Background and purpose Despite significant technical advances, recanalization rates after endovascular therapy of ruptured intracranial aneurysms (IAs) remain a clinical challenge. A histopathological hallmark of ruptured human IA walls is mural cell loss. Mural smooth muscle cells (SMCs) are known to promote intraluminal healing in thrombosed experimental aneurysms. In this rat model we assess the natural history and healing process after coil embolization in SMC-rich and decellularized aneurysms.
Methods Saccular aneurysms were created by end-to-side anastomosis of an arterial graft from the descending thoracic aorta of a syngeneic donor rat to the infrarenal abdominal aorta of recipient male Wistar rats. Untreated arterial grafts were immediately transplanted, whereas aneurysms with loss of mural cells were chemically decellularized before implantation. Aneurysms underwent coil implantation during aneurysm anastomosis. Animals were randomly assigned either to the non-decellularized or decellularized group and underwent macroscopic and histological analyses on days 3, 7, 21, or 90 post-coil implantation.
Results A total of 55 rats underwent macroscopic and histologic analysis. After coil embolization, aneurysms with SMC-rich walls showed a linear course of thrombosis and neointima formation whereas decellularized aneurysms showed marked inflammatory wall degeneration with increased recanalization rates 21 days (p=0.002) and 90 days (p=0.037) later. The SMCs showed the ability to actively migrate into the intra-aneurysmal thrombus and participate in thrombus organization.
Conclusions Coil embolization of aneurysms with highly degenerated walls is prone to further wall degeneration, increased inflammation, and recanalization compared with aneurysms with vital SMC-rich walls.
- vessel wall
Statistics from Altmetric.com
Since the introduction of detachable coils for the treatment of intracranial aneurysms (IAs) in 1991 by Guido Guglielmi,1 2 this technique has evolved in many neurosurgical centers as the most widely used therapy for IAs. Although it represents a less invasive approach than other modalities for many types of aneurysms, concerns regarding the durability of coil treatment have been raised. Angiographic recurrences have been detected in up to 20–30% of coiled aneurysms, with higher rates of recurrences in rupture cases.3–5 The course of an IA after coil embolization depends largely on a biological response in which the intraluminal thrombus transforms into fibrotic scar tissue, thus guaranteeing long-lasting exclusion of the aneurysm from circulation.6
This organization of the intraluminal thrombus and subsequent neointima formation appears to depend on the physiology and health of the aneurysmal wall.7–13 The walls of many IAs, especially ruptured ones, are characterized by loss of mural smooth muscle cells (SMCs), fibroblasts, and endothelial cells,14–16 which may explain the findings of impaired healing and recurrence after embolization.13
The concept that loss of healthy SMCs from the IA wall can impair the healing response after embolization and thus predispose to recanalization has not been studied before in a controlled experimental setting. We therefore investigated the effects of aneurysm wall cellularity related to the natural history and recurrence rates after coil embolization of saccular aneurysms in a rat model.
Materials and methods
Study design, animals, and anesthesia
Following an in vivo pilot series with six animals to test this protocol, we randomly assigned 57 male Wistar rats (Janvier, Saint-Berthevin Cedex, France) weighing 430±50 g to either the vital (non-decellularized) or decellularized group. Healing of the aneurysms was rated at follow-up days 3, 7, 21, or 90 postoperatively. At 21-day follow-up, the migration of aneurysm mural cells was assessed with two transgenic green fluorescence protein (GFP) mutant and two GFP wild type male Lewis rats (inbred, initially kindly provided by Professor Eiji Kobayashi, Center for Development of Advanced Medical Technology, Jichi Medical University, Tochigi, Japan). Figure 1 depicts the study design, dropout rates, and follow-up scheme.
The animals were housed in animal facilities at room temperatures of 22–24°C and 12 hours light/dark cycle. They had free access to tap water and pellet diet and were cared for in accordance with institutional guidelines. Rats were anesthetized by weight-adjusted intra-abdominal injection of medetomidine hydrochloride (0.5 mg/kg) and ketamine hydrochloride (50 mg/kg). Postoperative analgesia was performed by subcutaneous injections of buprenorphine (0.3 mg/kg). The experiments were approved by the Committee for Animal Welfare at the University of Bern, Switzerland (BE 103/13) and strictly followed the recommendations for good laboratory practice and guidelines for reporting animal research.
