Contributors
10 Animal Models of Intracerebral Hemorrhage
Kenneth R. Wagner, PhD , Research Associate Professor
Department of Neurology , University of Cincinnati College of Medicine , and Veterans Affairs Medical Center , Medical Research Service , Cincinnati , Ohio , USA Thomas G. Brott , MD, Professor
Department of Neurology , Mayo Clinic College of Medicine , Jacksonville , Florida , USA
INTRODUCTION
Experimental study of intracerebral hemorrhage (ICH) in animal models has increased dramatically in the last several years. A PubMed search on January 1, 2005, for articles with the keywords
“intracerebral hemorrhage” or “hematoma” and “animals” generated almost 300 references written since 1968. Two-thirds of these reports appeared in the last 5 years. These citations by researchers worldwide demonstrate a wide range of investigations examining all aspects of the disorder. The importance of experimental animal models for ICH research is that they permit a detailed examination of the pathophysiologic, biochemical, and molecular processes, as well as the mechanisms underlying brain tissue injury. They also enable testing of new pharmacologic and surgical therapies.
A comprehensive review of experimental ICH models was published by Kaufman and Schochet in 1992 ( 1 ). In 2002, we published an updated review ( 2 ). Since then, several reports that describe new models of ICH in the mouse have appeared ( 3,4 ). In addition, recent reports have descri bed interesting new fi ndings using knockout mice, such as the matrix metalloproteinase (MMP)-9-defi cient mouse ( 5 ). Thiex et al. ( 6 ) examined the extent of edema formation in a murine model of collagenase-induced ICH, which included recombinant tissue plasminogen activator (rtPA)-defi cient and wild-type mice. Interestingly, a new mouse model has been reported in which mouse embryos genetically null for all α-V-integrins develop ICH due to defective interactions between blood vessels and brain parenchymal cells ( 7 ). Several other recent studies have reported on new ICH treatments in animal models that test local ( 8 ) and global ( 9 ) brain hypothermia and the use of glutamate receptor antagonists ( 10 ), a statin ( 11 ), a cyclooxygenase-2 inhibitor ( 12 ), and metalloporphyrin heme oxygenase inhibitors ( 13–15 ). Another group ( 16 ) reported that intra-ventricular transplantation of embryonic stem (ES) cell-derived neural stem cells in rats with ICH generated ES-derived neurons and astrocytes around the hematoma cavities. The transplants were performed 7 days after the ICH, and 28 days later, ES-derived neurons and astrocytes could be detected in all 10 rats that received grafts.
MODELS AND SPECIES: OVERVIEW
The classic method of inducing ICH in experimental animals has been to directly infuse autologous blood into a specifi c brain region, usually the basal ganglia. This method has been employed in a variety of species, including rat, rabbit, cat, dog, pig, primate (reviews in Refs. 1 and 2 ), and, most recently, mouse ( 3,4 ). We have developed a large animal lobar ICH model in the pig, in which we infuse up to 3.0 mL of arterial blood into the frontal hemispheric white matter ( 17 ), and we have used this model to examine ICH pathophysiology and pathochemistry and surgical clot evacuation ( 17–19 ) (reviews in Refs. 20 and 21 ). A second commonly used model employs bacterial collagenase usually injected locally into the basal ganglia ( 22,23 ). Collagenase dissolves the extracellular matrix, leading to blood vessel rupture and an intracerebral bleed.
Overall, these models have signifi cantly contributed to our knowledge of ICH-induced injury. Specifi cally, they have provided information on the roles of mass effect and elevated intracranial pressure (ICP), alterations in blood fl ow and metabolism, and the impact of specifi c blood components on brain edema formation and blood-brain barrier (BBB) disruption. Along
with providing details of ICH-induced biochemical and molecular events, these models also enable testing of potential therapies, both surgical and pharmacologic.
In this chapter, the classic blood infusion models are described, and the collagenase model is reviewed. Each model section contains a detailed description of the fi ndings that have been obtained with individual species. The brain pathologic responses to ICH that are present in the various animal models are described and compared with observations in human ICH. Lastly, the limitations of animal models and their ability to fully address various aspects of human ICH are discussed.
INTRACEREBRAL BLOOD INFUSION ICH MODELS
Intraparenchymal infusion (or injection) of autologous blood is the classic technique by which to create an intracerebral hematoma. Clearly, this method does not reproduce the bleeding event of spontaneous ICH in humans, i.e., arterial vessel rupture. However, this model does enable the infused blood volume to be controlled and, therefore, generates reasonably reproducible hematoma sizes and mass effects. These models have been very useful for studying the pathophysiologic and biochemical events that result from the presence of blood in the brain tissue. Disadvan-tages of blood infusion models are the potential for ventricular rupture and for backfl ow of the infused blood along the needle track ( 24,25 ). Such events during blood infusion can lead to intraventricular and/or subarachnoid leakage of blood. This problem was addressed by a dou-ble hemorrhage method, in which a small volume of blood is initially infused into the caudate at a slow rate, followed by a 7-min pause to allow the blood to clot along the needle track, and then infusion of the remaining blood to produce a hematoma ( 24 ). The clotting of blood around the shaft of the needle prevents the backfl ow of blood into the subarachnoid space during the subsequent infusion, thereby enhancing the production of reproducible hematoma volumes.
