Activation of an Innate Immune Receptor, Nod1, Accelerates Atherogenesis in Apoe^−/− Mice

Atherosclerosis is a chronic inflammatory disease of vessel walls, characterized by the accumulation of leukocytes and their subsequent differentiation into cholesterol-laden foam cells (1). Innate and acquired immune systems are considered to be associated with the development of atherosclerosis (2–4). Concerning the innate immunity, epidemiological studies and animal experiments showed that infectious agents and their components contribute to the local chronic inflammatory process underlying atherosclerosis (4–7). The innate immune receptors recognize structurally conserved moieties and work as pattern recognition receptors (PRRs) such as TLRs, nucleotide oligomerization domain–like receptors (NLRs), retinoic acid–inducible gene-I-like receptors or C-type lectin receptors (8). Among PRR families, there is a line of evidence that TLRs, especially TLR4, TLR2, and their signaling molecule, MyD88 (7, 9, 10), play important roles in the development of atherosclerosis by initiating an inflammatory response to several pathogenic bacteria such as Chlamydophila pneumoniae or Porphyromonas gingivalis (5, 6).

With respect to NLRs, only a limited number of studies have shown that NLRs might be involved in atherosclerosis (11, 12). More than 20 NLRs have been identified in humans (8). Nucleotide-binding oligomerization domain (NOD)-1 is an intracellular sensor of bacterial peptidoglycan (PGN), and it is constitutively expressed in many types of cells, including endothelial cells (13, 14). NOD1 specifically recognizes a diaminopimelic acid–containing dipeptide, derived mostly from bacteria (13), and activates the signal pathway via the transcription factor NF-κB (14, 15). Recently, pure synthetic ligands for NLRs, such as FK156 and FK565 for NOD1 or MDP for NOD2, have been available for studies on the biological significance of these receptors. We have recently reported that oral administration of a synthetic Nod1 ligand FK565 efficiently induced acute vasculitis in mice (16). Therefore, it would make sense to investigate the contribution of a long-term exposure of a small dose of NOD1 ligand to the development of atherosclerosis, a chronic vascular inflammatory disease.

In this study, we show that oral administration of FK565 accelerated the development of atherosclerosis in apolipoprotein E knockout (Apoe^−/−) mice, and the effect was dependent on Nod1 in non–bone marrow-derived cells by transplantation experiments. By microarray analysis, we found that Ccl5 expression was significantly upregulated by FK565 administration in aortic roots during early stage of atherogenesis, and the treatment with Ccl5 antagonist significantly inhibited the acceleration of atherosclerosis in FK565-administered Apoe^−/− mice. Finally, we demonstrated that Nod1 deficiency resulted in reduced development of atherosclerotic lesions as well as their delayed progression in Apoe^−/− mice, indicating the contribution of Nod1 ligand to the development of atherosclerosis.

C57BL/6J Apoe^−/− mice were purchased from The Jackson Laboratory. Nod1^−/− mice in the C57BL/6 background were a gift from Tak Mak, University Health Network. Nod1^−/− mice were crossed with Apoe^−/− C57BL/6 mice. Heterozygous mice were intercrossed to generate homozygous Apoe^−/− mice bearing combinations of Nod1^+/+ and Nod1^−/− mice. The genotype for Apoe or Nod1 was confirmed using primers and conditions described in the The Jackson Laboratory Web site or the previous report (17), respectively. All mice were fed a normal chow diet and housed in a specific pathogen-free environment throughout the experiment. Drinking water and food for mice did not contain Nod1-stimulatory activity by a bioassay using HEK-Blue murine NOD1 cells (InvivoGen, San Diego, CA), which were HEK293 cells carrying a NF-κB reporter and transfected with a Nod1 construct. The study protocol was reviewed and approved by the Animal Care and Treatment Committee of Kyushu University.

For bone marrow transplantation studies, 9-wk-old male Apoe^−/−Nod1^−/− mice and Apoe^−/−Nod1^+/+ mice received 8 Gy total body irradiation to eliminate endogenous bone marrow stem cells and most of the bone marrow–derived cells, including macrophages. Bone marrow cells for transplantation into the irradiated mice were prepared by flushing both femurs of male Apoe^−/−Nod1^−/− or Apoe^−/−Nod1^+/+ mice. Donor cells were washed, suspended in sterile RPMI 1640 medium with 2% FCS, and concentrated to 1 × 10⁸ cells/ml. Three hours after irradiation, 1 × 10⁷ bone marrow cells were injected into the tail vein of a mouse. Successful reconstitution was confirmed by PCR genotyping of recipient mouse peripheral blood cells.

