MCP1/CCR2 Axis Promotes Vein Graft Neointimal Hyperplasia through Its Signaling in Graft Extrinsic Cell Populations

Abstract

Objective

To evaluate direct versus indirect MCP-1/CCR2 signaling and identify the cellular producers and effectors for MCP-1 during neointimal hyperplasia (NIH) development in vein grafts (VG).

Methods and Results

Genomic analysis revealed an over-representation of 13 inflammatory pathways in WT VGs compared to CCR2KO VGs. Further investigation with various VG-host combinations of MCP-1 and CCR2 deficient mice were used to modify the genotype of cells both inside (graft intrinsic group) and outside of the vein wall (graft extrinsic group). CCR2 deficiency inhibited NIH only when present in cells extrinsic to the graft wall, MCP-1 deficiency required its effectiveness in cells both intrinsic and extrinsic to the graft wall to suppress NIH. Deletion of either MCP-1 or CCR2 was equally effective in inhibiting NIH. CCR2 deficiency in the predominant neointimal cell population had no impact on NIH. Direct MCP-1 stimulation of primary neointimal SMCs had minimal influence on cell proliferation and matrix turnover, confirming an indirect mechanism of action.

Conclusions

MCP-1/CCR2 axis accelerates NIH via its signaling in graft extrinsic cells, particularly circulating inflammatory cells, with cells both intrinsic and extrinsic to the graft wall being critical MCP-1 producers. These findings underscore the importance of systemic treatment for anti MCP-1/CCR2 therapies.

Neointimal hyperplasia is the primary culprit that causes early vein graft failure. A large body of evidence has demonstrated circulating inflammatory cells, and in particular monocytes, as the critical drivers for this pathology¹. In a freshly created vein graft, monocytes follow the chemokine gradient established by the local inflammatory cascades, homing to the injured vessel wall. While various mediators may be involved in this process and marked functional redundancy may exist among them, monocyte chemoattractant protein (MCP) -1 has been shown to be critical for monocyte trafficking². Underscoring this concept, several groups, including our laboratory, have observed that blocking MCP-1/CCR2 signaling reduces monocyte content and attenuates neointimal growth in vein grafts^3–6 and injured arteries^{7, 8}.

MCP-1/CCR2 signaling has traditionally been considered to function exclusively in the regulation of monocyte biology². However, this long standing paradigm has been challenged by recent studies showing that smooth muscle cells are able to respond to MCP-1 stimulation with enhanced migration and upregulation of the pro-inflammatory genes^{5, 9–11}. While these studies suggest a novel mechanism that vascular cells may serve as direct MCP-1 effectors, translating these observations to the complex events such as vein graft failure has raised a critical issue concerning the precise cellular elements that drive NIH. This issue is further complicated by the dynamic alteration in the neointimal cell population, where the composition varies temporally due to the constant in and out flux of cells from local (e.g. the graft wall) and remote (e.g. bone marrow, circulation, and perivascular tissue) locations^{12, 13}.

MCP-1 is a secreted and soluble protein⁷. Once produced, it functions in an autocrine as well as paracrine manner. Although the location of MCP-1 production is not essential to understanding the general tissue response, it is critical for targeted anti MCP-1 therapy. During NIH development, the composition and the relative proportion of various cell types are in constant flux. Targeted therapy requires both identification of the MCP-1 producers and the temporal patterns during which these cells are most active. For example, monocytes secrete a large amount of MCP-1 in the vein graft wall^{7, 14}, yet are recruited primarily in early remodeling phases, with few cells present following resolution of the acute inflammation. Vascular cells such as SMCs and ECs then take over the task to maintain a persistent MCP-1 production in the neointimal lesion^{3, 5, 14}. This temporal switch in the predominant MCP-1 producer has raised uncertainties as to the most critical cell type that propagates the hyperplastic response. Clarification of this issue will not only provide mechanistic insights into the MCP-1/CCR2 signaling biology in the vein graft wall, but also direct how MPC-1 may be targeted to inhibit NIH.

A well-known phenomenon for chemokine signaling is the promiscuous ligand-receptor recognition. In addition to MCP-1, other MCPs (such as MCP- 2 to 5 and 16^{15, 16}) can also activate CCR2 at physiologic concentrations. This ligand redundancy has questioned the direct linkage between MCP-1 and CCR2, and whether MCP-1 production and/or CCR2 receptor expression is the dominant driver for neointimal thickening.

