Bilateral carotid occlusion coupled with systemic hypotension produces global brain ischemia in the rat, resulting in damage to the hippocampus with reproducible severity. Animal subjects are impaired with predictable patterns of brain damage, they recover expediently, and mortality rates are comparatively low.
Cardiac arrest followed by resuscitation often results in dramatic brain damage caused by ischemia and subsequent reperfusion of the brain. Global brain ischemia produces damage to specific brain regions shown to be highly sensitive to ischemia 1. Hippocampal neurons have higher sensitivity to ischemic insults compared to other cell populations, and specifically, the CA1 region of the hippocampus is particularly vulnerable to ischemia/reperfusion 2.
The design of therapeutic interventions, or study of mechanisms involved in cerebral damage, requires a model that produces damage similar to the clinical condition and in a reproducible manner. Bilateral carotid vessel occlusion with hypotension (2VOH) is a model that produces reversible forebrain ischemia, emulating the cerebral events that can occur during cardiac arrest and resuscitation. We describe a model modified from Smith et al. (1984) 2, as first presented in its current form in Sanderson, et al. (2008) 3, which produces reproducible injury to selectively vulnerable brain regions 3-6. The reliability of this model is dictated by precise control of systemic blood pressure during applied hypotension, the duration of ischemia, close temperature control, a specific anesthesia regimen, and diligent post-operative care. An 8-minute ischemic insult produces cell death of CA1 hippocampal neurons that progresses over the course of 6 to 24 hr of reperfusion, while less vulnerable brain regions are spared. This progressive cell death is easily quantified after 7-14 days of reperfusion, as a near complete loss of CA1 neurons is evident at this time.
In addition to this brain injury model, we present a method for CA1 damage quantification using a simple, yet thorough, methodology. Importantly, quantification can be accomplished using a simple camera-mounted microscope, and a free ImageJ (NIH) software plugin, obviating the need for cost-prohibitive stereology software programs and a motorized microscopic stage for damage assessment.
Brain damage as a consequence of cardiac arrest and stroke is a leading cause of death and long-term disability. While cardiopulmonary resuscitation for victims of cardiac arrest succeeds in restoring spontaneous circulation in about 70,000 patients per year in the US 7,8 at least 60% of these patients subsequently die in the hospital as a result of extensive brain damage and only 3-10% of resuscitated patients can resume their former lifestyles 9,10 . Clearly, understanding the mechanisms that lead to brain damage following global brain ischemia and designing therapeutic interventions to minimize neurologic trauma is of critical importance.
Brain ischemia can be modeled utilizing multiple methods. Most commonly, brain ischemia is produced in the rodent by occluding a major blood vessel in the brain, the middle cerebral artery, thus producing a focal ischemic stroke 11,12 . While clinically important, focal brain ischemia is not an accurate method to study brain damage produced by cardiac arrest/resuscitation. To model this clinical paradigm the entire brain must be made ischemic followed by reintroduction of blood flow. To closely mimic this clinical presentation, investigators experimentally induce cardiac arrest followed by resuscitation with CPR and defibrillation 13,14 . This model is clinically relevant, however unpredictable resuscitation times can increase variability and may make data analysis difficult to interpret. Additionally, this model is associated with a high mortality rate, further increasing animal number necessary to test a hypothesis. Investigating the cerebral response to global ischemia and/or reperfusion in a more reproducible, consistent, and survivable insult may be preferred.
Global ischemia can be induced in the brain while preserving some blood flow systemically. This reduces mortality, while allowing investigation of the mechanisms of tissue damage in the brain 2. To produce global brain ischemia, it is necessary to interrupt or greatly limit flow in all four vessels that supply the brain, the internal carotid arteries and vertebral arteries. These vessels supply the brain with blood flow through a vascular structure called the Circle of Willis, which forms an anastomotic loop. This vascular architecture allows the brain to retain perfusion in the event of proximal vascular occlusion. Therefore, to induce complete ischemia of the brain, blood flow through all contributory vessels must occur. Carotid artery occlusion can be accomplished using a minimally invasive ventral neck cut-down and application of aneurism clips for a desired period. Interruption of blood flow via the vertebral arteries can be difficult, as they are incased in the transverse foramina of the vertebral column. Investigators have addressed this by electrocauterizing the vertebral arteries 24-48 hr prior to carotid occlusion and brain ischemia (4VO model) 15. In contrast to this approach, Smith, et al. developed a method of inducing global brain ischemia by reducing mean arterial blood pressure (MAP) systemically to 40 mmHg to reduce perfusion through the vertebral arteries to a point where blood flow is lost or greatly reduced 2. When coupled with carotid occlusion, this method produces ischemia throughout the forebrain, resulting in a pattern of brain damage that closely mimics that of cardiac arrest survivors. In a further refinement of this method, the model we present here requires tight MAP regulation at 30 mmHg ± 1mHg during the entire 8 min of ischemia. We found this alteration improves the reproducibility of brain damage induced by this model while preserving the low mortality rate of the original technique designed by Smith et al.