Aneurysm model, surgical technique, and graft harvesting
According to the Helsinki rat microsurgical sidewall aneurysm model,17 the aneurysms were created under the operating microscope (Zeiss, OPMI pico, Germany) by end-to-side anastomosis of an arterial graft from the descending thoracic aorta of a syngeneic donor rat to the infrarenal abdominal aorta of a recipient male Wistar rat. Vital (non-decellularized) arterial grafts were immediately transplanted, whereas aneurysms with loss of mural cells were decellularized by sodium dodecyl sulfate before implantation.11 Coil implantation was performed with a predefined coil volume of 2 cm of 360° 3 mm coils (Target Detachable Coils 360, Stryker, USA) during aneurysm anastomosis (see online supplementary video I).7
The operated animals were then randomly allocated to either the vital aneurysm cohort or decellularized aneurysm cohort.18 Both the Wistar rats and GFP transgenic Lewis rats were operated in the same fashion. To assess the migration capability of aneurysm wall cells into the thrombus, a GFP mutant arterial graft was transplanted in a GFP wild type recipient rat. To assess the impact of a possible cell migration from the carrying vessel into the intra-aneurysmal thrombus, an arterial graft of a GFP wild type Lewis rat was decellularized and transplanted into a GFP mutant recipient rat.
All surgical procedures were video recorded (HXR-MC1P, Sony, Japan). Aneurysm dimensions were measured and photographed before and after implantation (including endoluminal assessment of the aneurysm orifice). Surgical characteristics including total operating time, vessel clamping time, anastomosis suturing time, and graft ischemia time were summarized (see online supplementary table I).
At follow-up the animals underwent laparotomy and microsurgical harvesting of the aneurysm. With harvesting of the aneurysm and its carrying vessel, the posterior wall of the fragmented abdominal aorta was opened and the aneurysm ostium was photographed for neointima evaluation. We rated the macroscopic neointima maturity on a 6-point scale in a blinded manner (see online supplementary table II).12 Perfusion fixation of the tissue was performed with 4% paraformaldehyde.
Aneurysm volume, neointima evaluation, and histology
Aneurysm volume was calculated at baseline (day of surgery) and follow-up using the cylinder formula (V = π × width/2 × depth/2 × length).19 Coil volume was determined using the cylinder formula, while considering the coil helix outer diameter according to the manufacturer's information as volume = (π (outer diameter)2 × length)/4. Aneurysm percent packing volume was calculated by dividing the volume of the coils delivered by the volume of the aneurysm. An online aneurysm volume calculator supported the measurements (http://www.angiocalc.com/percent_volume.php).
All aneurysm samples were stored in formaldehyde until they were embedded in paraffin blocks. The blocks were then sectioned in half along the longitudinal axis with a diamond saw (cut-grinder charly, Walter Messner GmbH, Germany), and the coils were carefully removed with a 26G hollow needle and microforceps under the operating microscope.20 Aneurysms were re-embedded and cut into 2 µm sections for hematoxylin-eosin (HE), monoclonal anti-α muscle actin (α-SMA), and Masson-Goldner’s (MT) staining. The probes for immunohistological assessment of GFP were boiled in citrate buffer (pH 6.0; 2.04 g tri-Na-Citrat-Dihydrat, 1 L distilled water, 500 μL hydrochloric acid) for 4 min at 350 W, then defrosted for a further 4 min. In the next step, the primary antibody (rabbit anti-GFP antibody, 1:250, Cell Signalling) was added and slides were incubated at 4°C overnight. Counterstaining with the secondary red fluorescent antibody (Alexa Anti-Rabbit 594, 1:250) as well as with 4′,6-diamidino-2-phenylindole (DAPI) (DAPI-Methanol, Roche Diagnostics AG, Basel, Switzerland) to identify GFP mutant cells was then performed. Slides were assessed by fluorescence microscopy with an exposure time of 50 ms for DAPI and 90–130 ms for TXRED. Distribution of GFP mutant cells was recorded in regions showing positive staining for both GFP and DAPI. Negative controls were performed in GFP wild type and GFP mutant muscle cells from hind limbs (see online supplementary material 1, the online supplementary figures 1 and 2).