Built upon the fundamentals outlined above, the current study sought to resolve these issues and elucidate the mechanisms whereby the MCP-1/CCR2 axis regulates NIH development in vein grafts via: 1) identifying the cellular producers and effectors for MCP-1, 2) determining the relative importance of promiscuous CCR2 binding to MCPs other than MCP-1, and 3) evaluating the in vivo significance of direct MCP-1/CCR2 signaling in vascular cells on growth of the neointima.

Methods

Animal Model of the Vein Graft

This study conforms to the American Physiological Society’s Guiding Principles for the Care and Use of Vertebrate Animals, and the Guide for the Care and Use of Laboratory Animals (National Research Council, Revised 2010). Murine vein grafts were created by placing an inferior vena cava (IVC) from a donor animal to the common carotid artery of a host animal. This procedure provides the opportunity to vary the genotype of cells inside the vein (graft intrinsic cell group) as well as those from locations outside of the vein wall (graft extrinsic cell group). The external branch of the common carotid artery was ligated to reduce blood flow through the vein graft. Details of this technique have been described in our previous reports^{3, 17}. All mice used in this study were adult (9–11 week old) males. MCP-1 null (MCP-1^−/−), CCR2 null (CCR2^−/−), CAG-EGFP (a ubiquitous EGFP reporter and abbreviated as EGFP⁺ in the rest of the text), and the wild type (WT) control (C57BL/6) mice were purchased from Jackson Laboratory. CCR2^−/−;EGFP⁺ mice were produced in our institution by crossing CCR2^−/− to EGFP⁺ strains. The genotype of the implanted vein graft and recipient host will be presented in the form of Vein→host in this report. Grafting a WT IVC to an EGFP host, for instance, will be noted as WT→EGFP⁺.

Microarray Gene Expression

The global gene expression in vein grafts with (n=5) or without (n=5) CCR2, was profiled using Agilent mouse genomic array chip, with age-matched WT non-implanted IVCs (n=5) serving as a baseline reference. Samples were collected seven days after implantation, with this time point chosen based on our previous investigation of the temporal pattern of gene expression and identification of the dominant response to injury (unpublished data).

Total RNA was extracted from vein grafts and IVCs with RNeasy® Mini columns (Qiagen, Valencia, CA) and cleaned with DNase I. Following reverse-transcription of RNA to cDNA, cRNA was generated with 20 µg cDNA template and labeled with Cy3 and Cy5 for reference (pooled IVCs) and experimental samples (vein grafts), respectively. Equal amount of labeled cRNA (100 ng) derived from each vein graft and the reference was mixed, fragmented, and hybridized to microarray chips (Agilent) at 60°C for 17 hours. After acquiring the raw data by scanning the arrays with an Agilent G2505 B Scanner, corrections for both background and variations in signal intensity were performed. The processed data was further analyzed with BRB Array Tools (http://linus.nci.nih.gov/BRB-ArrayTools.html) and FatiScan¹⁸. These programs apply completely different principles to their analyses, with the BRB Array Tools taking each gene as an independent variable and FatiScan considering genes functioning on a modular fashion¹⁸. For identification of genes differentially expressed between CCR2^−/− vein grafts and the WT controls, a random variance model with a nominal significance level of 0.001 was used. For FatiScan analyses, genes passed the initial raw data process was sorted with their statistics (two tailed Fisher’s exact test). KEGG pathway analysis was then performed with a partition of 30 and an adjusted p value of 0.05 (the Benjamini and Hochberg false discovery rate¹⁹).

Selective Deletion of MCP-1 or CCR2 in Graft Intrinsic or Extrinsic Cell Group

Vein grafts with and without MCP-1/CCR2 signaling in either graft intrinsic or extrinsic cells (n=9 to 16 per group) were created by varying the genotype of the venous conduits and host animals. The impact of MCP-1 or CCR2 signaling on neointimal thickening was evaluated separately, with the concept that deficiency in MCP-1 identifies the cellular producers while deficiency of CCR2 identifies cellular effectors for MCP-1. Comparison between the two evaluates the relative in vivo significance of the MCP-1/CCR2 axis and the promiscuous binding of CCR2 to other MCPs. Four weeks after graft implantation, surgical samples were perfusion fixed in 10% buffered formalin, paraffin embedded, histologically sectioned, and processed with Masson for the assessment of neointimal thickness²⁰. As what we have shown previously with rodent vein graft models^{3, 17}, the tunica media of the vein is only two to three cell layers thick and becomes indistinguishable from the neointimal tissue four weeks after graft implantation. The residual media might have been included in our measurements of the neointimal thickness.