The precise phenotype of cell death and overall extent of tissue damage caused by the model presented here are directly dependent on ischemic duration 16. Following 8 min of ischemia, CA1 neurons exhibit delayed cell death, suggesting that there is a temporal window for therapeutic intervention during the reperfusion phase 15,17 . At the onset of reperfusion, neurons quickly regain function and no immediate cell death is detectable 18. However this insult causes induction of cell death cascades (apoptosis) that culminate in release of apoptogenic proteins from mitochondria, including cytochrome c, between 4-6 hr of reperfusion 3,19. Between 6 and 24 hr of reperfusion, neurons of the CA1 hippocampus have committed to cell demise, and the apoptotic cell death program is executed 19. It should be noted that cell death phenotype responsible for ischemic injury is highly controversial. Early studies have suggested necrosis is the primary cell death phenotype 20,21 , while other others report apoptosis as the principal mechanism 22,23 . In total, current evidence suggest that cells die of a spectrum of cell death phenotypes ranging from classic apoptosis to necrosis. The specific mode of cell death is dependent on many factors, with the degree of contribution of each phenotype depending on the severity of the insult, amongst other factors 24,25. By 24 hr of reperfusion, dying cells possess pyknotic nuclei, condensed cytosol with clear evidence of aggregated cellular contents, and loss of functional mitochondrial morphology. Dead cells are further broken down, engulfed by immune cells such as macrophages and/or microglia, and cleared from the CA1 hippocampal region. By 4-7 days of reperfusion, dead cells are removed, and all that remains are inflammatory cells and activated glial cells 17,26 . Therefore, 7 days of reperfusion represents an optimal time where CA1 hippocampal neuronal death can be quantified using simple, non-specific cell stains including Cresyl violet or hemotoxylin-eosin and counted based on morphologic inclusion criteria. Cells remaining at this late reperfusion interval can be counted as surviving cells, thus providing an index of brain damage.
If this model is to be utilized to test therapeutic interventions, it is suggested that the experimental design follow STAIR criteria (Stroke Therapy Academic Industry Roundtable) 27. These guidelines should be followed when designing and conducting a study, however are not discussed here.
1. Preparation
All animal experiments must conform to institutional guidelines and receive approval by a respective animal care committee prior to initiation. All procedures presented here have been approved by the Wayne State University Institutional Animal Care and Use Committee and follow the guidelines on the ethical treatment of animals as put forth in the Guide for the Care and Use of Laboratory Animals and to the US Government Principles for the Utilization and Care of Vertebrate Animals Used in Testing, Research, and Training. Before beginning surgery, prepare necessary surgical material and a surgery recovery cage. This procedure is a survival surgery and it is therefore necessary to practice sterile technique.
2. Surgical Preparation
Sprague-Dawley rats (300-350 g) are used in the 2VOH model of global forebrain ischemia. Anesthesia is induced and maintained with isoflurane and supplemented with a sub-dissociative dose of ketamine. We have observed that anesthesia maintained by isoflurane alone causes an increased occurrence of seizure during post-op recovery. Seizures can contribute to neural damage and mortality, which will impact the reproducibility of this model. Supplementing anesthesia with ketamine allows a much lower dose of isoflurane to reach a surgical grade of anesthesia. Using a homeostatic thermal blanket allows strict regulation of core temperature. The surgical procedure entails carotid artery cut-down and isolation, and femoral artery cut-down and cannulation.
Note: It is important to keep detailed records during each individual surgical procedure. Changes in core temperature, blood pressure, or other physiologic variables that occur prior to or during surgery may drastically alter results, and records can be analyzed to ensure there is procedural consistency between individuals and groups. Consistency in the physiologic parameters of the animal can be improved by fasting the animals prior to surgery. Experimenters are encouraged to determine the method that results in the most consistent pre-operated hemodynamics and serum glucose concentration for their studies.