Stained slides were evaluated by two observers (JR and EN) under a light microscope (Olympus BX51, Olympus AG, Switzerland) equipped with a digital camera (Olympus DP72, Olympus AG, Switzerland). Histological slides were rated for changes affecting the aneurysm wall as well as intra-aneurysmal thrombus using a previously defined scoring method (see online supplementary table III)12 by reviewers blinded to the treatment group. Post-processing was performed using Adobe Photoshop 6 (V13.0, Adobe Systems) and Adobe Illustrator CC (V17.0, Adobe System).
Scores for histological characteristics and macroscopic neointima formation of vital and decellularized aneurysms were compared with the Mann–Whitney U test. Non-parametric values are expressed as median (IQR). Differences in aneurysm volumes and percent packing volumes were determined with the Student's t-test. Parametric values are expressed as mean±SD and 95% CI. A p value of <0.05 was considered statistically significant.
Data analyses and visualizations were performed using Graph Pad Prism statistical software 7 (GraphPad Software, San Diego, California, USA) for Mac.
Aneurysm coiling was successful in all 59 aneurysms with no coil displacements observed during follow-up. Four animals were lost to follow-up. Three animals in the vital group and one animal in the decellularized group died preterm. One animal died with induction of anesthesia, resulting in a total mortality rate of 7% associated with this surgical procedure. Autopsy was performed in all five animals.
Over time, volume measurements showed that coiled aneurysms remained stable in size in both vital and decellularized aneurysms (see online supplementary figure 3). However, a clear trend towards growth versus baseline values was seen on day 7 in the decellularized group (p=0.056), with 2/6 of these aneurysms demonstrating volume increases up to 260% on day 7. This trend became less obvious with advanced maturing aneurysms at later time points on days 21 and 90. Baseline aneurysm volumes were similar between vital and decellularized aneurysms at all time points. There were no differences in aneurysm coil percent packing volume between the groups at baseline: means for vital and decellularized aneurysms were 6.5±1.6% and 6.7±1.4%, respectively (see online supplementary figure 4).
Assessment of macroscopic neointima
Significant differences in the macroscopic evaluation of neointima between vital and decellularized aneurysms first became apparent at 21 days (p=0.002) and remained significant at 90 days (p=0.037). As the distribution in vital aneurysms tapered toward a complete and near-complete neointima formation at day 90, the decellularized aneurysms showed the largest diversification—that is, either complete recanalization (4/6 aneurysms) or complete neointima formation (2/6 aneurysms) (figure 2). A case-by-case analysis of neointima evolution is shown at various time points in vital and decellularized aneurysms (see online supplementary figure 5).
Histological changes of aneurysm wall
Compared with vital aneurysm walls, detailed histological analysis of the decellularized aneurysm wall demonstrated increased inflammation (p=0.022), periadventitial inflammation (p=0.007), and fibrosis (p=0.033) (see) with significant changes starting at day 7. Vital aneurysms showed a continuous decline in its wall and periadventitial inflammation over time, with missing or hardly detectable inflammation at day 90. Decellularized aneurysms maintained an elevated inflammation response in this compartment with poor inflammation suppression up to day 90. Both dissection and hematoma of the aneurysm wall were almost exclusively seen in the decellularized group at all time points. Physiologic wall cellularity remained similar in vital aneurysms throughout the observation period, while no aneurysm wall cells could be detected in the decellularized group (see online supplementary figure 6).
Histological thrombus organization
Detailed histological analysis of the intra-aneurysmal compartment revealed similar levels of neutrophils in the thrombus of vital and decellularized aneurysms on days 3 and 7. The number of neutrophils within the thrombus remained unchanged from day 3 to day 7, declined progressively by day 21, and disappeared at day 90 in vital aneurysms, while it had significantly increased on day 21 (p=0.004) in decellularized aneurysms.
