Interventional magnetic resonance imaging-guided cell transplantation into the brain with radially branched deployment.

Intracerebral cell transplantation is being pursued as a treatment for many neurological diseases, and effective cell delivery is critical for clinical success. To facilitate intracerebral cell transplantation at the scale and complexity of the human brain, we developed a platform technology that enables radially branched deployment (RBD) of cells to multiple target locations at variable radial distances and depths along the initial brain penetration tract with real-time interventional magnetic resonance image (iMRI) guidance. iMRI-guided RBD functioned as an "add-on" to standard neurosurgical and imaging workflows, and procedures were performed in a commonly available clinical MRI scanner. Multiple deposits of super paramagnetic iron oxide beads were safely delivered to the striatum of live swine, and distribution to the entire putamen was achieved via a single cannula insertion in human cadaveric heads. Human embryonic stem cell-derived dopaminergic neurons were biocompatible with the iMRI-guided RBD platform and successfully delivered with iMRI guidance into the swine striatum. Thus, iMRI-guided RBD overcomes some of the technical limitations inherent to the use of straight cannulas and standard stereotactic targeting. This platform technology could have a major impact on the clinical translation of a wide range of cell therapeutics for the treatment of many neurological diseases.

Sustained induction of heme oxygenase-1 in the traumatized spinal cord.

Oxidative stress contributes to secondary injury after spinal cord trauma. Among the consequences of oxidative stress is the induction of heme oxygenase-1 (HO-1), an inducible isozyme that metabolizes heme to iron, biliverdin, and carbon monoxide. Here we examine the induction of HO-1 in the hemisected spinal cord, a model that results in reproducible degeneration in the ipsilateral white matter. HO-1 was induced in microglia and macrophages from 24 h to at least 42 days after injury. Within the first week after injury, HO-1 was induced in both the gray and the white matter. Thereafter, HO-1 expression was limited to degenerating fiber tracts. HSP70, a heat shock protein induced mainly by the presence of denatured proteins, was consistently colocalized with HO-1 in the microglia and macrophages. This study to demonstrates long-term induction of HO-1 and HSP70 in microglia and macrophages after traumatic injury and an association between induction of HO-1 and Wallerian degeneration. White matter degeneration is characterized by phagocytosis of cellular debris and remodeling of surviving tissue. This results in the metabolism, synthesis, and turnover of heme and heme proteins. Thus, sustained induction of HO-1 and HSP70 in microglia and macrophages suggests that tissue degeneration is an ongoing process, lasting 6 weeks and perhaps even longer.

Bacterial endotoxin (lipopolysaccharide) stimulates the rate of iron oxidation.

Bacterial endotoxin (lipopolysaccharide) has affinity for a number of cations, including iron. Previous investigations have demonstrated that lipopolysaccharide can affect the oxidation rate of iron; heme-bound ferrous iron in hemoglobin is oxidized to ferric iron when hemoglobin binds lipopolysaccharide. In the present study, we directly examined the interaction between lipopolysaccharide and iron. Lipopolysaccharide caused a concentration-dependent increase in the rate of iron oxidation, with up to a 23-fold increase in oxidation in the presence of 200 microg/ml Escherichia coli lipopolysaccharide. This effect was seen both with several carbohydrate-rich smooth lipopolysaccharides and also with carbohydrate-poor rough lipopolysaccharide. Extensively deacylated rough lipopolysaccharide had no effect, suggesting a role of the fatty acid components of lipopolysaccharide in this process. Purified lipid A produced inconsistent results: some preparations stimulated iron oxidation and others did not. A series of sugars, starches and a preparation of purified O-chain polysaccharide (the carbohydrate portion of the lipopolysaccharide macro-molecule) had no effect on the rate of iron oxidation, whereas phospholipid-enriched brain tissue extracts (similar to the lipid A component of lipopolysaccharide) stimulated oxidation. We conclude that the lipid moiety of bacterial lipopolysaccharide is responsible for the stimulation of iron oxidation. This process may contribute to the ability of lipopolysaccharide to cause oxidation of heme-bound iron in hemoglobin.

Anti-oxidants prevent focal rat brain injury as assessed by induction of heat shock proteins (HSP70, HO-1/HSP32, HSP47) following subarachnoid injections of lysed blood.

