Local Delivery Is Critical for Monocyte Chemotactic Protein-1 Mediated Site-Specific Murine Aneurysm Healing

Background Local delivery of monocyte chemotactic protein-1 (MCP-1/CCL2) via our drug-eluting coil has been shown to promote intrasaccular aneurysm healing via an inflammatory pathway. Objective In this study, we validate the importance of local MCP-1 in murine aneurysm healing. Whether systemic, rather than local, delivery of MCP-1 can direct site-specific aneurysm healing has significant translational implications. If systemic MCP-1 is effective, then MCP-1 could be administered as a pill rather than by endovascular procedure. Furthermore, we confirm that MCP-1 is the primary effector in our MCP-1 eluting coil-mediated murine aneurysm healing model. Methods We compare aneurysm healing with repeated intraperitoneal MCP-1 versus vehicle injection, in animals with control poly(lactic-co-glycolic) acid (PLGA)-coated coils. We demonstrate elimination of the MCP-1-associated tissue-healing response by knockout of MCP-1 or CCR2 (MCP-1 receptor) and by selectively inhibiting MCP-1 or CCR2. Using immunofluorescent probing, we explore the cell populations found in healed aneurysm tissue following each intervention. Results Systemically administered MCP-1 with PLGA coil control does not produce comparable aneurysm healing, as seen with MCP-1 eluting coils. MCP-1-directed aneurysm healing is eliminated by selective inhibition of MCP-1 or CCR2 and in MCP-1-deficient or CCR2-deficient mice. No difference was detected in M2 macrophage and myofibroblast/smooth muscle cell staining with systemic MCP-1 versus vehicle in aneurysm wall, but a significant increase in these cell types was observed with MCP-1 eluting coil implant and attenuated by MCP-1/CCR2 blockade or deficiency. Conclusion We show that systemic MCP-1 concurrent with PLGA-coated platinum coil implant is not sufficient to produce site-specific aneurysm healing. MCP-1 is a critical, not merely complementary, actor in the aneurysm healing pathway.

iNtroDUctioN Intracranial aneurysms are estimated to occur in 1 in every 20-50 individuals, with about 50% mortality within 30 days if the aneurysm ruptures (1). Endovascular intrasaccular coiling is used widely in the management of unruptured and ruptured intracranial aneurysms. However, recanalization remains a sig nificant concern (2). While current endovascular treatments aim to prevent unruptured aneurysms from rupturing, or a ruptured aneurysm from rebleeding, they still carry inherent recurrence risk.
Complete and durable endovascular coil embolization requires complete wound healing to occur within the aneurysm sac. With the best possible results with platinum coil embolization, there is about 20-30% coil packing density of the intrasaccular space (3). The remaining 70-80% of intrasaccular space is filled with thrombus. As thrombus dissipates, there is "a race against time" for wound healing to occur. If wound healing is completed before the thrombus completely dissipates, then there is likely chance for durable cure. If complete wound healing is not achieved, any space between the coil and intraluminal tissue risk aneurysm recurrence.
In this study, we ask whether systemic MCP1, as opposed to local delivery, can produce biologically significant aneurysm tissuehealing. If systemic MCP1 can direct sitespecific aneurysm healing, then MCP1 could be administered via a systemic route rather than coated on a device that requires an endovascular procedure. Previous studies on MCP1 in other models have shown systemic MCP1 can direct sitespecific neutrophil infiltration, mesenchymal stem cell recruitment, and inflammatory and nociceptive mediators in various organs, such as in lung, heart, kidney, and postsurgical wound healing (10)(11)(12)(13).
We have previously shown local delivery of MCP1 to the aneurysm promotes inflammatory tissue ingrowth composed of macrophages and vSMCs (14). While we show intrasaccular MCP1 delivery promotes aneurysm healing, we need to validate this finding to determine that MCP1 is the critical component in the pathway. Otherwise, the observed aneurysm tissue healing may be due to another yet unidentified aspect of the MCP1 polymercoil construct in our experimental model. Or perhaps, MCP1 may have a complementary, not critical role in aneurysm healing.
In contrast to surgically implanted local MCP1eluting coil, we measure the ability of systemic injection of MCP1 to direct sitespecific tissuehealing within the aneurysm. In addition, we validate that MCP1 is in fact a critical cytokine in the aneurysm healing cascade by evaluating tissuehealing response with knockout (KO) or blockade of either MCP1 or its receptor CCR2.

