Osteoclast-Like Cells in Aneurysmal Disease Exhibit an Enhanced Proteolytic Phenotype

Abdominal aortic aneurysm (AAA) is among the top 20 causes of death in the United States. Surgical repair is the gold standard for AAA treatment, therefore, there is a need for non-invasive therapeutic interventions. Aneurysms are more closely associated with the osteoclast-like catabolic degradation of the artery, rather than the osteoblast-like anabolic processes of arterial calcification. We have reported the presence of osteoclast-like cells (OLCs) in human and mouse aneurysmal tissues. The aim of this study was to examine OLCs from aneurysmal tissues as a source of degenerative proteases. Aneurysmal and control tissues from humans, and from the mouse CaPO4 and angiotensin II (AngII) disease models, were analyzed via flow cytometry and immunofluorescence for the expression of osteoclast markers. We found higher expression of the osteoclast markers tartrate-resistant acid phosphatase (TRAP), matrix metalloproteinase-9 (MMP-9), and cathepsin K, and the signaling molecule, hypoxia-inducible factor-1α (HIF-1α), in aneurysmal tissue compared to controls. Aneurysmal tissues also contained more OLCs than controls. Additionally, more OLCs from aneurysms express HIF-1α, and produce more MMP-9 and cathepsin K, than myeloid cells from the same tissue. These data indicate that OLCs are a significant source of proteases known to be involved in aortic degradation, in which the HIF-1α signaling pathway may play an important role. Our findings suggest that OLCs may be an attractive target for non-surgical suppression of aneurysm formation due to their expression of degradative proteases.


Introduction
There are an estimated 1.1 million abdominal aortic aneurysms (AAAs) in the United States according to a comprehensive analysis of risk factors for AAA in more than three million individuals evaluated by ultrasound imaging [1]. In 2015, AAA was associated with 16,522 deaths in the US, and this number is expected to grow due to an aging population demographic [2]. The risk factors for the development of AAA include age, male gender, family history, cardiovascular disease, and smoking [1,3,4]. Unless detected via imaging scans, often for other reasons, AAAs frequently remain undiagnosed until rupture. Rupture is a highly fatal event. About half of patients survive long enough to make it to surgery, and of those that do, the mortality rate approaches 50% [5]. Existing guidelines suggest the option of elective surgical repair for aneurysms >5-5.5 cm in diameter [6]. Currently, open surgical repair and endovascular placement of a stent graft are the only proven treatments for AAA [7,8]. The significant invasiveness and morbidity and mortality associated with surgery stresses the need for alternative non-surgical therapeutic strategies [9,10].

TRAP-Positive Macrophages Produce More Cathepsin K and MMP-9 than TRAP-Negative Macrophages
We previously demonstrated that treatment of the RAW 264.7 mouse macrophage cell line with CaPO 4 and TNFα can induce osteoclastogenesis, as indicated by increased TRAP, cathepsin K, and MMP-9 expression [18]. Next, we were interested in examining the differential expression of the markers of osteoclastogenesis within the stimulated cell population. We applied flow cytometry to further analyze the in vitro osteoclastogenic stimulation of RAW 264.7 cells, allowing us to examine protease expression in subsets of the stimulated population through the gating strategy depicted in Figure 1.