FK565, a synthetic and pure Nod1 ligand, was supplied by Astellas Pharma (Tokyo, Japan). At 5 wk of age, mice were randomized to FK565-administered and nonadministered groups. The administration protocol was as follows: two doses of FK565 solution (10 or 50 μg), reconstituted at 10 μg/μl in sterile distilled H₂O, was orally administered once a day for 2 consecutive days and then observed for the following 5 d per week. This course of the intermittent administration was continued for 4 wk, and mice were euthanized 6 d after the last administration. For the transplanted mice, FK565 was administered for four courses from 7 wk after bone marrow transplantation and euthanized at 20 wk of age. After euthanasia, serum was obtained from the vena cava after an overnight fast.

Because the Ccl5 antagonist Met-RANTES has been used in vivo in animal models of atherosclerosis, we determined the protocol of Met-RANTES treatment according to previous reports (18–20) with minor modifications. Met-RANTES (R&D Systems, Minneapolis, MN) was administered i.p. in a single dose of 50 μg diluted in PBS 30 min before FK565 administration. In parallel, control mice received a similar volume of sterile PBS.

Mice were treated with 1 g/l ampicillin (Wako Pure Chemicals, Osaka, Japan) dissolved in drinking water, as well as an antibiotic concoction consisting of 5 mg/ml vancomycin, 10 mg/ml neomycin, and 10 mg/ml metronidazole (all from Wako Pure Chemicals) by oral gavage every 24 h according to the method as described (21). Gavage volume of 20 ml/kg body weight was delivered with a polytetrafluoroethylene tube with prior sedation of the mice. The antibiotic administration was initiated at 5 wk of age and continued for 5 wk from 1 wk before 4 wk FK565 administration.

Intestinal depletion was assessed by collecting feces, homogenizing in 1 ml sterile PBS, and serially diluting and plating on trypticase soy agar with 5% sheep blood (BD Biosciences, Franklin Lakes, NJ) for 48 h at 37°C aerobically or in an anaerobic chamber (AnaeroPack system; Mitsubishi Gas Chemical, Tokyo, Japan). The number of bacteria per milligram of feces was calculated based on the CFU counted in each serial dilution.

After mice were euthanized, the hearts were removed rapidly after perfusion with PBS. The hearts were embedded in optimal cutting temperature compound (Sakura Finetek Japan, Tokyo, Japan) and quickly frozen in liquid nitrogen or 4% paraformaldehyde-fixed and paraffin-embedded for histological and immunohistological analyses. Sixty serial cross-sections (6 μm thick) of the aortic root were prepared from the site where the three aortic valves first appeared according to the method as described (22). The atherosclerotic lesions in the aortic root were investigated at six locations, each separated by 60 μm. Three serial frozen sections prepared from each location were conventionally stained with Sudan IV (Tokyo Kasei Kogyo, Tokyo, Japan), Elastica–van Gieson (EVG), and H&E stains. Similarly, seven serial paraffin-embedded sections prepared from each location were stained with H&E, EVG, MOMA-2 (macrophage)–specific Ab (1:100; AbD Serotec, Raleigh, NC), CD3 (T cell)–specific Ab (1:500; Abcam, Cambridge, U.K.), NIMP-R14 (neutrophil)–specific Ab (1:500; Abcam), α–smooth muscle actin (α-SMA)–specific Ab (1:100; Dako, Glostrup, Denmark), and Ccl5-specific Ab (1:10; R&D Systems). As a negative control, nonimmune IgG or isotype control was used. Detection and visualization of primary Ab binding was done as described (16). For quantitative estimation of the plaque contents, we analyzed absolute areas or numbers of cells staining positive for a respective marker within the plaque area by using Adobe Photoshop CS5 and National Institutes of Health ImageJ software. The average value for the six locations for each animal was used for analysis.

