Female and male C57BL/6 mice (8–12 weeks, Charles River, Massachusetts, USA) were used for all experiments. Animal studies were approved by the Institutional Animal Care and Use Committee and complied with both the Guide for the Care and Use of Laboratory Animals and the ARRIVE guidelines (National Research Council, 1996). Animals received food and water ad libitum and were housed in a room with 12-h light/dark cycling. A total of 75 female mice and 75 male mice were used across all experiments. Eight female mice and 10 male mice were excluded due to insufficient aneurysm formation.

Mice were randomized to undergo either elastase surgery to create a saccular carotid artery aneurysm or sham surgery as previously described (Hoh et al., 2010). Briefly, mice were anesthetized using intraperitoneal injections of ketamine/xylazine. Through a midline neck incision, the right common carotid artery (RCCA) was isolated. The RCCA was bathed extravascularly in elastase (10 U/mL in phosphate-buffered saline (PBS); Worthington Biochemical Corporation, New Jersey, USA) for 20 min and was then occluded distally via cauterization. In the sham group, the RCCA was bathed in PBS and the distal end was not cauterized. Twenty one days later, mice were euthanized via cardiac perfusion of PBS followed by 4% paraformaldehyde (PFA) (Santa Cruz Biotechnology, Texas, USA). The RCCA was harvested, soaked in 4% PFA overnight, transferred to 18% sucrose overnight, and then frozen in Optimal Cutting Temperature (OCT) using dry ice and 2-methylbutane. A cryostat was then used to collect 5 μm sections. Immunohistochemistry was used to assess CXCL1 expression (see “Immunohistochemistry” section). Sections were imaged using a Leica DM2500 optical microscope (Leica Microsystems Inc., Illinois, USA); 40x images were used for quantification. The mean gray value was measured using ImageJ software and results were reported in relative fluorescence units (RFU) (Schneider et al., 2012).

We used our aneurysm coiling model to study CXCL1-mediated aneurysm healing. Saccular aneurysms were created in female and male mice as above. Twenty one days later, aneurysms were coiled using platinum coils coated in 50:50 poly (lactic-co-glycolic acid) (PLGA) (Sigma-Aldrich, Missouri, USA) (Schneider et al., 2012). Briefly, the RCCA aneurysm was isolated as in the initial aneurysm creation surgery. Blood flow through the aneurysm was occluded both proximally and distally using 3-0 sutures (Ethicon, New Jersey, USA). Then, an arteriotomy was made in the proximal RCCA, a PLGA coil was implanted, and the arteriotomy was closed using cauterization. Finally, the proximal and distal sutures were removed.

Once we confirmed increased CXCL1 expression in aneurysms compared to normal vasculature, we investigated the effect of systemic CXCL1 neutralization on tissue ingrowth after aneurysm coiling. Saccular aneurysms were created in female and male mice as above. Twenty one days later, aneurysms were coiled using platinum coils coated in 50:50 poly (lactic-co-glycolic acid) (PLGA) (Sigma-Aldrich, Missouri, USA) (Hoh et al., 2011). Mice were randomly assigned to receive blinded intraperitoneal injections of either CXCL1 neutralizing antibody (100 μg/mL, R&D Systems, Minnesota, USA) or isotype-matched IgG control (100 μg/mL, R&D Systems, Minnesota, USA). An intraperitoneal route was selected given its technical feasibility, high frequency of injections required for our study design, and well-established protocols for intraperitoneal injections that both we and others have used (Hourani et al., 2018). While this does subject the antibodies to first pass metabolism, we have previously demonstrated intraperitoneal injections to be feasible in our aneurysm model (Hourani et al., 2018). Injections were started 2 days prior to coiling and continued every 2 days, including the day of coiling, until the experimental endpoint. The rationale for initiating CXCL1 neutralization 2 days prior to aneurysm coiling was to ensure therapeutic effect by the time of aneurysm coiling surgeries. This dosing regimen was based on previous studies of systemic CXCL1 neutralization in mice (Nowicki et al., 2014, 2018). CXCL1 neutralization or IgG control was given for an additional 7, 14, or 21 days post-coiling. These mice were sacrificed 21 days after coiling and coiled aneurysms were harvested and sectioned as above.

To assess the efficacy of CXCL1 neutralization, intrasaccular tissue ingrowth was measured using histology. Three sections were chosen at random from each animal in mice treated for 7, 14, or 21 days and sacrificed at 21 days post-coiling. These sections were stained with hematoxylin and eosin and imaged. Tissue ingrowth was then measured using Image Pro (Media Cybernetics, Maryland, USA). To calculate the percentage of tissue ingrowth, the area of the lumen not filled by tissue ingrowth was subtracted from the total area inside the aneurysm wall. This calculation can be summarized as: ingrowth (%) = outer area-inner areaouter areax 100. The average of three distinct sections per animal was used for statistical analyses.