Cell Tracing

An assumption made in this experiment is that if the MCP-1/CCR2 signaling in neointimal cells contributes significantly to neointimal growth, genetic interruption of the MCP-1/CCR2 signaling in the cell group that dominates neointimal cell re-population would attenuate neointimal thickening. To test whether it can be the graft intrinsic or extrinsic cell group, we selectively traced the destination of cells from each group with EGFP reporter. Vein grafts with the restriction of EGFP to graft intrinsic (EGFP⁺→EGFP⁻), extrinsic (EGFP⁻→EGFP⁺), or both (EGFP⁺→EGFP⁺) cell groups (n=6 for each combination) were created with WT and EGFP mice. The same experiment was repeated under conditions with and without CCR2 deficiency to examine whether the primary source of the neointimal cell population is influenced by selective CCR2 deletion. All vein grafts were perfusion-fixed with 2% paraformaldehyde + 0.2% glutaraldehyde in PBS (pH 7.4) 4wks after placement. Frozen sections were counterstained with propidium iodide (PI) and EGFP positive cells evaluated with confocal microscopy and differential interference contrast (DIC) imaging.

Immunohistochemistry Assay

Endothelium, leukocytes, and macrophages were identified by their expression of CD31 (abcam), CD45 (R&D), and F4/80 (Serotec) and visualized with enzymatic substrate (DAB) or fluorescent conjugates (Texas-Red). Re-endothelialization by host derived cells was evaluated on cross sections. The extent of EGFP positive endothelial cells populating the luminal surface was measured and expressed as percent of the circumference.

Culture of the Primary Medial and Neointimal Smooth Muscle Cells (mSMCs, neoSMCs)

Medial SMCs were harvested from the outgrowth of aortic explants using methods described by others²¹. Neointimal tissue was selectively isolated using microsurgical dissection techniques, as we previously described²⁰. Both medial and neointimal explants were cultured in DMEM plus 10% FBS. Cell outgrowths were passaged in 2–3 weeks and the purity of SMCs evaluated using a flow cytometry-based assay for α-actin. Cultures with more than 95% α-actin positive cells were expanded by 5 – 10 passages for in vitro assays.

MCP-1 Stimulation

Proliferation of the neoSMCs and primary mSMCs was evaluated with BrdU incorporation and cell cycle progression in triplicate. Cells were seeded at a density of 10,000 cells/cm² and cultured in DMEM with 10% FBS. Upon reaching 90% of confluence, cells were serum deprived for 24 hours, followed by treatment with serum-free media, TGF-β1 (1ng/ml, R&D), MCP-1 (0.1, 1.0, 10, and 100ng/ml, GenWay), TGF-β1 + MCP-1 (10ng/ml), or serum for 24 hours. For BrdU incorporation assay, BrdU (10µM) was added to the medium 2 hours prior to the assay, and the amount of incorporated BrdU was quantified with a Cell Proliferation Assay kit (Roche Applied Science). For cell cycle assay, treated cells were washed with PBS, fixed with 70% ethanol, and incubated in propedium iodide (20µg/ml), Triton X-100 (0.1%), and RNase A (0.2mg/ml). DNA content of the individual cells was then quantified with flow cytometry.

Real Time RT-PCR

In attempt to evaluate the impact of direct MCP-1/CCR2 signaling on matrix synthesis and degradation, we measured the expression of type I collagen (COL1A2), connective tissue growth factor (CTGF), MMP- 2, and MMP-9 in neoSMCs following MCP-1 treatment. Quantitative RT-PCR was performed following the protocol described previously²⁰. The primer and probe sets applied to the assays are purchased from Applied Biosystems with the code of Mm00483937-m1, Mm01192933-g1, Mm00439498-m1, and Mm00442991-m1 for COL1A2, CTGF, MMP-2, and MMP-9 respectively.

NanoString Assay

The nCounter GX Mouse Inflammation Kit was purchased from Nanostring Technologies (http://www.nanostring.com/). Each assay includes 179 inflammation related genes (Table S1). WT, CCR2^−/−, and MPC1^−/− vein grafts (n=4 for each group) were created as described above. After evaluation of the quality (Bioanalyzer, Agilent) and concentration (Nanodrop, Agilent), 100ng of total RNA was hybridized onto the capture and reporter probes. Following removal of the unbound probes, the tripartite molecules were captured by fluidics devices and imaged with nCounter Digital Analyzer²². Raw Data was processed with nSolver analysis software v1.0 and further analyzed with BRB Tools.

