In Situ Decellularlization of a Large Animal Saccular Aneurysm Model: Sustained Inflammation and Active Aneurysm Wall Remodeling

Abstract

Background

We investigated decellularization as a strategy of achieving aneurysmal growth and lasting inflammation in a large animal model of saccular aneurysms which does not require grafting procedures.

Methods

18 New Zealand White rabbits were randomized 2:1 to receive endoluminal sodium dodecyl sulfate (SDS) treatment following elastase (infusion with 1% solution, 45 min) or elastase-only treatment (control). Every two weeks three of the rabbits (2 elastase+SDS, 1 elastase), were subjected to MRI, followed by contrast injection with myeloperoxidase (MPO)-sensing contrast agent. The MRI was repeated 3 hours post contrast injection and the enhancement ratio (ER) was calculated. All aneurysms were measured by digital subtraction angiography every 2 weeks. Following MRI, aneurysms were explanted and subjected to immunohistopathology.

Results

During follow-up MRI the average ER for SDS-treated animals was 1.63±0.20, compared to 1.02±0.06 for controls (p<0.001). The width of SDS-treated aneurysms increased significantly over the elastase aneurysms (47% vs. 20%, p<0.001). Image analysis of thin sections showed infiltration of MPO-positive cells in decellularized aneurysms and surroundings up through the 12 week observation period while control tissue had 5–6 times less cells present 2 weeks after creation. Immunohistochemistry demonstrated the presence of MPO-positive cells surrounding decellularized lesions at early time points, MPO-positive cells were found in the adventitia and in the thrombi adherent to the aneurysm wall at later timepoints.

Conclusions

In situ decellularization of a large animal model of saccular aneurysms recapitulates features of unstable aneurysms such as chronic inflammation (up to 12 weeks), active aneurysm wall remodeling, and growth (up to 8 weeks).

Introduction

Contemporary large animal models of saccular aneurysms include surgical anastomosis of a vein pouch to an arteriotomy in pigs, rabbits and dogs or elastinolysis of the common carotid artery in rabbits (reviewed in¹ and²). However, following a short period after aneurysm creation,³ these models develop into stable aneurysms that do not demonstrate characteristics of vulnerable aneurysms in humans; namely, absence of inflammation, intramural hemorrhage, and ultimately rupture (reviewed in⁴), only showing some areas of thin, acellular aneurysm wall.⁵ To address these model limitations, an innovative strategy has been developed to implant decellularized arterial grafts in rodents,⁶ which have shown both growth and rupture of these abdominal aortic aneurysms. The objective of our study is to explore the feasibility of inducing in situ decellularization in a common rabbit model of saccular aneurysms⁷ to recapitulate pathological features of unstable aneurysms and to perform comparative longitudinal non-invasive magnetic resonance imaging (MRI) of inflammation marker myeloperoxidase (MPO). We report the growth characteristics, inflammation-induced MRI signal changes and pathology of this modified saccular aneurysm model as compared to the unmodified model in a longitudinal study.

Methods

In-Vivo Feasibility Study

All animal research activities were approved by the Institutional Animal Care and Use Committee (IACUC). Eight New Zealand White rabbits (sex: male, weight range 3.0 – 3.5 kg) were used for a pilot study to assess the optimal sodium dodecyl sulfate (SDS) infusion protocol and to determine if the animals could tolerate the procedure. All procedures were performed under general anesthesia, prior to all surgical procedures, the animals were pre-anesthetized by a subcuticular injection of glycopyrrolate (0.01mg/kg). Anesthesia was induced by an intramuscular injection of ketamine (35mg/kg) and xylazine (5mg/kg) and maintained with mechanical ventilation of 1–3% isoflurane. The physiologic status of the animal was assessed using continuous monitoring of respiration rate, heart rate, oxygen saturation level, end-tidal CO₂ level, and temperature. At the conclusion of the procedure the animals were euthanized by overdose of sodium pentobarbital (150mg/kg). Similar to elastase incubation, all SDS infusions were performed with balloon occlusion of the right common carotid artery origin to prevent systemic administration.