The initial aim of this study was to determine if the HSP70 (the main inducible heat shock protein), HO-1 (heme oxygenase-1, HSP32) and HSP47 (a collagen chaperone) stress proteins were induced in the same focal regions of rat brain following experimental subarachnoid hemorrhage (SAH). The next objective was to determine whether anti-oxidants prevented the stress gene expression in the focal regions. Lysed blood (150 microliter) was injected into the subarachnoid space of adult, female Sprague-Dawley rats via the cisterna magna. Animals were sacrificed 24 h later. Immunocytochemistry showed focal regions of stress gene induction in most animals (13/21), HSP70 and HO-1 proteins being expressed in neurons, microglia and astrocytes and HSP47 being expressed in microglia. Co-induction of the same three stress proteins was observed in focal areas in the striatum and cerebellum as well. In the 13 animals with focal regions of stress gene induction there were 8.1+/-1.8 foci in cortex, 5.5+/-0.9 foci in striatum, and 11.7+/-7.3 foci in cerebellum in the brain of each animal. The focal regions of stress gene induction varied in size from 200 micrometer to 7 mm in diameter. Systemic administration of the tirilazad-like anti-oxidants U101033E (n=8) and U74389G (n=7) completely blocked stress protein induction in focal brain regions normally produced by cisternal injections of lysed blood. There were fewer drug treated animals (0/15) with focal areas of stress gene induction compared to non-drug (13/21) treated animals following the cisternal lysed blood injections (p<0.01 using Fisher's probability test). This study shows that anti-oxidants prevent focal regions of injury as assessed by heat shock protein expression in a rat model of SAH.

Heme oxygenase-1 is induced in glia throughout brain by subarachnoid hemoglobin.

The heme oxygenase-1 gene, HO-1, induced by heme, ischemia, and heat shock, metabolizes heme to biliverdin, free iron, and carbon monoxide. Though the distribution of HO-1 has been described in normal rat brain, little is known about how extracellular heme proteins in the subarachnoid space distribute in brain. To address this issue, hemoglobin was injected into the cisterna magna of adult rats. Expression of HO-1 in these animals was compared with saline-injected, BSA-injected, and uninjected controls. Western blot analysis showed that 24 hours after injection oxyhemoglobin increased HO-1 levels approximately four- to fivefold in all brain regions studied compared with saline-injected and BSA-injected controls. In the brain, HO-1 immunoreactivity was evident at 4 hours and peaked at 24 hours after oxyhemoglobin injections, returning to control levels by 4 to 8 days. This HO-1 induction was detected mainly in cells with small, rounded somas bearing two to four truncated processes, a morphology consistent with that of microglia. These cells were double-stained with the microglial marker, OX42, in every brain region examined. It is proposed that subarachnoid hemoglobin may be taken up into microglia wherein heme induces HO-1.

Induction of heme oxygenase-1 after hyperosmotic opening of the blood-brain barrier.

The induction of the stress protein heme oxygenase-1 (HO-1) was studied in the rat brain after intracarotid administration of hyperosmolar mannitol. HO-1 was immunolocalized in fixed sections of brain 24 h to 7 days after injection. Immunoglobulin G (IgG) was immunolocalized in adjacent sections to demonstrate areas of breakdown of the blood-brain barrier. Induction of HO-1 was also evaluated by Western immunoblots, performed at 24 h after the insult. Immunofluorescent double labelling with monoclonal antibodies to HO-1 and either glial fibrillary acidic protein or the complement C3bi receptor was used to determine if glia/macrophages expressed HO-1. There was pronounced, widespread induction of HO-1 in the ipsilateral hemisphere and cerebellum by 24 h both by immunocytochemistry and by Western blots. This induction was markedly attenuated at later times. HO-1 was induced in astrocytes and microglia/macrophages in the ipsilateral hemisphere. In addition, the protein was induced in Bergmann glia and scattered microglia/macrophages in the cerebellum. The mechanism of induction of HO-1 in glia after opening of the blood-brain barrier could include exposure to heme proteins, denatured proteins and other plasma constituents known to induce HO-1. This glial induction may reflect a protective response of these cells.

The effect of magnesium on oxidative neuronal injury in vitro.

