A Novel Rabbit Thromboembolic Occlusion Model

Abstract

Background:

To develop preclinical thromboembolic occlusion model for studying revascularization strategies.

Methods:

Clot analog with barium sulfate was injected into distal aorta in 9 New Zealand white rabbits. Situation of aorta occlusion was compared among fibrin rich (n = 4), red blood cell (RBC) rich (n = 3), and whole blood clot analogs (n = 2) using digital subtraction angiography (DSA). Arterial geometries, histologic features and circumferential stretch of distal aorta in rabbits were compared to common carotid artery in swine and distal internal carotid artery (ICA) in human. Aspiration thrombectomy and mechanical thrombectomy using stent retriever were performed in two rabbits.

Results:

The aortic bifurcation was occluded after a single delivery of clot in 4 cases. It was occluded after second clot injection in 5 remaining rabbits. Fragmentation of RBC-rich clots occurred during clot injection in two cases. Mean diameter of distal aorta and right common iliac artery in rabbits was 3.7 ± 0.4, and 2.8 ± 0.3 mm; mean diameter of human ICA, first and second segment of middle cerebral artery (M1, M2) was 3.6 ± 0.4, 3.1 ± 0.4, and 2.4 ± 0.4 mm, respectively. Arterial revascularization was achieved in both rabbits. Geometric, mechanical and histological factors of distal aorta in rabbit were more close to human distal ICA than swine carotid artery.

Conclusion:

Arterial occlusion can be achieved at the aortic bifurcation in rabbits, which is comparable to human ICA bifurcation. This thrombectomy model has the potential to be used for testing of thrombectomy devices.

Introduction

Achieving single-pass complete revascularization as quickly and safely as possible is the goal of acute ischemic stroke (AIS) treatment. However, less than 40% of patients regain functional independence when treated with fibrinolytic agents [1], and 20% of patients receiving mechanical thrombectomy for large vessel occlusion (LVO) are not revascularized [2]. One major factor that can significantly influence the success of the thrombectomy procedure is the composition of the occlusive clot [3].

In order to improve the effectiveness of current treatment methods for AIS, novel endovascular technologies and pharmacological treatment strategies need to be developed and tested, and related animal models in order to assess device efficacy and safety in an environment which, as much as possible, replicates the biological, physiological and biomechanical situation of an intracranial LVO. In this study, we evaluated a thromboembolic occlusion model in rabbits which has the potential for thrombectomy device and thrombolytic pharmaceuticals testing.

Materials and Methods

Blood clot preparation

Whole blood, fibrin-rich (90% plasma: 10% RBCs) and RBC-rich (10% plasma: 90% RBCs) clot analogs were prepared as previously described[4]. The clot analog was prepared using homogeneous blood from another rabbit was used in 7 rabbits to occlude the distal aortic bifurcation. Autologous blood was used to make clots and occlude distal aorta in other two rabbits after gaining experience from the above 7 rabbits. Barium sulfate (10% w/v) was added to the clot mixture to improve radiopacity [5]. The blood clot mixtures were quickly loaded in 3 cc Luer Lock syringes, which were then spun at 20 revolutions per minute (RPM) in a hybridization incubator for 1 hour at 37°C to prevent clot in-homogeneity. Clot compositions were confirmed from Hematoxylin and Eosin (H&E) staining (Figures 1 A–F).

Figure 1.

A-F. Clot gross photos and histological composition confirmation. Gross photographs of RBC-rich (A), fibrin-rich (B), and whole blood (C) clot with barium sulfate. H&E staining images of RBC-rich (D, 10x magnification), fibrin-rich (E, 10x magnification) and whole blood (F, 10x magnification) clot. Areas of barium sulfate are marked with green arrow. G. Radiographic images before clot injection, showing the aortic bifurcation (white arrow). H. DSA image, showing occlusion of distal aorta after fibrin-rich clot injection (white arrow). I. Plain film after RBC-rich clot injection, indicating opacified clot fragments in bilateral iliac arteries (white arrow). J-L. Same subject before fibrin-rich clot injection (white arrow) (J); road-mapping image showing clot in right iliac artery (white arrow) (K); DSA image after clot removal (L), showing completely patent right iliac artery (white arrow).

