Preclinical models of intracerebral hemorrhage are utilized to mimic certain aspects of clinical disease. Thus, mechanisms of injury and potential therapeutic strategies may be explored. In this protocol, two models of intracerebral hemorrhage are described, intrastriatal (basal ganglia) injections of autologous blood or collagenase.
Intracerebral hemorrhage (ICH) is a common form of cerebrovascular disease and is associated with significant morbidity and mortality. Lack of effective treatment and failure of large clinical trials aimed at hemostasis and clot removal demonstrate the need for further mechanism-driven investigation of ICH. This research may be performed through the framework provided by preclinical models. Two murine models in popular use include intrastriatal (basal ganglia) injection of either autologous whole blood or clostridial collagenase. Since, each model represents distinctly different pathophysiological features related to ICH, use of a particular model may be selected based on what aspect of the disease is to be studied. For example, autologous blood injection most accurately represents the brain's response to the presence of intraparenchymal blood, and may most closely replicate lobar hemorrhage. Clostridial collagenase injection most accurately represents the small vessel rupture and hematoma evolution characteristic of deep hemorrhages. Thus, each model results in different hematoma formation, neuroinflammatory response, cerebral edema development, and neurobehavioral outcomes. Robustness of a purported therapeutic intervention can be best assessed using both models. In this protocol, induction of ICH using both models, immediate post-operative demonstration of injury, and early post-operative care techniques are demonstrated. Both models result in reproducible injuries, hematoma volumes, and neurobehavioral deficits. Because of the heterogeneity of human ICH, multiple preclinical models are needed to thoroughly explore pathophysiologic mechanisms and test potential therapeutic strategies.
Intracerebral hemorrhage (ICH) is a relatively common form of cerebrovascular disease with approximately 40–50% of afflicted patients dying within 30 days 1. Unfortunately, little improvement has been made in the mortality rate over the last 20 years 2. Reports from the National Institutes of Health 3 and guidelines from the American Heart Association 4 stressed the importance of developing clinically relevant models of ICH to extend the understanding of pathophysiology and develop targets for new therapeutic approaches.
Several models exist to mimic human ICH 5. As understanding of ICH pathophysiology matures, it has become evident that a variety of models may be used to examine different aspects of the disease. Previously used models include murine amyloid angiopathy 6, intraparenchymal microballoon insertion and inflation 7, and direct arterial blood infiltration 8,9. Lobar hemorrhage from amyloid angiopathy has been modeled with the use of transgenic mice and represents a distinct ICH subtype. Microballoon models mimic acute mass effect from hematoma formation but fail to capture the brain’s cellular response to the presence of blood. Finally, direct arterial blood infiltration subjects the brain to arterial pressures from the femoral artery. Thus, this model mimics arterial pressures and the presence of blood but does not subject the brain to microvascular injury from small blood vessel rupture. Further, this model has inherently high variability. Interestingly, spontaneously hypertensive rats 10 develop spontaneous ICH as they age. Study of these animals after ICH development may mimic the disease in the presence of one of the major comorbidities predisposing humans to ICH. While these other models exist, intrastriatal injection of Clostridial collagenase 11 or instrastiatal injection of autologous whole blood 12 are, currently, the two most common models used in preclinical ICH research.
ICH model selection should be made based on the objective of the experimental question, including species selection and method of inducing hematoma formation. For instance, pigs are large animals with relatively large white matter brain volumes compared to mice. Thus, porcine models are suited to study white matter pathophysiology following ICH. In contrast, rodent brains are largely gray matter, but transgenic systems make rodents useful to assess molecular mechanisms of injury and recovery after ICH. Each model has its inherent strengths and weaknesses (Table 1), which should be carefully considered prior to experimentation.
The following protocols demonstrate the autologous blood and collagenase injection models in mice. These models have each been translated from models originally developed in rats 13,14 and allow the use of widely available transgenic technology to explore molecular mechanisms associated with cell death after ICH. Both represent distinctly different injury mechanisms from human ICH, and both have distinctly different expected outcome in terms of behavioral and histological measures. Thus, certain hypotheses may lend themselves to one model over the other, but many ideas may require validation in both models.
Table 1. Comparison of characteristics of collagenase- and autologous blood injection intracerebral hemorrhage models.
Collagenase Injection | Blood Injection | |
Ease of Use | +++ | ++ |
Reproducibility | ++ | ++ |
Control of Hemorrhage Size | ++ | +++ |
Blood Reflux | + | ++ |
Simulates Human Disease | + | – |
Simplicity | ++ | ++ |
Use in Multiple Species | ++ | ++ |
Ethics Statement: This protocol has been approved by the Duke University Institutional Animal Care and Use Committee and follows all the guidelines for the ethical use of animals.