Three rabbits underwent standard elastase incubation, followed by infusion of 1% SDS for 15 minutes, 60 minutes, or 2% SDS for 60 minutes. After two weeks, the animals were euthanized. To optimize the SDS infusion protocol, five rabbits underwent bilateral common carotid artery (CCA) exposure followed by placement of a 6F hemostatic introducer, allowing for n=10 tests. All arteries were exposed to 50U of elastase for 15 minutes. In three of the rabbits once the elastase incubation was completed, a 1% solution of SDS was infused slowly (30mL/hr) via the microcatheter under balloon protections for 15, 30 or 45 minutes (Figure 1). The SDS was continuously infused via the microcatheter under balloon occlusion of the right CCA origin and circulated distally where it exited the introducer. To assess if the order of elastase and SDS treatment influenced the overall decellularization, the last two animals underwent acute aneurysm creation just as before, except SDS treatment was done prior to elastase incubation.

Figure 1:

Schematic of isolation of the right common carotid artery (CCA) from the circulating blood pool. SDS is infused slowly through the microcatheter, into the RCCA, and cleared via the introducer side-port, following the pathway of the blue arrows.

After completion of elastase/SDS incubation, each artery was flushed with 10mL of saline. The animals were euthanized, CCAs were excised and placed in 10% formalin overnight. The specimens were embedded in paraffin, thin sectioned, stained with H&E and then imaged with a standard microscope. Images were analyzed for the density of cell nuclei within the smooth muscle layer. Comparisons were made with the innominate artery, serving as control.

Longitudinal In-Vivo Study

After optimizing the SDS infusion protocol (1% SDS infusion for 45 minutes), 18 New Zealand White rabbits (sex: either, weight range 3.0 – 3.5 kg) were randomized 2:1 to receive elastase and SDS versus just elastase during aneurysm creation. The diameter and height of the aneurysm was measured every 2 weeks using non-invasive DSA and cone beam CT (CBCT). In order to measure effect of the SDS treatment on the size of the aneurysm, only the width change was considered, since the apex of the aneurysm is not a true wall. The baseline width of the aneurysm was chosen as the width of the CCA approximately 4 mm from the origin. Every two weeks, three of the rabbits (2 elastase+SDS, 1 elastase), were subjected to MR imaging. This protocol consisted of Time of Flight (ToF) angiography (TR/TE: 21/4ms, FA: 20°, FoV: 110×110mm, matrix: 326×164, thickness: 1mm), coronal T2w-TSE (TR/TE: 3000/80ms, FA: 90°, FoV: 120×120mm, matrix: 220×176, thickness: 0.9mm, ETL: 16), and axial respiratory gating MSDE (TR/TE: 2500/6ms, FA: 90°, FoV: 108×108mm, matrix: 216×216, thickness: 1.8mm, ETL: 9, V_enc: 1). After pre-contrast imaging a macrocyclic paramagnetic contrast agent (Gd-5HT-DOTAGA, dose: 0.1 mmolGd/kg) that has been shown to selectively enhance in areas of MPO accumulation,^8,9 was injected IV and the MR was then repeated at 3 hours post contrast injection, and the enhancement ratio was measured. At the eight-week timepoint two rabbits, one elastase+SDS and one elastase, were given Gadavist (0.1 mmolGd/kg, Bayer HealthCare, NJ), instead of the MPO-sensitive contrast agent; the same rabbits were then re-imaged four weeks later with the MPO-sensitive contrast agent. Following MR imaging the animals were euthanized, the aneurysm was exposed, flushed with cold saline, filled with OCT compound, and snap frozen in liquid nitrogen. Detailed methods for histology and microscopy can be found in the online supplement.

Quantification and Statistical analysis

For quantification of MR signal pre and post contrast injection, an annular region of interest (ROI) was drawn to include the wall of the aneurysm but exclude the center of the lumen. A second ROI was drawn in the muscle tissue to normalize the signal between scans. Then the Enhancement Ratio (ER) was calculated by $ER= \left(\raisebox{1ex}{$An{x}_{\mathit{\text{post}}}$}\!\left/ \!\raisebox{-1ex}{$Mu{s}_{\mathit{\text{post}}}$}\right.\right)/\left(\raisebox{1ex}{$An{x}_{pre}$}\!\left/ \!\raisebox{-1ex}{$Mu{s}_{pre}$}\right.\right)$, this allows for direct comparison between animals and timepoints. All ER analysis was performed on the axial MSDE imaging, ranging from 1 to 3 slices depending on the size of the aneurysm and the imaging planes there were selected.

All statistical calculations were performed using R3.5.2 using publicly available packages. For those data that were normally distributed either a Welch’s t-test, or an ANOVA to determine if a factor was significant followed by a Tukey post hoc test were used. For those that were not normally distributed, a Mann-Whitney-U test was used. An alpha of 0.05 was significant for all tests.