3. Ischemia Protocol
As mentioned previously, four vessels supply perfusion to the brain. Proximal occlusion of the two carotid arteries will not result in brain ischemia because the vertebral arteries will compensate through the Circle of Willis. It has been shown that carotid occlusion, coupled with induced systemic hypotension, will limit perfusion through the vertebral arteries resulting in brain ischemia. Here we describe the protocol for withdrawing blood to produce hypotension and clamping the carotid arteries to produce controlled, reversible ischemia. See Figure 1.
Blood oxygen and carbon dioxide, glucose concentration and pH can vary among individuals in the sample group and cause variation in the injury. Monitoring these physiological variables can reduce variability of the infarction. As mentioned, temperature is another parameter which is important to regulate, however brain temperature may not correspond to core temperature, as perfusion is greatly restricted during experimental ischemia28. Our specific surgical setup results in temperature changes in the brain that mirror changes in core temperature during ischemia. However, it is important for the experimenter to determine this empirically by monitoring brain temperature with thermocouples or by other means to standardizing the procedure. If core temperature does not mirror brain temperature, it is important to maintain the brain temperature normothermic during the entire procedure, irrespective of core temperature.
Randomization dictates that the surgeon randomizes the animal to ischemia or sham-operated control group at this point. Sham surgery will follow all procedures exactly the same as ischemic animals excluding any reduction in blood pressure and carotid artery occlusion. It is important that the sham-operated controls be under the same dose of anesthesia for a duration that matches that of the ischemic animals. If there is a treatment group included, which requires anesthesia before or after ischemia, sham and untreated ischemia groups should match the anesthesia dose and duration of the treatment group.
4. Analgesia and Recovery
Animal welfare is the priority during recovery as well as the surgical procedure. Post-operative care should follow the specific guidelines of the institution. Animals that undergo global brain ischemia are susceptible to seizures. To minimize seizure occurrence it is important that the recovery room have minimal stimulation, i.e. noises and visual disturbances. Under our animal use protocol, we house post-operative animals for 72 hr in a secluded room for this purpose.
Note: Any signs of pain or distress are immediately alleviated with analgesics. If sympotoms are not alleviated with one dose of analgesics or persist after the fourth consecutive dose of analgesia, the animal will be euthanized and excluded from the study. Surgical wounds should be inspected for proper healing for eight days post-op.
5. Tissue Collection and Processing
The brain is fixed by transcardial infusion of paraformaldehyde. After gross sectioning and cryoprotection, we snap freeze the brain and section on a cryostat. These brain sections are stained with Cresyl violet and imaged on a light microscope.
6. Brain Damage Quantification
ImageJ, a free program offered by the National Institute of Health, can be used to count and quantify the viable neurons, which will represent the extent of brain damage.
The 2VOH model of global brain ischemia/reperfusion causes neuronal death in the CA1 region of the hippocampus. Figure 2 represents the injury produced by 8 min of global brain ischemia, processed 14 days after reperfusion. Figures 2A and 2B compare the hippocampi from sham and post-ischemic brains, stained with Cresyl violet. Figure 2A shows a hippocampus from a sham-operated rat which exhibits normal morphology, including an intact CA1. Figure 2B demonstrates the hippocampus subject to ischemia/reperfusion, where the dentate gyrus, CA2 and CA3 have minimally affected morphology. The selectively vulnerable neurons of the CA1 hippocampus do not survive ischemia/reperfusion injury and only scattered neurons remain. However, dramatic increases in infiltrating macrophages and/or microglia (Iba-1-positive cells), and reactive astrocytes (GFAP-positive cells) are evident in the CA1 hippocampal field. Specific immunofluorescent labeling of neurons with NeuN, macrophage/microglia labeling with Iba-1, and astrocyte labeling with GFAP corroborate these findings.
Figure 2C presents sample neuron counting windows for microscope images of the CA1 at 40x magnification. The top panel shows CA1 from a sham-operated control and the lower panel shows the CA1 following ischemia/reperfusion. The injured CA1 displays a staining that includes microglia and astrocytes, which appear small and tubular or teardrop shaped. These can be distinguished from neurons by their shape. Intact hippocampal neurons, whether in the dentate gyrus, CA2 or the CA1, are large with a circular or pyramidal shape.