Immediately after mice were euthanized, the aorta was dissected from the proximal ascending aorta to the bifurcation of the iliac artery. The adventitial tissue was carefully removed, and then the aorta was opened longitudinally, fixed in 10% buffered formalin overnight, stained with Sudan IV, pinned on a black wax surface, and photographed for quantification of en face plaque areas. En face images were obtained by a stereomicroscope (SteREO Lumar V12; Carl Zeiss, Oberkochen, Germany) equipped with a digital camera (AxioCam MRc5; Carl Zeiss) and analyzed by Adobe Photoshop CS5 and National Institutes of Health ImageJ software. Lipid lesion formation was analyzed by the determination of the percentage area stained with Sudan IV to the total aortic area. Quantification of atherosclerotic lesions was performed by a single observer blinded to the experimental protocol.

To obtain sufficient amount of RNA for microarray analysis, the aortic roots removed from three animals were mixed, homogenized, and used as one sample. Two independent experiments were performed. Total RNA was extracted with an RNeasy fibrous tissue kit (Qiagen, Hilden, Germany) and amplified using an amino allyl MessageAmp II aRNA amplification kit (Ambion, Austin, TX). Double-stranded cDNA was synthesized from total RNA using oligo(dT) primer with a T7 RNA polymerase promoter site added to the 3′ end. Then, in vitro transcription was performed in the presence of amino allyl uridine-5′-triphosphate to produce multiple copies of amino allyl-labeled cRNA. Amino allyl–labeled cRNA was purified, reacted with N-hydroxy succinimide esters of Cy3 (GE Healthcare, Little Chalfont, U.K.) using NimbleGen’s protocol and hybridized for 19 h at 42°C to the Mouse Gene Expression 12 × 135K array (100718_MM9_EXP_HX12; Roche NimbleGen, Madison, WI) consisting of 44,170 genes. The arrays were scanned on GenePix 4000B (Molecular Devices, Sunnyvale, CA). The averages of triplicate spot intensities were extracted using NimbleScan v2.5 (Roche NimbleGen) and processed using robust multiarray analysis method (16). The scaled gene expression values were imported into GeneSpring 11.5.1 software (Agilent Technologies, Santa Clara, CA) for preprocessing and data analysis (16). The expression value of each gene was normalized to the 75th percentile shift expression of all genes. Probe sets were deleted from subsequent analysis when they were displayed an absolute value <30 in all experiments. The fold change was calculated as the ratio of the two group means based on the observed signal values. Microarray data were deposited in Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo/) under accession no. GSE48947.

Total RNA was extracted from the aortic root using RNeasy fibrous tissue, followed by cDNA synthesis using a high-capacity RNA-to-cDNA kit (Applied Biosystems, Foster City, CA). Mouse Ccl5, Cxcl16, Ccl8, and Gapdh expression levels were analyzed by TaqMan gene expression assays Mm01302427_m1, Mm00469712_m1, Mm01297183_m1, and Mn-99999915_g1 (Applied Biosystems), consisting of a 20× mix of unlabeled PCR primers and a TaqMan MGB probe (FAM dye-labeled), and TaqMan gene expression master mix (Applied Biosystems). Mouse Gapdh was used as internal controls. These TaqMan probes were labeled with the quencher TAMRA (emission I, 582 nm) at the 3′ end through a linker-arm nucleotide. The mRNA expression levels of the targeted genes were quantified by the StepOne real-time PCR system and the StepOne software v2.1 (Applied Biosystems). To calculate the relative expression level, the level of gene expression was divided by that of Gapdh. All experiments were carried out in duplicate and repeated three times for confirmation.

Total cholesterol and triglyceride concentrations in the murine sera were measured with a LabAssay cholesterol kit and a LabAssay triglyceride kit (Wako Pure Chemicals), respectively. Blood glucose levels were determined with a Medisafe-Mini glucose meter (Terumo, Tokyo, Japan). Serum insulin levels were measured with a mouse insulin ELISA kit (Mercodia, Uppsala, Sweden). The homeostasis model assessment of insulin resistance index ([glucose (mg/dl) × insulin (μU/ml)]/405) was calculated.

Data were analyzed by Student t test, Dunnett test, or Tukey–Kramer honestly significant difference test using JMP version 8.0 (SAS Institute, Cary, NC). Values of p < 0.05 were considered statistically significant.