TRAP-Positive Macrophages Produce More Cathepsin K and MMP-9 than TRAP-Negative Macrophages
We previously demonstrated that treatment of the RAW 264.7 mouse macrophage cell line with CaPO4 and TNFα can induce osteoclastogenesis, as indicated by increased TRAP, cathepsin K, and MMP-9 expression [18]. Next, we were interested in examining the differential expression of the markers of osteoclastogenesis within the stimulated cell population. We applied flow cytometry to further analyze the in vitro osteoclastogenic stimulation of RAW 264.7 cells, allowing us to examine protease expression in subsets of the stimulated population through the gating strategy depicted in Figure 1. Flow cytometry gating strategy. Cells derived from culture or tissue samples were gated for size, singlets, and viability. Live cells were then analyzed for their expression of CD11b, TRAP, cathepsin K, MMP-9, and HIF-1α. The CD11b + population was divided into TRAP − and TRAP + populations which were analyzed for their expression of cathepsin K and MMP- 9. In vitro stimulation of macrophages with CaPO4 and TNFα results in increased TRAP (0.9633% ± 0.008819% vs. 7.733% ± 0.1556%, p < 0.001), cathepsin K (0.2233% + 0.04702% vs. 3.913% + 0.07364%, p < 0.0001, and MMP-9 (0.4267% ± 0.1415% vs. 4.057% ± 0.1648%, p < 0.0001) expression (Figure 2A-C). We also examined the expression of HIF-1α for its potential role in osteoclastogenic signaling and found elevated expression in stimulated cells compared to control (0.2333% ± 0.05487% vs. 1.207% ± 0.1328%, p < 0.01) ( Figure 2D). Furthermore, among stimulated cells, we demonstrated that osteoclasts (TRAP-positive) produced more cathepsin K (5121 ± 277.6 vs. 17185 ± 513, p < 0.0001) and MMP-9 (6291 ± 617.3 vs. 20702 ± 961.5, p < 0.001) than nonactivated macrophages (TRAP-negative) ( Figure 2E,F, respectively). Flow cytometry gating strategy. Cells derived from culture or tissue samples were gated for size, singlets, and viability. Live cells were then analyzed for their expression of CD11b, TRAP, cathepsin K, MMP-9, and HIF-1α. The CD11b + population was divided into TRAP − and TRAP + populations which were analyzed for their expression of cathepsin K and MMP-9.

Aneurysmal Tissues Exhibit Increased Expression of Osteoclastogenic Markers
Mice treated via the CaPO 4 and AngII models exhibited aneurysm formation in the carotid arteries and abdominal aortae, respectively, as defined by a 50% or greater increase in vessel diameter ( Figure 3A,B, respectively). The treatment of carotid arteries in the CaPO 4 model resulted in increased vessel diameter compared to untreated contralateral carotids (0.508 ± 0.01319 mm vs. 1.018 ± 0.03597 mm, p < 0.0001) ( Figure 3C). Similarly, we found increases in the diameters of abdominal aortae upon administration of AngII compared to PBS-only controls (0.9267 ± 0.01764 mm vs. 1.942 ± 0.2552 mm, p < 0.05) ( Figure 3D).

Aneurysmal Tissues Exhibit Increased Infiltration of TPMs
We evaluated aneurysmal tissues for the presence of TPMs via flow cytometry and immunofluorescence. In the mouse CaPO4-induced aneurysm model, we found that aneurysmal tissues expressed significantly higher levels of TPMs compared to controls as demonstrated by flow Additionally, we compared HIF-1α expression between myeloid and TPM cell populations from mouse and human aneurysms. The expression of TRAP is a long-established marker of osteoclastogenically activated macrophages [15]. Monocytes/macrophages are derived via the myeloid lineage, and CD11b is a reliable marker for these cells, which are the precursors to osteoclasts. We observed increased HIF-1α expression in TPMs from mouse CaPO 4 -induced aneurysms (21.73% ± 6.202% vs. 67.7% ± 2.195%, p < 0.001), AngII-induced aneurysms (5.572% ± 1.356% vs. 56.22% ± 5.458%, p < 0.001), and human AAAs (20% ± 4.952% vs. 82.8% ± 11.58%, p < 0.01) ( Figure 5D-F, respectively).