This double-infusion approach has been employed by several groups ( 3,26 ).
Rats
The most frequently used species for experimental ICH studies has been the rat, and the most commonly injected site has been the basal ganglia. Some of the earliest ICH studies in rat, conducted in the mid-to-late 1980s, examined relationships between mass effect, perihemato-mal blood fl ow, and ICP ( 27–31 ). In addition, microballoons were employed to examine the relationships between mass volumes, elevations of ICP, and local perfusion ( 32 ). Overall, inves-tigators concluded that perihematomal blood fl ow was markedly reduced in their model. Based on their fi ndings, it was suggested that ischemia was responsible for secondary damage after ICH (reviews in Refs. 33 and 34 ).
However, not all investigators have concluded that ischemia is responsible for peri-hematomal tissue injury after ICH. Although initial ischemia of 20% to 30% below baseline has been noted in a rat model, recovery of fl ow and hyperemia in the hours following ICH were described ( 35 ). More recently, Yang et al. measured local cerebral blood fl ow (CBF) using [14C]-iodoantipyrine ( 25 ). They reported that CBF was reduced to 50% of control at 1 hr after ICH, returned to control values by 4 hr, but then decreased to <50% of control in the subacute phase between 24 and 48 hr. They concluded that although some degree of ischemia occurs in the early minutes to a few hours following ICH, the degree of ischemia is neither severe nor the basis for the development of perihematomal edema ( 25 ). Furthermore, in human ICH ( 36,37 ), Positron Emission Tomography studies demonstrated that, although blood fl ow might be reduced, this reduction is coupled with a reduction in metabolism in perihematomal tissue;
therefore, cerebral ischemia is not present.
Some investigators have reported that hyperemia is present following ICH in rats ( 30 ), and we observed hyperemia in our porcine ICH model ( 20,38 ). A recent report that increased glucose metabolism in perihematomal rat brain in the early hours after ICH is due to glutamate receptor activation ( 10 ) might explain the hyperemia and the marked increases in perihemato-mal lactate previously reported in a porcine ICH model ( 18 ).
Studies in rats over the past decade have established that activation of the coagulation cascade and specifi c plasma proteins are required for both acute and delayed development of perihematomal edema. The plasma protein thrombin, when infused into brain, produces
edema that is comparable to that generated by infusions of whole blood ( 39 ). The importance of the coagulation cascade in ICH-induced perihematomal edema formation was demonstrated both in rat and in porcine ICH models ( 40 ). Heparinized blood infusions generated very little edema when compared to infusions of unheparinized blood. As demonstrated in models in which packed red cells alone were infused ( 40 ), the contribution by erythrocytes to edema for-mation is delayed. Red cells infused into the basal ganglia in rats did not produce a dramatic increase in edema until 3 days postinfusion ( 40 ). In contrast, infusions of lysed autologous erythrocytes into the rat brain produced marked edema 24 hr after infusion. In the rat model, whereas hemoglobin infusions produce severe edema, packed erythrocytes do not. Similarly, in a porcine ICH model, lysed blood causes severe brain edema and death, presumably as a result of high concentrations of released hemoglobin ( 41 ).
Complement activation and membrane attack–complex formation also appear to con tribute to perihematomal edema formation, as N-acetylheparin, which inhibits complement activation, diminished this edema ( 42 ). Importantly, these results in an animal ICH model suggest that the complement system could be targeted for future ICH treatment.
Lastly, studies in a rat cortical ICH model demonstrate the additional toxicity of hemor-rhage, as compared to cerebral ischemic insults ( 43 ). These investigators showed that extravasated whole blood causes a greater degree of cell death and infl ammation than do ischemic lesions of similar size ( 44 ).
Cats
Autologous blood infusions have been used to produce experimental ICH in cats. An early experimental ICH study demonstrated an important relationship between the size and loca-tion of an intracerebral hematoma, funcloca-tional defi cits, and ICP elevaloca-tions ( 45 ). Other fi ndings demonstrated that increased ICP was the main cause of blood volume/fl ow reductions shortly after hematoma induction in the basal ganglia ( 46 ). The relationship between neurologic defi cits and hematoma volume was also observed, and it was found that urokinase-induced resolution of internal capsule hematomas also improved neurologic outcome ( 47 ). These fi ndings in cats support those in human ICH, which reported a strong relationship between hematoma size and clinical outcome ( 48 ).