The effect of magnesium on the oxidative neuronal injury induced by hemoglobin was assessed in murine cortical cell cultures. Exposure to 5 microM hemoglobin in physiologic (1 mM) magnesium for 26 h resulted in the death of about one-half the neurons and a sixfold increase in malondialdehyde production; glia were not injured. Increasing medium magnesium to 3 mM reduced neuronal death by about one-half and malondialdehyde production by about two-thirds; neuronal death and lipid peroxidation were approximately doubled in 0.3 mM magnesium. Comparable results were observed in spinal cord cultures. The NMDA antagonist MK-801 weakly attenuated hemoglobin neurotoxicity in low-magnesium medium, but tended to potentiate injury in physiologic magnesium. Incubation in low-magnesium medium alone for 24 h reduced cellular glutathione by approximately 50% in mixed neuronal and glial cultures but by only 10% in pure glial cultures. The iron-dependent oxidation of phosphatidylethanolamine liposomes was attenuated in a concentration-dependent fashion by 2.5-10 mM magnesium; a similar effect was provided by 0.01-0.1 mM cobalt. However, oxidation was weakly enhanced by 0.5-1 mM magnesium. These results suggest that the vulnerability of neurons to iron-dependent oxidative injury is an inverse function of the extracellular magnesium concentration. At high concentrations, magnesium inhibits lipid peroxidation directly, perhaps by competing with iron for phospholipid binding sites. At low concentrations, enhancement of cell death may be due to the combined effect of increased NMDA receptor activity, glutathione depletion, and direct potentiation of lipid peroxidation.

Focal hyperexpression of hemeoxygenase-1 protein and messenger RNA in rat brain caused by cellular stress following subarachnoid injections of lysed blood.

Induction of the hemeoxygenase-1 (ho-1) stress gene is of importance for rapid heme metabolism and protection against oxidative injury in vitro and in vivo. Although ho-1 expression is observed in glia following exposure to whole blood and oxyhemoglobin, expression is mild, and other stress genes are not induced simultaneously in this setting. Hemeoxygenase-1 can be induced by several other physiological stresses in addition to heme. In the brain, ho-1 induction has been observed in the penumbra following focal cerebral ischemia. Because lysed blood is a spasmogen, the authors studied focal hyperexpression of the ho-1 gene after injection of lysed blood, whole blood, or saline into the cisterna magna of adult rats. Immunocytochemical analysis of HO-1 was performed at 1, 2, 3, and 4 days after the injections. Because the 70-kD inducible heat shock protein (HSP70) is induced by cellular stress, alternate sections were immunostained for HSP70 to assess whether focal hyperexpression was a stress phenomenon. An oligonucleotide probe was also used for in situ hybridization to demonstrate that ho-1 messenger (m)RNA was present. Focal HO-1 immunostained areas were observed after lysed blood injection only and were located mainly in the basal cortex and cerebellar hemisphere, although focal hyperexpression was also found in many other regions. The intensity of staining and the number of regions were maximum at 1 day. Double-labeled immunofluorescence revealed that many HO-1-immunoreactive cells were microglia. The HSP70 immunostaining of adjacent sections from the same animals demonstrated focal regions of immunoreactivity whose topography corresponded exactly with the topography of the HO-1-immunostained areas. Conventional histology in regions of HO-1 hyperexpression was often normal. In situ hybridization using the same oligonucleotide demonstrated that ho-1 mRNA was induced in focal areas of forebrain and in large regions of cerebellum within 6 hours of injection. These results demonstrate that focal hyperexpression of the ho-1 stress gene occurs after lysed blood injection and appears to be an indicator of cellular stress and injury in regions in which infarction does not occur. These results also suggest that cellular injury that occurs after injection of lysed blood may go undetected using conventional histology. Although direct heme metabolism was not investigated, our results indicate that rapid metabolism of heme, both intracellular and extracellular, may prove to be beneficial after subarachnoid hemorrhage.

Induction of heme oxygenase-1 (HO-1) in glia after traumatic brain injury.