Discussion
Nearly 25 years ago, Thompson et al. demonstrated increased production of the gelatinase, MMP-9, localized to macrophages in AAA tissues compared to those from normal and occlusive arteries [27]. Longo et al. elucidated the role of MMP-9 in aneurysm development in MMP-9 knockout (KO) mice, in which KO attenuated disease formation [28]. More specifically, they revealed that macrophage-derived MMP-9 and mesenchymal MMP-2 are both required for aneurysm formation in the mouse elastase model. In our previous work, we identified the presence of OLCs in aneurysmal disease and demonstrated a concurrent increase in MMP-9 production, which corroborates previous evidence of the importance of MMP-9 in AAA formation [18,22,28]. We also demonstrated that targeting of OLCs via administration of bisphosphonates, which are commonly used to treat osteoporosis, or RANKL-neutralizing antibody, decreased aneurysmal dilation in the mouse CaPO4 and AngII models, respectively [18,22]. In addition to MMP-9, cathepsin K has been identified as another important protease upregulated in AAA [38]. Expression of cathepsin K, a strong elastase which is highly expressed in osteoclasts and vital for their function in bone degradation, was of particular interest in our hypothesis of OLC involvement in aneurysm formation [39][40][41][42]. Cathepsin K deficiency was shown by Sun et al. to result in decreased aneurysm formation in the elastaseinduced mouse model [43]. However, the AngII-induced aneurysm model, in apoE −/− , cathepsin K −/− double KO mice, showed no attenuation of aneurysm formation [44]. A possible explanation for this unexpected result could be the compensatory action of other elastinolytic cathepsins, such as S and L, as described previously [45]. Furthermore, AngII treatment of macrophages stimulated increased production of cathepsin F, which may specifically compensate for cathepsin K in AngII-induced aneurysms compared to the elastase model [46].
In this study, we sought to further investigate whether TPMs found in aneurysmal tissue are significant sources of the vessel-degrading proteases MMP-9 and cathepsin K, similar to osteoclasts.