Neointima formation was already more complete and thicker in vital aneurysms than in decellularized aneurysms on day 21 and more clearly differed on day 90 (p=0.001). The α-SMA staining in vital aneurysms showed the tendency of circumferential distribution of α-SMA positive cells along the aneurysm wall or broader spaced coverage along the dome compared with the thrombus organization of decellularized aneurysms. With MT staining, vital aneurysms revealed a denser collagen deposition in the neointima and detectable collagen depositions at the dome by day 21. In contrast, decellularized aneurysms showed insufficient collagen depositions in the neointima and no evidence of collagen at the dome (figure 3).
In α-SMA staining, vital (non-decellularized) aneurysms showed infiltration of the luminal thrombus by α-SMA positive cells (figure 3). This infiltration and subsequent thrombus organization and neointima formation started from the distal thrombus adjacent to the aneurysm wall in the vital aneurysms and followed a centripetal progression at consecutive time points. In contrast, thrombus organization and neointima formation was concentrated at the aneurysm orifice and neck in decellularized aneurysms. Experiments with GFP mutant aneurysms implanted in GFP wild type recipients, and vice versa, demonstrated that the cells organizing the luminal thrombus and neointima originated from the aneurysm wall in vital (non-decellularized) aneurysms and originated from the parent artery in decellularized aneurysms (figure 4).
The results of our study show that preservation of aneurysm wall cellularity significantly benefits healing of aneurysms after coil treatment. Thus, loss of mural cells from the aneurysm wall may explain post-embolization recurrence. Early histological changes after coil embolization are characterized by inflammatory responses that affect the aneurysm wall and intra-aneurysmal compartment in aneurysms with a vital (non-decellularized) and decellularized wall. Aneurysms with vital walls demonstrate the ability to attenuate the inflammatory response, organize the thrombus into a mature neointima, and thereby exclude the aneurysm from the circulation. Aneurysms with decellularized walls show a prolonged inflammatory state with the propensity to undergo wall remodeling toward an increased wall vulnerability, insufficient thrombus organization, and failure of neointima formation.
Ruptured intracranial aneurysms following coil embolization have higher rates of recanalization compared with their unruptured counterparts.3–5 Wall degeneration and loss of mural cells are histopathological findings characteristic of ruptured intracranial aneurysms.14 Our results underline the eminent role of preserved aneurysm wall cellularity toward mature neointima formation following coil treatment, whereas aneurysms with a decellularized wall show a significant tendency in macro- and microscopic neointima transformation failure. Aneurysms with preserved cytoarchitecture demonstrated the ability of SMCs to migrate into the thrombus, proliferate, and participate in collagen production and thereby directly contribute to neointima formation.
Intraluminal cell transplantation of myocytes compensated for the inability of decellularized aneurysms to organize the thrombus.12 Similar results have been described with autologous mesenchymal stem cells attached to platinum coils, which significantly improve histological healing.21 Our histopathological findings of aneurysm healing show similarities with human histological studies after Guglielmi detachable coil treatment of intracranial aneurysms—specifically, an initial migration of inflammatory cells into the intra-aneurysmal thrombus is followed by a centripetal invasion of fibroblasts and transformation of the thrombus into scar tissue.22–24 The importance of the aneurysm wall’s cell contribution to aneurysm thrombus organization with neointima formation has been previously demonstrated in animal studies7 and aligns with our observations in GFP transgenic animals in which SMCs from the aneurysm wall migrated into the thrombus and participated in neointima formation.