1. Preparation of Equipment
2. Collagenase Injection Model
3. Autologous Blood Injection Model
4. Sham Operation
5. Post-surgery Care
Due to differences in hematoma formation (Figure 1), ipsilateral turning is shown immediately after wake up for autologous blood injected mice and within 2 – 4 hr after collagenase injection, as hematoma expansion occurs (Figure 2). Absence of ipsilateral turning should raise concern for absence of significant injury. On the first post injury day, mice in both models should demonstrate significant neurological deficits (Figure 3). At 24 hr after injection, ipsilateral hemispheres show stable hematoma volumes (Figure 4); further, at 24 hours after injection, brain water content should be expected to be 79.8 + 0.34% in collagenase-injected mice and 79.3 + 0.23% in autologous blood-injected mice. Mortality should be expected to occur between 10 – 25% of collagenase-injected mice and less than 10% of autologous blood-injected mice. Unavoidable death due to hematoma volume, cerebral edema, and increased intracranial pressure usually occurs within the first 24 – 48 hr after intrastriatal injection. Death occurring after 72 hr may often be avoided with proper post-injury care (e.g., ready access to softened food and water). Functional recovery generally begins by post injury day 2 with autologous blood-injected mice recovering significantly faster than collagenase-injected mice.
Figure 1. Serial magnetic resonance imaging of mouse brains comparing autologous blood and collagenase injection models of intracerebral hemorrhage. After intracerebral hemorrhage induction via left intrastriatal injection of 35 µl autologous blood (A) or 0.075 U Type IV-S clostridial collagenase (B) in 10 – 12 week old C57/Bl6 male mice, serial magnetic resonance imaging demonstrates hematoma expansion in the collagenase-injected mice compared to stable hematoma formation in the autologous blood-injected mice. Hematoma volumes are 10.1, 23.1, 29.9 mm3 at 1, 6, & 12 hr after collagenase injection, respectively, and 7.0, 5.8, 3.2 mm3 at 1, 6, and 24 hr after whole blood injection, respectively. Please click here to view a larger version of this figure.
Figure 2. Corner turn test in mice 24 hr after intracerebral hemorrhage. Taken immediately after intrastriatal collagenase injection into the left basal ganglia, presence of expected ipsilateral turning response in 10 – 12 week old C57/Bl6 male mice denotes adequate injury. This turning should occur immediately after significant injury in mice injected with autologous blood and within 2 – 4 hr in collagenase-injected mice. Mice in both models showed more left turns after injury compared to uninjured mice (**p <0.01; one-way ANOVA with post-hoc Scheffe’s test; n = 10/group).
Figure 3. Rotarod performance after intracerebral hemorrhage in mice. Baseline and post injury rotorod latencies of 10 – 12 week old C57/Bl6 male mice for one week after left intrastriatal 35 µl autologous blood-, 0.075 U Type IV-S clostridial collagenase-injection, or sham operation (*p = 0.022; repeated measures ANOVA with post-hoc Scheffe’s test; F-value = 12.726; n = 10/group). Mice are assessed via rotorod testing every other day after injury to avoid significant training bias.
Figure 4. Hematoxylin and eosin stains of mouse brain after intracerebral hemorrhage. Photomicrographs of 10 – 12 week old C57/Bl6 male mouse brains at 24 hr after left intrastriatal injection of 35 µl autologous blood (right) or 0.075 U Type IV-S Clostridial collagenase (left). Hematoma volumes are 20.2 mm3 after collagenase injection and 6.4 mm3 after whole blood injection. Please click here to view a larger version of this figure.
Despite emerging preclinical research and resultant large clinical trials for promising therapeutics 15-18, there are no pharmacological interventions demonstrated to improve outcome in ICH, and care remains largely supportive. Lists of possible therapies may be generated by high throughput technologies, such as transcriptomic and proteomic work. While these technologies continue to advance our knowledge of potential therapeutic targets, forward and backward translation of promising targets may be best examined through use of clinically relevant preclinical models 19-22. Such models are useful because they allow rapid throughput of selected candidates, examination of mechanisms in vivo, inexpensive investigation of dosing, therapeutic window, and other parameters germane to developing clinical trials 23-25. While obvious advantages exist in using preclinical models, modeling should occur in the most clinically relevant but logistically feasible system available. While models exist for 'higher' order animals such as primates, use of mice to model human disease provides an inexpensive, high-throughput, and powerful technology to examining pathologic mechanisms and therapeutic effects. Incorporating transgenic systems allows for an even more robust evaluation of mechanistic pathways and cell populations involved.