Discussion
Nearly 25 years ago, Thompson et al. demonstrated increased production of the gelatinase, MMP-9, localized to macrophages in AAA tissues compared to those from normal and occlusive arteries [27]. Longo et al. elucidated the role of MMP-9 in aneurysm development in MMP-9 knockout (KO) mice, in which KO attenuated disease formation [28]. More specifically, they revealed that macrophage-derived MMP-9 and mesenchymal MMP-2 are both required for aneurysm formation in the mouse elastase model. In our previous work, we identified the presence of OLCs in aneurysmal disease and demonstrated a concurrent increase in MMP-9 production, which corroborates previous evidence of the importance of MMP-9 in AAA formation [18,22,28]. We also demonstrated that targeting of OLCs via administration of bisphosphonates, which are commonly used to treat osteoporosis, or RANKL-neutralizing antibody, decreased aneurysmal dilation in the mouse CaPO 4 and AngII models, respectively [18,22]. In addition to MMP-9, cathepsin K has been identified as another important protease upregulated in AAA [38]. Expression of cathepsin K, a strong elastase which is highly expressed in osteoclasts and vital for their function in bone degradation, was of particular interest in our hypothesis of OLC involvement in aneurysm formation [39][40][41][42]. Cathepsin K deficiency was shown by Sun et al. to result in decreased aneurysm formation in the elastase-induced mouse model [43]. However, the AngII-induced aneurysm model, in apoE −/− , cathepsin K −/− double KO mice, showed no attenuation of aneurysm formation [44]. A possible explanation for this unexpected result could be the compensatory action of other elastinolytic cathepsins, such as S and L, as described previously [45]. Furthermore, AngII treatment of macrophages stimulated increased production of cathepsin F, which may specifically compensate for cathepsin K in AngII-induced aneurysms compared to the elastase model [46].
In this study, we sought to further investigate whether TPMs found in aneurysmal tissue are significant sources of the vessel-degrading proteases MMP-9 and cathepsin K, similar to osteoclasts. Using flow cytometry, we evaluated live cells from aneurysmal tissue for expression of TRAP, cathepsin K, and MMP-9. We chose to compare human AAA and carotid plaque tissues for these investigations due to the distinct differences in the pathology of the two diseases. Carotid plaque formation is an atherosclerotic disease which is characterized by increased calcification and narrowing of the artery, as opposed to aneurysm formation, which does not display high levels of calcification, but rather a degeneration of the arterial wall leading to vessel dilation. We expect the calcium anabolism that is characteristic of plaque formation does not involve high levels of osteoclastogenesis, which is a catabolic event, and thus characteristic of aneurysmal disease. In confirmation of our previous studies, and those of others, we found significant increases in TRAP, cathepsin K, and MMP-9 expression in mouse and human aneurysmal tissues compared to controls. In accordance with our previous studies, we demonstrated the presence of TPMs in aneurysmal tissues via flow cytometry and immunofluorescence staining. Notably, we revealed the presence of a distinctly multinucleate TPM, a characteristic morphology of osteoclasts, in a mouse CaPO 4 -induced aneurysm ( Figure 6G). Moreover, we showed that TPMs produce significantly more cathepsin K and MMP-9 than TRAP-negative myeloid cells. We demonstrated that TPMs, specifically, are producing significant amounts of proteases, and thus, functioning as OLCs. To our knowledge, this study is the first to examine discrete live cell populations from aneurysmal tissues for their differential expression of osteoclast-associated proteins.
The HIF-1α signaling pathway is involved in osteoclastogenesis and may play an important role in upregulating the proteolytic capacity of osteoclasts [36,[47][48][49]. Therefore, we were particularly interested in the role of HIF-1α signaling as it relates to the potential stimulation of OLCs in aneurysmal tissue. Importantly, as we demonstrated here, OLCs in aneurysmal tissue are a significant source of the degradative proteases cathepsin K and MMP-9. In these experiments, we demonstrated increased HIF-1α expression in human and mouse aneurysmal tissues compared to controls, which may be due, in part, to hypoxic conditions found in aneurysmal tissues [31,33,[50][51][52][53]. Moreover, we showed that HIF-1α expression in TPMs is significantly higher than myeloid cells. This is particularly interesting in light of the previously described role for HIF-1α in osteoclastogenesis and aneurysm formation. Further evidence of the potential importance of HIF-1α is derived from studies investigating factors related to the inhibition of HIF-1α, osteoclastogenesis, and aneurysm formation, such as female sex and diabetes. Sex is a significant risk factor for aneurysm formation, with women exhibiting approximately a 4-5-fold reduction in prevalence [54,55]. Mukundan et al. demonstrated that estradiol can inhibit hypoxia-induced HIF-1α expression in an estrogen-receptor mediated manner [56]. Likewise, estrogen deficiency served to stabilize HIF-1α in osteoclasts and lead to their activation [57]. Diabetes was identified to correlate negatively with aneurysm formation [4]. With respect to hyperglycemia, we previously showed that decreased MMP-9 expression and macrophage activation under hyperglycemic conditions correlated with suppression of aneurysm formation [58]. Subsequently, we showed that hyperglycemia suppressed macrophage activation via the downregulation of the GLUT1 receptor, which is a transcriptional target of HIF-1α [59,60]. In light of the important role of HIF-1α in osteoclast activation and aneurysm formation, in this study, we demonstrated increased HIF-1α expression in OLCs derived from aneurysmal tissues, concurrent with a significant increase in protease expression by this cell population.
The function of HIF-1α in aneurysms remains unclear, as multiple studies have demonstrated seemingly contradictory results. Previous studies have demonstrated that HIF-1α expression is negatively associated with aneurysm formation [61][62][63], while others demonstrate a positive association [32,37]. The conflicting results in these studies have not yet been sufficiently explained. However, the ambiguity with respect to the role of HIF-1α in AAA formation may be the result of variations in the models of AAA induction, or differential roles of HIF-1α with respect to temporal expression or the specific cell types investigated in these studies. In future experiments, we hope to elucidate the role of HIF-1α as it relates to aneurysm induction and progression, particularly concerning the activation and effector function of OLCs.
This study demonstrates that OLCs are resident in aneurysmal tissues. We identified these cells as significant sources of the vessel-degrading proteases cathepsin K and MMP-9, therefore, OLCs represent intriguing targets for therapeutic strategies to treat or prevent aneurysmal degeneration. However, this study has some limitations. First, we did not attempt to inhibit OLCs and evaluate tissues post-treatment for aneurysm formation, OLC infiltration, and OLC phenotype. Moreover, outside of HIF-1α expression, we did not examine the signaling pathways involved in OLC activation in aneurysmal tissues. Although previous work performed in our lab detailed the difference in signaling pathways important for osteoclastogenesis in the CaPO 4 and AngII mouse models, we are interested in examining the relative importance of TRAF2/TRAF6 signaling in human tissues. This may give us insight as to the relative contribution of the RANK/RANKL and CaPO 4 + TNFα osteoclastogenesis pathways in human aneurysmal disease, and thus inform future therapeutic interventions. These experiments are forthcoming. Future studies will examine multiple potential therapeutic targets in the axis of OLC activation as it relates to aneurysm formation. These therapies will target OLC activation at the levels of stimulus, signaling, and effector function through the application of neutralizing antibodies, HIF-1α inhibitors, and protease inhibitors or bisphosphonates, respectively. It is apparent that many human diseases are multifactorial in their etiology or progression, therefore, we hypothesize that combination therapies, incorporating multiple targets in the OLC activation axis, will be most effective for the prevention or treatment of human AAAs.