In this study we examined the induction of heme oxygenase-1 (HO-1) in glia in the traumatized rat brain. HO-1 was immunolocalized in fixed sections of brain 3 h to 5 days after injury. Induction of this enzyme in astrocytes, microglia/macrophages, and oligodendrocytes was evaluated using immunofluorescent double labeling with monoclonal antibodies to glial fibrillary acidic protein, complement C3bi receptor, and myelin basic protein. Induction of HO-1 was apparent in the injured hemisphere and cerebellum as early as 24 h postinjury. The protein was likewise noted in similar regions of the brain at 72 h postinjury but appeared to be more widespread in its distribution. At 5 days postinjury, there was a notable decline in the degree of immunostaining for HO-1. HO-1 was typically induced in astrocytes in the cerebral cortex at the site of impact, in the deep cortical layers adjacent to the hemorrhagic lesions, and in the hippocampus. HO-1 was induced in Bergmann glia in the vermis of cerebellum. In addition, HO-1 was also induced in microglia/macrophages scattered throughout the ipsilateral cerebral cortex, cerebellum and subarachnoid space. These findings demonstrate prolonged glial induction of HO-1 in the traumatized brain. Such a response may reflect a protective role of these cells against secondary insults including oxidative stress.

Hemoglobin potentiates excitotoxic injury in cortical cell culture.

Excessive activation of glutamate receptors may contribute to neuronal loss after a traumatic or ischemic central nervous system insult. Such injuries are often associated with hemorrhage and extravasation of hemoglobin, a prooxidant and putative neurotoxin. In this study, we investigated the effect of nontoxic concentrations of hemoglobin on the neurotoxicity of the synthetic glutamate receptor agonists NMDA, AMPA, and kainate in primary murine cortical cultures. Continuous exposure to each excitotoxin alone for 24-28 h produced concentration-dependent neuronal death (EC(50) about 12 mu M for AMP(+)A, 50 mu M for kainate, and 12 mu M for NMDA). Hemoglobin 0.25-1.0 mu M consistently potentiated the neurotoxicity of low concentrations of AMPA and kainate, increasing neuronal loss by about 150% at 6 mu M AMPA and by about 90% at 30 mu M kainate. This effect was attenuated by the iron chelator deferoxamine and the alpha-tocopherol analogue trolox. Hemoglobin coexposure had less impact on slowly triggered NMDA neurotoxicity, significantly increasing neuronal death only at agonist concentrations that alone produced little or no injury. Hemoglobin pretreatment had no effect on the rapidly triggered excitotoxicity induced by brief exposure to high concentrations of NMDA. These results suggest that hemoglobin may contribute to neuronal loss after CNS hemorrhage by exacerbating excitotoxicity. At moderate levels of agonist exposure, this effect may be somewhat selective for the AMPA/kainate component of injury.

Heme-oxygenase-1 induction in glia throughout rat brain following experimental subarachnoid hemorrhage.

The heme released following subarachnoid hemorrhage is metabolized by heme-oxygenase (HO) to biliverdin and carbon monoxide (CO) with the release of iron. The HO reaction is important since heme may contribute to vasospasm and increase oxidative stress in cells. HO is comprised of at least two isozymes, HO-2 and HO-1. HO-1, also known as heat shock protein HSP32, is inducible by many factors including heme and heat shock. HO-2 does not respond to these stresses. To begin to examine HO activity following subarachnoid hemorrhage (SAH), the expression of HO-1 and HO-2 was investigated after experimental SAH in adult rats. Immunocytochemistry for HO-1, HO-2 and HSP70 proteins was performed at 1, 2, 3 and 4 days after injections of lysed blood, whole blood, oxyhemoglobin and saline into the cisterna magna. A large increase in HO-1 immunoreactivity was seen in cells throughout brain following injections of lysed blood, whole blood, and oxyhemoglobin but not saline. Lysed blood, whole blood and oxyhemoglobin induced HO-1 in all of the cortex, hippocampus, striatum, thalamus, forebrain white matter and in cerebellar cortex. HO-1 immunoreactivity was greatest in those regions adjacent to the basal subarachnoid cisterns where blood and oxyhemoglobin concentrations were likely highest. Double immunofluorescence studies showed the HO-1 positive cells to be predominately microglia, though HO-1 was induced in some astrocytes. HO-1 expression resolved by 48 h. HO-2 immunoreactivity was abundant but did not change following injections of blood. A generalized induction of HSP70 heat shock protein was not observed following injections of lysed blood, whole blood, oxyhemoglobin, or saline. These results suggest that HO-1 is induced in microglia throughout rat brain as a general, parenchymal response to the presence of oxyhemoglobin in the subarachnoid space and not as a stress response. This microglial HO-1 response could be protective against the lipid peroxidation and vasospasm induced by hemoglobin, by increasing heme clearance and iron sequestration, and enhancing the production of the antioxidant bilirubin.

Ultrastructure of excitotoxic neuronal death in murine cortical culture.