Statistical Analyses

All data are expressed as mean ± SEM. Comparisons were done using ANOVA, Tukey’s post-hoc analysis, and unpaired t-test, as appropriate. P < 0.05 was considered significant

Results

Inflammatory Pathways Dominate the Global Gene Expression in Vein Grafts with and without Intact MCP-1/CCR2 Signaling

A week after graft implantation, 218 genes were differentially expressed, with the vast majority down-regulated in CCR2^−/− vein grafts compared to WT controls (Figure 1A). Functional annotation clustering analysis (DAVID) at the BP-3 level demonstrated enrichment of three major clusters (Table S2) and one KEGG pathway cluster (Table S3), with a high percentage of inflammatory response pathways characterizing the difference between WT and CCR2^−/− vein grafts (Figure S1). This finding was further confirmed with FatiScan analyses (Figure 1B). Fifteen KEGG pathways were differentially expressed between WT and CCR2^−/− vein grafts, with the majority (13 KEGG pathways) of them over-represented in WT vein grafts (red bars). Functional annotation demonstrates that all 13 pathways are well-recognized inflammatory pathways associated with innate and adaptive immune responses.

Figure 1.

Removal of CCR2 attenuates early (one week) inflammatory responses in vein grafts. Heat map (A) details 218 genes that were differentially expressed in WT and CCR2KO vein grafts (41,534 total, unsupervised clustering analysis with an FDR=0.001). Color-codes are on a bidirectional two-fold scale, with red representing upregulation and green depicting downregulation. Of the 15 KEGG pathways enriched by FatiScan analysis (B), 13 pathways are over-represented in WT vein grafts (red bars) and all associated with inflammatory responses.

MCP-1/CCR2 Signaling in Graft Extrinsic Cell Group Drives Neointimal Thickening

Using gene deletion, we have demonstrated that early reduction in monocyte infiltration correlates with a significant attenuation of the following neointimal thickening in CCR2 deficient vein grafts³. While this study along with the microarray data presented above suggest that the MCP-1/CCR2 axis regulates NIH via inflammatory mechanisms driven by monocytes (indirect effects), results from other groups show that MCP-1 can regulate SMC biology by initiating the CCR2 signaling directly in SMCs. Such direct effects may also play an important role in neointimal thickening^{5, 9, 10}. To address this issue, cells intrinsic or extrinsic to the vein graft were independently evaluated for the influence of MCP-1 and CCR2 on NIH development.

Robust NIH developed in WT vein grafts (Figure 2A, WT→WT) but was attenuated in the absence of MCP-1 or CCR2 (Figure 2A, KO→KO). Selective deletion of MCP-1 identified both graft intrinsic and extrinsic cell groups as significant MCP-1 producers in the vein graft wall. While removal of MCP-1 from both cell groups significantly inhibited NIH (p=0.026), leaving its production in either group boosted the hyperplastic response to the level same as preservation of MCP-1 production in both cell groups (Figure 2B), indicating cells either intrinsic or extrinsic to the graft wall are able to produce sufficient MCP-1 to accelerate NIH.

Figure 2.

Selective deletion of MCP-1 or CCR2 in cell group intrinsic or extrinsic to the graft wall demonstrates that MCP-1 drives NIH via both autocrine and paracrine effects on graft extrinsic cell group. Each combination is denoted in a form of Vein→Host. An overview of neointimal thickening at 4wk time point in WT→WT and KO→KO vein grafts is demonstrated in panel A (Masson’s staining), with the white dash lines highlighting the neointimal adventitial border. Quantitative results are summarized in charts (B) and (C), where symbols * and # indicate significant difference between the indexed group and the WT→WT controls while symbol @ denotes that the difference between vein grafts with complete loss of MCP-1 and CCR2 is insignificant (@p=0.42, unpaired t test).

Selective deletion of the MCP-1’s receptor CCR2 revealed that MCP-1 producers may not necessarily be the same cell group as MCP-1 effectors. As evidenced by the neointimal thickening in WT→WT and KO→WT groups in Figure 2C, vein grafts developed robust NIH as long as the CCR2 was preserved intact in graft extrinsic cell group,; otherwise, NIH was significantly inhibited, as reflected by NIH formed in WT→KO and KO→KO groups. A simultaneous removal of CCR2 from graft intrinsic cell group had no additive effect on the hyperplastic response regardless the presence or absence of CCR2 in graft extrinsic group (Figure 2C). These results demonstrate that cells extrinsic to the graft wall serve as the essential effectors for MCP-1/CCR2 signaling during NIH development. Cells intrinsic to the graft wall, though proved to be deleterious MCP-1 producers (Figure 2B), do not contribute, if any, to NIH via MCP-1 activation.