Rabbits
Hematoma volumes in rabbits were studied using autologous blood stereotactically injected into the thalamus, and it was found that these animals would tolerate clots with a volume of 3% to 5% of their brain volume, which approximates a 50-cc clot in humans ( 49 ). In one of the earliest studies of the effi cacy of hematoma removal with thrombolytics the authors demonstrated 86% effi cacy in lysing intracranial hematomas with urokinase in a rabbit ICH model, whereas only 3 of the 13 controls showed evidence of hematoma resolution with saline injections into the clot ( 50 ). Furthermore, histologically, no increased damage or infl ammation was noted between these animals and 22 additional rabbits treated with urokinase or saline 24 hr after clot injection.
The authors concluded that urokinase could be employed safely and effectively for the lysis of intracranial hematomas and that a delay in therapy of up to 24 hr does not signifi cantly compromise its effi cacy. The effi cacy of urokinase in stereotactic human ICH treatment was further demonstrated by Zuccarello et al. ( 51 ).
In a very detailed examination of the cellular activation in the perifocal reactive zone in a rabbit ICH model, Koeppen et al. (52) found that experimental hematomas resolved much more slowly after the injection of whole blood than after the injection of red blood cells, suggesting that proteins in the coagulated blood contribute to the injury process.
Imaging studies of ICH in rabbits demonstrated that susceptibility-weighted gradient-echo imaging at 1.5 T is highly sensitive in detecting hyperacute parenchymal, as well as subarachnoid and intraventricular hemorrhages ( 53 ). A rabbit lobar ICH model in which arterial pressure was used to infuse autologous blood into the deep frontal white matter, served to investigate the degree of neuronal injury inside and outside the hematoma at 24 hr ( 54 ).
Dogs
Canines were among the earliest-used animals for studying experimental ICH. A 1975 study of the tolerance of the dog’s brain to blood injection in different sites—the brain parenchyma and the ventricles—found different lethal volumes for each specifi c ICH site ( 55 ). The authors
found that 8 mL was lethal in the dog brain parenchyma and concluded that that death was not a random event but was due to the failure of vital functions as a result of elevated ICP.
Further detailed studies demonstrated the evolution of brain injury following ICH in a canine parietal lobe hematoma model by high-resolution sonography, CT, and neuropathologic exam-inations ( 56 ). These authors found a correlation between the sequence of changes on CT and sonography images in their ICH model and the fi ndings following ICH in patients. In early MRI studies in canine ICH models, venous and arterial blood infusions and intraventricular locations of blood were compared, and it was concluded that gradient-echo sequences would be highly useful in detecting and delineating hemorrhages in human ICH patients ( 57 ).
In histologic and CT studies of internal capsule hematomas conducted in dogs, 3 distinct stages were identifi ed on histology and CT: ( i ) in the acute stage (≤ 5 days), homogeneous high density was present at the periphery of the hematoma on CT, while histologically, a necrotic layer of tissue existed at the boundary of the clot; ( ii ) in the subacute stage (5–14 days), perihemato-mal density was decreased with ring enhancement after contrast injection and corresponded to the appearance of immature connective tissue with argentophilic fi bers; ( iii ) in the chronic stage (>15 days), contraction of the enhancing ring was noted, corresponding to the development of mature connective tissue with collagen fi bers ( 58 ).
Using a mongrel dog ICH model, another study ( 59 ) determined the effect of massive ICH on regional CBF (rCBF) and metabolism by testing the hypothesis put forward by Mendelow and coworkers from their studies in rats ( 27–34 ) regarding intracerebral bleed-induced peri-hematomal ischemia. Interestingly, these investigators found no evidence for an ischemic penumbra within the fi rst 5 hr after hemorrhage, despite prominent increases in ICP and mean arterial blood pressure (MABP) following hematoma induction, indicative of a Cushing response.
Other investigators found that hypertonic saline, at 3% and 23.4% concentrations, was as effective as mannitol in controlling intracranial hypertension in a dog model ( 60 ). In addition, 3% hypertonic saline appeared to have a longer effect than either 23.4% saline or mannitol. No effect on rCBF or cerebral metabolism was observed with any of the agents. In their study of the pharmacologic reduction of MABP, the same group using the same canine model found that reducing MABP with intravenous labetalol within the normal autoregulatory curve of CPP had no adverse effects on ICP and perihematomal or distant rCBF ( 54 ). They concluded that MABP reduction is safe within autoregulation limits in the acute period after ICH.
