|Home | About | Journals | Submit | Contact Us | Français|
Although oxidative stress has been implicated in the pathogenesis of numerous neurodegenerative conditions, the precise mechanisms by which reactive oxygen species (ROS) induce neuronal death are still being explored. The generation of reactive lipid peroxidation products is thought to contribute to ROS neurotoxicity. Isoprostanes (IsoPs), prostaglandin-like molecules formed in vivo via the ROS-mediated oxidation of arachidonic acid, have been previously demonstrated to be formed in increased amounts in the brains of patients with various neurodegenerative diseases. Recently, we have identified a new class of IsoPs, known as A2- and J2-IsoPs or cyclopentenone IsoPs, which are highly reactive electrophiles and form adducts with thiol-containing molecules, including cysteine residues in proteins and glutathione. Cyclopentenone IsoPs are favored products of the IsoP pathway in the brain and are formed abundantly after oxidant injury. These compounds also potently induce neuronal apoptosis by a mechanism which involves glutathione depletion, ROS generation, and activation of several redox-sensitive pathways that overlap with those involved in other forms of oxidative neurodegeneration. Cyclopentenone IsoPs also enhance neurodegeneration caused by other insults at biologically relevant concentrations. These data are reviewed, whereas new data demonstrating the neurotoxicity of J-ring IsoPs and a discussion of the possible role of cyclopentenone IsoPs as contributors to neurodegeneration are presented.
Oxidative stress has recently emerged as a major contributor to numerous neurodegenerative diseases (Beal 1995; Andersen 2004). Substantial research efforts have been targeted at identifying the mechanisms by which the excessive generation of reactive oxygen species (ROS) causes neuronal death and dysfunction. One leading hypothesis is that lipid peroxidation, the free radical-mediated oxidation of membrane lipids, contributes to the pathological effects of oxidative stress in the brain. In support of this theory, increased levels of bioactive lipid peroxidation products have been identified in affected brain regions from humans with various neurodegenerative diseases, as well as in corresponding animal models (reviewed in Keller and Mattson 1998; Montine et al. 2002, 2004).
Lipid peroxidation can lead to the formation of highly reactive electrophilic compounds, such as 4-hydroxynonenal, acrolein, and levuglandin-like compounds in the brain (Montine et al. 1997; Sayre et al. 1997; Lovell et al. 2001; Davies et al. 2004; Zagol-Ikapitte et al. 2005). Furthermore, cellular studies have shown that some of these lipid peroxidation products can independently cause neuronal dysfunction and death. Indeed, many of the pathogenic events associated with oxidative stress can be recapitulated by application of lipid peroxidation products to neuronal cultures (Kruman et al. 1997; Mattson et al. 1997; Neely et al. 1999; Picklo et al. 1999; Camandola et al. 2000; Shringarpure et al. 2000). Furthermore, overexpression of glutathione transferase enzymes, which metabolize electrophilic lipids, protects neurons from oxidative insults (Xie et al. 1998, 2001). Thus, lipid peroxidation likely contributes to the observed effects of ROS in neurons (Keller and Mattson 1998; Kruman et al. 1997), although many of the exact oxidized lipid species involved are unknown.