Animal procedure

All animal procedures were approved by our institutional animal care and use committee. Nine rabbits were used for this study. Throughout the procedure, the rabbits were intubated and maintained with 2.5-3% isoflurane carried by 100% oxygen after induction through intramuscular injection of ketamine/xylazine. The right common carotid artery (RCCA) was exposed and a 6 French sheath was inserted in a retrograde fashion. A 5 French SOFIA catheter (MicroVention Inc., Tustin, CA) was advanced to the distal abdominal aorta through the sheath. 0.5 cc clot analog was slowly injected into the distal aorta at one time. Abdominal DSA was performed to reveal distal aorta and bilateral iliac arteries before and after clot injection. The clot was removed using a Solitaire stent retriever (Medtronic, Inc., Minneapolis, Minnesota) in one rabbit. In another rabbit, Sofia catheter aspiration (MicroVention Inc., Tustin, CA) was performed for clot removal. Animals were then euthanized via intravenous Sodium Pentobarbital injection and aorto-iliac bifurcation was harvested.

Comparison of human cerebral and rabbit aortic bifurcation vessels

The diameters of distal aorta and bilateral common iliac and external iliac were measured from available DSA images before clot injection. Those diameters were compared to the previously published diameters of human internal carotid artery (ICA) as well as M1 and M2 segments of middle cerebral artery (MCA) [6]. To compare the histological structure of the wall of the distal aorta and MCA-M1, rings of tissues were fixed in formalin, paraffin embedded and sectioned in 5 microns thick slides. The slides were then stained with H&E and scanned with 1.25x magnification (Fig. 2A).

Figure 2.

Comparison of rabbit aorta to human and swine arteries. Histology of M1 (A1), rabbit aorta (A2), and swine carotid (A3). (B) The experimental setup to do tensile test to evaluate the circumferential stiffness and strength of the arteries. The arteries were pulled by two 20-gauge needles at 0.1 mm/s until fracture (yellow ellipses) and the pulling force was measured using a force sensor. (C) Force load-stretch curves of the ICA, M1, M2, rabbit aorta, and swine carotid artery. Scale bar = 1mm.

The mechanical properties (stiffness and strength) of arteries were also compared among human, rabbit, and swine by testing on a customized tensile test machine (Fig. 2B). Human cerebral arteries (ICA, M1, and M2) were harvested from the hospital autopsies of adult human cadavers within 24 hours of death following institutional approval. New Zealand white rabbit aorta and domestic swine carotid were harvested immediately after euthanization. All the arteries were submerged inside 0.9% saline solution and stored in 4 °C. Tensile test was carried out within 24 hours of artery harvest using a customized tensile test machine as described before (7, 8). The tensile test machine includes a translation stage (200cri; Siskiyou) to pull the arteries quasi-statically until fracture at 0.1 mm/s and the tensional force on the arteries was measured by a force transducer (Gamma; ATI Industrial Automation). The arteries were cut into ring segments measured about 6 mm in length and cannulated with two 20-gauge needles with one fixed to the translational stage and the other one fixed to the force transducer. The elongation of the artery rings was recorded using a video camera (α6000; Sony). The stretch and load were calculated from the force transducer and camera measurement using the equations proposed by another group in testing cerebral arteries (9). Briefly, the stretch measures the expansion of the arterial ring circumferentially and the load measures the force load per unit length of the arterial ring. The circumferential stiffness was calculated by finding the force load needed to achieve 10% radial expansion of the arteries. The stiffness quantifies how the arteries will deform upon thrombectomy device load, such as radial distension upon stent retriever expansion, vacuum pressure of the aspiration catheter, and/or the tractional force of the devices. The tensile strength is the peak force loading right before the artery was fractured by tensional load. The tensile strength quantifies how strong the artery is before vessel wall injuries are induced by thrombectomy devices.

For the human artery specimens, we harvested and tested 1 ICA, 1 M1, and 1 M2 from the same brain; for the swine artery specimens, we harvested 1 swine carotid and cut 7 artery rings at different locations along the carotid for testing; for the rabbit artery specimens, we harvested 1 rabbit aorta and cut 3 artery rings at different locations toward the distal aorta for testing.