One major difference in our series comparing groups of aneurysms with vital and decellularized aneurysm walls was the regulation of the inflammatory response after coil treatment. An increased inflammatory reaction following intracranial aneurysm coil embolization appears to be an initial physiologic response to the intraluminal thrombus.25 26 Initially, both vital and decellularized aneurysms had a similar inflammatory reaction after coil treatment that affected the intra-aneurysmal compartment and aneurysm wall. Although vital aneurysms showed the ability to attenuate the initial immunologic reaction, decellularized aneurysms demonstrated throughout the observation period significantly increased numbers of neutrophils in the aneurysm wall and intraluminal thrombus, with elevated prevalence of wall dissection and wall hematoma. This degenerative reverse wall remodeling predisposes the aneurysm to growth and rupture.11 Although a the increase in aneurysm volume was not statistically significant, a clear trend was observed in decellularized coiled aneurysms at the time when the inflammatory reaction peaked. Several aneurysms demonstrated volume increases of up to 260% of their baseline value. One can postulate that the coil mass provided mechanical protection and prevented further growth and ultimately rupture. Furthermore, in our series, the coil treatment of aneurysms may have stabilized the thrombus by influencing intra-aneurysmal flow dynamics and thereby preventing the aneurysm from repeat cycles of thrombolysis and thrombus formation. Both cycles are associated with the maintenance of a pro-inflammatory state27 and destabilization of the aneurysm wall.11 Histopathological assessment of endothelial cells may provide additional information on aneurysm healing, as early endothelial invasion of the intra-aneurysmal thrombus has been associated with recanalization and recurrences after coil embolization of aneurysms.28
Limitations of study
Although the model used in our study does not represent a true intracranial aneurysm such as the model described by Hashimoto et al,29 it permits the assessment of basic concepts of aneurysm wall biology7 11 12 while representing a highly standardized and reproducible aneurysm model.17 Furthermore, the low morbidity/mortality of the model and relatively low costs make it a suitable tool to test and refine embolization devices that will be tested later in other more complicated and expensive models.30–32 Another limitation of the study is the lack of imaging modalities. While some aneurysms show a stepwise thrombosis, other aneurysms demonstrate continually repeating cycles of clot formation, dissolution, and aneurysm recanalization.11 T2* gradient echo sequences on a 7T MRI scan show a significant correlation with the amount of thrombus formation identified histologically in a rat aneurysm model.33 Repeat imaging modalities using this technique would potentially provide additional information on the natural history of aneurysm healing. Furthermore, high frequency optical coherence tomography, a novel intravascular imaging technology, enables precise and accurate measurement of coverage gaps at the neck of aneurysms in vivo.34 There is evidence that uniform aneurysm neck coverage by intrasaccular devices is critical for aneurysm occlusion. This imaging modality would be beneficial in group stratification and potentially exclude recurrent aneurysms due to insufficient aneurysm neck coverage. In our series we used a standardized coil volume in conjunction with a very standardized aneurysm volume, thereby expecting a homogeneous distribution of orifice coverage between the different groups. Furthermore, we included a firm follow-up modality using macroscopic and histopathological assessment of recurrence and aneurysm healing. The follow-up periods in our series assessed the short-term (days 3 and 7) and long-term (days 21 and 90) aneurysm healing following coil embolization.
Experimental coil-treated saccular aneurysms missing mural cells are incapable of organizing an intraluminal thrombus, are marked by a prolonged inflammatory reaction that affects the intraluminal thrombus and aneurysm wall, and thereby leads to aneurysm wall instability with frequent aneurysm recurrences. Aneurysms with preserved mural cells show the ability of myocytes to migrate into the thrombus, attenuate the inflammatory response, organize the thrombus into a mature neointima, and thereby exclude the aneurysm from the circulation and prevent recurrence. online supplementary figure 6
We thank Pablo Faundez, Alexander Yang, and Zach Folzenlogen for their technical support, and Majlinda Kalanderi for the illustrations.
Contributors All authors made a substantial contribution to the concept and design or analysis and interpretation of data, drafting or critical revision of the manuscript, and final approval of the submitted version of the manuscript.
Funding This work was supported by a research grant from the Research Council of the Kantonsspital Aarau, grant number 1410.000.05.
Competing interests None declared.
Patient consent for publication Not required.
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement Data are available upon reasonable request. All data relevant to the study are included in the article or uploaded as supplementary information.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.