Raymond F. Regan

Estrogens attenuate neuronal injury due to hemoglobin, chemical hypoxia, and excitatory amino acids in murine cortical cultures

A growing body of evidence supports the hypothesis that estrogens may be beneficial in Alzheimers disease and other neurodegenerative processes. Less is known of their therapeutic potential in acute CNS insults. In this study, we assessed the effect of estrogens in three injury paradigms that may be relevant to CNS hemorrhage, trauma, and ischemia. Supraphysiologic concentrations of 17beta-estradiol, estrone, or equilin attenuated neuronal loss due to prolonged exposure to the pro-oxidant hemoglobin, with complete protection at 10 microM. Most of this effect persisted despite concomitant treatment with the antiestrogen ICI 182,780 or the protein synthesis inhibitor cycloheximide. In contrast, the non-estrogenic steroid methylprednisolone, which is currently in clinical use in spinal cord injury, reduced neuronal loss by only about 30%. High concentrations of equilin or estrone also attenuated the submaximal neuronal injury induced by 3.5-4.5 h exposure to the cytochrome oxidase inhibitor sodium azide, with near complete protection at 30 microM. Estrogens had a weaker and somewhat variable effect on pure excitotoxic injury, reducing neuronal loss due to 24 h kainate exposure by about half, and due to 24 h NMDA exposure by 15-65%; similar neuroprotection was provided by the antioxidant 21-aminosteroid U74500A. These results suggest that estrogens may be beneficial in acute CNS injuries associated with oxidative and excitotoxic stress. Investigation of high dose estrogen therapy in in vivo models of CNS hemorrhage, trauma, and ischemia is warranted.

Hemin induces an iron‐dependent, oxidative injury to human neuron‐like cells

Hemin is released from hemoglobin after CNS hemorrhage and is present at high micromolar concentrations in intracranial hematomas. This highly reactive compound is potentially cytotoxic via a variety of oxidative and nonoxidative mechanisms. However, despite its clinical relevance, little is known of its effect on neuronal cells. In this study, we tested the hypotheses that hemin is toxic to human neurons at physiologically relevant concentrations and that its toxicity is iron dependent and oxidative. A homogeneous population of neuron‐like cells was produced by sequential treatment of SH‐SY5Y cells with retinoic acid and brain‐derived neurotrophic factor, using the protocol of Encinas et al. Hemin exposure for 24 hr resulted in cell death that progressively increased between 3 and 30 μM (EC50 approximately 10 μM); protoporphyrin IX, the iron‐free congener of hemin, was not toxic. Cell death commenced at 14 hr and was preceded by a marked increase in cellular reactive oxygen species (ROS). Most injury and ROS production were prevented by concomitant treatment with an equimolar concentration of the lipid‐soluble iron chelator phenanthroline; the water‐soluble chelator deferoxamine was also effective at concentrations of 0.1 mM or higher. Heme oxygenase‐2 was constitutively expressed by these cells, and heme oxygenase‐1 was induced by hemin. Heme oxygenase inhibition attenuated ROS generation and reduced injury by about one‐third. Cell death was also prevented with the sulfhydryl reducing agents glutathione and mercaptoethanol. Nuclear morphology in the hours prior to cell lysis revealed a predominantly homogenous staining pattern; the percentage of fragmented nuclei was increased only at 4 hr and then accounted for only 1.45% ± 0.25% of cells. The general caspase inhibitor zVAD‐fmk had no effect on cell viability. These results suggest that hemin is toxic to human neuron‐like cells at concentrations that are less than 3% of those observed in intracranial hematomas. In this model, its toxicity is iron dependent, oxidative, and predominantly necrotic.

The effect of NMDA, AMPA/kainate, and calcium channel antagonists on traumatic cortical neuronal injury in culture

A traumatic insult was delivered to murine cortical neuronal and glial cell cultures by tearing the cell layer with a stylet in a grid pattern. Consistent with prior observations, neurons adjacent to a tear developed immediate swelling, and then went on to degenerate over the next several hours. Delivery of multiple tears produced enough cell death that measurable levels of lactate dehydrogenase accumulated in the bathing medium 24 h later, correlating well with the extent of cell death as assessed by Trypan blue exclusion and cell counts. 50-75% of this trauma-induced cell death was blocked by the NMDA receptor antagonist MK-801. 10-100 microM CNQX also attenuated neuronal degeneration, but this neuroprotective effect was likely due to attenuation of NMDA receptor-mediated toxicity, since the more specific AMPA/kainate antagonist NBQX was ineffective. CNQX also did not augment the protective effect of MK-801. High concentrations of nimodipine or nifedipine produced modest neuroprotective effects; either dihydropyridine when combined with MK-801 reduced injury more than MK-801 alone. These results suggest that traumatic neuronal death in this in vitro model is mediated in part by excessive activation of NMDA receptors, and in part by mechanisms sensitive to high concentrations of dihydropyridines, but not by AMPA/kainate receptors.

