Caspase-dependent and -independent neuronal death: two distinct pathways to neuronal injury.

Caspases are cysteine proteases that mediate apoptotic death in a variety of cellular systems, including neurons. Caspases are activated through extrinsic or intrinsic pathways. The latter is used by most neurons in most situations. In this pathway, release of mitochondrial cytochrome c into the cytoplasm induces formation of the apoptosome, which leads to the activation of caspase 9 and subsequently other caspases. Recent data demonstrate that when caspase activation is inhibited at or downstream of the apoptosome, neurons undergo a delayed, caspase-independent death. Furthermore, there are instances, most notably following excitotoxic injury and calcium overload, in which the direct cell death pathway elicited differs from classical apoptosis. The molecular and biochemical features of such caspase-independent, nonapoptotic forms of neuronal death are just beginning to be elucidated, but alterations at the level of the mitochondria and noncaspase proteases play significant roles. Mitochondrial alterations in caspase-independent death may include energy depletion, generation of free radicals, opening of the permeability transition pore, and release of cytotoxic proteins, such as apoptosis-inducing factor. The particular mechanisms employed can be context dependent. In disease states, in which a combination of apoptotic and nonapoptotic death occurs, therapeutic strategies need to take into account both caspase-dependent and -independent pathways.     

The metabotropic glutamate system promotes neuronal survival through distinct pathways of programmed cell death

Activation of the metabotropic glutamate receptor (mGluR) system can prevent free radical, nitric oxide (NO)-induced programmed cell death (PCD). To investigate the mechanisms utilized by the mGluR system to regulate the induction of PCD, we examined the course of PCD in real time in individual, living, primary hippocampal neurons. We assessed both phosphatidylserine (PS) externalization, an early event in PCD, and DNA fragmentation during NO toxicity and mGluR modulation to determine the individual contributions of PS externalization and genomic DNA fragmentation during neuronal PCD. Exposure to the NO donors (300 microM SNP or 300 microM NOC-9) induced PCD in approximately 75% of neurons over a 24-h period. The externalization of PS in neurons increased to 21 +/- 2% as early as 3 h following NO exposure and then increased to 80 +/- 2% over a 24-h period. The externalization of PS was independent of the loss of membrane integrity. Agonists for individual mGluR subgroups were equally able to prevent NO-induced neuronal death and DNA degradation, yet they possessed differential abilities to regulate PS externalization. The group I agonist DHPG (750 microM) and the group III agonist L-AP4 (750 microM) both prevented and reversed NO-induced PS externalization. In contrast, activation of group II subtypes using L-CCG-I (750 microM) did not prevent PS externalization. Employing an experimental model that independently led to the externalization of PS residues, we demonstrated that PS externalization does not immediately impact on neuronal survival. Yet, subsequent neuronal survival may ultimately depend upon preventing PS externalization to avoid neuronal tagging for phagocytosis. Since group I and III mGluR subtypes possess the unique ability to maintain genomic integrity and membrane PS asymmetry, these agents may provide superior overall protection against NO-induced neuronal injury.     

MAP kinase pathways in neuronal cell death

The signaling pathways which contribute to neuronal death during development, aging and disease have been extensively studied. While initial efforts focused on developmental death, increasing evidence suggests that mitogen-activated protein kinase pathways play a role in human pathology. In particular, the c-Jun N-terminal kinases (JNKs), mitogen-activated protein kinases activated by extracellular stimuli including stress, are a major focus. Knock-out mouse studies have demonstrated that removing particular JNK genes can reduce the severity in various disease scenarios, including those which are used to model Parkinson's disease and cerebral ischemia. In addition, activation of JNKs can be seen in human disease tissue. In this review we bring together the evidence for JNK being an important regulator of neuronal loss and outline the advancement of small molecule inhibitors for future therapeutic intervention.     

Prevention of nitric oxide-induced neuronal injury through the modulation of independent pathways of programmed cell death.