To investigate the effect of a Nod1 ligand on atherosclerosis, we orally administered FK565, a pure synthetic Nod1 ligand, to Apoe^−/− male mice fed a chow diet. After preliminary experiments with orally administered FK565 at different doses consecutively or intermittently in Apoe^−/− mice (data not shown), we analyzed the effect of orally administered FK565 on 9-wk-old Apoe^−/− mice at 10 or 50 μg twice a week for 4 wk from 5 wk of age (Fig. 1A). No significant differences between the body weight and serum cholesterol and triglyceride levels of both groups of mice were observed (Supplemental Table I). Compared with the nonadministered group, a dose-dependent effect of FK565 was observed in the development of atherosclerosis in aortic roots with the lesion volumes of 1.85-fold (85%) increase at 10 μg and 5.49-fold (449%) increase at 50 μg FK565 in the FK565-administered group (Fig. 1B, 1D). In the FK565-administered group, a similar effect was also observed on the plaque formation in aortas of 1.10-fold (10%) increase at 10 μg and 33.9-fold (3289%) increase at 50 μg FK565 (Fig. 1C, 1E). To confirm the specificity of the FK565 treatment for Nod1, we studied Apoe^−/− Nod1^−/− mice in the same manner. There were no differences in the body weight and serum cholesterol or triglyceride levels in both groups of mice (Supplemental Table I). The plaque formation results showed no significant differences between FK565-administered Apoe^−/−Nod1^−/− mice and nonadministered mice (Fig. 1D, 1E). These results clearly demonstrated that oral administration of a pure synthetic Nod1 ligand accelerated the development of atherosclerosis in Apoe^−/− mice in a Nod1-dependent manner.

To rule out the possibility that oral administration of FK565 alters the composition of gut commensal bacteria and the dysbiosis induces acceleration of atherosclerosis, we depleted the intestinal microbiota by providing antibiotics in drinking water and with oral gavage of the antibiotic concoction according to the previous report (21). Even in gut flora–depleted mice, FK565 administration accelerated the development of atherosclerosis, suggesting that the acceleration of atherosclerosis by orally administered FK565 does not depend on alteration of the composition of gut commensal bacteria (Supplemental Fig. 1).

To explore the effects of Nod1 ligand on the formation of atherosclerotic lesions, we performed immunohistochemical studies of vessel-wall constituents within the plaques in aortic roots. FK565-administered Apoe^−/− mice showed a significantly higher level of macrophage and T cell infiltration within the plaques than did nonadministered Apoe^−/− mice. Neutrophil numbers within the plaques in FK565-administered Apoe^−/− mice also showed an increasing tendency than did those in nonadministered Apoe^−/− mice, although statistically not significant (Fig. 2A–D).

The previous report demonstrated that smooth muscle cells were intermingled with foam cells or tended to form a cap at the top of lesion in Apoe^−/− mice at 10–15 wk of age (23). A marked increase of smooth muscle cell content in the plaques was observed in the FK565-administered group (p = 0.026; Fig. 2A, 2E). These results suggested that Nod1 ligand accelerated the progression of advanced atherosclerotic lesions, with a remarkable increase of macrophage and T cell infiltration in the plaque.

As mentioned above, the effect of FK565 in atherogenesis was solely mediated by Nod1. Accordingly, to further test whether non–bone marrow-derived cells expressing Nod1 would contribute to the development of atherosclerosis by FK565 administration, we performed bone marrow transplantation experiments at 20 wk of age. All mice received FK565 for four courses. As expected, the atherosclerotic lesions in aortic roots of Apoe^−/− mice with Apoe^−/− bone marrow cells were significantly larger than those of Apoe^−/−Nod1^−/− mice transplanted with Apoe^−/−Nod1^−/− bone marrow (Fig. 3). Similarly, in chimeric Apoe^−/− mice with Apoe^−/−Nod1^−/− bone marrow cells, the atherosclerotic lesions were significantly larger than those of Apoe^−/−Nod1^−/− mice transplanted with either Apoe^−/− or Apoe^−/−Nod1^−/− bone marrow. Conversely, chimeric Apoe^−/−Nod1^−/− mice with Apoe^−/− bone marrow cells had significantly smaller atherosclerotic lesions than did Apoe^−/− mice transplanted with either Apoe^−/− or Apoe^−/−Nod1^−/− bone marrow (Fig. 3). The data for total aortic plaque formation showed the same trends as did the data for aortic root lesion area discussed above (data not shown). These results indicated that non–bone marrow-derived cells had a pivotal role in acceleration of atherosclerosis induced by Nod1.