Results

In-Vivo Feasibility Study

Incubation of 2% SDS for 60 minutes led to rupture of the CCA (Supplemental Figure 1) and the animal expired. Incubation of 1% SDS for 15 minutes failed to achieve decellularization of the CCA, and 60 minutes led to animal health decline that required daily care. Continuous infusion of 1% SDS was well tolerated in all animals. The 45-minute infusion of 1% SDS protocol led to optimal decellularization without acute rupture or other sequelae and included histological observations of intramural hemorrhage (Figure 2 A–D). Quantitation of nuclear density showed that continuous infusion of 1% SDS for 45 minutes resulted in a near complete decellularization of the CCA having less than 10% of the innominate artery cell nuclei density (Figure 2 E). Finally, there was no difference in histological features as a function of the order of administration of elastase or SDS.

Figure 2:

Histology of decellularization. A) 15-minute SDS infusion, some nuclei removed, but overall no large areas of remove. B) 30-minute SDS, complete removal of nuclei in some areas, with patches in other areas. C) 45-minute SDS infusion almost complete nuclei removal and thinning of the wall. D) Intramural hemorrhage found after SDS infusion. E) Number of cell nuclei per mm² left after SDS treatment of rabbit model elastase-induced aneurysms, data shown as mean±SD.

Longitudinal In-Vivo Study

Of the 18 rabbits used in the longitude study, 17 of them reached designated endpoint. One animal in the elastase+SDS group died during recovery due to rupture of the CCA, leading to an overall morbidity rate of 0% and mortality rate of 8% in the elastase+SDS group, and 0% morbidity and mortality in the elastase only group. Aneurysms in the two groups were not significantly different at baseline (2.4 mm vs. 2.2 mm, p = 0.25). Addition of the SDS modification led to continuous increase of aneurysm width (Figure 3) during the entire observation period. Aneurysm widths increase in the control group stabilized after 2 weeks. This led to significant increase in the aneurysm width at 4, 6, and 8 weeks (p < 0.001) with an absolute width difference of 0.92 mm.

Figure 3:

The overall change in aneurysm width comparing the SDS-treated to the elastase-only control aneurysms. After the 2 week timepoint, control aneurysms stopped growing, while the SDS treated aneurysms continued to grow to the 8 week timepoint. *** - p< 0.001

For the 17 remaining rabbits that underwent MR imaging post aneurysm creation (Figure 4 A–D), 11 were treated with SDS while 6 were control. At all timepoints the SDS group showed enhancement over the control group with an average enhancement ratio of 1.63 vs 1.01 (p < 0.001, Figure 4 E). The only time that the control aneurysms showed an enhancement over 1.1 was at the two-week timepoint, which is known to be within the healing period following elastase incubation.^3,7 At the two-week timepoint enhancement for the SDS infused group was not limited to the wall of the aneurysm, with perivascular enhancement. Four weeks following aneurysm creation, enhancement was largely limited to the aneurysm wall. At 8 weeks, the two rabbits that were imaged using Gadavist showed minimal enhancement; however, once brought back at 12 weeks the SDS infused rabbit did show enhancement with the MPO specific contrast agent (Figure 4 D).

Figure 4:

MR imaging of rabbit aneurysm models. A) Standard elastase aneurysm, arrow indicates the wall of the artery. No enhancement was observed at 3 hours. B) SDS treated rabbit at the same timepoint as the elastase rabbit. Here clear enhancement of the aneurysm is seen post contrast. C) 8 week imaging of SDS rabbit post Gadavist^™ injection, almost no enhancemnet was seen. There was slight thickening of the wall (arrow). D) 12 week MRI of the same rabbit (see pannel C), this time with the MPO sensitive MR contrast agent. Clear enhancement of MR images was noted. Arrows indicate aneurysm. E) Enhancement Ratio of SDS-treated versus control (elastase treatment only). Data was averaged over all timepoints, with the elastase+SDS showing significantly higher ER compared to the elastase-only (1.63 vs 1.01, p <0.001).

Immunohistochemistry and fluorescence microscopy.