A well characterized pathway of lipid peroxidation is the formation of isoprostanes (IsoPs), prostanglandin (PG)-like molecules generated nonenzymatically via the ROS-dependent oxidation of arachidonic acid (Morrow et al. 1990; Roberts and Morrow 2002). IsoPs containing various prostane ring structures are formed in vivo, including F2-IsoPs, which are isomeric to PGF2α (Morrow et al. 1990), and D2/E2-IsoPs, which are isomers of PGD2 and PGE2, respectively (see Fig. 1; Morrow et al. 1994). Because of their stability, the measurement F2-IsoPs by mass spectrometry has been extensively employed as an marker of oxidant stress, and is widely considered to be the gold-standard index of lipid peroxidation in vivo (Morrow and Roberts 1998; Musiek et al. 2005). Increased cerebral F2-IsoP formation has been associated with numerous neurodegenerative diseases, including Huntington’s Disease (Montine 1999a, b, c), Creutzfeldt–Jakob disease (Minghetti et al. 2000), traumatic brain injury (Bayir et al. 2002), multiple sclerosis (Greco et al. 1999), and Alzheimer’s disease (AD; Reich et al. 2001). In the case of AD, exposure of neurons or synaptosomes to amyloidogenic Aβ peptide leads to increased F2-IsoP formation (Mark et al. 1999; Brunetti et al. 2004). Furthermore, significant increases in F2-IsoPs, which precede amyloid plaque formation, were observed in a mouse model of AD (Pratico et al. 2001). Several human studies have detected significantly elevated levels of F2-IsoPs in postmortem brain tissue and cerebrospinal fluid obtained from AD patients (Pratico et al. 1998; Reich et al. 2001), as well as in CSF from living patients with probable AD (Montine 1999a, b, c; Pratico et al. 2000). Increased F2-IsoPs formation occurred specifically in affected brain regions and correlated with the degree of neuronal degeneration (Montine 1999a, b, c). These studies suggest that oxidative stress and IsoP formation are conserved characteristics of the neurodegenerative process.
Whereas F2-IsoPs are the most thoroughly studied IsoP subtype, IsoPs with alternative ring structures are also formed by the same pathway. Indeed, E2/D2-IsoPs, which resemble PGE2 and D2, respectively, are formed in competition with F2-IsoPs from the same endoperoxide precursor (Fig. 1a,b; Morrow et al. 1994). Interestingly, the loss of reducing environment in the brain, manifested by depletion of glutathione (GSH) and vitamin E, shifts the IsoP pathway toward the formation of E/D-ring IsoPs and away from reduced F-ring IsoPs (Montine et al. 2003). In peroxidizing brain synaptosomes, E2/D2-IsoP are the favored products of the IsoP pathway, and their levels far exceed those of F-ring IsoPs (Montine et al. 2003). Moreover, E2/D2-IsoP levels are significantly elevated in the brains of human AD patients, and the ratio of E/D-ring to F-ring IsoPs is increased in this disease (Reich et al. 2001). Thus, these findings demonstrate that alternative IsoP subtypes are formed along with F2-IsoPs under neurodegenerative conditions.
E2/D2-IsoPs are unstable compounds and spontaneously dehydrate in vivo to form A2/J2-IsoPs, also known as cyclopentenone IsoPs (Fig. 1; Chen et al. 1999a). Cyclopentenone IsoPs are so named because they contain a highly reactive α,β-unsaturated cyclopentenone ring structure and are thus potent electrophiles (Milne et al. 2005). Cyclopentenone IsoPs rapidly form Michael adducts with cellular thiol groups, such as those found on cysteine residues in proteins or in GSH (Chen et al. 1999a, b). Because other electrophilic lipid peroxidation products can damage neurons, we hypothesized that cyclopentenone IsoPs might be neurotoxic products of the IsoP pathway and might contribute to the pathogenesis of oxidative neurodegeneration. In support of this hypothesis, we have recently demonstrated that cyclopentenone IsoPs are considerably more abundant than F2-IsoPs in both human and rat brain tissue (Musiek et al. 2006). Oxidative injury caused by exposure to the free radical initiator 2,2′-azobis(2-amidinopropane)hydrochloride (AAPH) caused a 12-fold increase in cyclopentenone IsoPs in brain homogenates, generating cyclopentenone IsoP levels as high as ~550 nM. Thus, cyclopentenone IsoPs are formed in biologically relevant concentration in the brain under conditions of oxidative stress and appear to be favored products of the IsoP pathway in the brain.