To explore the mechanisms underlying the progression of atherosclerosis provoked by Nod1 ligand stimulation, we performed microarray gene expression profiling of the aortic roots of 9-wk-old Apoe^−/− mice with or without four courses of FK565 administration. Comparative analysis revealed that 1777 genes were upregulated in the aortic roots of FK565-administered mice compared with those of nonadministered mice (Fig. 4A). We also analyzed gene expression in 6-wk-old Apoe^−/− mice after only one course of FK565 administration, when there was weak accumulation of macrophages in the lesions, which was recognized as the initial stage of atherosclerosis (23) (data not shown). The analysis revealed that 1401 genes were upregulated by one course of FK565 administration, and 100 genes were consistently upregulated by FK565 administration in both time points, indicating that the 100 genes contributed to the initial and early stages of atherosclerosis induced by Nod1 agonist. A gene ontology analysis of the 100 genes showed that a number of biological process terms were associated with immune response (Supplemental Table II). Therefore, we focused on chemokine/cytokine genes. As expected, we identified a number of chemokine/cytokine genes known to be linked to atherosclerosis. The genes upregulated by FK565 in 6-wk-old Apoe^−/− mice included Ccl4, Ccl5, Ccl8, Cxcl16, and Il2rb, which are involved in recruitment and activation of macrophage and lymphocyte. In contrast, in 9-wk-old Apoe^−/− mice, genes associated with recruitment and activation of inflammatory cells, including chemokine/chemokine receptor genes (Cx3cl1, Ccl3, Ccl5, Ccl8, Cxcl1, Cxcl12, Cxcl16, Ccr2, Ccr5, Cxcr3, and Cxcr6) and IL genes (Il1b and Il6) were upregulated. In both time points, we observed that three genes were elevated >2-fold (Fig. 4B). To validate the microarray data, we compared expression levels by real-time RT-PCR of these three genes in Apoe^−/− mice with or without FK565 administration after both one and four courses and found that only one gene, Ccl5, also known as RANTES, was significantly upregulated by FK565 administration at both time points (Fig. 4C). Immunoreactivity for Ccl5 was localized in the intimal areas (Fig. 4D), and Nod1 ligands slightly increased Ccl5 levels in the lesions of Apoe^−/− mice, but without a statistically significant difference (Fig. 4E). To determine a role of Ccl5 in the atherosclerosis progression exerted by Nod1 activation, we investigated whether injection of a Ccl5 antagonist, Met-RANTES, could inhibit the acceleration of atherosclerosis in FK565-administered Apoe^−/− mice. As expected, Met-RANTES–treated mice showed a significant decrease of atherosclerotic lesion area in aortic root, compared with PBS-treated mice (Fig. 4F), and they had a similar volume of plaques in aortic roots of Apoe^−/− mice without FK565 administration. These results suggested that Ccl5 plays a crucial role in the acceleration of atherosclerosis induced by FK565.

To determine which cell type frequencies, macrophages or T cells, were altered after Met-RANTES administration, we performed immunohistochemical studies in aortic roots of FK565-administered mice with or without Met-RANTES administration. Both MOMA-2⁺ macrophages and CD3⁺ T cells in atherosclerotic lesions significantly decreased after Met-RANTES administration, suggesting that Ccl5 contributes to accumulation of both cell types in the atherosclerotic lesions (Supplemental Fig. 2A, 2B).

