A Novel Rat Model of Embolic Cerebral Ischemia Using a Microcatheter and Zirconia Ball Under Fluoroscopic Guidance

Background and Purpose: The failure of neuroprotective treatment-related clinical trials may be partially caused by unestablished animal models. Existing animal models are less likely to provide occlusion conned to the middle cerebral artery (MCA), making transarterial intervention dicult. We aimed to develop a novel focal stroke model using a microcatheter and zirconium dioxide that is non-magnetic under uoroscopic guidance, which can monitor MCA occlusion and can improve hemorrhagic complications. Methods: Using male Sprague Dawley rats (n=8), a microcatheter was navigated from the caudal ventral artery to the left internal carotid artery using digital subtraction angiography to establish local occlusion. Results: All rat cerebral angiographies were successful. No rats had hemorrhagic complications. Six (75%) rats underwent occlusion of the MCA bifurcation by zirconium dioxide. Accidentally, the left posterior cerebral artery was failure embolized in 2 rats (25%) via the posterior communicating artery. The median operating time was 8.5 min. Moreover, 2% 2,3,5-triphenyl tetrazolium chloride staining showed that the median infarct volume (mm 3 ) was 280 (interquartile range; 267-333) 24 hours after the left MCA bifurcation occlusion. Conclusions: We present a novel rat model for focal stroke using a microcatheter and zirconium dioxide which does not affect the MRI. The model is predictable and reproducible, which is well conned within the territory supplied by the MCA. Fluoroscopy was able to identify which the MCA occlusion and model success while creating the model. It permitted exclusion of animals with complications from the experiment.


Introduction
Stroke is the second most common cause of death and the third most common cause of disabilityadjusted life years lost worldwide 1 . In most patients, ischemic stroke results from occlusion of the middle cerebral artery (MCA). As such, rodent models have been developed to mimic human focal ischemic stroke 2 . The most widely used rat model for focal cerebral ischemic stroke involves an intraarterial suture occlusion of the MCA (MCAO), in which a nylon thread is blindly inserted from the external carotid artery (ECA) 3,4 . The main limitation of this model is that blood ow to the posterior communicating artery (PcomA) and branches of the internal carotid artery (ICA) are obstructed, resulting in large and variable infarctions of the cortical and subcortical areas [5][6][7][8][9][10][11] . Besides, the procedure is very invasive because the ECA is ligated after the neck incision. Ligation of the ECA results in ischemic necrosis of the mastication and hypopharyngeal muscles, which negatively affects behavioral testing outcomes 12,13 . There are additional models for inducing stroke, all with merits and limitations (Table   1) 3,4,[14][15][16][17][18][19] .
There has been growing interest regarding the use of regenerative medicine for sequelae of cerebral infarctions, and clinical trials on intravenous administration of bone marrow mesenchymal cells have commenced. However, intravenous administration may result in the medicine being taken into organs other than the target, such as the lungs, thereby reducing the therapeutic effect. Therefore, it is necessary to urgently develop a highly reproducible animal model that can verify intra-arterial cell administration for cerebral infarctions.
Here, we present a highly selective rat MCAO model by percutaneous caudal arterial puncture using a microcatheter and zirconia ball under uoroscopic guidance. The model is predictable and reproducible, which is well con ned within the territory supplied by the MCA, and the surgical procedures can be typically completed within 10 min. Furthermore, minimally invasive and highly reproducible procedure lead to a "reduction" in the suffering and "re nement" of the welfare of laboratory animals.

Materials And Methods
The authors declare that all supporting data are available within the article.