Monkeys
Hematomas were generated in the caudate nucleus of vervet monkeys by allowing femoral arterial blood to enter the structure via a stereotactically implanted needle ( 61 ). ICP peaked at 51 ± 8 mmHg at 3 min after the bleed and remained high throughout the 3 hr procedure.
rCBF was signifi cantly reduced in all brain regions for 1 hr after the ictus, with the lowest values in the periphery of the hematoma. Some rCBF values were below the ischemic threshold for 90 min after the hemorrhage. Another group ( 62 ) reported an early thrombolytic treatment study in Macaque monkeys in abstract form in 1982. They found that urokinase promoted basal ganglia hematoma resorption that correlated with improvement in the clinical exam.
Pigs
Our group has developed and extensively studied the pathophysiology, pathochemistry, and treatment of ICH in a porcine white-matter (lobar ICH) model ( 17 ) (reviews in Refs. 20 and 21 ).
The advantages of using the pig for an ICH model include its large gyrated brain and large amount of hemispheric white matter, its relatively low cost, and its noncompanion animal status.
The pig’s large brain enables the production of hematoma volumes up to approximately 3 cc by slowly (10–15 mins) infusing autologous arterial blood through a plastic catheter into the frontal white matter. We have used this model to investigate ICH pathophysiology and pathochem-istry, edema development, the role of blood components, and metabolism ( 18,20,63 ) (reviews in Refs. 21 and 64 ). The large hematoma volumes that can be generated in this model make it useful to investigate studies of neurosurgical clot evacuation ( 19,65 ).
This lobar ICH model in the pig is clinically relevant for several reasons: ( i ) bleeds into the white matter are common in human ICH and occur with almost similar frequency to basal ganglia bleeds ( 66 ); ( ii ) lobar ICH is the most frequent hemorrhage site in the young ( 67 );
( iii ) white-matter damage is an important contributor to long-term morbidity following ICH
( 68,69 ). Because white matter is more vulnerable to vasogenic edema development than is gray matter ( 70 ), this model is especially applicable for studying edema-associated injury.
Studies in the porcine ICH model have demonstrated the important role of clot formation, retraction, and plasma protein accumulation in perihematomal edema development ( 17,20 ).
The importance of coagulation in animal models is translatable to patients who developed ICH after thrombolytic treatment but failed to develop signifi cant edema despite large intrace-rebral masses ( 71 ). Although whole blood is responsible for the majority of the hematoma’s mass effect, infusions of packed red cells alone fail to generate perihe matomal edema. As described above, nonclotting blood also produces minimal perihematomal edema in both rat and porcine models. Thus, these results support the conclusion that the early and substantial perihema-tomal edema that follows ICH does not result from the mass effect and potentially reduced perfusion that are induced by the hematoma. Rather, the fi ndings suggest that this very early edema results primarily from the coagulation cascade and clot retraction. The cellular elements of the hematoma are concentrated at the core as clotting proceeds. The fl uid/serum compo nents of the whole blood are extruded to the perimeter ( 72 ). Interestingly, it was recently reported ( 73 ) that increased rates of water diffusion measured in the perihematomal region by diffusion-weighted MRI in ICH patients suggested that this edema development is plasma derived. These results also are consistent with the experimental and human ICH studies, which show relatively more early edema in the setting of normal coagulation compared with the lesser edema seen in the setting of anticoagulants or thrombolytics ( 71 ).
This porcine ICH model has been especially useful for clot evacuation studies. We have demonstrated that early (3.5 hr) clot aspiration after rtPA-induced lysis markedly reduced (by
>70%) both clot volume and perihematomal edema and protected the BBB at 24 hr following ICH ( 19 ). This reduction in clot volume, achieved with rtPA liquifi cation of the clot, was signifi cantly (>37%) greater than the reduction obtained by mechanical aspiration without rtPA.
Lastly, we have used the porcine ICH model to test the Possis AngioJet rheolytic thrombectomy catheter ( 74 ). This surgical clot removal study showed that the device was very effective for rap-idly removing intracerebral hematomas, producing an average 61% decrease in clot volumes in approximately 30 sec. Other treatments studied in this model include inhibiting heme oxygenase by a metalloporphyrin ( 13 ).
BACTERIAL COLLAGENASE MODEL
The bacterial collagenase model employs the local injection of bacterial collagenase into the basal ganglia to induce an intracerebral bleed ( 22,23,75 ). This model mimics spontaneous ICH in humans by dissolving the extracellular matrix around capillaries, resulting in active intraparenchymal bleeding. The hemorrhage is simple to produce. The animals develop spontane -ous, reproducible hemorrhages, with volumes that correlate with the amount of collagenase injected, and signifi cant blood leakage does not develop along the needle track. A disadvantage is that bacte rial collagenase introduces a signifi cant infl ammatory reaction ( 44,76,77 ) that is more intense than that observed in experimental ICH models that employ blood infusion. The infl ammatory response is also more intense than that observed following human ICH ( 78 ). The collagenase model also differs from the punctate arterial rupture that produces human ICH, as the collage-nase dissolves the extracellular matrix around capillaries to produce hemorrhage.
