|Home | About | Journals | Submit | Contact Us | Français|
Intracerebral hemorrhage (ICH) is a devastating form of stroke. In this study, we examined the efficacy of deferoxamine (DFX), an iron chelator, after collagenase-induced ICH in 12-month-old mice. Intracerebral hemorrhage was induced by intrastriatal injection of collagenase. Deferoxamine (200mg/kg, intraperitoneal) or vehicle was administrated 6hours after ICH and then every 12hours for up to 3 days. Neurologic deficits were examined on days 1 and 3 after ICH. Mice were killed after 1 or 3 days of DFX treatment for examination of iron deposition, neuronal death, oxidative stress, microglia/astrocyte activation, neutrophil infiltration, brain injury volume, and brain edema and swelling. Collagenase-induced ICH resulted in iron overload in the perihematomal region on day 3. Systemic administration of DFX decreased iron accumulation and neuronal death, attenuated production of reactive oxygen species, and reduced microglial activation and neutrophil infiltration without affecting astrocytes. Although DFX did not reduce brain injury volume, edema, or swelling, it improved neurologic function. Results of our study indicate that iron toxicity contributes to collagenase-induced hemorrhagic brain injury and that reducing iron accumulation can reduce neuronal death and modestly improve functional outcome after ICH in mice.
Intracerebral hemorrhage (ICH) is a common and devastating subtype of stroke that is associated with high morbidity and mortality (Qureshi et al, 2009). A better understanding of the pathophysiology of ICH may lead to better clinical management of patients with ICH (Wang, 2010; Wang and Doré, 2007b). It is well known that hemorrhaged blood is highly toxic to brain tissue, as is the heme derived from hemoglobin and other hemoproteins. Heme is degraded by heme oxygenase into iron, biliverdin, and carbon monoxide (Wang and Doré, 2007a). Research suggests that hemoglobin breakdown and subsequent iron accumulation within the brain mediate secondary brain injury after ICH (Wagner et al, 2003; Wang, 2010; Xi et al, 2006) and that iron chelation can be a potential therapy for ICH (Selim, 2009). However, conflicting results have been reported regarding the efficacy of the iron chelator deferoxamine (DFX) in different animal models of ICH (Gu et al, 2009; Okauchi et al, 2009; Song et al, 2007; Warkentin et al, 2010). Moreover, the effect of DFX on iron-mediated oxidative stress and inflammation after ICH has not been examined.
The aged population is vulnerable to ICH, and aging might affect ICH-induced brain injury and functional outcome. Studies have shown that aging exacerbates ICH-induced brain injury in rats and that the cause is most likely overactivation of microglia/macrophages and greater induction of heme oxygenase-1 in the perihematomal region (Gong et al, 2005; Lee et al, 2009). As most preclinical studies of ICH have been carried out in young animals, which limit the direct translation of preclinical studies into clinical trials, we investigated iron toxicity and iron-mediated oxidative damage in aged mice subjected to a collagenase-induced ICH model and assessed the efficacy of DFX on ICH outcomes.
This study was conducted in accordance with the National Institutes of Health guidelines for the use of experimental animals. Experimental protocols were approved by the Johns Hopkins University Animal Care and Use Committee. Twelve-month-old C57BL/6 male mice (weighing 25 to 35g) were obtained from Charles River Laboratories (Wilmington, MA, USA). This age was selected because cerebrovascular effects of aging are well developed in mice at 12 months (Park et al, 2007).
The procedure for modeling ICH by intrastriatal injection of collagenase was adapted to mice from an established rat protocol (Rosenberg et al, 1990) and has been described previously (Grossetete and Rosenberg, 2008; Wang et al, 2003). We injected mice in the left striatum with collagenase VII-S (0.075Units in 500nL saline; Sigma, St Louis, MO, USA) at the following stereotactic coordinates: 1.0mm anterior and 2.2mm lateral of the bregma, 2.7mm in depth. This procedure resulted in reproducible lesions that were mostly restricted to the striatum.
Mice were randomly assigned to receive either DFX (200mg/kg; Sigma) or vehicle (saline). All injections were administrated intraperitoneally at 6hours after collagenase injection and then every 12hours for 3 days. The dosing regimens were based on previous work performed in the blood model of ICH in rats (Okauchi et al, 2009). The percentage change in body weight was calculated according to the formula: change in body weight (%)=(posthemorrhage weight at each time point−prehemorrhage weight)/prehemorrhage weight.
Mice (n=5 per group) were anesthetized and underwent intracardiac perfusion with 4% paraformaldehyde in 0.1mol/L phosphate-buffered saline (pH 7.4). The brains were removed and kept in 4% paraformaldehyde for 24hours and then immersed in 30% sucrose for 3 days at 4°C. The brains were then cut into 30-μm-thick coronal sections with a cryostat.