Figure 2D is a graphical analysis of the 2VOH induced injury quantified by CA1 neuronal counts. It can be concluded that 8 min of ischemia produced by the 2VOH model can cause a consistent and reproducible injury that is primarily isolated to the CA1 hippocampus.
Figure 1. Experimental Timeline. The experimental timeline is a representation of the surgical steps and tissue processing.
Figure 2. Representative Results. Hippocampal damage in an untreated rat subjected to I/R (B) compared to a sham-operated control rat (A). Cresyl violet stained hippocampus (10X) [Surround] 40X magnifications of the CA1, CA2, CA3, hilus, and DG (top leftà clockwise). (B) Damage is localized to the CA1 hippocampus. (C) Representative counting windows used for neuron counting and damage quantification in a sham-operated control animal (top, control) and an animal exposed to global brain ischemia (bottom, I/R). (D) Representative quantitation of CA1 hippocampal neuron counts from an unpublished study (Mean + Standard Deviation; n = 8, p<0.05).
The model described here produces an ischemic insult to the brain that can occur as a result of cardiac arrest and resuscitation, providing an injury similar to that found in humans. This method for producing global brain ischemia is one of multiple protocols. We utilize this protocol foremost for its comparably low mortality rate, rapid recovery, and reproducible results. The cardiac arrest/resuscitation model is arguably the most clinically relevant model, however technically the most difficult to constantly reproduce. The 4VO model of global brain ischemia is another commonly used protocol, valuable in the fact that all four contributing vessels are occluded during ischemia. This model also has drawbacks, which include multiple surgeries and possibility of neuroprotective preconditioning. The 4VO model has been adapted, allowing transient occlusion of all four vessels during one procedure to produce ischemia, however the surgery remains technically demanding 30.
Variations on 2VOH have been shown to produce a range of results 31,32 . Duration of ischemia and degree of hypotension are the major variables that can affect outcome. Reports have shown that if blood pressure is not reduced enough the injury produced can be inconsistent or unilateral in rats 30. We have found 30 mmHg to be an optimal degree of hypotension because it produces reliable ischemia without causing notable peripheral injury. This paradigm produces ischemia characterized as severe, however following our protocol will reduce the otherwise increased chance of complications. The severe nature of the insult may be the reason slight variation in physiological parameters between animals does not typically affect injury consistency.
The authors have nothing to disclose.
Material Name | |||
5-0 VICRYL suture, reverse cutting | Ethicon | J391H | |
Scalpel, No.10 | Swann-Morton | 6601 | |
Gauze Sponges | Fisher | 22-362-178 | |
18G x 1 ½ in needle | BD | 305201 | |
23G x 1 in needle | BD | 305145 | |
26 G x 3/8 in needle | BD | 305110 | |
18 G x 1 ¼ catheter | EXEL | 26735 | |
1 ml syringe | BD | 309659 | |
10 ml syringe | BD | 309604 | |
60 ml syringe | BD | 309653 | |
Surgilube | Henry Schein | 1152666 | |
.9% Saline, plastic IV bag | Henry Schein | 1537468 | |
Suture 3-0 Silk | Henry Schein | 1007842 | |
Puralube Ophthalmic Ointment | Henry Schein | 3390017 | |
Betadine | Henry Schein | 6903564 | |
Sterile Towel Drape | Moore Medical | 14170 | |
Polyethylene Tubing, 50 | Intramedic | 427411 | |
Stopcock, 3 way | Smiths medical | MX9311L | |
Drug Name | |||
AERRANE (isoflurane) | Henry Schein | 2091966 | |
Mapap Liquid (Tylenol) | Major Pharmaceuticals | 1556 | |
Kedavet (ketamine) | Ketathesia Butney | NDC 50989-996-06 | |
Butorphic (butorphanol) | Lloyd Labs | 4881 | |
Heparin | APP Pharmaceuticals | 504011 | |
Chemical Name | |||
Paraformaldehyde prills | Elecron Microscopy Sci. | 19202 | |
2-methylbutane | Sigma | 270342 | |
Cresyl Violet Acetate | Sigma | C5042 | |
Sucrose | Sigma | S9378 | |
Software | |||
ImageJ | NIH |