Currently two murine models are in common use: intrastriatal autologous blood or collagenase injection. Both models are versatile and easy to use, relative to other stroke models. Both models can induce ICH in different brain areas 26, allowing evaluation of regional responses; hematoma volume can be controlled and changed, allowing for evaluation of mild, moderate, and severe injury; and clinically relevant physiology (e.g., blood pressure, temperature, etc.) can be controlled. Finally, while each model was originally developed in the rat, both have since been translated into mice to allow the use of transgenic systems 21,24,25,27. However, each model lends itself to the study of different aspects of ICH, as each represents distinctly different components of ICH. Autologous blood-injection may recreate the brain's response to intraparenchymal blood exposure. Thus, initial mass affect and shear forces, mild inflammatory changes, apoptosis, and blood resorption may all be studied 10,28. Further, recent modifications to this model have resulted in the ability to mimic hematoma expansion 29,30. However, this model does not invoke the component of vascular injury and/or hematoma expansion found in the human disease. In contrast, collagenase-injection adds the elements of vascular rupture, early hematoma expansion, and enhanced neuroinflammatory effect. While obvious concerns exist about artifactual contribution of collagenase to this inflammatory effect, there is a lack of hard evidence for this 31, and our own (unpublished) data suggest that collagenase in isolation does not induce a marked inflammatory response in cell culture.
From a procedural standpoint, both models require limited skill with microsurgery and, thus, are easily learned so as to obtain reproducible effects. Pitfalls to be avoided include: 1) invasion of the dura or creating thermal brain injury when drilling, 2) or penetration of the ventricular system with needle insertion. Dural injury allows for reflux of injectant, and intraventricular injection results in little to no intraparenchymal hematoma formation. Further, care must be taken upon needle withdrawal to not disrupt newly formed/forming hematoma. Mortality is to be expected in a certain percentage of mice but is directly related to hematoma size and degree of injury desired; thus, this outcome may be titrated by injectant volume/concentration.
As with all models, protocols will be optimized for use by specific operators. Due to inherent variability in all in vivo systems, experience with a particular model as a key factor for success cannot be overstated. A model's distinctive characteristics, operator experience with a given model, outcome metrics of interest, and logistical factors must all be taken into account when selecting the best possible experimental model.
The authors have nothing to disclose.
This work was supported by the American Heart Association Scientist Development Grant and the Foundation for Anesthesia Education and Research (MLJ). We would like to thank Talaignair N. Venkatraman PhD for his assistance with magnetic resonance imaging.
Name | Company | Catalog Number | Comments |
Stereotactic frame | Stoelting Co. | 51603 | |
Probe holder with corner clamp | Stoelting Co. | 51631 | |
Mini grinder | Power Glide | Model 60100002 | |
0.5 µl Hamilton syringe | Hamilton Co. | 86259 | 25 gauge needle |
50 µl Hamilton syringe | Hamilton Co | 7637-01 | |
26G Hamilton needle | Hamilton Co | 7804-03 | |
Syringe pump | KD Scientific | Model 100 | |
Heat therapy water pump | Gaymar Industries, Inc. | Model# TP650 | |
Circulating waterbed | CMS Tool & Die, Inc. | ||
Rodent ventilator | Harvard Apparatus | Model 683 | |
Isoflurane vaporizer | Drager | Vapor 19.1 | |
Air flowmeter | Cole Parmer | Model PMR1-010295 | |
Induction chamber | Self made | ||
Otoscope | Welch Allyn | 22820 | |
intravenous catheter | Becton-Dickinson | 381534 | 20-gauge, 1.16 inch Insyte-W |
Isoflurane | Baxter Healthcare Corporation | NDC10019-360-69 | |
Collagenase Type IV-S | Sigma | C1889 | |
Polyethylene tubing PE20 | Becton-Dickinson | 427406 | |
Polyethylene tubing PE10 | Becton-Dickinson | 427401 | |
30G 1 inch needle | Becton-Dickinson | 305128 | |
27G 1 1/4 inch needle | Becton-Dickinson | 305136 | |
Surgical scissors | Miltex | 21-539 | |
Forceps | Miltex | 17-307 | |
Needle holder | Boboz | RS-7840 | |
Monofilament suture | Ethicon | 8698 | Size 5-0 |
Indicating controller | YSI | 73ATD |