We have performed experiments examining the effects of synthetic 15-A2t-IsoP (Fig. 1b; Zanoni et al. 2002), a major cyclopentenone IsoP isomer formed in vivo (Chen 1999b), in primary cortical neuronal cultures. 15-A2t-IsoP is highly neurotoxic, inducing neuronal death with an LD50 of 950 nM (Musiek et al. 2006). 15-A2t-IsoP-induced cell death is apoptotic, as cells exposed to this IsoP exhibited condensed, fragmented nuclei, as well as increased caspase-3 cleavage, and are completely protected by the pan-caspase inhibitors. 15-J2t-IsoPs, a cyclopentenone IsoPs which is formed similarly to 15-A2t-IsoPs, was recently synthesized (Fig. 1b; Zanoni et al. 2003). This J-ring IsoP induces cell death in HT22 hippocampal cells with slightly greater potency than 15-A2t-IsoP (Fig. 1c), suggesting that many if not all cyclopentenone IsoPs are neurotoxic.
Because of their reactivity, cyclopentenone IsoPs rapidly form adducts with GSH (Chen et al. 1999a, b) and are metabolized in cells via conjugation to GSH by glutathione transferases (GSTs, see Fig. 2; Milne et al. 2004). Exposure of neurons to 15-A2t-IsoP overwhelms the GST system, causing rapid GSH depletion. 15-A2t-IsoP also independently induces mitochondrial free radical production, resulting in membrane lipid peroxidation (Musiek et al. 2006). These two processes set in motion a feed-forward cycle of oxidation in neurons, as cyclopentenone IsoPs are initially generated as products of oxidative stress, then in turn promote further ROS production, thereby generating more of themselves (Fig. 2). Accordingly, 15-A2t-IsoP toxicity can be mitigated by free radical scavenging agents, suggesting that redox alterations caused by 15-A2t-IsoP contribute to its toxicity.
Cyclopentenone IsoPs induce apoptosis in neurons through activation of specific signaling pathways, all of which have been implicated in other forms of oxidative neurodegeneration (Fig. 2). The stress-responsive kinase ERK1/2 is phosphorylated in a redox-dependent manner in response to cerebral ischemia (Alessandrini et al. 1999; Namura et al. 2001; Noshita et al. 2002), and inhibition of ERK attenuates ischemic neurodegeneration, as well as neuronal death in response to other oxidative stimuli (Stanciu et al. 2000; Du et al. 2002; Lee et al. 2003; Chu et al. 2004). Further, dysregulated ERK phosphorylation is observed in neurons of patients with several neurodegenerative diseases, including AD (Perry et al. 1999) and Parkinson’s disease (Zhu et al. 2002). Similarly, ERK is transiently phosphorylated in neurons exposed to 15-A2t-IsoP downstream of mitochondrial ROS production, whereas inhibition of ERK phosphorylation partially protects from IsoP neurotoxicity (Musiek et al. 2006).
Human and rodent 12/15-lipoxygenase (12/15-LOX), an enzyme which is expressed in neurons and can oxidize arachidonic acid (Nishiyama et al. 1993), has also been shown to participate in several forms of neurodegeneration associated with oxidative stress (Li et al. 1997; Canals et al. 2003; Zhang et al. 2004). 12/15-LOX has been linked to AD, as inhibition of 12/15-LOX protects neurons from β-amyloid toxicity, and increased 12/15-LOX expression has been described in neurons of patients with AD (Lebeau et al. 2004; Pratico et al. 2004). We observed that 15-A2t-IsoP causes 12/15-LOX activation and that the 12/15-LOX inhibitor baicalein protects neurons from IsoP toxicity (Musiek et al. 2006), providing a potential mechanism linking oxidative stress, 12/15-LOX activation, and cell death.