The Promiscuous Recognition of Multiple MCPs by CCR2 Receptor Is Not a Significant Determinant to the Eventual Neointimal Thickening

A well-known phenomenon for transgenic mice is that the resultant defects are not limited to the function of the ablated gene. The CCR2^−/− mice, for instance, suffer attenuated Th1 responses in addition to the diminished monocyte trafficking^{23, 24}. Using the Nanostring assay, we characterized the expression profile of the inflammatory genes in vein grafts with and without MCP-1 or CCR2. Unsupervised clustering analysis (Figure 3A) clearly defined separate expression patterns in WT, MCP-1^−/−, and CCR2^−/− vein grafts. Notable is the pronounced difference of the CCR2^−/− vein grafts from their MCP-1^−/− and WT counterparts. Comparison between CCR2^−/− than MCP-1^−/− groups at a significance level of 0.01 identified that 27 genes were differentially expressed, with the majority of them up-regulated following CCR2 deletion (Figure 3B). GO analysis assigned these genes to two clusters at the BP-4, with annotations in the primary cluster demonstrating significant differences in complement activation (GO:0006956, GO:0006958), leukocyte mediated immunity (GO:0002443, GO:0002449), and regulation of immune response (GO:0050778, GO:0002684) (Figure 3C).

Figure 3.

Vein grafts defer early (day 3) gene signatures following genetic ablation of MCP-1 and CCR2. With an emphasis on inflammatory genes (total 179), heat map A demonstrates the disparity of the gene expression profiles between WT, MCP-1KO, and CCR2KO vein grafts (unsupervised clustering analysis). Genes (total 27) characterizing the differentially (P<0.001) expression profile for MCP-1KO and CCR2KO vein grafts are shown in heat map (B). Functional annotation at biological process level 4 indicates that these genes are specialized in innate immune responses (C). Heat maps are color-coded on a two-fold scale, with green or light blue indicating the lowest level and red or dark blue being the highest level.

Despite these differences in inflammatory response and the potential that other CC chemokines (e.g. MCP- 2 to 5) may bind to CCR2 and play a compensatory role , neointimal thickness was similar following either MCP-1 or CCR2 deletion(Figure 2B and 2C, p=0.42). Since CCR2 is the only functional receptor established for MCP-1²⁵, this data suggests that CCR2 primarily binds to MCP-1 to proceed NIH development and its promiscuous binding to other MCPs plays a modest role in this process.

MCP-1/CCR2 Signaling in Vascular Intrinsic Cells Holds Limited in vivo Significance in Neointimal Thickening

Previous reports detailing the response of SMCs to MCP-1 have been conflicting^{5, 10, 26, 27}. While our data would support that the direct CCR2 binding in cells intrinsic to the graft wall has limited influence on driving the hyperplastic process, this conclusion may be confounded by an extensive repopulation of the vein wall by cells recruited from locations outside the conduit. Using the EGFP model to trace cell lineage, the contribution of cells from the vein wall to the developing neointimal was evaluated. As shown in Figure 4, all neointimal cells show a strong EGFP signal in vein grafts created with EGFP veins and EGFP hosts (EGFP⁺→EGFP⁺; Figure 4A). In sharp contrast, only a few EGFP⁺ cells were detected in the neointimal region in vein grafts created by implanting wild type veins to EGFP hosts (EGFP⁻→EGFP⁺; Figure 4B). Switching the donor-host relationship reversed the composition of the neointimal cell population, with the EGFP⁺ cells again being the predominant group (EGFP⁺→EGFP⁻; Figure 4C). Similar patterns were also observed in the absence of the CCR2 receptor (Figure 4D–F). Selective removal of CCR2 exhibited no significant effects on the pattern, where cells intrinsic to the vein graft (e.g. EGFP⁻ cells) remain dominated the developing neointima (Figure 6A–B). Taken in context of the graft morphology data (Figure 2C), these observations indicate that direct MCP-1/CCR2 signaling in vein graft derived SMCs holds a limited importance in vivo.

Figure 4.