Animal model and protocol
Male normotensive adult Sprague Dawley rats (n = 8; 11-27 weeks of age; 366-587 g on the day of surgery) obtained from Nihon SLC (Japan SLC, Inc. Shizuoka, Japan) were housed in polycarbonate cages under temperature-controlled conditions (temperature: 24-25°C; relative humidity: 50-60%) with a 12-h light-dark cycle. All rats had free access to water and pelleted food (CE-2, CLEA Japan, Inc., Gifu, Tokyo, Japan). Our study was approved by the Institutional Animal Care and Use Committee of the Jikei University School of Medicine (protocol number: 2016-105). All procedures were conducted according to the Fundamental Guidelines for Proper Conduct of Animal Experiments and Related Activities in Academic Research Institutions issued by the Japanese Ministry of Education, Culture, Sports, Science, and Technology. This study doesn't need the control group because of making stroke model study.
Anesthesia was maintained with 1-3% iso urane through a vaporizer for small experimental animals and a facial mask. Body temperature was maintained during surgery at 36-37.5°C using a multi-panel heater (Vivaria, Osaka, Japan). The animal's body, including the entire tail, must be kept warm, as previously reported. 20 No preparation, such as shaving, disinfection, or antibiotics, is required.
Major arteries of rat with microcatheter and zirconia ball and a simpli ed operation scheme are shown in Cord, Tokyo, Japan) was rst ushed with heparinized physiological saline. The rat was placed in a supine position with a stretched tail. Indwelling needle (venous indwelling needle for human, 22 gauge, SR-OT2225C, TERUMO, Tokyo) was inserted through the ventral midline artery approximately 5 cm from the root of the tail of the rat ( Fig. 2A). After con rmation of arterial blood back ow from the indwelling needle, a microcatheter and wire (OD 0.4mm, FGW16-AG18S30, Toray Medical, Tokyo, Japan) were inserted as long as no resistance was felt. Upon resistance, the insertion of the microcatheter was stopped immediately, and the position of the microcatheter was con rmed using uoroscopic imaging.
The wire and microcatheter were guided from the caudal ventral artery to the abdominal aorta (Fig. 2B), the aortic arch, and the left common carotid artery (CCA) (Fig. 2C) using uoroscopic imaging. Then, the wire was carefully removed so as not to change the position of the microcatheter. The contrast media (Iohexol, Daiichi Sankyo, Tokyo, Japan) at half concentration was rapidly injected from the microcatheter. Then, left CCA angiography was performed, in which 0.1 ml of the contrast media was injected using a digital subtraction angiography unit (Artis Zee, Siemens, Germany; Fig. 2D). The bifurcation of the ICA branch was visualized by tilting the C-arm at an LAO angle of 80° (Fig. 2D). The contrast images were overlaid for improving visualization, and the microcatheter was guided to the branch of the left ICA. Left ICA selective angiography could then detect the left anterior cerebral artery (ACA), MCA, and PcomA (Fig.  3A). Then, one zirconia ball (zirconium dioxide, 0.4 mm in diameter; Nikkato, Tokyo, Japan. Fig. 4A) was advanced in the microcatheter by slow injection of approximately 0.1 ml of heparinized physiological saline (Supplemental File -Videos 1 and 2), and the ACA-MCA bifurcation was selectively embolized (Figs. 3B, 4B). The zirconium dioxide does not affect the MRI because of non-magnetic material. The high visibility of zirconia ball during uoroscopy was found to be comparable to that of wire. The microcatheter was inserted into the right CCA and contrasted to con rm that the left ACA had retrograde blood ow from the right ACA and only obstructed the left MCA in the frontal view (LAO angle 0°) (Fig.  3C). The time from the puncture to model creation was measured. The plastic outer indwelling needle was xed with sealing lm (Para lm, Bemis, United States of America) and gaffer tape and was not removed until 24 hours after the embolization. Then, cerebral angiography was performed to con rm whether spontaneous recanalization had occurred. No postoperative pain relief was required.
Cerebral angiography 24 hours after the creation of the stroke model Cerebral angiography was performed 24 hours after the creation of the stroke model to determine whether spontaneous recanalization had occurred. Cerebral angiography could be performed in the same way as when the model was created without re-puncture by using the plastic outer indwelling needle that was left behind.
Magnetic resonance imaging (MRI) analysis MRI was performed using a 9.4-T BioSpec 94/20 (Biospin GmbH, Ettlingen, Germany) device and a transmitting coil with a 72-mm inner diameter and a rat brain size 4 channel receiver coil immediately before the rats were euthanized. T2-weighted imaging and MR angiography were performed using the time of ight theory. T2-weighted imaging was performed using the rapid acquisition with a relaxation enhancement pulse sequence 21  166×148×180, and image resolution: 122×122×122 μm. The total scan time of the three-dimensional fast low angle shot pulse sequence was 36 minutes 20 seconds. MR angiography was reconstructed using the maximum intensity projection method in ParaVision version 6.0.1 software (Bruker, Ettlingen, Germany). To acquire MRI data, the rats were scanned in the prone position on an imaging stretcher and administered a mixture of air and 1.5-3.0 % concentrated iso urane (Abbott Laboratories, Abbott Park, IL, USA). Respiration was regularly monitored during the scanning to manage the animal's physical condition.