Microscopy of thin sections of rabbit aneurysm samples performed at early time point (2 weeks post SDS-treatment) revealed major differences between the SDS-treated and control (elastase only) groups of rabbits. On aneurysm samples sectioned longitudinally the differences on regular H&E-stained sections were apparent in overall morphology, remnant vessel thickness, the presence of lesions and the number of infiltrating cells (See Supplemental Figure 2). We observed more frequent mural thrombus of aneurysms associated with SDS-treatment at earlier time points (2–6 weeks) and multiple lymphocyte infiltrations in the area of decellularization which showed gradual healing with fibroblast proliferation alongside with infiltration of MPO-positive cells (see Supplemental histology). Identification of MPO-positive cells by using red-fluorescent substrate (Figure 5) showed that both control (elastase-only, Figure 5A) and elastase+SDS treated aneurysms (Figure 5B) experienced an influx of red fluorescent MPO-positive inflammatory cells. However, in control aneurysm sections MPO-positive cells were less numerous and this difference was more pronounced at the later observation time points (4–12 weeks). To track the cell numbers and their MPO-positive area on sections we performed AI-assisted segmentation and quantification of cell-associated vs. background fluorescence (Supplemental Figure 3) using AI-generated probability maps that reflected the distribution of MPO activity (Supplemental Figure 3 C and D). On fluorescence images there were two types of MPO-positive cells that could be roughly classified as “large” cells (area> 45 μm²) which were morphologically similar to monocytes/macrophages and “small” cells (area<45 μm²), morphologically identical to neutrophils. Over time the proportion of “large” cells that morphologically resemble macrophages in total population present in SDS-treated aneurysms remained high throughout the experiment. The numbers “large” cell type were decreasing in control aneurysms group together with the total cell number per section area as shown in Figure 6 A,B. Both type of cells were present as MPO-antigen positive cells (Figure 6 A), some of which were monocytes/macrophages, while many more cells showed strong staining with anti-calprotectin (S100A8/A9 complex) antibodies, which react with both monocytes/macrophages and granulocytes on frozen sections (Supplemental Figure 4). The decellularized regions of vessel wall showed extensive network of collagen fibrils that had strong autofluorescence under exciting blue light. These areas were surrounded by “swarms” of MPO- positive neutrophils (Figure 6 C, Supplemental Figure 4 A). Some of the neutrophils were observed infiltrating decellularized collagen and forming MPO-positive neutrophil extracellular traps (NETs, Figure 5 D,E).

Figure 5:

Immunofluorescence (IF) and immunohistochemistry (IHC) of rabbit aneurysms at 2 week post creation. A- IF image of control rabbit aneurysm; B- IF - image of SDS-treated rabbit aneurysm. Red – MPO substrate (5HT-Cy3), green – autofluorescence of collagen, blue – DAPI (nuclei); C – IHC of SDS-treated rabbit aneurysm. Blue- DIG-labeled anti-MPO mAb/anti-DIG-AP conjugate. D – IHC of neutrophils and NETs in decellularized area, E- control IHC image (no anti-MPO mAb). L- indicates lumen, *- decellularized area of the vessel wall.

Figure 6:

Distribution of MPO positive cell numbers and cell areas per section of SDS/elastase-treated rabbit aneurysm and control elastase-only models and the measurements of time-dependent changes in MPO-positive cell numbers over time. A- distribution of MPO-positive cells derived from 95% probability maps obtained by TWS processing of fluorescent images of rabbit sections obtained at different time points, insets show representative enlarged (250x) color fluorescent images of “large” cells (area> 45 μm²) and “small” cells (area<45 μm²); B- cumulative MPO-positive cell number change per section area over time as measured in animals terminated at the indicated time points. The data is shown as mean±SD, (n=2–3 sections/time point).

Discussion

Treatment of rabbit carotid arteries with elastase results in vascular models with geometry and hemodynamic characteristics resembling human saccular cerebral aneurysms.¹⁰ The rabbit model of saccular aneurysms created through elastinolysis is well-suited for neurointerventional device testing. Recient literature has pointed to the intact nature of the wall of aneurysm models might lead to increased rates of healing¹¹ in response to endovascular treatment. This could explain the differences seen between meta-analysis of human aneurysm coiling data showing occlusion rates of less than 60%^12,13 and recurrence rates as high as 28%¹⁴ compared to a more limited dataset of coiled rabbit elastase aneurysms with between 0–17% recurrence rates.^15,16 The ability of the smooth muscle cells within the lining of the rabbit aneurysm wall to affect and increase the rate of coiled aneurysm healing, could explain why this disparity exists, and an aneurysm model that more closely mimics that of the more dangerous acellular type D¹⁷ would allow for a closer comparison to healing rates in humans. Interestingly, there does not appear to be a difference in the healing rate for flow diverting stents; 60–80% occlusion rate^12,18 in humans compared to 74%¹⁹ for rabbits, suggesting that the healing is due to the parent artery and not that of the aneurysm wall. Overall, SDS treatment of elastase induced aneurysm model could allow for better understanding of the biological response to coiling and other intrasaccular devices.