Heme oxygenase-2 knockout neurons are less vulnerable to hemoglobin toxicity

When cortical neurons are exposed to hemoglobin, they undergo oxidative stress that ultimately results in iron-dependent cell death. Heme oxygenase (HO)-2 is constitutively expressed in neurons and catalyzes heme breakdown. Its role in the cellular response to hemoglobin is unclear. We tested the hypothesis that HO-2 attenuates hemoglobin neurotoxicity by comparing reactive oxygen species (ROS) formation and cell death in wild-type and HO-2 knockout cortical cultures. Consistent with prior observations, hemoglobin increased ROS generation, detected by fluorescence intensity after dihydrorhodamine 123 or dichlorofluorescin-diacetate loading, in wild-type neurons. This fluorescence was significantly attenuated in cultures prepared from HO-2 knockout mice, and cell death as determined by propidium iodide staining was decreased. In other experiments, hemoglobin exposure was continued for 19 h; cell death as quantified by LDH release was decreased in knockout cultures, and was further diminished by treatment with the HO inhibitor tin protoporphyrin IX. In contrast, HO-2 knockout neurons were more vulnerable than wild-type neurons to inorganic iron. HO-1, ferritin, and superoxide dismutase expression in HO-2 -/- cultures did not differ significantly from that observed in HO-2 +/+ cultures; cellular glutathione levels were slightly higher in knockout cultures. These results suggest that heme breakdown by heme oxygenase accelerates the oxidative neurotoxicity of hemoglobin, and may contribute to neuronal injury after CNS hemorrhage.

The Effect of Magnesium on Oxidative Neuronal Injury In Vitro

Abstract: The effect of magnesium on the oxidative neuronal injury induced by hemoglobin was assessed in murine cortical cell cultures. Exposure to 5 µM 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 ∼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.

Heme oxygenase-1 induction protects murine cortical astrocytes from hemoglobin toxicity

Exposure to micromolar concentrations of hemoglobin (Hb) results in the oxidative death of cultured cortical neurons, but glia are resistant. The role of heme oxygenase-1 (HO-1) induction on this glial resistance was investigated. Within two hours of exposure to 5 microM Hb, immunoblotting demonstrated an increase in HO-1 in confluent glial cultures. Consistent with prior observations, 23-30 h Hb exposure had little or no effect on glial viability, as assessed by lactate dehydrogenase release. Concomitant treatment with the HO inhibitors tin protoporphyrin IX or the D-amino acid peptide rvnlrialry resulted in release of 40-71% of glial lactate dehydrogenase; protein synthesis inhibition with cycloheximide produced a similar effect. These results are consistent with the hypothesis that HO-1 induction protects cortical astrocytes from Hb toxicity.

Heme Oxygenase-2 Deficiency Contributes to Diabetes-Mediated Increase in Superoxide Anion and Renal Dysfunction

Heme oxygenase-1 (HO-1) and -2 play an important role in cytoprotection and are physiologic regulators of heme-dependent protein synthesis in renal tissues. The impact of HO-2 deletion comparing hyperglycemic HO-2 (+/+) mice and HO-2 knockout (-/-) mice was examined. Hyperglycemia was induced by streptozotocin (STZ) injection, and its effect on renal HO-1/HO-2 protein, HO activity, and creatinine levels were assessed. The effect of HO induction using systemic administration of the HO inducers heme or cobalt protoporphyrin and the effect of HO inhibition using systemic administration of the HO inhibitor tin mesoporphyrin also were assessed in STZ-treated mice. In STZ-treated HO-2 (-/-) mice, there was marked renal functional impairment as reflected by an increase in plasma creatinine, associated with acute tubular damage and microvascular pathology as compared with HO-2 (+/+). In these animals, HO activity was decreased with a concomitant increase in superoxide anion. Upregulation of HO-1 in HO-2 (-/-) mice by weekly administration of cobalt protoporphyrin prevented the increase in plasma creatinine levels and tubulointerstitial and microvascular pathology. Inhibition of HO activity by administration of tin mesoporphyrin accentuated superoxide production and increased creatinine levels in hyperglycemic HO-2 (-/-) mice. In conclusion, HO-2 deficiency enhanced STZ-induced renal dysfunction and morphologic injury and HO-1 upregulation in HO-2 (-/-) mouse rescue and prevented the morphologic damage. These observations indicate that HO activity is essential in preserving renal function and morphology in STZ-induced diabetic mice probably via mitigation of concomitant oxidative stress.

Heme oxygenase-2 gene deletion attenuates oxidative stress in neurons exposed to extracellular hemin.