Statistical analysis

Standard analyses to assess the artery diameter, thickness, stiffness, maximum stretch, and strength and the results are presented with mean and SD. To compare the mechanical properties among human cerebral arteries, swine carotid and rabbit aorta, the Shapiro–Wilk test was first carried out to check if the variables followed a normal distribution. The Wilcoxon rank sum test was used since none of the samples was normally distributed. Statistical significance is indicated by p<0.05.

Results

Anatomical similarities between rabbit aortic bifurcation and human cerebral circulation and thickness

Mean diameters of rabbit distal abdominal aorta, right and left common iliac artery, right and left external iliac artery were comparable to that of human ICA, MCA-M1 and MCA-M2, respectively (Table 1). The swine carotid had a dimeter of 2.2-3 mm. The human cerebral artery had a thickness of 0.44 mm (SD 0.06 mm) while the arterial thickness was 0.30 mm (SD 0.06 mm) and 0.65 mm (SD 0.07 mm) for the rabbit distal abdominal aorta and swine aorta, respectively.

Table 1.

Comparison of Sizes of Human Cerebral Vessels and Rabbit Distal Abdominal Aorta

Human Vessel Diameter (mean ± SD, mm) [6]		Rabbit Vessel Diameter (mean ± SD, mm) [n = 5]
Internal Carotid Artery (ICA)	3.6 ± 0.4	Distal Aorta	3.7 ± 0.4
Middle Cerebral Artery (MCA)-M1 segment	3.1 ± 0.4	Right Common Iliac Artery	2.8 ± 0.3
		Left Common Iliac Artery	2.4 ± 0.3
Middle Cerebral Artery (MCA)-M2 segment	2.4 ± 0.4	Right External Iliac Artery	2.3 ± 0.3
		Left External Iliac Artery	2.1 ± 0.3

Mechanical properties

As the arteries were stretched, the force load on the arteries increased and the slope increased as the arteries were further stretched (Fig. 2C). The load required to achieve a stretch of 1.1 or a radial expansion of 10% was 0.017 - 0.047 N/mm (0.023 ± 0.022 N/mm) for the human cerebral arteries, 0.002 - 0.011 N/mm (0.008 ± 0.005 N/mm) for the rabbit aorta, and 0.009 - 0.028 N/mm (0.017 ± 0.006 N/mm) for the swine carotid. Compared to the stiffness of human cerebral arteries, the differences were not statistically significant for the swine aorta (p = 1) or the rabbit aorta (p = 0.40).

The rabbit aorta had similar maximum stretch and strength compared to human cerebral arteries while the swine carotid was more stretchable and had higher strength than the human cerebral arteries. The maximum stretch of the swine aorta was 2.06 ± 0.13, significantly (p=0.01) higher than the human cerebral arteries (1.42 ± 0.15), while the difference between the human cerebral arteries and the rabbit aorta (maximum stretch of 2.06 ± 0.32) was not significant (p = 0.10). The swine aorta could withstand over three times higher force load before fracture. The tensile strength of the swine aorta was 2.23 ± 0.50 N/mm, significantly (p = 0.01) higher than that of the human cerebral arteries (0.55 ± 0.12 N/mm) and the rabbit aorta (0.86 ± 0.24 N/mm). The tensile strength of the rabbit aorta and human cerebral arteries are not statistically different (p = 0.20).

Clot placement

The aortic bifurcation was completely occluded after single clot injection in 4 rabbits (2 with fibrin rich clot, 1 with whole blood clot, and 1 with RBC rich clot). For the remaining 5 rabbits, the aortic bifurcation was completely occluded following second clot injection (2 with fibrin rich clot, 1 with whole blood clot, and 2 with RBC rich clot). Autologous clot was used for aortic bifurcation occlusion in 2 rabbits (1 with fibrin rich clot, 1 with RBC rich clot). Aortic occlusion was achieved in both rabbits after single clot injection. Fragmentation of the RBC-rich clot occurred during first injection in 2 rabbits (Figures 1 G–I).

Clot removal testing

Clot retrieval was attempted in two rabbits after autologous clot injection. One fibrin-rich clot was removed completely using the stent retriever at the first-pass (Figures 1 J–L) in the first rabbit. RBC-rich clot was completely removed at the first-pass with aspiration in another rabbit.