The collagenase ICH model has been used by several investigators in mouse, rat, and pig.
Their reports shed light on the pathochemical events following ICH and describe several new experimental treatments for ICH that have not been examined in the blood infusion model.
Recent studies have demonstrated: ( i ) the role of MMPs in BBB opening and edema development following collagenase-induced ICH and the effectiveness of MMP inhibitors ( 79,80 ), ( ii ) that select MMPs exhibit increased expression after ICH and that minocycline is neuroprotective by suppressing monocytoid cell activation and downregulating MMP-12 expression ( 81 ), and ( iii ) that the tripeptide macrophage/microglial inhibitory factor inhibits microglial activation and results in functional improvement when given before, as well as after, the onset of collagenase-induced ICH ( 82,83 ). Various other reports using this model have described detailed studies of the collagenase dose effect ( 84 ), imaging features and histopathology ( 76,85 ), neurobehavioral results and therapy ( 86–88 ), and infl uence of hyperglycemia ( 89 ). Several drug treatments aimed at different molecular mechanisms of injury have also been studied, including free radical
scavengers/spin traps ( 90,91 ), neurotransmitter receptor agonists ( 92,93 ) and antagonists ( 94 ), cytokines and infl ammation ( 77,95–98 ), and neuroprotectives ( 99 ).
The collagenase model in rodents has also been extended to pigs ( 100,101 ), and the investi-gators have reported studies of somatosensory-evoked potentials elicited by electrical stimulation of the contralateral snout, as well as changes in DC-coupled potential, which was monitored in the somatosensory region following induction of ICH into the primary sensory cortex.
ISCHEMIA-REPERFUSION HEMORRHAGE MODEL
An interesting and possibly clinically relevant ICH model has been described in rhesus monkeys, in which hematomas were induced during the vasoproliferative stages of a maturing ischemic infarct ( 102 ). In this model, elevating MABP at 5 days after permanent middle cerebral artery occlusion caused hemorrhagic infarct conversion. Interestingly, previously middle cerebral artery-occluded animals that were made hypercarbic with 5% CO 2 air at 5 days postischemia had slowly progressive elevation in ICP and MABP and developed intracerebral hematoma involving the putamen, external capsule, and claustrum, occasionally dissecting through to ipsilateral ventricle. Several clinical reports have cited this model, but no further work has been done using it.
BRAIN PATHOLOGIC RESPONSE TO ICH IN ANIMAL MODELS
Overall, the brain pathologic responses to an intracerebral hematoma in experimental animal models are comparable to those seen in human ICH ( 20,56,58,78,103–105 ). The three stages of perihematomal tissue injury defi ned by Spatz in 1939 ( 104 )—initial deformation, edema and necrosis, and clot absorption and scar or cavity formation—have also been regularly described in animal models, albeit at a faster rate. An excellent description of these changes in rat was provided, in which regions of pallor and spongiform change due to edema formation devel-oped adjacent to clots within 2 hr ( 105 ). By 6 to 15 hr, disrupted myelinated nerve fi bers and degeneration bulbs were present, along with increasing swelling of the corona radiata as edema fl uid continued to accumulate. At 24 hr, white-matter edema was more marked and extensive.
By 48 hr, hematomas in rat and dog ICH models were surrounded by edema, vacuolation, and acellular plasma accumulations, with astrocytic swelling present adjacent to, and distant from, the hematoma.
In our porcine ICH model, marked, rapidly developing perihematomal edema with a high water content is already present in white matter by 1 hr after ICH ( 17 ). This edema produces perihematomal hyperintensity on T2-weighted MRI ( 20,38 ) similar to that in ICH patients ( 106 ). We observed increases in edema volumes by 50% over the fi rst 24 hr in the porcine ICH model due to delayed BBB opening ( 17 ). In the collagenase ICH model, similar hyperintensi-ties on T2-weighted imaging surround hematomas and extend along posterior white-matter fi ber tracts ( 85 ). Histologically, by 3 days, we observed decreased Luxol fast blue staining in edematous white matter, suggestive of myelin injury, and markedly increased glial fi brillary acidic protein immunoreactivity, indicative of reactive astrocytosis ( 20 ). In our porcine model, neovascularization is present at 7 days. By 2 weeks, continued hematoma resolution and glial scar and cyst formation are similar to both rodent and human ICH pathologies. A similar brain pathologic response occurs in porcine white matter, in which only plasma is infused, thereby demonstrating the signifi cance of the blood’s plasma protein component in ICH-induced brain injury ( 20,107 ).