Cells intrinsic to the venous conduit dominate neointimal cell repopulation under conditions with and without intact CCR2 signaling. Green represents EGFP signal while red depicts nuclei (PI counterstain). The EGFP genotype of vein grafts is noted on top of each column while the CCR2 genotype of vein grafts is given on the left of each row, with an arrow to designate the vein-host relationship. White dash lines indicate the border between neointimal and adventitial layers. Symbols of “+” and “-” indicate presence and absence of the labeled gene, respectively.

Figure 6.

Loss of MCP-1/CCR2 signaling in cells extrinsic to the graft wall does not influence their contribution to endothelial repair. To simplify the nomenclature, vein grafts created by implanting WT veins (e.g. CCR2^+/+/EGFP^0/0) to EGFP hosts with or without CCR2 are labeled as CCR2^+/+ and CCR2^−/− respectively, in top panel (A), (B), and (C). Accumulation of host-origin cells (green) is demonstrated in panel (A) and (B), where doted white lines indicate the neointimal adventitial boarder. Coverage of the luminal surface by host-origin endothelial cells (ECs) is summarized in panel (C), where NS stands for “not significant”. Representative images demonstrating host-origin ECs (white arrows) in a CCR2^−/− vein graft are provided in panel (D), (E), and (F).

MCP-1 is Neither Mitogenic nor Stimulatory to NeoSMCs in vitro

To further define the pathways whereby MCP-1/CCR2 signaling drives neointimal thickening, we evaluated the effects of MCP-1 stimulation on neoSMC proliferation and gene expression. The neoSMCs treated with escalating dose of MCP-1 demonstated no difference in BrdU incorporation compared with non-treated controls (Figure 5A). Further analysis of the mitogenic response using cell cycle progression analysis revealed that MCP-1 induced few (1%) of the treated cells to undergo mitosis, while serum stimulated 22% of the cells to enter into S phase (Figure 5B). A direct impact of MCP-1 on matrix metabolism was also not detected in neoSMC cultures. While TGF-β1 stimulated significant changes in CTGF and MMP-9 expression, administration of MCP-1 did not alter the expression of CTGF, COL1A2, MMP-2, and MMP-9 in either direction (Figure 4C). Bioactivity of the MCP-1 applied to this experiment was verified with a membrane fluidity assay (data not shown).

Figure 5.

Direct MCP-1 stimulation lacks biologic effects on neoSMCs. Mitogenic effects of the MCP-1 on primary neoSMCs are summarized in panel (A) and (B), while its effects on the expression of genes known to regulate matrix metabolism in primary neoSMCs are shown in panel (C). Differences among groups treated with escalating dose of MCP-1 including vehicle controls are not statistically significant for all tested parameters (One-way ANOVA).

CCR2 Is Not Required for Re-endothelialization of the Denuded Luminal Surface

Under fluorescent microscopy, a combination of both EGFP⁺ and EGFP^- cells lining the lumen surface were observed in both WT→CCR2^+/+/EGFP⁺ and WT→CCR2^−/−/EGFP⁺ vein grafts (Figure 6A–B). Immunohistochemistry assay for CD31, a lineage marker specific for endothelial cells, confirmed this monolayer of cells to be of endothelial origin (Figure 6E). On the same section that had undergone immunohistochemistry staining, the EGFP signal remained well preserved (Figure 6D). Confocal imaging demonstrated a co-localization of the EGFP and CD31 signal on the luminal surface (Figure 6F), indicating a contribution of cells of host origin to endothelial repair. Quantitatively, cells of host origin covered nearly 40% of the luminal surface and this capacity was independent of the status of MCP-1/CCR2 signaling, as evidenced by the insignificant difference in the coverage attributed to host cells with or without CCR2 (Figure 6C). The accumulation of cells with host origin in both the neointimal and adventitial regions varied widely in WT→CCR2^+/+/EGFP⁺ and WT→CCR2^−/−/EGFP⁺ vein grafts and further evaluation on these regions was not performed.

Discussion

The critical role of the MCP-1/CCR2 signaling in NIH development in vein grafts has been clearly demonstrated in previous studies from our group and others^3–5. In the current study, we have provided mechanistic insights into how MCP-1/CCR2 axis regulates NIH. Our results demonstrate that: 1) an inflammatory gene signature discriminates vein grafts with and without MCP-1/CCR2 signaling; 2) the graft extrinsic cell group serves as the primary effectors for MCP-1 while both graft intrinsic and extrinsic cell groups are critical MCP-1 producers; 3) removal of MCP-1 or the CCR2 receptor is equally efficient in inhibiting NIH; and 4) abolishing the direct MCP-1/CCR2 signaling in neoSMC does not impact neointimal growth. These findings suggest that MCP-1 signals through CCR2 of the graft extrinsic cell group, most likely circulating monocytes, to accelerate NIH. The direct MCP-1/CCR2 signaling in vascular cells and the promiscuous binding of CCR2 to other MCPs hold minimal in vivo significance in MCP-1/CCR2 signaling regulated NIH development in vein grafts.