Discussion

A reliable in vivo model of acute ischemic stroke is essential for testing of thrombectomy devices and pharmaceuticals. Here we offer a novel rabbit thrombectomy model using the aortic bifurcation that demonstrates occlusion of the target vessel and revascularization by standard mechanical thrombectomy approaches. This study corroborates similarities between the diameters of arteries at the distal aortic bifurcation in rabbits and the distal internal carotid artery bifurcation in humans, which measure very close to measurements from a previous study [6]. The occlusion with our blood clot analogs in the aortic bifurcation in rabbits mimics the clinical scenario of LVO in AIS. In addition, our rabbit thrombectomy model is reliable, easy to access from an endovascular approach, is amenable to 6 French guide catheters. Furthermore, it is possible to achieve successful revascularization of both RBC-rich and fibrin-rich clots in this rabbit model. Importantly, our study found that the rabbit aortic bifurcation bared close resemblance to the human MCA bifurcation in its biomechanical properties, more so than the swine carotid.

A number of models have been proposed for the study of mechanical thrombectomy in AIS. The most widely used is the swine model (10–11). However, the presence of a large rete mirabile limits catheter access to the cerebral circulation. In addition, our mechanical testing also showed that the widely used swine aorta is more stretchable and stronger compared to human cerebral arteries, which could underestimate the vessel damage and reduce the merit of the testing results. In comparison, the rabbit aorta, although more stretchable as well, has similar tensile strength compared to human cerebral arteries. This indicates the rabbit aorta might more realistically replicate vessel wall injury during thrombectomy procedures. In addition, the rabbit aorta also has non-significantly different stiffness compared to the human cerebral arteries, making it an ideal candidate to replicate the vessel deformation under vacuum suction and catheter/stent pull during thrombectomy. Researchers also reported a canine MCA model [12], which can accommodate large bore catheters and can be used for device testing. However, the canine model does have limitations including cost, the presence of profound arterial tortuosity and the smaller size of canine MCA. Therefore, it may not be the most suitable model for in vivo study of large bore aspiration catheters or stent-retriever devices. Furthermore, other thrombectomy models such as the in silicon and in vitro models do not recapitulate hemodynamic, physiologic and biologic characteristics of the vascular system including the biological and physical interactions between occlusive emboli and the vessel wall.

Researcher from other groups reported the use of clot to occlude the carotid artery in rabbits to create vessel occlusion model (13–14). However, since the diameter of the carotid artery is much smaller than distal aorta, the carotid occlusion rabbit model may not resemble human ICA bifurcation occlusion as well as the aortic bifurcation model in this study. SOFIA Plus catheter cannot be put in the rabbit carotid artery for thrombus aspiration. Importantly, we were able to demonstrate the feasibility of performing revascularization experiments with different devices (stent retrievers and aspiration catheters) and with different types of clot analogues (RBC rich, fibrin rich). This is important given the growing interest in studying the interaction between clot composition and physical characteristics and revascularization outcomes with different devices (15–17). A recent publication indicates that large artery atherosclerosis (LAA) clots are larger, soft and associated with a large red blood cell-rich extracted clot area; while cardioembolic clots are smaller and stiffer with higher fibrin and platelets/ other content than LAA clots. Cryptogenic clots are similar to those of cardioembolic clots (18). Clot perviousness is associated with first pass angiographic success in patients treated with the aspiration first approach but not with stent retriever for thrombectomy. Higher perviousness value, which is associated with higher RBC density and lower fibrin density (17), was correlated to higher revascularization rate and clinical outcome (19). This model has potential for the study of blood clots with various compositions and perviousness.

There are several limitations of this model. Clots needed to be injected repeatedly in order to achieve stable occlusion in some cases. Also, the addition of Barium Sulfate might impact the biomechanical properties of the clot analogs. Surgical procedure is required for access, and there is limitation of access diameter. It will be challenging to put 8F balloon guide catheter (BGC) for access, which makes it difficult to test large bore aspiration systems. There is only 1 location for thrombectomy per experiment. The aorta is relatively straight and not representative of the human intracranial circulation (or aortic arch), which might lead to overestimation of efficacy in this model.

Conclusion

Arterial occlusion can be achieved at aortic bifurcation in rabbits, which has similar geometric, mechanical, and histological features as human ICA bifurcation. Our animal model has the potential to be used for testing of thrombectomy devices.