The time course of infl ammation and cell death following infusion of whole blood into the rat striatum has been carefully examined by several groups ( 44,108 ). They have also characterized the cellular perihematomal infl ammatory response, including the infi ltration of immune cells and activation of microglia. Several others have examined DNA fragmentation using terminal deoxynucleotidyl transferase dUTP nick-end labeling staining ( 107,109–111 ). In addition, molecular analyses of the proinfl ammatory transcription factor, nuclear factor-κ B, and cytokine responses to ICH have been conducted ( 26,64,112–114 ).
An interesting concept of the mechanism of cell death after ICH has been proposed from studies of ICH in rat, i.e., the “black hole” model of hemorrhagic damage by Felberg and cowork-ers ( 115 ). These investigators showed that histologic damage from ICH is very prominent in the immediate perihematomal region. Except for substantia nigra pars reticulata , they found no evidence of neuronal loss in distal regions. The term “black hole” refers to this continued destruction of neurons, which occurs over at least 3 days, as the neurons come into proximity to the hematoma.
LIMITATIONS OF ANIMAL MODELS
Several prominent characteristics of human spontaneous ICH are not well mimicked by current animal models. Human ICH is fundamentally linked to advancing age. The incidence of spontaneous ICH is about 25 times higher for those who are age ≥75 years compared to those who are age ≥45 years ( 116 ). Thus, ICH models in young animals do not reproduce the pre-existing degenerative changes in small arteries, arterioles, neurovascular units, or surrounding brain tissue. In addition, the human genetic response capability to brain injury is now known to change with advancing age ( 117 ). Recently, in an effort to address this problem, the response to ICH was studied in young (3 months) versus aged (18 months) rats, and more severe brain and neurologic defi cits were reported in old rats that persisted for 4 weeks after ICH ( 118 ). Addi-tionally, older rats had stronger microglial activation and a greater perihematomal induction of the heat-shock proteins HSP-27 and HSP-32. A goal for future research is to determine the degree to which brain tissue responses to ICH in aging animal models mimic those in aging humans.
Human ICH usually occurs in the setting of longstanding comorbidities (e.g., tobacco use, diabetes, hypertension) and commonly in the presence of active drugs (e.g., antiplatelet, anti-coagulant, statin). These conditions cannot be easily reproduced in animal models. For example, even spontaneously hypertensive rats are not likely to reproduce the decades-long effects in humans of elevated arterial pressure. Human ICH also varies by race, in incidence overall, and in incidence by age epoch, suggesting important and yet-to-be-discovered variations in genetic susceptibility. Inferences regarding treatment must also be drawn with caution. Findings have demonstrated signifi cant benefi ts from mechanical and pharmacologic interventions that have not been reproduced in human trials, whether small and focused ( 51,119 ) or large and inclusive ( 120 ).
Among the several potential explanations for these discrepancies, delay in treating ICH patients is the likeliest. However, toward improving the design of future ICH models, it is important that conductors of future studies in animal models consider the limitations in translating the fi ndings to human ICH treatment.
SUMMARY OF ANIMAL SPECIES AND ICH INDUCTION METHODS
In this review we have described the various animal species that have been employed in ICH research. In addition, we have discussed the several methodologies that have been employed to produce intracerebral hematomas, presenting the pros and cons of the individual species and the ICH induction techniques. In this section, we have summarized these advantages and disadvantages and have suggested the “best” models and methods based on the goals of the study, the experimental plan, the desired hematoma volumes, and the expense.
Rodents have the advantage of being the most commonly used species in ICH research.
The literature on neurobehavioral testing is well developed and the reagents for immunocy-tochemistry and molecular biology have been extensively studied. The recent development of mouse ICH models enables the study of transgenic and knockout animals, which is a clear advantage for uncovering the detailed molecular pathophysiologic events underlying the devel-opment of tissue injury following ICH.
Large animals (pigs, dogs, and primates) have certain advantages over rodents in ICH research. These include their large gyrated brains with a signifi cant amount of white matter.
Large animals enable the induction of greater hematoma volumes to test the effi cacy of surgical evacuation techniques or combined surgery and drug treatments. The well-developed frontal white matter in the pig has been especially useful for pathophysiologic studies of ICH-induced
white-matter injury as well as surgical clot evacuation studies. In addition, pigs have the advantage as compared to dogs and cats that they are less expensive to purchase and are considered non-companion animals. Primates are exceedingly expensive to purchase and house and require special facilities and veterinary care.
Regarding the methods for inducing an intracerebral hematoma, as described above, i.e., the two commonly used methods are the classical blood infusion method and the collagenase injection method. Neither method exactly models the human event, i.e., sudden arterial rupture with a rapid intraparenchymal accumulation of blood. Currently, there is no model of intra-cerebral blood vessel rupture to induce ICH. Although both the direct blood infusion model and the collagenase model have their artifi cialities, the arterial blood infusion through an indwell-ing catheter described throughout the review is generally considered to be the method of choice for inducing experimental ICH. The use of the bacterial collagenase enzyme to “dissolve” the extracellular matrix has been considered to be more artifi cial due to its severe infl ammatory response and secondary pathophysiology that occurs in the setting of an already damaged brain parenchyma.