To determine whether Nod1 signaling pathway contributes to the early step of atherosclerosis, we studied Apoe^−/− and Apoe^−/−Nod1^−/−male mice at the ages of 9, 12, 20, and 40 wk. Apoe^−/−Nod1^−/− mice showed modest elevations of serum cholesterol and triglyceride levels compared with those in Apoe^−/− mice, but there were no significant differences (Supplemental Table I). Additionally, no difference was observed in the fasting blood glucose levels, serum insulin levels, or homeostasis model assessment of insulin resistance values between Apoe^−/− and Apoe^−/−Nod1^−/−mice at 20 wk of age (data not shown). Nonetheless, Apoe^−/−Nod1^−/− mice showed 43, 63, and 26% decreases of atherosclerotic lesion areas in aortic root, compared with those in their age-matched Apoe^−/− mice, at the ages of 12, 20, and 40 wk, respectively (Fig. 5A, 5C). By analysis of aortic lesion areas, Apoe^−/− and Apoe^−/−Nod1^−/− mice had plaques from 20 wk of age, and they had similar distribution patterns of plaques, with the highest density occurring in the lesser curvature of the aortic arch at 40 wk of age (Fig. 5B). However, plaque formation areas in aortas were significantly reduced (54 and 25%) by Nod1 deficiency at the age of 20 and 40 wk, respectively (Fig. 5B, 5D). These results demonstrated that Nod1 deficiency not only attenuated early atherogenesis but also decelerated the progression of atherosclerosis.

To determine the cellular mechanism for the decelerated atherogenesis by Nod1 deficiency, we analyzed the cell composition of atherosclerotic lesions of Apoe^−/− mice and Apoe^−/−Nod1^−/− mice by immunohistochemical studies. The areas for MOMA-2⁺ macrophages or α-SMA⁺ smooth muscle cells as well as the numbers of CD3⁺ T cells or NIMP-R14⁺ neutrophils in atherosclerotic lesions of Apoe^−/−Nod1^−/− mice were smaller than those in Apoe^−/− mice, but the differences were not statistically significant (Supplemental Fig. 2C–F).

In microarray and RT-PCR analyses, Nod1 ligands increased Ccl5 mRNA levels in aortic root of Apoe^−/− mice compared with nonadministered mice (Fig. 4B, 4C). Ccl5 was predominantly upregulated in early plaques (24) and is considered to be important in early lesion formation. Therefore, we examined Ccl5 gene expression levels in aortic root of Apoe^−/− and Apoe^−/−Nod1^−/− mice in early stages of atherosclerosis at 6, 9, 12, and 20 wk of age. At 6–12 wk of age, Ccl5 mRNA levels in aortic roots of Apoe^−/−Nod1^−/− mice were lower than those in their age-matched Apoe^−/− mice, and the difference was statistically significant at 9 wk of age (Fig. 5E). These results suggested that Ccl5 plays crucial roles in the development of atherosclerosis during the early phase induced by Nod1 ligand.

The present study has demonstrated that long-term oral administration of a pure synthetic Nod1 ligand accelerated the development and progression of atherosclerosis in Apoe^−/− mice, independent of cholesterol or triglyceride levels in blood. Additionally, the complete loss of this receptor significantly decreased the size of atherosclerotic lesions, providing evidence of a solid relationship between NOD1 and atherosclerosis. Previously, only limited studies have shown the effect of NOD1 in the cardiovascular field (25, 26) or the involvement of NOD1 in vascular inflammation induced by C. pneumoniae (12). No prior study, however, has directly proven the association between NOD1 and atherosclerosis in vivo.

By transplantation experiments, the major effector cells of the Nod1 ligand were nonhematopoietic cells. This is consistent with the observation that NOD1 in nonhematopoietic cells plays a key role in activation of human endothelial cells, mediated by C. pneumoniae (12), the chronic or recurrent infection of which was reported to have a close relationship to the development of atherosclerosis (27). Alternatively, Levin et al. (28) showed that increased lipid intake in receptor-interacting protein 2 (Rip2)^−/− macrophages resulted in increased atherosclerotic lesions in apolipoprotein B^−/−low-density lipoprotein receptor^−/− mice transplanted with Rip2^−/− bone marrow. Myeloid-specific depletion of Nod1 showed no significant difference in plaque formation in our study, and myeloid-specific depletion of Nod2, but not Rip2, showed significant reduction in the lipid-rich necrotic area in low-density lipoprotein receptor^−/− mice (29). Therefore, it is possible that Rip2^−/− macrophages exert their effects mainly through the alteration of lipid metabolism rather than Nod1 or Nod2 signaling pathways.