Iron deposition was evaluated with Perls staining. Fluoro-Jade B (FJB) was used to quantify neuronal death (Wang and Tsirka, 2005a; Xue et al, 2009). To quantify iron- or FJB-positive cells, three sections per mouse with similar areas of hematoma were selected. The numbers of iron- or FJB-positive cells from 9 locations per mouse (3 fields per section × 3 sections per mouse) were averaged and expressed as positive cells per square millimeter. Sections (n=5 per group) were analyzed by an observer blinded to the experimental cohort.
Immunofluorescence was carried out as described previously (Wang and Tsirka, 2005a). The primary antibodies used were rabbit anti-myeloperoxidase (neutrophil marker; 1:500; Dako, Carpinteria, CA, USA), rabbit anti-Iba 1 (microglial marker; 1:500; Wako Chemicals, Richmond, VA, USA), and rabbit anti-glial fibrillary acidic protein (astrocyte marker; 1:500; Dako). Stained sections were examined using a fluorescence microscope (Eclipse TE2000-E, Nikon, Tokyo, Japan). The numbers of immunoreactive cells over a microscopic field of × 40 from 12 locations per mouse (4 fields per section × 3 sections per mouse) were averaged and expressed as positive cells per square millimeter. Sections (n=7 per group) were analyzed by an investigator blinded to the experimental cohort.
Production of reactive oxygen species (ROS) after ICH was investigated by in situ detection of oxidized hydroethidine (Kamada et al, 2007; Wang and Tsirka, 2005a). Hydroethidine, a cell-permeable oxidative fluorescent dye, is oxidized by superoxide to ethidium (Bindokas et al, 1996), which intercalates within DNA and emits a red fluorescent signal. Fluorescence intensity was quantified in predefined areas of the hemorrhagic striatum (at the injection site and at 360μm on each side) after subtraction of the color density on the contralateral striatum. Sections (n=5 per group) were analyzed by an observer blinded to the experimental cohort using ImageJ software (version 1.42q; NIH, Bethesda, MD, USA).
Neurologic deficits were assessed on days 1 and 3 after ICH. An investigator blinded to the experimental cohort scored all mice on six neurologic tests, including body symmetry, gait, climbing, circling behavior, front limb symmetry, and compulsory circling (Wang and Doré, 2007a). Each test was graded from 0 to 4, establishing a maximum deficit score of 24.
Mice were killed after the neurologic examination on day 3 after ICH. The entire brain of each mouse was cut into 50-μm-thick sections with a cryostat. Sections were stained with Cresyl violet (for neurons) and Luxol fast blue (for myelin) before being quantified for gray and white matter injury using SigmaScan Pro software (version 5.0.0 for Windows; Systat, Port Richmond, CA, USA). The injury volume in cubic millimeters was calculated by multiplying the thickness by the sum of the damaged areas of each section (Wang et al, 2003). Sections (n=8 per group) were analyzed by an observer blinded to the experimental cohort.
Brain edema was determined by the wet–dry weight ratio method as described previously (Wang and Tsirka, 2005a). Brain water content was expressed as (wet weight−dry weight)/wet weight of brain tissue × 100%.
Brain swelling was quantified by calculating the percentage of hemispheric enlargement at 3 days after ICH (Wang et al, 2008). Hemisphere enlargement (%) was expressed as: ((ipsilateral hemisphere volume−contralateral hemisphere volume)/contralateral hemisphere volume) × 100. Sections (n=8 per group) were analyzed by an observer blinded to the experimental cohort.
All data are expressed as means±s.d. Differences between two groups were determined by two-tailed Student's t-test. Statistical significance was set at P<0.05.
Using Perls staining, we found iron-positive cells in the perihematomal region on day 3 after ICH. On the basis of cell morphology, most of the Perls-positive cells were activated amoeboid microglia/macrophages with cytoplasmic iron deposits. Iron deposition was decreased on day 3 after ICH in mice administered DFX twice daily (n=5 per group, P<0.01; Figures 1A and 1B).
Fluoro-Jade B staining was used to detect neuronal degeneration. As with iron deposition, DFX treatment significantly reduced the number of FJB-positive cells in the perihematomal region on day 3 after ICH (n=5 per group, P<0.01; Figures 1C and 1D).
Neuronal death can be caused by iron-mediated oxidative stress. We used the hydroethidine technique to study the effect of DFX on posthemorrhagic ROS production. The presence of ROS was observed as red particles in the cell and was evident in the perihematomal region on day 1 after ICH. In the contralateral side, ROS signals were present but relatively weak. In DFX-treated mice, ROS production was attenuated in the perihematomal region (n=5 per group, P<0.01 versus vehicle-treated group; Figures 2A and 2B).