Neuronal injury may be dependent upon the generation of the free radical nitric oxide (NO) and the subsequent induction of programmed cell death (PCD). Although the nature of this injury may be both preventable and reversible, the underlying mechanisms that mediate PCD are not well understood. Using the agent nicotinamide as an investigative tool in primary rat hippocampal neurons, the authors examined the ability to modulate two independent components of PCD, namely the degradation of genomic DNA and the early exposure of membrane phosphatidylserine (PS) residues. Neuronal injury was determined through trypan blue dye exclusion, DNA fragmentation, externalization of membrane PS residues, cysteine protease activation, and the measurement of intracellular pH (pHi). Exposure to the NO donors SIN-1 and NOC-9 (300 micromol/L) alone rapidly increased genomic DNA fragmentation from 20 +/- 4% to 71 +/- 5% and membrane PS exposure from 14 +/- 3% to 76 +/- 9% over a 24-hour period. Administration of a neuroprotective concentration of nicotinamide (12.5 mmol/L) consistently maintained DNA integrity and prevented the progression of membrane PS exposure. Posttreatment paradigms with nicotinamide at 2, 4, and 6 hours after NO exposure further demonstrated the ability of this agent to prevent and reverse neuronal PCD. Although not dependent upon pHi, neuroprotection by nicotinamide was linked to the modulation of two independent components of neuronal PCD through the regulation of caspase 1 and caspase 3-like activities and the DNA repair enzyme poly(ADP-ribose) polymerase. The current work lays the foundation for the development of therapeutic strategies that may not only prevent the course of PCD, but may also offer the ability for the repair of neurons that have been identified through the loss of membrane asymmetry for subsequent destruction.     

Expression time course and spatial distribution of activated caspase-3 after experimental status epilepticus: contribution of delayed neuronal cell death to seizure-induced neuronal injury

Pilocarpine-induced status epilepticus (PCSE) is a widely used model to study neurodegeneration in limbic structures after prolonged epileptic seizures. However, mechanisms mediating neuronal cell death in this model require further characterization. Examining the expression time course and spatial distribution of activated caspase-3, we sought to determine the role of apoptosis in PCSE-mediated neuronal cell death. Expression of activated caspase-3, predominantly located in neurons, was detected 24 h (amygdala; piriform and temporal cortex) and 7 days (hippocampus; amygdala; piriform, temporal and parietal cortex; thalamus) after PCSE with strongest induction being observed in the amygdala, the piriform cortex, and the hippocampus. Further analysis revealed TUNEL positivity (24 h and 7 days after SE) and a significant, progressive neuronal cell loss in all brain regions displaying caspase-3 activation. Corresponding to high levels of activated caspase-3 expression, neuronal cell loss was most pronounced in the amygdala, piriform cortex, and dorsal CA-1 hippocampus. These results demonstrate that apoptosis contributes significantly to PCSE-induced neuronal cell death.     

Clusterin interaction with Bcl-xL is associated with seizure-induced neuronal death

Status epilepticus causes significant damage to the brain, and cellular injury due to prolonged seizures may cause the pathogenesis of epilepsy or cognitive deficits. Clusterin mediates several cell signaling pathways, including cell death or survival pathways in the brain. A nuclear form of clusterin protein has been suggested to have pro-apoptotic properties. Bcl-x(L) functions as a dominant-negative modulator of the pro-apoptotic protein Bax. However, the relationship between clusterin and Bcl-x(L) in cell death signaling in the brain remains unknown. Therefore, we examined whether clusterin interacts with Bcl-x(L) after seizures or whether this interaction is related to neuronal death. We found increased levels of nuclear clusterin and cleaved caspase-3 in CA3 neurons after prolonged seizures induced by systemic kainic acid, along with extensive hippocampal cell death, as evidenced by terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate nick-end labeling (TUNEL) and anti-active caspase-3 staining. Furthermore, co-immunoprecipitation and double immunofluorescence analyses revealed that clusterin interacted with Bcl-x(L) in dying CA3 neurons while the levels of Bcl-x(L), Bad or Bax remained constant. These findings provide evidence that nuclear clusterin signals cell death at least via an interaction with Bcl-x(L) in the hippocampus after seizures, suggesting that targeting nuclear clusterin may be a promising novel strategy to protect against seizure-induced neuronal injury.     

Possible involvement of programmed cell death pathways in the neuroprotective action of polyphenols