Postmortem analysis
Twenty-four hours after embolization, the animals were deeply anesthetized and euthanized. The brains were removed and were sectioned coronally into 6 slices (thickness, 2 mm) from the olfactory bulb to the cerebellum and then stained with 2% 2,3,5-triphenyl tetrazolium chloride (TTC) for 20 minutes. Using image analysis software (ImageJ, version 1.53a), the non-infarcted area of the ipsilateral hemispheres and the areas of the contralateral hemisphere were calculated on the rostral side and caudal side, respectively.
Infarct volume on each section (mm 3 ) = (areas of the contralateral hemisphere -the non-infarcted area of the ipsilateral hemisphere) on the rostral side (mm 2 ) × 1 (mm) + (areas of the contralateral hemispherethe non-infarcted area of the ipsilateral hemisphere) on the caudal side (mm 2 ) × 1 (mm) The sum of each section volume is considered to be the total volume of infarction 23,24 . A postmortem analysis was performed by one researcher (H.M.), who was blinded from details of the surgical procedures.

Results
All rat cerebral angiographies were successful. There no exclusions. No rats had hemorrhagic complications including a subarachnoid hemorrhage (SAH). Six (75%) rats underwent occlusion of the ACA-MCA bifurcation by zirconia ball (Figs. 3B, 3C, 4B, Supplemental File -Video 1). Accidentally, the left posterior cerebral artery was failure embolized in 2 rats (25%) through the PcomA (Fig. 3D, Supplemental File -Video 2). Fluoroscopy was able to identify which the MCAO and model success while creating the model. The median operating time was 8.5 min (interquartile range [IQR]; 6-14.25 min). The maximum contrast material dose was up to 1 ml, and the intraoperative blood loss was less than 1 ml. No cases of spontaneous recanalization 24 hours after the embolization occurred. No tails had ischemia because the 2 lateral caudal arteries took over the supply through numerous anastomoses. There were no deaths within the 24 hours before the postmortem analysis. The zirconia ball did not affect the MRI (Fig. 5).
Furthermore, 2% TTC staining showed that the median infarct volumes (mm 3 ) were 280 (IQR; 267-333) 24 hours after the left MCA bifurcation occlusion (Fig. 6A). Twenty-four hours after the left PcomA occlusion, the stroke lesion volume was 251 mm 3 and 446 mm 3 (Fig. 6B). There was no infarction of the thalamus and hypothalamus in MCAO model.