Microglia/macrophage activation occurs within 1hour after ICH (Wang and Doré, 2007a) and contributes to ICH-induced brain injury (Wang and Doré, 2007b). Activated microglia/macrophages in the hemorrhagic striatum were characterized as cells with an amoeboid or spherical appearance and a cell body usually >7.5μm in diameter with short, thick processes and intense immunoreactivity. Resting microglia on the contralateral side were characterized by small cell bodies, long processes, and weak immunoreactivity. To examine the effect of DFX on the state of microglia/macrophages after ICH, we used immunofluorescence staining of Iba1, a marker for microglia/macrophages (Shehadah et al, 2010; Wang and Doré, 2007a). Using a combination of morphologic criteria and a cell body diameter cutoff of 7.5μm, microglia/macrophages were classified as either resting or activated (Wang et al, 2008). Activated microglia/macrophages were evident in the perihematomal region on day 3 after ICH, but DFX treatment significantly reduced their numbers (n=7 per group, P<0.05 versus vehicle-treated group; Figures 3A and 3B).
Compared with resting astrocytes, reactive astrocytes in the hemorrhagic striatum exhibit more intense glial fibrillary acidic protein immunoreactivity and a greater number, length, and thickness of glial fibrillary acidic protein-positive processes. We used glial fibrillary acidic protein immunofluorescence labeling to examine the effect of DFX on astrocyte reactivity. We found that the number of reactive astrocytes in the perihematomal region on day 3 after ICH was unchanged by DFX treatment (n=5 per group, P>0.05 versus vehicle-treated group; Figures 3A and 3B).
Neutrophils infiltrate the hemorrhagic striatum starting as early as 4hours after ICH (Wang and Tsirka, 2005a). To examine the effect of DFX on neutrophil infiltration, we used myeloperoxidase immunofluorescence labeling. Significantly more myeloperoxidase-immunoreactive neutrophils were present in the hemorrhagic striatum of vehicle-treated mice than in that of DFX-treated mice on day 3 after ICH (n=5 per group, P<0.05 versus vehicle-treated group; Figures 3A and 3B).
To examine the effect of DFX on neurobehavioral deficits after ICH, we assessed neurologic function at baseline and on days 1 and 3 after ICH. Compared with baseline values, neurologic deficits were apparent in all mice subjected to ICH. Mice treated with DFX had neurologic deficit scores that were significantly lower than those of vehicle-treated mice on day 3 but not on day 1 after ICH (P<0.05); however, neurologic scores for the individual tests were not significantly different between treatment groups (all P>0.05; n=10 per group, Figure 4).
As DFX treatment improved neurologic function, we assessed its effect on brain injury volume, edema, and swelling. Brain injury volume was measured by Luxol fast blue/Cresyl violet staining 3 days after ICH. Contrary to our expectations, DFX treatment did not reduce brain injury volume on day 3 after ICH (n=8, P>0.05; Figures 5A and 5B). Similarly, ICH caused a marked increase in brain water content in the ipsilateral striatum 3 days after ICH that was not affected by DFX treatment (n=4, P>0.05; Figure 5C).
We next measured the percentage of hemispheric enlargement to evaluate brain swelling, which is believed to contribute to brain damage and death that result from severe stroke. Hemispheric enlargement was quantified by digitizing serial brain sections and calculating the cumulative area. Like brain water content, brain swelling was not significantly different between vehicle- and DFX-treated groups (n=8, P>0.05; Figure 5D).
Mouse body weight was decreased compared with baseline on days 1 and 3 after ICH. Surprisingly, DFX treatment increased body weight loss by 5.0% 3 days after ICH (n=10, P<0.05 versus vehicle-treated group; Figure 6).
In this study, we found that collagenase-induced ICH in mice results in iron overload, oxidative damage, glial activation, and neutrophil infiltration, consequences that may contribute to secondary brain damage. Using this model, we have confirmed the potential neuroprotective properties of the iron chelator DFX in 12-month-old mice with collagenase-induced ICH. We showed that DFX decreased iron accumulation and neuronal death, attenuated production of ROS, and reduced microglial activation and neutrophil infiltration. Although DFX treatment improved neurologic function, it did not reduce brain injury volume, edema, or swelling. These findings, collectively, provide novel evidence that iron toxicity contributes to collagenase-induced hemorrhagic brain injury and that reducing iron accumulation might reduce neuronal death and improve functional outcome after ICH.
It has been suggested that ICH-associated secondary brain injury is mediated by the generation of ferrous iron during heme degradation. Iron is toxic to neurons because it catalyzes the Fenton reaction, which produces highly reactive hydroxyl radicals that lead to oxidative stress and cell death (Gaasch et al, 2007). Moreover, iron could induce neuronal death even after it has bound to ferritin because it can be locally released in its ferrous form under the acidic conditions that follow stroke (Bishop and Robinson, 2001). Iron-mediated neurotoxicity and the benefits of DFX have been reported in a whole-blood model in rats and piglets (Gu et al, 2009; Okauchi et al, 2009; Song et al, 2007), but we are the first to show iron accumulation after collagenase-induced ICH in mice. Consistent with the results obtained from the whole-blood model in piglets (Gu et al, 2009), we showed in this study that iron chelation with DFX decreases iron accumulation and neuronal death after collagenase-induced ICH in mice.