We have previously demonstrated that early reduction in monocyte recruitment correlates with attenuated NIH development in CCR2 deficient vein grafts³. Consistent with this report, the current study further identified cells extrinsic to the venous conduit as the primary effectors for MCP-1. Although graft extrinsic cells are a mixed cell population that may include leukocytes and vascular cell precursors (e.g. circulating progenitors and peri-vascular fibroblasts)²⁸, a major population in this group is leukocytes. Among the leukocyte population, both monocytes and neutrophils have been demonstrated as critical drivers for NIH²⁹. However, previous investigations have shown that neutrophils are generally unresponsive to MCP-1 signaling due to the absence of the CCR2 receptor in this subpopulation^{25, 30}. These observations together with results from the current study have led us to conclude that monocytes are the main effector for MCP-1/CCR2 signaling regulated NIH development in vein grafts.

Monocytes in the circulating blood comprise a heterogeneous population. Recent studies have identified two distinct subsets, termed inflammatory and patrolling and defined as Gr1⁺CCR2⁺CX3CR1^low and Gr1⁻CCR2⁻CX3CR1^high, respectively³¹. Once penetrated into tissue, these cells may be further polarized to M1 and M2 macrophages, with the M1 specializing at inflammation and M2 functioning to maintain tissue homeostasis³². In addition to monocytes, other inflammatory cell lineages, such as T-cells and dentritic cells may also participate in NIH process³³. While the current study suggests monocytes as the cellular effector for MCP-1/CCR2 signaling in vein grafts, identification of the exact subset that drives NIH remains an open question pending complete characterization and quantification of all inflammatory infiltrates in the wall. As reported by our group³ and others⁷, inflammatory cells take different mechanisms to adhere and migrate in vessels, in part dependent on the presence of an intact endothelial monolayer. Although phenotypic characterization of the inflammatory infiltrates has been successful in the studies for atherosclerosis³⁴, application of similar strategy to assay cells in the vein graft wall remains challenging. Specific difficulties include the preservation of lineage markers and collection of cells in a spatial-specific fashion (e.g. luminal surface vs. adventitia).

Several groups have reported that SMCs express CCR2 and MCP-1 can regulate SMC biology, such as proliferation^{5, 9} and production of inflammatory cytokines¹¹, through the direct binding of MCP-1 to its CCR2, as opposed to the resultant effects secondary to monocyte activation. Using an in vivo approach, the current study demonstrates that removal of the CCR2 receptor from the predominant neointimal cell population has no significant impact on neointimal thickening, suggesting that the in vivo significance of the direct MCP-1 simulation of vascular SMCs is of limited importance. Schepers et al⁵ has reported that CCR2 antagonist inhibits NIH development in human veins cultured in a system without monocytes, which differs from our observations. While the pre-existing pathology in the “normal” human saphenous vein³⁵ may be a contributor, further studies are required to identify the differences between these ex vivo and in vivo models.

Consistent with the observations in our murine vein graft model, our in vitro experiments demonstrate that neoSMCs are refractory to MCP-1 stimulation. No significant changes in proliferation or the expression of CTGF, collagen, and MMP-9 were observed following treatment of neoSMC with various doses of MCP-1. Similar responses were also observed for medial SMCs (data not shown). While our results from both in vitro and in vivo experiments support the concept that the direct MCP-1 regulation of SMC biology is not essential to the hyperplastic vein graft response, mixed results and potential controversies exist in the field. Some groups have documented stimulatory effects of MCP-1 on SMC proliferation and gene expression^{5, 9, 10}, while others have reported that MCP-1 treatment leads to negligible or even slightly inhibitory effects on these processes^{26, 27}. Inconsistency among laboratories is not uncommon and maybe magnified by non-physiologic SMC cell culture parameters such as the seeding density, quiescence, and phenotype (e.g. contractile vs dedifferentiated) , all of which can impact the response of SMCs to the imposed stimulant^{36, 37}. Although not comprehensive, the in vivo evidence provided in this report provides important clarification of these conflicting results.