Ischemic and traumatic brain injury are likely to involve neuronal injury triggered by glutamate receptor overactivation. Although excitotoxic neuronal injury has been widely studied in the setting of primary culture, the extent to which these in vitro injury paradigms resemble in vivo ischemic injury morphologically has not previously been well studied. We studied glutamate receptor mediated neuronal death by transmission electron microscopy and light microscopy. Morphologic characteristics of neurons injured by 10 min exposure to 500 microM glutamate include rapid swelling of mitochondria and endoplasmic reticulum, and cytoplasmic and nuclear lucency. Both alpha-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid and kainic acid caused vacuolation, dilatation of the endoplasmic reticulum, cytoplasmic condensation and random condensation of chromatin with preserved mitochondria. None of these injuries was ameliorated by cycloheximide or actinomycin D; all were significantly lessened by aurintricarboxylic acid. Gel electrophoresis showed no increase in DNA fragmentation over control. The morphologic changes seen with alpha-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid and kainate are distinct from the changes induced by glutamate. Excitotoxic injury in this system due to high concentrations of glutamate resembles necrosis while the other agonists produce a different form of cell death which is neither necrosis nor apoptosis.

Induction of heme oxygenase-1 (HO-1) after traumatic brain injury in the rat.

The induction of heme oxygenase-1 (HO-1), the 32 kDa heat shock protein, was examined in the traumatized rat brain. At 24 h after either mild or severe brain injury or sham surgery, HO-1 was immunolocalized in fixed sections of brain. After mild brain injury, hemorrhage was apparent in the subarachnoid space, external capsule and cerebellum. HO-1 was induced in similar areas in macrophages in the subarachnoid space and in glia in the cortex adjacent to the site of impact, the ipsilateral hippocampus, external capsule and cerebellum. After severe brain injury, extensive hemorrhage occurred in the external capsule, hippocampus and cerebellum. HO-1 was induced in glia in these areas of hemorrhage but was more extensive than that seen after mild injury and included the contralateral external capsule and hippocampus. These findings demonstrate remarkable induction of HO-1 in glia in the injured brain. Since heme is a potent inducer of HO-1, it is likely that the subarachnoid and/or intraparenchymal blood induced HO-1 in the glia where the heme was metabolized to biliverdin, iron, and carbon monoxide.

Traumatic neuronal injury in cortical cell culture is attenuated by 21-aminosteroids.

The effect of the 21-aminosteroids U74500A and U74389F, alone and in combination with the NMDA receptor antagonist MK-801, on traumatic neuronal injury was quantitatively assessed in murine neocortical cell cultures. Consistent with prior observations, a mechanical insult to the culture monolayer resulted in widespread neuronal death over the following 24 h. Treatment with either U74500A or U74389F provided moderate protection, reducing neuronal death as measured by lactate dehydrogenase release by 25-50%. This effect was most consistent when these agents were preincubated for 2 h prior to injury. Combined treatment with a 21-aminosteroid plus the NMDA receptor antagonist MK-801 reduced injury more than either drug alone. Approximately 40% of the neuronal death occurring in the presence of MK-801 was blocked by concomitant treatment with 10 microM U74500A or U74389F. These results suggest that free radicals may contribute to cell death in this in vitro model of traumatic neuronal injury.

Assessment of hemoglobin-dependent neurotoxicity: alpha-alpha crosslinked hemoglobin.

Adult human hemoglobin A0 (HbA0) has been shown to be neurotoxic, and we wish to report on similar studies conducted using a modified hemoglobin, which has been crosslinked between the alpha subunits (alpha-alpha Hb). Cortical cell cultures were prepared from fetal Swiss-Webster mice at 15-16 days gestation. Mature cultures (days in vitro, 12-16) were exposed to alpha-alpha Hb in a defined medium for 24-48 hours at 37 degrees C. Low micromolar amounts of alpha-alpha Hb were neurotoxic in a concentration-dependent fashion. This toxicity was attenuated by the antioxidants Trolox and U-74500A and by the iron chelator deferoxamine. The hemoglobin-binding protein, haptoglobin, also completely blocked alpha-alpha Hb-dependent neurotoxicity. The latter result was unexpected because complex formation between alpha-alpha Hb and haptoglobin was not detected using assays of haptoglobin fluorescence and hemoglobin peroxidase activity.