One of the hallmarks of Alzheimer's disease is the accumulation of senile plaques composed of extra-cellular aggregates of beta-amyloid (Aβ) peptides. It is well established that at least in vitro, Aβ triggers apoptotic cell death via the activation of caspase-dependent and -independent cell death effectors, namely caspase-3 and apoptosis inducing factor (AIF), respectively. Epidemiological studies have reported that elderly people have a lower risk (up to 50%) of developing dementia if they regularly eat fruits and vegetables and drink tea and red wine (in moderation). Numerous studies indicate that polyphenols derived from these foods and beverages account for the observed neuroprotective effects. In particular, we have reported that polyphenols extracted from green tea (i.e. epigallocatechin gallate or EGCG) and red wine (i.e. resveratrol) block Aβ-induced hippocampal cell death, by at least partially inhibiting Aβ fibrillisation. It has been shown that polyphenols may also modulate caspase-dependent and -independent programmed cell death (PCD) pathways. Indeed, polyphenols including resveratrol, EGCG and luteolin significantly inhibit the activation of the key apoptotic executioner, caspase-3 and are able to modulate mitogen-activated protein kinases known to play an important role in neuronal apoptosis. Moreover, it has been reported that polyphenols may exert their anti-apoptotic action by inhibiting AIF release from mitochondria, thus providing new mechanism of action for polyphenols. This review aims to update the current knowledge regarding the differential effects of polyphenols on PCD pathways and discuss their putative neuroprotective action resulting from their capacity to modulate these pathways.     

Strain differences in seizure-induced cell death following pilocarpine-induced status epilepticus

Mouse strains differ from one another in their susceptibility to seizure-induced excitotoxic cell death. Previously, we have demonstrated that mature inbred strains of mice show remarkable genetic differences in susceptibility to the neuropathological consequences of seizures in the kainate model of status epilepticus. At present, while the cellular mechanisms underlying strain-dependent differences in susceptibility remain unclear, some of this variation is assumed to have a genetic basis. However, it remains unclear whether strain differences in susceptibility to seizure-induced cell death observed following kainate administration are observed following systemic administration of other chemoconvulsants. In rodents, the cholinomimetic convulsant pilocarpine is widely used to induce status epilepticus (SE), followed by hippocampal damage and spontaneous recurrent seizures, resembling temporal lobe epilepsy. This model has initially been described in rats, but is increasingly used in mice. We characterized neuronal pathologies after pilocarpine-induced status epilepticus (SE) in eight inbred strains of mice focusing on the hippocampus. A ramping-up dose protocol for pilocarpine was used and behavior was monitored for 4-5 h. While we did not observe any significant differences in seizure latency or duration to pilocarpine among the inbred strains, we did observe a significant difference in susceptibility to the neuropathological consequences of pilocarpine-induced SE. Of the eight genetically diverse mouse strains screened for pilocarpine-induced status, BALB/cJ and BALB/cByJ were the only two strains that were resistant to the neuropathological consequences of seizure-induced cell death. Additional studies of these murine strains may be useful for investigating genetic influences on pilocarpine-induced status epilepticus.     

Development of a model of seizure-induced hippocampal injury with features of programmed cell death in the BALB/c mouse

Although mice are amenable to gene knockout, they have not been exploited in the setting of seizure-induced neurodegeneration due to the resistance to injury of key mouse strains. We refined and developed models of seizure-induced neuronal death in the C57BL/6 and BALB/c strains by focally evoking seizures using intra-amygdala kainic acid. Seizures in adult male BALB/c mice, or C57BL/6 mice as reference, caused ipsilateral death of CA1 and CA3 neurons within the hippocampus. Termination of seizures by lorazepam was more effective than diazepam in both strains, largely restricting neuronal loss to the CA3 sector. Electroencephalography (EEG) recordings defined injurious and non-injurious seizure patterns, which could not be separated adequately by behavioral observation alone. Degenerating neurons in the hippocampus were positive for DNA fragmentation and approximately a third of these exhibited morphologic features of programmed cell death. Western blot analysis revealed the cleavage of caspase-8 after seizures in both strains. These data refine our C57BL/6 model and establish a companion model of focally evoked limbic seizures in the BALB/c mouse that provides further evidence for activation of programmed cell death after seizures.     

Physiological and induced neuronal death are not affected in NSE-bax transgenic mice

Bax, a family member of the survival protein Bcl-2, is expressed in the nervous system during development and throughout adulthood. Bax deficiency has been demonstrated to prevent developmental and trophic factor deprivation-induced neuronal death. To further clarify the role of Bax in naturally occurring neuronal death and in neuronal death following apoptotic stimuli, we generated several lines of transgenic mice expressing the human Bax protein specifically in neurons, under the control of the neuron-specific enolase promoter. Transgene expression was first detected around E10.5 and E12.5, depending on the transgenic line. The total number of ganglion cells in the retina and of pyramidal cells in the hippocampus, both expressing the transgene, was similar in control and transgenic mice. In addition, in our model system, Bax overexpression did not appear to influence the in vitro survival of sensory neurons isolated from dorsal root ganglia after nerve grwoth factor (NFG) deprivation or the apoptotic death of motor neurons following axotomy. J. Neurosci. Res. 52:247–259, 1998. © 1998 Wiley-Liss, Inc.     