MateriaLs aND MethoDs
All animal experiments were performed in accordance with approved protocol #201604771 from the University of Florida Institutional Animal Care and Use Committee and comply with Animal Research: Reporting of In Vivo Experiments guidelines.
Detailed materials and methods are included in Supplemental Material.

resULts effect of systemic administration of Mcp-1 on aneurysm healing
We initially performed a dose response trial of systemic intraperi toneal MCP1 at doses of 0.1, 1.0, and 10 μg/dose. Animal health and survival rates did not differ by group as a result of systemic MCP1 treatment, and no difference in ingrowth was detected between dose response groups (data not shown). Thus, 100 µL of 100 µg/mL MCP1 in PBS was administered every other day over 3 weeks, the same concentration used in solution to create our previously assayed coated coils (14). To verify that the 100 µg/mL dose achieves a systemic therapeutic level, we measured systemic soluble levels of MCP1 in mice that received systemic MCP1 versus control PBS. Serum MCP1 level 6 h postMCP1 injection is >4 μg/mL versus vehicle <60 pg/mL (p < 0.01, n = 5 per group, data not shown). In a separate cohort, we then compared aneurysm tissue ingrowth in mice implanted with poly (lacticcoglycolic) acid (PLGA) coils and 100 µg/mL systemic MCP1 versus PBS. Tissue ingrowth with systemic MCP1 was 5 versus 16% with PBS vehicle (p = 0.0144, n = 6 and 7, respectively; Figures 1A,B).

Cell-Specific Populations
PLGA coil with systemic MCP1 versus vehicle.
ingrowth compared with PLGA control, as shown in our previous study (9) and characteristics of aneurysms per group do not differ significantly ( Figures S1A,B in Supplementary Material) despite differences in luminal ingrowth ( Figure S1C in Supplementary Material). MCP1 eluting coils were implanted in our murine saccular aneurysm model in MCP1 KO or CCR2 KO mice to determine the role of MCP1 or CCR2 depletion on aneurysm tissue healing. There was significantly decreased tissue ingrowth in MCP1 KO compared with control, expressed as percent crosssectional area of aneurysm sac: aneurysm tissue ingrowth in MCP1 KO was 11% versus WT control 56% (p < 0.0001, n = 8 and 10, respectively; Figure 1C). Furthermore, there was sig nificantly decreased tissue ingrowth in CCR2 KO mice compared with control: aneurysm tissue ingrowth in CCR2 KO was 4.6% versus WT control 56% (p < 0.0001, n = 6 and 10, respectively; Figure 1D).

effect of Mcp-1 antibody or ccr2 antagonist on aneurysm healing
Monocyte chemotactic protein1 or CCR2 were selectively inhib ited to determine their effect on aneurysm tissue healing. Tissue ingrowth significantly decreased with MCP1 blockade compared with control: tissue ingrowth in animals with MCP1 eluting coil and antiMCP1 was 0.8% versus control 56% (p < 0.0001, n = 4 and 10, respectively; Figure 1E). Furthermore, there was significantly decreased ingrowth with CCR2 blockade compared with control: ingrowth in animals with MCP1 eluting coil and CCR2 antagonist is 9.2% versus vehicle 56% (p < 0.0001, n = 5 and 10, respectively; Figure 1F). Representative H&E stained images of aneurysm ingrowth are depicted ( Figure 1G).