NMDA neurotoxicity in murine cortical cell cultures is not attenuated by hemoglobin or inhibition of nitric oxide synthesis.

The role of nitric oxide in N-methyl-D-aspartate (NMDA) neurotoxicity was investigated in murine cortical cell cultures. Exposure of cultures to 300 microM NMDA for 5 min resulted in death of 50-80% of neurons over the subsequent 24 h. This injury was not attenuated by hemoglobin, the nitric oxide synthase (NOS) inhibitors NG monomethyl-L-arginine (MMA) or N omega-nitro-L-arginine (NA), or L-arginine depletion. Hemoglobin and NOS inhibitors consistently prevented the increase in cyclic guanosine monophosphate (cGMP) seen after NMDA exposure. These results suggest that NMDA neurotoxicity in this cell culture system is mediated, at least in part, by mechanisms other than NOS activation.

Neurotoxicity of hemoglobin in cortical cell culture.

Hemoglobin (Hb) has been demonstrated to be neurotoxic when injected into the cerebral cortex in vivo. However, associated systemic factors such as ischemia and epileptogenesis have limited investigations of Hb toxicity in the intact central nervous system (CNS). In this study, the neurotoxicity of human Hb was assessed in mixed neuronal and glial neocortical cell cultures derived from fetal mice. Exposure of cultures to Hb for 24-28 h produced widespread and concentration-dependent neuronal death (EC50 1-2.5 microM), without injuring glia. Brief exposures (1-2 h) were not toxic. Neuronal death was completely blocked by the 21-aminosteroid U74500A, the antioxidant Trolox, and the ferric iron chelator deferoxamine. The results of these experiments suggest that, in this system, Hb is a potent neurotoxin, and that Hb neurotoxicity may contribute to secondary injury processes after trauma and intracranial hemorrhage.

Pretreatment with NMDA antagonists limits release of excitatory amino acids following traumatic brain injury.

After central nervous system (CNS) trauma, there are marked elevations in the extracellular levels of excitatory amino acids (EAA), which are believed to contribute to delayed tissue damage. Administration of N-methyl-D-aspartate (NMDA) receptor antagonists reduces injury severity after brain or spinal cord trauma, presumably by blocking the postsynaptic NMDA receptor. In the present studies, levels of extracellular amino acids were monitored by microdialysis during, and after, a moderately severe fluid-percussion brain injury to rats. Pretreatment (15 min prior to injury) with the non-competitive NMDA antagonist dextrorphan or the competitive NMDA antagonist CGS 19755 significantly attenuated the post-traumatic increase in extracellular glutamate. Pretreatment with dextrorphan attenuated the post-traumatic increase in extracellular levels of aspartate; although these differences did not reach significance when examined as absolute values, they were significant when analyzed as percent increase over pre-trauma baseline levels. These results are consistent with recent experiments and suggest that NMDA antagonists may limit the release of glutamate and aspartate after trauma through a presynaptic mechanism.

Dextran-coupled deferoxamine improves outcome in a murine model of head injury.

Tissue damage involving oxygen-derived free radicals may be greatly exacerbated by free, reactive iron, which acts as a catalyst in oxidative reactions. The effects of free iron can be attenuated by the administration of deferoxamine (DFO), an iron chelator. However, DFO has limited therapeutic utility because it has a short plasma half-life (approximately 5.5 min in mice) and produces profound hypotension upon intravenous infusion. These negative attributes have been circumvented by the covalent attachment of DFO to large polymers, such as dextran or hydroxyethyl starch. The ability of the dextran-conjugated DFO (DEX-DFO) to inhibit iron-catalyzed reactions with lipids was compared to that of the native molecule in an in vitro model of CNS lipid degradation in the presence of 200 microM ferrous iron. There was no difference between native DFO and the modified form. Modified and unmodified DFO were also compared for therapeutic efficacy in a murine model of head injury. Using a previously described "grip test" as a measure of neurologic impairment following injury, DEX-DFO, native DFO, and dextran were administered intravenously 3-5 min after injury. Dextran-DFO significantly decreased the incidence of severe neurologic impairment at dosage levels of 0.1 (n = 92), 1.0 (n = 76), and 10.0 (n = 80) mg/kg. Administration of native DFO or dextran had no effect at the same dosages and concentrations. These results suggest that the murine model of head injury contains a significant iron-dependent component that should be assessed in other models of neural injury.