OVERALL SUMMARY AND CONCLUSIONS
As described in this review, experimental animal ICH models reproduce important pathophys-iologic events that develop in human ICH, including perihematomal edema, markedly reduced metabolism, and comparable brain tissue pathologic responses. Thus, these animal ICH models are important tools for new understanding of the mechanisms underlying brain injury after an intracerebral bleed. The recent publication from several laboratories describing ICH models in the mouse will enable new investigations into secondary infl ammatory responses, intracel-lular signaling, and molecular events that are expected to provide future therapeutic targets for treating ICH. The continued use of a large noncompanion animal, such as the pig, enables studies of ICH-induced white-matter injury—an important contributor to patient morbidity.
The large animal model also permits studies of surgical treatments that could be combined with pharmacologic approaches. A future animal model that would have considerable clinical applicability would be one that mimics the spontaneous enlarging hematoma with continued bleeding that is observed in about 30% of human ICH patients ( 121 ).
ACKNOWLEDGMENTS
These studies were supported by funding from the National Institute of Neurological Diseases and Stroke (R01NS-30652) and the Department of Veterans Affairs Medical Research Service.
REFERENCES
1. Kaufman HH , Schochet SS . Pathology, pathophysiology and modeling . In: Kaufman HH , ed.
Intracerebral Hematomas: Etiology, Pathophysiology, Clinical Presentation and Treatment . New York : Raven Press , 1992 : 13 – 20 .
2. Andaluz N , Zuccarello M , Wagner KR . Experimental animal models of intracerebral hemorrhage . Neurosurg Clin N Am 2002 ; 13 : 385 – 393 .
3. Belayev L , Saul I , Curbelo K , et al. Experimental intracerebral hemorrhage in the mouse: histo-logical, behavioral, and hemodynamic characterization of a double-injection model . Stroke 2003 ; 34 : 2221 – 2227 .
4. Nakamura T , Xi G , Hua Y , et al. Intracerebral hemorrhage in mice: model characterization and application for genetically modifi ed mice . J Cereb Blood Flow Metab 2004 ; 24 : 487 – 494 .
5. Tang J , Liu J , Zhou C , et al. MMP-9 defi ciency enhances collagenase-induced intracerebral hemor-rhage and brain injury in mutant mice . J Cereb Blood Flow Metab 2004 ; 24 : 1133 – 1145 .
6. Thiex R , Mayfrank L , Rohde V , et al. The role of endogenous versus exogenous tPA on edema formation in murine ICH . Exp Neurol 2004 ; 189 : 25 – 32 .
7. McCarty JH , Lacy-Hulbert A , Charest A , et al. Selective ablation of alpha v integrins in the central nervous system leads to cerebral hemorrhage, seizures, axonal degeneration and premature death . Development 2005 ; 132 : 165 – 176 .
8. Wagner KR , Zuccarello M . Focal brain hypothermia for neuroprotection in stroke treatment and aneurysm repair . Neurol Res . 2005; 27:238–245 .
9. MacLellan CL , Girgis J , Colbourne F . Delayed onset of prolonged hypothermia improves outcome after intracerebral hemorrhage in rats . J Cereb Blood Flow Metab 2004 ; 24 : 432 – 440 .
10. Ardizzone TD , Lu A , Wagner KR , et al. Glutamate receptor blockade attenuates glucose hyperme-tabolism in perihematomal brain after experimental intracerebral hemorrhage in rat . Stroke 2004 ; 35 : 2587 – 2591 .
11. Seyfried D , Han Y , Lu D , et al. Improvement in neurological outcome after administration of ator-vastatin following experimental intracerebral hemorrhage in rats . J Neurosurg 2004 ; 101 : 104 – 107 . 12. Chu K , Jeong SW , Jung KH , et al. Celecoxib induces functional recovery after intracerebral
hemor-rhage with reduction of brain edema and perihematomal cell death . J Cereb Blood Flow Metab 2004 ; 24 : 926 – 933 .
13. Wagner KR , Hua Y , de Courten-Myers GM , et al. Tin-mesoporphyrin, a potent heme oxygenase inhibitor, for treatment of intracerebral hemorrhage: in vivo and in vitro studies . Cell Mol Biol (Noisy.-le-grand) 2000 ; 46 : 597 – 608 .
14. Huang FP , Xi G , Keep RF , et al. Brain edema after experimental intracerebral hemorrhage: role of hemoglobin degradation products . J Neurosurg 2002 ; 96 : 287 – 293 .