DiscUssioN
Taking advantage of commonly recruited immune cell line ages in aneurysms has been suggested as a means to provoke inflammatory tissuehealing (15). Wound healing and vascular repair are marked by inflammatory, degradative and prolifera tive stages, with eventual M2like polarization of macrophages to support tissue healing (16). The locally activated tissue upregulates cytokines which perpetuate chronic remodeling until the intraluminal aneurysm sac stabilizes (17). These stages of intracranial aneurysm healing after coiling may be similar to peripheral wound healing, which progresses through hemo stasis, inflammation, proliferation of granulation tissue, and resolution (18).
Monocyte chemotactic protein1 has been investigated in its chemotactic and proinflammatory properties, whereas study ing its role in vessel healing has been limited to other vascular environments, such as regression of atherosclerotic plaques and systemically inflamed endothelial cellmonocyte adhesion (19,20). Previous studies using MCP1 KO mice in diabetic wound healing have shown reduced macrophage recruitment into the local environment (21). We have previously shown that recruit ment of this macrophage axis is instrumental in our MCP1 mediated aneurysm healing model (22). We have also previously shown that MCP1 contributes to aneurysm healing by means of its downstream mediators (23). Presently, we emphasize the importance of local activity provided by MCP1 eluting coils.
Using our elastaseinduced carotid aneurysm model, we used PLGA coil control with systemic injection of MCP1 compared with PBS vehicle to allow for a experimentally analogous control to MCP1 locally eluting coil. Interestingly, the degree of tis sue ingrowth in animals implanted with PLGA coil following systemic IP injection of MCP1 attenuated PLGA coilinduced tissuehealing response in the local aneurysm environment com pared with PBS vehicle (Figure 1A). This suggests that, in our hands, a chemotactic effect to the locally inflamed environment is critical for murine aneurysm tissue ingrowth, as administra tion of systemic MCP1 may detract from any endogenous local chemotactic gradient to the aneurysm site. Tissue ingrowth is also attenuated by systemic deficiency of both MCP1 and CCR2 and delivery of systemic MCP1 neutralizing antibody and CCR2 antagonist, further supporting the mechanistic role of MCP1 mediated aneurysm healing.
We crossvalidate that MCP1 is critical in the aneurysm tissue healing pathway. By inhibiting MCP1 or its receptor CCR2, or by genetic KO of MCP1 or CCR2, we eliminate the aneurysm tissuehealing response. Both MCP1 and receptor CCR2 KO mice exhibit decreased tissue ingrowth into the aneurysm lumen (Figures 1C,D). Deficiency of MCP1 and CCR2 exhibited sig nificant decrease in ingrowth from control, which may be attrib utable to systemic inhibition of macrophage migration from their bone marrow or circulating monocyte source (9,24). However, it has been shown that for KO animals, similar or redundant path ways can often be upregulated to compensate from the expected deficient phenotype (24). Therefore, we further crossvalidated MCP1's critical role in aneurysm healing by studying the effect of selective inhibitors. One caveat is that CCR2 receptor antago nism would affect CCL7, CCL8, CCL11, and CCL13 in addition to baseline MCP1/CCL2 signaling (25,26). An additional limitation of our study is the sole use of female animals, as our previous studies and preliminary data exclusively were in females for purposes of bone marrow transplanted chimeras. However, it is possible that gender may have an effect on healing responses and this is a direction for future study in this model. Although MCP1 inhibitor group was minimally underpowered, the mean difference from control was greater than expected. Application of a systemic MCP1 neutralizing antibody and CCR2 antagonist also exhibits decreased tissue ingrowth with respect to control (Figures 1E,F).
The induction of inflammatory cell migration is also encour aged by and dependent on local elution of MCP1 from our coated platinum coils. Specifically, restorative M2like macrophages and vSMC or myofibroblast differentiation (both visualized with αsmooth muscle actin) are activated in aneurysms treated with MCP1 eluting coils, whereas this profile is not achieved in sys temic MCP1 administered cohorts (Figures 2B,C). Meanwhile, the cellular phenotype of ingrowth induced by systemic MCP1 injection does not differ from that of PBS vehicle when using PLGAcoated platinum coils. This enhancement is attenuated in MCP1/CCR2 KO and blockade groups compared with local MCP1 administration (Figure 2). M2like macrophages and myofibroblasts are found to promote aneurysm healing by intrasaccular fibrosis, thereby ameliorating aneurysm rupture (17,27).
A persistent inflammatory phase of aneurysm wound healing is associated with increased extracellular matrix deposition and progression to fibrotic proliferation (2). Chronic inflammation has also been described as the prodrome of a fibrotic response, as seen in pathogenic processes in other tissues (28). To isolate the aneurysm dome from the parent vessel, ingrowth into the aneu rysm should persist beyond the inflammatory and remodeling stages of wound healing such that recanalization does not occur. Therefore, a fibrotic response is desired in aneurysm healing to attenuate rebleeding events after endovascular coiling.
In conclusion, we find that durable aneurysm healing, as a consequence of inflammatory cell chemotaxis, is dependent on local, and not systemic, administration of MCP1. The role of MCP1 is essential for an inflammatory response, which elicits aneurysm healing.
aUthor coNtriBUtioNs SH and KM equally contributed to acquisition and analysis of the data, drafting and revisions of the manuscript. DW, HF, and CJ greatly contributed to data acquisition. SD contributed intellec tual content and critical revisions. KH contributed to acquisition and analysis of data and critical revisions to the manuscript. BH contributed to the concept of the work, interpretation of data, and revisions to the manuscript.

acKNoWLeDgMeNts
We thank the contributions of Paul Kubilis for assisting with statistical analysis, and Bridget Kelley, Warren Rehrer, and Rachel Penumudi for blinded observation and assistance in animal procedures.
fUNDiNg This work is supported by the National Institutes of Health award number R01 NS083673 to BH and Brain Aneurysm Research Grants to KH and BH.