Discussion
We have presented a novel zirconia MCAO rat model that is minimally invasive and highly reproducible. To the best of our knowledge, this is the rst study to present a highly selective MCAO model by percutaneous caudal arterial puncture using zirconia ball and microcatheter under uoroscopic guidance. Fluoroscopy was able to identify which the MCAO and model success while creating the model. It permitted exclusion of animals with complications from the experiment. The advantages of the zirconia MCAO rat model for experimental studies are discussed below and are summarized in Table 1.
First, this novel model can create an infarct only in the MCA territory without ischemia of the ECA, the PcomA, and branches of the ICA, and can also identify the site of vessel occlusion intraoperatively. All existing models of MCAO except caudal artery approach share the same disadvantages. Ligation of the ECA and insertion of the suture occludes the entire course of the ICA, leading to obstruction of the ECA, the PcomA, and branches of the ICA, such as the hypothalamic and anterior choroidal arteries. Ligation of the ECA results in ischemic necrosis of the mastication and hypopharyngeal muscles, which negatively affects behavioral testing outcomes 12 . The PcomA and branches of the ICA are the main arteries to the thalamus, hypothalamus, hippocampus, and substantia nigra [25][26][27] . As such, because these important areas are infarcted, the animal have a variable level of increased body temperature, disturbed water homeostasis, severe paresis, and high postoperative mortality, which altogether can result in bias in experimental studies [5][6][7][8][9][10][11] . The macrosphere and suture models lead to reduced blood ow in the ECA and anterior choroidal artery 14,15 . The zirconia MCAO model can reduce inter-animal variability and mortality by selectively causing ischemia only in the MCA blood ow area.
Second, the zirconia MCAO model is minimally invasive and safe as it only requires a caudal artery puncture and does not require a craniotomy and cervical incision (Table 1). Moreover, in our study, there was not a single case of hemorrhagic complications. SAH occurred in 8%-30% of animals in the intraluminal suture MCAO model (Table 1). [16][17][18] SAH can lead to serious complications such as vasospasm and intracerebral hemorrhage, and thus can confound the results of experimental stroke studies. If SAH occurs, the zirconia MCAO model can detect SAH while creating the model using uoroscopy. Minimally invasive and safe methods can prevent the possible loss of the experimental animal, and lead to a "reduction" in the suffering and "re nement" of the welfare of laboratory animals. 28 Third, the zirconia MCAO model uses uoroscopy, which means that vessel occlusion and model success can be checked for in real-time while creating the model. This is particularly important for treatment intervention studies using animal models because it avoids performing an intervention with failed models.
Finally, by placing a plastic outer indwelling needle in the caudal artery, cerebral angiography can be repeated. This method can also be used for arterial drug administration, which can be repeated in a highly selective and direct manner to the target artery. Multiple doses of the drug or cells can be administered to the target organ at the time of therapeutic intervention, thereby enabling the establishment of more sophisticated and complex therapeutic intervention studies. Endovascular therapy is already commonly used in humans and is easy to apply clinically.
Our study has several limitations. First, the model failure rate was 25% due to the occlusion of PcomA instead of MCA. Second, left ACA blood ow was dependent on the right ACA instead of the left ICA. However, no infarcts in the ACA region appeared, so this does not appear to be an issue. Third, some physiological data, such as blood gases and animal body temperatures, were not measured. Finally, the model cannot induce reperfusion. Any of the current preclinical model of stroke including the current one in this study has caveats. We think multidirectional approach using different preclinical models can help to better understand the pathophysiology of stroke.

Conclusions
In conclusion, the zirconia MCAO model presented here has several advantages. It is minimally invasive, and uoroscopic guidance allows for the selective endovascular occlusion of the MCA. The technique also avoids ischemia of the ECA, anterior choroidal artery, hypothalamic artery, and PcomA. This is important because the confounding effects of tissue damage outside the MCA territory in the MCAO model are considerable. From an ethical standpoint, this model signi cantly limits the suffering caused by complications secondary to large, unspeci c infarct lesions, surgical exploration of the neck, and ligation of the ECA. The microcatheter that carries the microwire can also serve as a working channel wherein therapeutic agents may be administered in a targeted fashion. We hope that this approach will be carried out by many researchers, leading to a "reduction" in the suffering and "re nement" of the welfare of laboratory animals. The development of the zirconia MCAO model also helps reduce costs associated with investing in unnecessary clinical trials, the death of laboratory animals, and ineffective drug discovery.