BackgroundHemin, the oxidized form of heme, accumulates in intracranial hematomas and is a potent oxidant. Growing evidence suggests that it contributes to delayed injury to surrounding tissue, and that this process is affected by the heme oxygenase enzymes. In a prior study, heme oxygenase-2 gene deletion increased the vulnerability of cultured cortical astrocytes to hemin. The present study tested the effect of HO-2 gene deletion on protein oxidation, reactive oxygen species formation, and cell viability after mixed cortical neuron/astrocyte cultures were incubated with neurotoxic concentrations of hemin.ResultsContinuous exposure of wild-type cultures to 1–10 μM hemin for 14 h produced concentration-dependent neuronal death, as detected by both LDH release and fluorescence intensity after propidium iodide staining, with an EC50 of 1–2 μM; astrocytes were not injured by these low hemin concentrations. Cell death was consistently reduced by at least 60% in knockout cultures. Exposure to hemin for 4 hours, a time point that preceded cell lysis, increased protein oxidation in wild-type cultures, as detected by staining of immunoblots for protein carbonyl groups. At 10 μM hemin, carbonylation was increased 2.3-fold compared with control sister cultures subjected to medium exchanges only; this effect was reduced by about two-thirds in knockout cultures. Cellular reactive oxygen species, detected by fluorescence intensity after dihydrorhodamine 123 (DHR) staining, was markedly increased by hemin in wild-type cultures and was localized to neuronal cell bodies and processes. In contrast, DHR fluorescence intensity in knockout cultures did not differ from that of sham-washed controls. Neuronal death in wild-type cultures was almost completely prevented by the lipid-soluble iron chelator phenanthroline; deferoxamine had a weaker but significant effect.ConclusionsThese results suggest that HO-2 gene deletion protects neurons in mixed neuron-astrocyte cultures from heme-mediated oxidative injury. Selective inhibition of neuronal HO-2 may have a beneficial effect after CNS hemorrhage.

Ferritin induction protects cortical astrocytes from heme-mediated oxidative injury

Hemin is released from hemoglobin after CNS hemorrhage and may contribute to its cytotoxic effect. In a prior study, we demonstrated that heme oxygenase-1 induction protected murine cortical astrocytes from hemoglobin toxicity. Since heme metabolism releases iron, this observation suggested that these cells are able to effectively sequester and detoxify free iron. In this study, we tested the hypotheses that astrocytes increased ferritin synthesis after exposure to heme-bound iron, and that this induction protected cells from subsequent exposure to toxic concentrations of hemin. Incubation with low micromolar concentrations of hemin, hemoglobin, or ferrous sulfate increased ferritin expression, as detected on immunoblots stained with a polyclonal antibody that was raised against horse spleen ferritin. Time course studies demonstrated an increase in ferritin levels within 2 h. Weak and scattered cellular staining was detected by immunohistochemistry in control, untreated cultures, while diffuse immunoreactivity was observed in cultures exposed to heme-bound iron. An enhanced ferritin band was detected on immunoblots from cultures that were treated with purified apoferritin, consistent with astrocytic ferritin uptake. Immunoreactivity after apoferritin treatment was not altered by concomitant treatment with cycloheximide. Pretreatment with apoferritin protected astrocytes from hemin toxicity in a concentration-dependent fashion between 1 and 4 mg/ml. At the highest concentration, cell death due to a 6-h exposure to 30 microM hemin was decreased by about 85%. A protective effect was also produced by induction of endogenous ferritin with nontoxic concentrations of ferrous sulfate, hemoglobin, or hemin. These results suggest that cortical astrocytes respond to exogenous heme-bound or free iron by rapidly increasing ferritin synthesis. The combined action of heme oxygenase-1 and ferritin may be a primary astrocytic defense against heme-mediated injury.

Increasing Expression of Heme Oxygenase-1 by Proteasome Inhibition Protects Astrocytes from Heme-mediated Oxidative Injury

Hemin is released from hemoglobin after CNS hemorrhage, and may contribute to cell loss in surrounding tissue. Heme oxygenase-1 (HO-1) is induced by these injuries, and may have an effect on cell viability. In a prior study, we reported that increasing HO-1 expression by adenoviral gene transfer prior to hemin exposure attenuated oxidative stress and cell death in astrocytes. However, rapid gene transfer to the CNS may not be feasible. HO-1 expression is controlled by a stress-responsive transcription factor, Nrf2, which is a labile protein that is subject to proteasomal degradation. In this study, we hypothesized that preventing degradation of Nrf2 with a lipid-soluble proteasome inhibitor would increase HO-1 expression and protect astrocytes from hemin. Treatment of cortical astrocyte cultures with 1 microM MG-132 resulted in a rapid increase in Nrf2, to a level that was five-fold that of vehicle-treated cultures by 2 h. This was followed by a three to six-fold increase in HO-1 expression that persisted through the 16 h observation period. Exposure of cultures to 30 microM or 60 microM hemin for 8 h resulted in death, as measured by LDH release, of 39+/-3.0 or 67.5+/-5.9% of astrocytes. Pre-treatment with MG-132 prevented approximately half of this injury. Cytoprotection persisted at 24 h, and was also observed when injury was assessed via the MTT assay. Astrocyte protein oxidation produced by hemin was also significantly attenuated by MG-132 pre-treatment. These results suggest that increasing HO-1 expression with a proteasome inhibitor protects astrocytes from heme-mediated oxidative injury. This pharmacological approach may provide a mechanism for rapidly upregulating HO-1 in astrocytes after CNS hemorrhage.