Cell Culture and Treatments
The

Human Tissue
Aneurysmal tissues were obtained from patients with a diagnosis of AAA, and an indication of maximum aortic diameter exceeding 5.5 cm, undergoing elective open repair surgery. Carotid plaque tissues were obtained from patients undergoing endarterectomy. All surgeries were performed at the University of Wisconsin Hospital. All normally discarded tissue obtained for this study was approved for use by the Minimal Risk-Health Sciences Institutional Review Board at the University of Wisconsin-Madison (submission ID 2018-0327, 25 April 2018). This tissue is not subject to IRB review because, in accordance with federal regulations, this project does not involve human subjects as defined under 45 CFR 46.102(e)(1).

Mouse Arterial Aneurysm Models and Treatments
Ten-week-old male C57BL/6 mice were obtained from the Jackson Laboratory (Bar Harbor, ME, USA). The procedures for creating our CaPO 4 -induced mouse model of arterial aneurysm were previously described [64]. Briefly, 0.5 M CaCl 2 -soaked gauze was applied perivascularly for 20 min on the carotid artery, as indicated. The gauze was replaced with another PBS-soaked gauze for 10 min, and the incised area was sutured. Untreated contralateral arteries served as controls. The mice were sacrificed seven days after surgery. Treated and control arteries were measured with an electronic digital caliper (VWR International, West Chester, PA, USA). Arteries collected for histological examinations were fixed by perfusion with 4% paraformaldehyde. To obtain live cells for flow cytometry, the arteries were not fixed, but perfused with DMEM prior to resection. AngII-induced aneurysms were produced in retired male breeder apoE −/− mice (>6 months of age) obtained from the Jackson Laboratory, as previously described [22]. Mice received either Ang II (1000 ng/min/kg) (Sigma-Aldrich) or PBS via continuous infusion via micro-osmotic pump (Alzet model 1004, Durect Corporation, Cupertino, CA, USA) implanted in the back of the mouse. Mice were sacrificed 28 days post-implantation. Aortae collected for histological examinations were fixed by perfusion with 4% paraformaldehyde (Santa Cruz Biotechnology, Dallas, TX, USA). To obtain live cells for flow cytometry, the aortae were perfused with DMEM prior to resection. The maximum diameter of the abdominal aortae was measured ex vivo with digital calipers to evaluate aneurysm formation. All animal procedures were conducted in accordance with experimental protocols that were approved by the Institutional Animal Care and Use Committee at the University of Wisconsin-Madison (Protocol M005383, 8 April 2016).