CXCL1 expression and inflammatory cell infiltration of neutrophils, M1 macrophages, and M2 macrophages was assessed using immunohistochemistry. Tissue sections underwent heat-based antigen retrieval using DAKO (Agilent, California, USA) followed by blocking with 4% horse serum. Primary antibodies included: CXCL1 (ab86436, abcam, Massachusetts, USA), NIMP-R14 (ab2557, abcam, Massachusetts, USA), F4/80 (MCA497, Bio-Rad, California, USA), iNOS (ab15323, abcam, Massachusetts, USA), and arginase 1 (ab60176, abcam, Massachusetts, USA). Secondary antibodies include the following Alexa Fluor antibodies: donkey anti-rat 594, donkey anti-rabbit 594, donkey anti-goat 488, and donkey anti-rabbit 488 (ThermoFisher, Massachusetts, USA). Following secondary antibody incubation, slides were mounted using VECTASHIELD with 4^‘,6-diamidino-2-phenylindole (DAPI) (Vector Laboratories, California, USA) as a nuclear counterstain. For inflammatory cell quantification, three non-overlapping regions were imaged from each animal at 40x. NIMP-R14 was used as a neutrophil marker and F4/80 was used to identify macrophages. M1 macrophages were identified by co-staining for F4/80 and iNOS and M2 macrophages were identified by co-staining for F4/80 and arginase 1. To quantify total macrophage infiltration, we recorded the total number of F4/80⁺ cells, including those without co-staining with iNOS or arginase 1, when counting M1 and M2 macrophages and used the average for statistical analyses. Positive controls were included with all staining experiments.

Animals were randomly assigned to different experimental groups prior to initiating any surgeries. Ingrowth measurements and cell counts were performed by two blinded observers and all image files were deidentified and randomized. The Mann-Whitney U test used to detect differences in CXCL1 expression, tissue ingrowth, and inflammatory cell counts. One-way ANOVA with Newman-Keuls post hoc test was used to compare tissue ingrowth between different lengths of treatment within an experimental group. Analyses were performed using GraphPad Prism and results are reported with standard error of the mean. The threshold for statistical significance was set to p < 0.05.

Using immunohistochemistry, we first confirmed that both female and male mice had increased CXCL1 expression in carotid artery aneurysms compared to the normal carotid arteries in sham-operated mice (females: 26.8 ± 3.22 vs. 10.6 ± 2.31 RFU, n = 4 each, p = 0.03; males: 32.3 ± 4.42 vs. 8.1 ± 2.55 RFU, n = 4 each, p = 0.03) (Figures 1A, B). There was no difference in CXCL1 expression between females and males with aneurysms (26.8 ± 3.22 vs. 32.3 ± 4.42 RFU, n = 4 each, p = 0.49) (Figure 1B).

In female mice, CXCL1 neutralization increases tissue ingrowth compared to IgG when given for 14 days (66.4% ± 2.60 vs. 51.2% ± 5.01; n = 5 each; p = 0.03) and 21 days (71.9% ± 5.79 vs. 44.3% ± 5.10; n = 7 and 8; p = 0.002) but not when given for 7 days (48.1% ± 5.71 vs. 49.4% ± 7.03; n = 9 and 7; p = 0.92) (Figures 2A, B). In male mice, CXCL1 neutralization increased tissue ingrowth compared to IgG when given for 14 days (69.8% ± 5.13 vs. 47.0% ± 5.29; n = 6 each; p = 0.004) and 21 days (67.8% ± 3.93 vs. 53.8% ± 4.25; n = 7 and 8; p = 0.02) but not when given for 7 days (52.3% ± 3.68 vs. 50.4% ± 6.33; n = 5 each; p = 1.00) (Figures 2A, B).

There was a significant reduction in neutrophil infiltration at 14 days post-coiling with CXCL1 neutralization compared to IgG in females (0.43 ± 0.28 vs. 5.21 ± 2.40 cells/hpf, n = 7 and 8, p = 0.04) and males (0.00 ± 0.00 vs. 4.42 ± 2.41 cells/hpf, n = 5 and 4, p = 0.04) (Figures 3A, B). Similarly, with 21 days of CXCL1 neutralization, both female (0.79 ± 0.43 vs. 5.96 ± 2.28 cells/hpf; n = 7 and 4, p = 0.03) and male (0.38 ± 0.38 vs. 4.27 ± 1.54 cells/hpf, n = 7 and 8, p = 0.03) aneurysms had significantly fewer neutrophils compared to control (Figures 3A, B). At 7 days post-coiling, however, CXCL1 neutralization did not affect neutrophil filtration in females (5.50 ± 1.31 vs. 6.56 ± 1.79 cells/hpf, n = 4 and 3, p = 0.72) or males (6.25 ± 1.92 vs. 8.89 ± 4.45 cells/hpf, n = 4 and 3, p = 0.63) (Figures 3A, B).