Chemokine networks possess a great deal of functional redundancy that is supported by the promiscuous ligand/receptor recognition and overlap in intracellular signaling pathways that are inherent in the system. For example, MCP-1 binds to receptor CCR2 to regulate monocyte trafficking, but the same CCR2 signaling pathway may be activated by other MCPs such as MCP- 2 to 5^{2, 38}. In addition to MCP-1, several other members such as MCPs, IP-10, and Fractakine¹⁶ can guide monocytes migrating to inflammatory lesions. . It seems that none of the individual members would be essential to monocyte trafficking. However, it has been demonstrated that each of these chemokine axis can induce a robust and non-overlap in vivo output³⁹. For example, mice deficient in MCP-1 or CCR2 demonstrate a dramatic reduction in monocyte content in atherosclerotic lesion^{40, 41}. This functional robustness of each chemokine axis has posed a critical question as to the signaling axis that dominates NIH development in vein grafts. Our results demonstrate that deletion of MCP-1 or CCR2 is equally efficient in inhibiting NIH. Since CCR2 is the only receptor identified for the MCP-1²⁵, the equivalent MCP-1/CCR2 effects on NIH suggest that MCP-1 acts as the primary ligand for CCR2 to drive neointimal growth and the impact of CCR2 binding to other MCPs on this process is marginal.

Vein bypass procedures offer an opportunity to treat the conduit prior to implantation, thus minimizing the off-target effects that would otherwise be caused by systemic interventions. Exploring this opportunity, a wide-range of preclinical studies has focused on ex vivo approaches for delivery of therapeutic agents directly into the wall^42–44. Unfortunately, expansion of these strategies into clinical trials has been unsuccessful^{45, 46}. In the current study, we have identified the graft extrinsic cell group as the primary effector for MCP-1/CCR2 signaling. This finding supports the concept that these effectors need be targeted prior to their arrival at the graft wall and points toward the need for systemic delivery of anti MCP-1/CCR2 treatments. In addition, we show that cell groups either intrinsic or extrinsic to the graft wall can produce sufficient MCP-1 to accelerate NIH, indicating both cell groups should be treated when targeting MCP-1. Taken together, systemic delivery of anti MCP-1/CCR2 is an essential component in the design of effective therapies to reduce intimal hyperplasia and improve vein graft survival.

NIH, though may lead to luminal narrowing or complete occlusion, is also part of the physiologic adaptation that repairs the injured venous conduit. Recent observations suggest the homing of vascular progenitors to the injured vessel is a critical component for structural stabilization and re-endothelialization of the wall. Studies in the field have identified several key molecules that may guide stem cell homing. Among them is MCP-1 that has been shown to mediate the recruitment of hematopoietic stem cells to repair the injured liver⁴⁷ and heart⁴⁸. Using EGFP cell tracing approach, we examined the importance of MCP-1 in repairing endothelial denudation in vein grafts. The results show that cells, independent of CCR2 expression, were recruited to the luminal surface and contributed to endothelial repair, suggesting that MCP-1/CCR2 signaling is not pivotal for re-endothelialization in our model system.

Genetic tools have provided powerful approaches to isolate specific elements for detailed investigation; however, inherent flaws in these models have called for caution in interpretation of the experimental results. The CCR2^−/− strain, for instance, displays a skew from Th1 to Th2 response²³, with decreased production of IFN-γ and monocyte/macrophage phagocytosis²⁴. A reduced capacity of producing MCP- 3 and 5 has also been documented for MCP-1^−/− strain⁴⁹. In addition to these various compensatory changes, a recent study has suggested the existence of an un-identified receptor for MCP-1¹⁰, which further questioned the linear MCP-1/CCR2 recognition and the phenotypic similarity of the MCP-1^−/−, and CCR2^−/− vein grafts. In the current study, we profiled the expression of inflammatory genes in WT, MCP-1^−/−, and CCR2^−/− vein grafts. Unsupervised cluster analysis detected significant differences among these groups, with the majority of the differentially expressed genes up-regulated in CCR2^−/− vein grafts. Although limited to inflammatory gene set, this “tubular view” clearly demonstrates that the phenotype of the MCP-1^−/− vein graft is different from that of the CCR2^−/− vein graft. While attributed partially to the promiscuous ligand/receptor recognition, this disparity may also be the result of the defects beyond the function of the target gene in these strains. Therefore, the cause-effect relationship established with these transgenic strains for the pathogenesis of diseases needs to be verified with other approaches such as pharmaceutical intervention and siRNA gene silencing.