Flow Cytometry
Mouse and human tissue samples were incubated at 37 • C with shaking for 1.5 h in digestion buffer containing 450 U/mL collagenase I (Sigma-Aldrich), 125 U/ml collagenase XI (Sigma-Aldrich), 60 U/mL hyaluronidase (Worthington Biochemical, Lakewood, NJ, USA), and 60 U/ml deoxyribonuclease I (Worthington), as previously described [65]. The digested tissue was then dissociated to obtain a single-cell suspension by passage through a 70 µm cell strainer (Falcon, Corning, NY, USA), and 2.5 × 10 5 cells were stained for flow cytometric analysis as follows. First, the cells were blocked with 100 µg/mL of purified mouse IgG (Jackson Immuno-Research, West Grove, PA, USA) for 20 min at 4 • C. Then the cells were washed with wash buffer (PBS, 3% FBS) and stained for surface expression of CD11b and viability with anti-CD11b-BV711 antibody (563168, BD Biosciences, Franklin Lakes, NJ, USA) and Ghost 510 (Tonbo, San Diego, CA, USA), respectively, for 30 min at 4 • C. The cells were then washed and resuspended in fixation buffer (Invitrogen, San Diego, CA, USA), and allowed to fix for 15 min at room temperature. The cells were washed in permeabilization buffer (Invitrogen), and resuspended in permeabilization buffer containing the following primary antibodies: Anti-TRAP-Alexa Fluor 488 (ab216934, Abcam, Cambridge, UK), anti-cathepsin K-Alexa Fluor 647 (sc-48353, Santa Cruz Biotechnology), anti-MMP-9-PE (sc-13520, Santa Cruz Biotechnology), and anti-HIF-1α-Alexa 594 (sc-53546, Santa Cruz Biotechnology). The antibodies and cells were incubated for 30 min at room temperature. Following incubation, the cells were washed with permeabilization buffer and resuspended in wash buffer. The cells were analyzed on an Attune NxT Flow Cytometer (Invitrogen, Carlsbad, CA, USA). Fluorescence minus one controls were prepared and analyzed for each fluorochrome used.

Immunofluorescence Staining
Perfusion fixed arteries from mice treated with CaPO 4 or AngII, and human carotid plaque and AAA tissue fixed overnight in 4% paraformaldehyde, were embedded with optimal cutting temperature compound, frozen, and sectioned at 8 µm. The sections were then rinsed in DI water and blocked with 5% donkey serum (Santa Cruz Biotechnology) for one hour. After blocking, the tissue sections were washed twice in 1X Tris-buffered saline-Tween 20 (TBST) (Affymetrix, Cleveland, OH, USA) and surface-stained with anti-CD11b primary antibody (ab133357, Abcam) overnight at 4 • C. After washing with TBST, the sections were incubated with donkey anti-rabbit Alexa Fluor 594 secondary antibody for one hour (Life Technologies, San Diego, CA, USA). After washing, the sections were permeabilized for 20 min in 10× TBST for intracellular staining. The sections were then washed and blocked with 1% bovine serum albumin (BSA) (Santa Cruz Biotechnology) for one hour. After washing, the sections were incubated overnight with anti-TRAP Alexa Fluor 488 antibody (ab216934, Abcam) at 4 • C. After overnight incubation, the sections were washed and counterstained with 4,6-diamidino-2-phenylindole dihydrochloride (DAPI) (Life Technologies) for 20 min. The sections were washed for a final time and mounted with Vectashield hardset antifade mounting medium (Vector Labs, Burlingame, CA, USA). The stained tissue sections were then examined with a Nikon E600 fluorescent microscope (Nikon, Tokyo, Japan).

Statistical Analysis
Data are reported as the means ± standard error of the means (SEM). The means of two groups were compared using the unpaired, two-tailed Student's t-test. Statistical analysis was performed with the GraphPad Prism program, version 7.0 (GraphPad Software. Inc. San Diego, CA, USA). p-values < 0.05 were accepted as statistically significant.

AAA
Abdominal aortic aneurysm OLC Osteoclast-like cell TRAP Tartrate-resistant acid phosphatase NF-κB Nuclear factor kappa-B RANKL Receptor activator of nuclear factor kappa-B ligand MMP-9 Matrix metalloproteinase-9 TRAF TNF receptor-associated factor MAPK mitogen-activated protein kinase NFATc1 Nuclear factor of activated