Ischemia-induced programmed cell death in astrocytes

Astrocytes are essential for neuronal survival and function, neurogenesis, and neural repair. Although astrocytes are more resistant than neurons to most stress conditions in vitro, certain astrocyte subtypes, such as the glial fibrillary acidic protein (GFAP)-negative protoplasmic astrocytes that predominate in gray matter structures, may be equally or more sensitive than neurons to ischemia in vivo. Programmed cell death differs from passive, necrotic death in that cell constituents actively participate in cell demise. Like neurons, astrocytes undergo programmed cell death during normal development. Cell culture studies have shown that astrocytes can be induced to undergo apoptosis and other forms of programmed cell death by many factors relevant to ischemia, including acidosis, oxidative stress, substrate deprivation, and cytokines. Animal models of cerebral ischemia have confirmed nuclear condensation and upregulation of Bax and caspases in a subset of astrocytes exposed to ischemia, especially in immature brain. A causal role for these events in astrocyte death is supported by improved astrocyte survival after inhibition of caspase-dependent cell death pathways. Astrocyte survival is also improved by blocking the poly(ADP-ribose)-1 cell death pathway. Markers of programmed cell death are generally less evident and less widespread in astrocytes than in neighboring neurons. However, most studies to date have relied only on markers of classical apoptosis. In addition, these studies have relied almost exclusively on GFAP to identify astrocytes. Since most protoplasmic astrocytes are poorly immunoreactive for GFAP, the extent of ischemia-induced programmed cell death in this cell type remains uncertain.     

SK2 channels are neuroprotective for ischemia-induced neuronal cell death

In mouse hippocampal CA1 pyramidal neurons, the activity of synaptic small-conductance Ca²⁺-activated K⁺ channels type 2 (SK2 channels) provides a negative feedback on N-methyl--aspartate receptors (NMDARs), reestablishing Mg²⁺ block that reduces Ca²⁺ influx. The well-established role of NMDARs in ischemia-induced excitotoxicity led us to test the neuroprotective effect of modulating SK2 channel activity following cerebral ischemia induced by cardiac arrest and cardiopulmonary resuscitation (CA/CPR). Administration of the SK channel positive modulator, 1-ethyl-benzimidazolinone (1-EBIO), significantly reduced CA1 neuron cell death and improved CA/CPR-induced cognitive outcome. Electrophysiological recordings showed that CA/CPR-induced ischemia caused delayed and sustained reduction of synaptic SK channel activity, and immunoelectron microscopy showed that this is associated with internalization of synaptic SK2 channels, which was prevented by 1-EBIO treatment. These results suggest that increasing SK2 channel activity, or preventing ischemia-induced loss of synaptic SK2 channels, are promising and novel approaches to neuroprotection following cerebral ischemia.     

Proliferation and programmed cell death of neuronal precursors in the mushroom bodies of the honeybee

We have studied proliferation and programmed cell death in the brain of the honeybee during metamorphosis. DNA fragmentation detection using the TUNEL method combined with 5-bromodeoxyuridine incorporation experiments reveal that in the mushroom bodies neurogenesis is terminated by extensive apoptosis. Proliferation of mushroom body neuroblasts is active until the fourth day of pupal development, ceasing abruptly within 1 day after the onset of apoptosis in the mushroom body proliferative clusters. Inside the mushroom bodies, apoptosis spreads from the apical ends of proliferative clusters, beneath the brain's surface, toward the basal ones. The distributions of apoptotic cells and those in the S phase of the cell cycle overlap significantly. Electron microscopic analysis gives further evidence that mushroom body neuroblasts themselves undergo programmed cell death. We suggest that programmed cell death may be the main factor controlling the final number of Kenyon cells produced during metamorphosis. The overlap in time and space between proliferation and apoptosis raises the question of whether the neuronal precursors switch to programmed cell death during the progression of the cell cycle, or afterwards.     

