a commercial kit (BCA protein assay kit, Pierce, Rockford, IL). Data were expressed as CCL3 (pg) per total protein (mg) for each sample.
In vivo administration of p38 MAPK inhibitor. In a separate CaCl 2 model experiment, mice were further injected intraperitoneally with 1 mg/kg of SB239063 (a p38 inhibitor, Tocris bioscience), dissolved in 3% dimethylsulfoxide (DMSO, resolved with PBS) or vehicle only, at multiple time points, 1 h before CaCl 2 application and thereafter daily for 6 week. At the indicated time intervals after the CaCl 2 application, mice were sacrificed to measure the diameter of the aorta.
In vivo administration of CCR5 antagonist. In a separate CaCl 2 model experiment, mice were intraperitoneally injected with 10 mg/kg of maraviroc (a CCR5 antagonist, MedChem Express, Monmouth Junction, NJ) dissolved in PBS or PBS, at multiple time points, 1 h before CaCl 2 application and thereafter daily for 3 week. At the indicated time intervals after the CaCl 2 application, mice were sacrificed to measure the diameter of the aorta.
MMP-9 inhibitor administration in vivo. In a separate CaCl 2 -model experiment, mice were intraperitoneally injected with 75 μg/mouse of an MMP-9 inhibitor (Enzo Life Science, Farmingdale, NY) dissolved in 3% DMSO and resolved with PBS or vehicle only, at multiple time points, 1 h before CaCl 2 application, and 4, 8, 12, and 16 day after CaCl 2 application. At the indicated time intervals after the CaCl 2 application, mice were sacrificed to measure the diameter of the aorta.