The standard elastase induced saccular aneurysm model has been recently shown to recapitulate several characteristics of human aneurysms pathology.⁵ Over several weeks the aneurysm models may develop multiple human-like pathology phenotypes and some of them bear signatures of potential instability. However, an aneurysm model with predictable growth patterns and instability-linked molecular marker expression is still lacking. We hypothesized that in situ decellularization of a commonly used large animal model of saccular aneurysms may lead to active wall remodeling. Decellularization leaves primarily collagen network and elastin matrix as available basement for potential new growth (Figure 2 and Supplemental Figure 2). During the period that the aneurysm is healing, there appears to be active remodeling and continued migration of cells containing myeloperoxidase, i.e. neutrophils and macrophages, into the vessel wall (Figure 5). Consistent with inflammatory cell presence, as determined by cells positive by MPO activity (fluorescent substrate) and MPO antigen (anti-MPO antibody staining) were MPO activity imaging by using in vivo MRI performed during the period of up to 12 weeks following SDS-modified aneurysm creation. This is a long term increase in comparison to standard elastase model,³ that only showed elevated inflammatory infiltrates within the first 3 weeks of creation. It has been shown that progressive amounts of cell loss, inflammation, and mural thrombus are associated with rupture status,^17,20 suggesting cellular mechanisms of aneurysm instability. Moreover, the standard rabbit elastase model can demonstrate all four of the human aneurysm wall states described in⁵. However, of three categories only the cell loss was seen to match the histological categories. Thrombus formation was only seen in the apex of the aneurysm, where over time the previously patent carotid artery undergoes thrombosis. Inflammation was observed only at two weeks thereafter. The use of decellularization in surgical graft aneurysms had been recently reported in rabbits,²¹ with similar finding of increased inflammation in the decellularized aneurysms. However, unlike the surgical model, this endovascular model does not involve vascular response to suturing at the neck. The addition of SDS in the aneurysm creation process showed evidence of all three categories, hypocellular walls, continued inflammation over 12 weeks, and mural thrombus within the aneurysm lumen at numerous locations.

Recently the use of MR vessel wall imaging for the screening and follow-up of aneurysms has increased in prevalence. The goal is to identify “at risk” patients that show enhancement within the wall of the aneurysm after a contrast injection; however, the exact mechanism and interpretation of enhancement is under debate^22–26. What is known is ruptured aneurysms show higher levels of inflammation, neutrophil and macrophage presence, and secreted enzymes such as MPO,^9,27,28 and the signatures of inflammation could be a collectively or individually considered as bio-markers of said “at risk” aneurysms. The development of small-molecule MR contrast agents that selectively accumulate in area of MPO, and thus inflammation,^8,9 added more clarification on this subject; with the ability to help filtering out many of the associated artifacts. Namely, the accumulation of inflammation-specific low molecular weight imaging agent in the aneurysm assurance that the MR signal enhancement seen is not simply a flow artifact, or extensive vasa vasorum.²⁹ Here we have clearly demonstrated that an aneurysm that is undergoing active inflammation; 2 weeks for elastase only, or 12 weeks for elastase+SDS, shows selective enhancement with an MPO-specific MRI contrast agent. Further we have presented evidence that standard clinical MR contrast agent (gadobutrol) does no show the same level of continued enhancement.

The model of in situ decellularization is still limited by some of the aspects common to all extracranial models. Since the perivascular tissues surrounding carotid arteries are nothing like that of the brain, the source of the neutrophils and, consequently, MPO are not that same as in a human scenario. There is a lymph node in the proximity of the CCA, and this lymph node is an additional potential source of macrophages and neutrophils,³⁰ which are migrating into the area caused by SDS damage to the vessel wall. However, our goal was to create aneurysms that harbor on-going inflammation, and though the traffic of neutrophils and macrophages may be dissimilar to that of a human, the outcomes may be comparable. Furthermore, we were not able to observe the aneurysm rupture during relatively short observation period, other than in the acute case.

Conclusions

We describe a method of in situ decellularlization in a large animal model of saccular aneurysms. This model included active remodeling and aneurysm growth, persistent inflammation over at least a 12-week period, and at time of creation, intramural thrombus was observed. Although no ruptures occurred during follow-up, this model does show many more of the hallmarks of an unstable aneurysm.