The marked expression of Ccl5 mRNA gene in the Nod1-induced atherosclerotic lesions and the inhibitory effect of Met-RANTES on Nod1-induced acceleration of atherosclerosis demonstrated that Ccl5 expression might be pathophysiologically important for the development and/or advancement of atherosclerosis. Consistent with the previous report that Ccl5 expression was most marked in early plaque (24), the effect of Nod1 for plaque formation appeared to be dominant at early stage (<12 wk) because there was a marked downregulation of Ccl5 expression at that stage in Apoe^−/−Nod1^−/− mice. Ccl5 is a proinflammatory chemokine that regulates the trafficking of leukocytes such as macrophages and Th1 T cells, mediated by activation of the receptors Ccr1, Ccr3, Ccr4, and Ccr5 (30). In atherosclerosis, Ccr5 is known to be involved, not only in the recruitment of mononuclear cell, but also in the modulation of the immune balance (18, 31–33). Similarly, in FK565-administered Apoe^−/− mice, Met-RANTES administration attenuated the infiltration of MOMA-2⁺ macrophages and CD3⁺ T cells in atherosclerotic lesions, suggesting that accumulation of both cell types by Ccl5 plays a role in Nod1-induced acceleration of atherosclerosis.

It has been reported that chronic or repeated infection was a strong risk factor for the development and aggravation of atherosclerosis (27, 34–36). However, it was difficult to show direct evidence of existing microorganisms in atherosclerotic plaques. Several studies showed the bacterial PGNs or DNA fragments existing in atherosclerotic plaques (37–39), and the metagenomic analysis of gut microbiome in patients with symptomatic atherosclerotic plaques and healthy controls demonstrated that patient metagenomes were enriched in genes encoding PGN biosynthesis (40). Furthermore, a recent report demonstrated that PGNs from microbiota were detected in systemically circulating blood, and they modulated the innate immune system through Nod1 (41). In the present study, we have demonstrated that Nod1 ligand directly contributed to the development of atherosclerosis. Taken together, systemically circulating PGN fragments from endogenous microbiota may activate Nod1-expressing vascular cells to produce various chemokines, including Ccl5, to promote the accumulation of inflammatory cells in the plaques and eventually accelerate the atherosclerotic formation in vessels. Although endogenous Nod1 ligands are still unknown, it is possible that yet-undetermined endogenous Nod1 ligands contribute to the development of atherosclerosis, because there are several studies that endogenous ligands toward other PRRs such as Tlr2, Tlr4, and Nlrp3 were involved in the acceleration of atherosclerosis (11, 42–45).

A recent study showed that gut flora influenced the nutrient processing and the metabolism in mice, and microbial processing of dietary choline was significantly correlated with atherosclerosis (46). Although it is still controversial whether Nod1 deficiency contributes to alteration of the composition of microflora (47, 48), alteration of the microbiota by Nod1 deficiency might play a role in the decreased development of atherosclerosis in Apoe^−/−Nod1^−/− mice. Because oral administration of FK565 might influence microbiota in the long-term observation, further investigation is needed on a role of microbiota in the acceleration of atherosclerosis through Nod1 activation.

In the previous studies, endogenous and synthetic ligands for Tlr2 and Tlr4 influenced the formation of atherosclerosis in addition to Nod1 (7, 9, 49). We found that Nod1 deficiency significantly reduced plaque formation by 50% in the aortas of 20-wk-old Apoe^−/− mice, indicating that the contribution of Nod1 on atherogenesis is as strong as those of Tlr2 and Tlr4 suggested by the previous study (49). Interestingly, the effect of Nod1 for plaque formation seemed to be dominant at early stage (<12 wk) as shown by marked downregulation of Ccl5 expression in Nod1^−/− mice. Hence, it is possible that the ligand-specific effects of innate immune-mediated atherosclerosis formation might exist, and various types of ligands from microbiota would independently or synergistically work together for the development of atherosclerosis.

In conclusion, Nod1 activation plays one of the key roles in the development and progression of atherosclerosis. Further comprehensive studies on the innate immune ligand-specific contribution are necessary to understand the pathogenesis of atherosclerosis.

We are thankful to H. Fujii and C. Arimatsu for technical assistance, J. Kishimoto for statistical analyses, and Y. Nakashima for planning advice.

The authors have no financial conflicts of interest.