In both females and males, there was no difference in the number of M1 macrophages between CXCL1 neutralization and IgG control 7 days after coiling (female; 1.17 ± 0.32 vs. 1.67 ± 0.38 cells/hpf, n = 4 and 3, p = 0.37: male; 0.93 ± 0.46 vs. 0.89 ± 0.48 cells/hpf, n = 5 and 3, p = 1.00) and 14 days after coiling (female; 1.52 ± 0.86 vs. 1.54 ± 0.73 cells/hpf, n = 4 and 4, p = 1.00: male; 0.47 ± 0.29 vs. 1.13 ± 0.50 cells/hpf, n = 5 each, p = 0.44) (Figures 3C, D). At 21 days post-coiling, CXCL1 neutralization significantly reduced M1 macrophages in males (0.73 ± 0.28 vs. 2.03 ± 0.45 cells/hpf, n = 5 each, p = 0.045), but not females (1.21 ± 0.43 vs. 2.53 ± 0.50 cells/hpf, n = 7 and 4, p = 0.22) (Figures 3C, D).

At 14 days post-coiling, female (2.25 ± 0.39 vs. 0.79 ± 0.31 cells/hpf, n = 4 each, p = 0.03) and male (2.00 ± 0.38 vs. 0.27 ± 0.12 cells/hpf, n = 5 each, p = 0.02) mice treated with CXCL1 neutralization had more M2 macrophages compared to IgG control-treated mice (Figures 3C, D). CXCL1 neutralization did not significantly affect the number of M2 macrophages at 7 days (female; 0.33 ± 0.14 vs. 0.50 ± 0.50 cells/hpf, n = 4 and 3, p = 0.85: male; 0.73 ± 0.44 vs. 0.33 ± 0.33 cells/hpf, n = 5 and 3, p = 0.63) or 21 days post-coiling (female; 1.10 ± 0.59 vs. 1.46 ± 0.34 cells/hpf, n = 7 and 4, p = 0.34: male; 1.81 ± 0.42 vs. 0.79 ± 0.41 cells/hpf, n = 6 and 8, p = 0.13) ( fig3Figures 3C, D).

There was no significant change in total macrophage infiltration between CXCL1 neutralization and IgG control at 7 (female; 3.83 ± 0.97 vs. 5.08 ± 0.90 cells/hpf, n = 4 and 3, p = 0.24: male 6.23 ± 1.09 vs. 4.00 ± 1.60 cells/hpf, n = 5 and 3, p = 0.19), 14 (female; 4.35 ± 0.96 vs. 3.10 ± 1.26 cells/hpf, n = 4 each, p = 0.32: male; 5.85 ± 1.31 vs. 4.41 ± 1.31 cells/hpf, n = 5 each, p = 0.38), or 21 days (female; 4.02 ± 0.82 vs. 6.43 ± 1.06 cells/hpf, n = 7 and 4, p = 0.12: male; 4.59 ± 0.77 vs. 3.81 ± 0.88 cells/hpf, n = 5 each, p = 0.52) post-coiling (Figures 3C, D).

This study had some limitations. First, our coiling model uses an extracranial vessel. Nonetheless, the purpose of this model is to study intraluminal tissue healing. Next, our CXCL1 blockade vs. IgG control dosing regimen started 2 days prior to aneurysm coiling. Our results demonstrate that neutrophil cells remain elevated and M2 macrophages remain scarce until 14 days of CXCL1 blockade. From a clinical translational perspective, future studies are required to determine if persistent CXCL1 blockade is needed from the time of aneurysm treatment or if treatment can be started during latter parts of the inflammatory cascade (e.g., 3–5 days after aneurysm treatment).

Further, while we investigated the effect of CXCL1 neutralization on neutrophil infiltration and macrophage phenotypes, additional studies are needed to directly study the relative contributions of each. Future studies in which neutrophils and M2 macrophages are directly depleted would provide more definitive proof of causality. In addition, we have previously investigated the effects of local modulation of inflammatory pathways (Hoh et al., 2011, 2018). Future studies targeting local CXCL1 blockade through drug-eluting coils will be of significant translational value. Additionally, there is growing evidence that IA pathophysiology is unique between females and males. While the intent of this study was not direct comparison between females and males, future studies in this area would be of great clinical significance.