Spatiotemporal patterns of neuronal programmed cell death during postnatal development of the gerbil cochlea

During early postnatal development, afferent neurons of the cochlear (spiral) ganglion progressively refine their projections to auditory hair cells so that, by hearing onset, most cochlear nerve fibers innervate a single hearing receptor. One mechanism that might contribute to these changes in cochlear innervation is the programmed cell death (apoptosis) of developing neurons within the spiral ganglion. In the present study, we used the TUNEL method and morphological criteria to identify apoptotic cells within the spiral ganglion of the Mongolian gerbil during the first week of postnatal life when afferent projections to the cochlea are actively refined in this species. The locations of individual apoptotic spiral ganglion cells were mapped onto three-dimensional reconstructions of the entire ganglion for an age-graded series of gerbils to produce the first high-resolution, spatiotemporal maps of apoptotic ganglion cell death for the postnatal cochlea. We observed a significant increase in apoptosis in the spiral ganglion from postnatal day (P) 4 through P6. During this time, the most intense apoptotic activity occurred in regions of the spiral ganglion providing innervation to the lower middle and apical turns of the cochlea. The time course and regional variation of programmed cell death within the developing gerbil spiral ganglion are discussed in terms of the postnatal refinement of cochlear innervation and its possible functional significance for hearing in gerbils.     

Membrane asymmetry and DNA degradation: functionally distinct determinants of neuronal programmed cell death

The ability to elucidate the molecular mechanisms that modulate programmed cell death (PCD) may provide the crucial clues to unravel the cellular basis of neurodegenerative disorders. Employing both a novel assay to follow serially PCD in individual living neurons and the neuroprotective agent lubeluzole as an investigative tool, we examined the development of nitric oxide (NO)-induced PCD over time through the reversible annexin V labelling of membrane phosphatidylserine (PS) exposure and the electron microscopy of genomic DNA in primary rat hippocampal neurons. Exposure to the NO generators SNP (300 microM) or NOC-9 (300 microM) alone increased annexin V-positive neurons in the population from 7% +/- 4% in untreated cultures to 13% +/- 4% at 1 hr and to 61% +/- 5% at 24 hr. Administration of a neuroprotective concentration of lubeluzole (750 nM) at the time of NO exposure initially prevented the exposure of PS residues, but consistently maintained DNA integrity over a 24 hr period. During posttreatment paradigms of lubeluzole (750 nM) at 2, 4, and 6 hr following NO exposure, progression of membrane PS inversion was reversed and subsequently suppressed over a 24 hr course. Our work illustrates that neuronal PCD is composed of at least two physiologically distinct and separate pathways that consist of the externalization of membrane PS residues and the independent maintenance of genomic DNA integrity. In addition, neuronal injury is fluid and reversible in nature, suggesting a "window of opportunity" for the repair and reversal of neurons yet to be committed to PCD.     

Identification of neuronal programmed cell death in situ in the striatum of normal adult rat brain and its relationship to neuronal death during aging

Apoptotic neurons have been identified in normal adult rat striatum by terminal deoxynucleotidyl transferase mediated dUTP-biotin nick end labeling technique. This observation suggests that neuronal programmed cell death starts at an early stage of adult life and may contribute to the aging associated neuronal loss. In addition, the frequency of apoptotic cells was found to significantly increase in old rats, which implies that aging itself accelerates the process.     

Tumour necrosis factor-mediated cell death pathways do not contribute to muscle fibre death in dystrophinopathies

There is evidence that apoptotic cell death mechanisms contribute to muscle fibre loss in dystrophinopathies, but little knowledge about the activators of the final degrading caspase cascade in muscle fibre apoptosis. As mitochondria-related activation of this caspase cascade, through e.g. APAF-1, could not be proven in dystrophin-deficient muscle, this study searches for other prospective candidates that may directly trigger apoptotic cell degradation by mitochondria-independent pathways involving the interaction of tumour necrosis factor- (TNF-) and TRAIL with death receptors and subsequent activation of caspase-8. The expression of TNF-, TNF-R1, TRAIL, NF-B and caspase-8 were studied in muscle biopsy specimens from 14 patients with a dystrophinopathy [10 Duchenne muscular dystrophies (DMD), 2 Becker MD, and 2 DMD carriers] by immunohistochemistry and Western blotting. In all types of dystrophinopathies, necrotic muscle fibres undergoing myophagocytosis displayed strong expression of TNF-, TNF-R1, and TRAIL, which, however, was attributed to phagocytosing cells and not to the muscle fibres themselves. There was no up-regulation in normal-shaped or atrophic non-necrotic muscle fibres, or in intact muscle fibre segments adjacent to segmental necrosis and myophagocytosis. The expression profiles of caspase-8 and NF-B resembled that of normal control muscle. There were likewise no significant differences in the Western blot analyses between normal control and dystrophin-deficient muscle. Based on these findings, a contribution of TNF- or TRAIL-mediated cell death pathways to muscle fibre apoptosis or necrosis in dystrophinopathies could not be confirmed.