Finally, 15-A2t-IsoP also induces the phosphorylation of the redox-sensitive adaptor protein p66shc in neurons (Musiek et al. 2006). P66shc phosphorylation contributes to oxidative stress-induced cell death, as mice lacking this protein are resistant to oxidant injury and have increased lifespan (Migliaccio et al. 1999; Trinei et al. 2002). Furthermore, p66shc phosphorylation has recently been shown to play a role in β-amyloid neurotoxicity (Smith et al. 2005), revealing another important neurodegenerative pathway activated by cyclopentenone IsoPs. Thus, cyclopentenone IsoPs induce neuronal death through activation of several signaling proteins that have been previously but disparately implicated in oxidative neurodegeneration.
As products of oxidative stress, cyclopentenone IsoPs are generated in response to an initial neurotoxic insult. Thus, determining the effect of these IsoPs in combination with other oxidative insults is of utmost importance in understanding their potential impact in vivo. One such oxidative insult is the application of extracellular glutamate to HT22 hippocampal cells or immature primary neurons. Because neither of these cell types express mature N-methyl-D-aspartate receptors, millimolar concentrations of glutamate inhibits cellular uptake of cystine in this system, causing rapid GSH depletion and oxidative cell death (Murphy et al. 1989; Tan et al. 2001). Interestingly, previous studies have shown that loss of GSH in these cells induces death via induction of mitochondrial ROS production (Tan et al. 1998) and activation of 12/15-LOX (Li et al. 1997) and ERK (Stanciu et al. 2000), just as in 15-A2t-IsoP-treated neurons. Accordingly, we found that 15-A2t-IsoP strongly potentiates neurodegeneration caused by sublethal concentrations of glutamate in both HT22 cells and primary neurons at concentrations as low as 100 nM (Musiek et al. 2006). Taken together, the data demonstrate that cyclopentenone IsoPs are neurotoxic products of lipid peroxidation that activate several discreet pathogenic signaling pathways associated with oxidative neurodegeneration, and which enhance neuronal death caused by other oxidative insults.
Although the initial studies are intriguing, much remains unknown about the potential role of cyclopentenone IsoPs in neurodegeneration. It is likely that these compounds modulate numerous other signaling cascades in neurons that have not yet been identified. Furthermore, the exact protein targets and cysteine adducts that are generated have not yet been elucidated. Proteomic approaches are currently being employed to identify these adducts in neurons and other cell types. The mechanism by which cyclopentenone IsoPs promote mitochondrial ROS production is also an area of continued study. Finally, the quantification of cyclopentenone IsoPs in human postmortem brain samples from patients with various neurodegenerative diseases will provide important information about the involvement of these molecules in neurologic disease in vivo. Novel mass spectrometric approaches have been developed in our lab to reliably and accurately quantify 15-A2- and 15-J2-IsoPs in brain tissue (Musiek et al. 2006) and are currently being applied to tissue samples from human patients as well as from animal models of neurodegenerative diseases. The findings that cyclopentenone IsoPs are formed via the same pathway and more abundantly than F2-IsoPs in brain and that F2-IsoP levels are increased in the brains of patients with various neurodegenerative conditions (Montine et al. 2004) suggest that cyclopentenone IsoPs are likely also formed abundantly in these diseases. Thus, therapeutic interventions that prevent cyclopentenone IsoP formation, adduction, or downstream signaling might hold promise for the prevention and treatment of neurodegenerative diseases.
This work was supported by NIH grants DK48831, GM15431, CA77839, RR00095, ES13125, HD15052, and NS050396. ESM was supported by a grant from the PhRMA Foundation.
Erik S. Musiek, Division of Clinical Pharmacology, Departments of Medicine and Pharmacology, Vanderbilt University School of Medicine, 526 RRB, 23rd and Pierce Aves, Nashville, TN 37232-6602, USA.
BethAnn McLaughlin, Departments of Neurology and Pharmacology, Vanderbilt University School of Medicine, MRB III Room 8110, 465 21st Avenue South, Nashville, TN 37232-8548, USA.
Jason D. Morrow, Division of Clinical Pharmacology, Departments of Medicine and Pharmacology, Vanderbilt University School of Medicine, 526 RRB, 23rd and Pierce Aves, Nashville, TN 37232-6602, USA.