It is well known that ROS are produced during normal oxidative metabolism, but high ROS levels can damage neurons and cause death. Abnormal iron overload is believed to participate in the induction of toxic ROS. Substantial evidence indicates that ROS are critical to the oxidative brain damage that occurs after ICH (Wang and Doré, 2007b). After ICH, the extracellular spaces of the brain are exposed to high concentrations of hemoglobin and its breakdown products. The fact that iron chelators and free radical scavengers can block hemoglobin-induced neurotoxicity (Wang et al, 2002) suggests that iron and iron-mediated oxidative stress contribute to hemoglobin-induced neurotoxicity. Indeed, high levels of oxidative stress, as measured by protein carbonyl formation or increased ethidium (oxidized hydroethidine), have been shown after intrastriatal injection of blood or collagenase (Qu et al, 2007; Wang and Tsirka, 2005a). Neurons in the hemorrhagic brain show a decreased ability to respond to oxidative stress (Wang and Doré, 2007a), particularly with regard to their low levels of glutathione and glutathione peroxidase. Consequently, excess iron released during heme degradation after ICH may predispose neurons to iron-induced oxidative stress. Deferoxamine binds ferric iron and prevents the formation of hydroxyl radicals through the Fenton reaction. Moreover, DFX reduces hemoglobin-induced brain Na+/K+ ATPase inhibition and neuronal toxicity (Song et al, 2007; Wan et al, 2006). In our study, ROS production measured by increased ethidium was markedly increased in the hemorrhagic striatum 1 day after ICH. Deferoxamine attenuated ROS production, further supporting the idea that iron toxicity underlies the increase in oxidative damage after ICH.
In addition to iron-mediated ROS production, inflammation has also been shown to be associated with neuronal death (Wang, 2010; Wang and Doré, 2007b). We along with others have shown that activated microglia and astrocytes and infiltrating neutrophils are present in the hemorrhagic striatum early after ICH (Wang and Doré, 2007b). Increasing evidence supports the premise that activated microglia/macrophages and infiltrating neutrophils are the major sources of proinflammatory mediators (Wang, 2010; Wang and Doré, 2007b). Consistent with this notion, inhibition of microglial activation before or after ICH decreased neuronal death and improved neurologic function (Wang et al, 2003; Wang and Tsirka, 2005b). We showed for the first time that DFX reduced the number of activated microglia/macrophages and infiltrating neutrophils but not of astrocytes 3 days after ICH. Although the underlying mechanisms are not clear, microglia have a larger capacity to take up free iron than do neurons and astrocytes (Bishop et al, 2010). Iron accumulation in microglia might stimulate the activation of these cells in the hemorrhagic brain. Deferoxamine may suppress inflammation by reducing iron-mediated oxidative damage or by blocking the signals that activate microglia and recruit neutrophils. Astrocytes provide little iron storage (Zecca et al, 2004) and hence are highly resistant to iron-induced toxicity (Kress et al, 2002).
Studies have shown that DFX improves neurologic function in rats subjected to the whole-blood ICH model (Okauchi et al, 2009, 2010). Consistent with these studies, we found that DFX improved neurologic function after a 3-day treatment when begun 6hours after ICH. However, the improvement was relatively modest. Furthermore, DFX did not reduce lesion volume, edema, or swelling, results consistent with those reported in the rat collagenase model (Warkentin et al, 2010). In addition, we found that DFX slightly enhanced weight loss. Although the underlying mechanisms are not clear, DFX produces adverse effects in the digestive system that include abdominal discomfort and nausea (Okauchi et al, 2010). Others have not reported weight loss in rats exposed to the whole-blood or collagenase model (Okauchi et al, 2009; Warkentin et al, 2010), but the discordant findings may be explained by differences in animal models and species and the age of the animals studied.
In conclusion, systemic administration of DFX decreased perihematomal iron accumulation and neuronal death, attenuated production of ROS, and reduced microglial activation and neutrophil infiltration. Although DFX treatment did not reduce lesion volume, brain edema, or brain swelling, it modestly improved neurologic function. These findings provide novel evidence that iron chelation with DFX could offer some benefit to patients with ICH.
The authors thank Dr Raymond C Koehler, Dr Adam Sapirstein, Sarah Busse, and Claire Levine for their important contributions.
The authors declare no conflict of interest.
This work was supported by AHA 09BGIA2080137 and NIH K01AG031926 (JW).