15. Koeppen AH , Dickson AC , Smith J . Heme oxygenase in experimental intracerebral hemorrhage: the benefi t of tin-mesoporphyrin . J Neuropathol Exp Neurol 2004 ; 63 : 587 – 597 .
16. Nonaka M , Yoshikawa M , Nishimura F , et al. Intraventricular transplantation of embryonic stem cell-derived neural stem cells in intracerebral hemorrhage rats . Neurol Res 2004 ; 26 : 265 – 272 . 17. Wagner KR , Xi G , Hua Y , et al. Lobar intracerebral hemorrhage model in pigs: rapid edema
develop-ment in perihematomal white matter . Stroke 1996 ; 27 : 490 – 497 .
18. Wagner KR , Xi G , Hua Y , et al. Early metabolic alterations in edematous perihematomal brain regions following experimental intracerebral hemorrhage . J Neurosurg 1998 ; 88 : 1058 – 1065 .
19. Wagner KR , Xi G , Hua Y , et al. Ultra-early clot aspiration after lysis with tissue plasminogen activator in a porcine model of intracerebral hemorrhage: edema reduction and blood-brain barrier protection . J Neurosurg 1999 ; 90 : 491 – 498 .
20. Wagner KR , Broderick JP . Hemorrhagic stroke: pathophysiological mechanisms and neuroprotective treatments . In: Lo EH , Marwah J , eds. Neuroprotection . Scottsdale : Prominent Press , 2001 : 471 – 508 . 21. Wagner KR , Sharp FR , Ardizzone TD , et al. Heme and iron metabolism: role in cerebral hemorrhage .
J Cereb Blood Flow Metab 2003 ; 23 : 629 – 652 .
22. Rosenberg GA , Mun-Bryce S , Wesley M , et al. Collagenase-induced intracerebral hemorrhage in rats . Stroke 1990 ; 21 : 801 – 807 .
23. Rosenberg GA , Estrada E , Kelley RO , et al. Bacterial collagenase disrupts extracellular matrix and opens blood-brain barrier in rat . Neurosci Lett 1993 ; 160 : 117 – 119 .
24. Deinsberger W , Vogel J , Kuschinsky W , et al. Experimental intracerebral hemorrhage: description of a double injection model in rats . Neurol Res 1996 ; 18 : 475 – 477 .
25. Yang GY , Betz AL , Chenevert TL , et al. Experimental intracerebral hemorrhage: relationship between brain edema, blood fl ow, and blood-brain barrier permeability in rats . J Neurosurg 1994 ; 81 : 93 – 102 . 26. Hickenbottom SL , Grotta JC , Strong R , et al. Nuclear factor-kappaB and cell death after experimental
intracerebral hemorrhage in rats . Stroke 1999 ; 30 : 2472 – 2477 .
27. Bullock R , Mendelow AD , Teasdale GM , et al. Intracranial haemorrhage induced at arterial pressure in the rat . Part 1: Description of technique, ICP changes and neuropathological fi ndings . Neurol Res 1984 ; 6 : 184 – 188 .
28. Mendelow AD , Bullock R , Teasdale GM , et al. Intracranial haemorrhage induced at arterial pressure in the rat . Part 2: Short term changes in local cerebral blood fl ow measured by autoradiography . Neurol Res 1984 ; 6 : 189 – 193 .
29. Nath FP , Jenkins A , Mendelow AD , et al. Early hemodynamic changes in experimental intracerebral hemorrhage . J Neurosurg 1986 ; 65 : 697 – 703 .
30. Nath FP , Kelly PT , Jenkins A , et al. Effects of experimental intracerebral hemorrhage on blood fl ow, capillary permeability, and histochemistry . J Neurosurg 1987 ; 66 : 555 – 562 .
31. Kingman TA , Mendelow AD , Graham DI , et al. Experimental intracerebral mass: description of model, intracranial pressure changes and neuropathology . J Neuropathol Exp Neurol 1988 ; 47 : 128 – 137 . 32. Kingman TA , Mendelow AD , Graham DI , et al. Experimental intracerebral mass: time-related effects
on local cerebral blood fl ow . J Neurosurg 1987 ; 67 : 732 – 738 .
33. Mendelow AD . Spontaneous intracerebral haemorrhage . J Neurol Neurosurg Psychiatr 1991 ; 54 : 193 – 195 .
34. Mendelow AD . Mechanisms of ischemic brain damage with intracerebral hemorrhage . Stroke 1993 ; 24 , I115 – I117 .
35. Ropper AH , Zervas NT . Cerebral blood fl ow after experimental basal ganglia hemorrhage . Ann Neurol 1982 ; 11 : 266 – 271 .
36. Powers WJ , Zazulia AR , Videen TO , et al. Autoregulation of cerebral blood fl ow surrounding acute (6 to 22 hours) intracerebral hemorrhage . Neurology 2001 ; 57 : 18 – 24 .