Correspondence: Stephan N. Witt, Department of Biochemistry and Molecular Biology, Louisiana State University Health Sciences Center, 1501 Kings Highway, Shreveport, LA 71130-3932, USA. Tel.: +1 318 675 7891; fax +1 318 675 5180; e-mail: firstname.lastname@example.org
The neuronal protein α-synuclein (α-syn) has been suggested to be one of the factors linked to Parkinson's disease (PD). Several organisms, including the rat, mouse, worm, and fruit fly, are being used to study α-syn pathobiology. A new model organism was recently added to this armamentarium: the budding yeast Saccharomyces cerevisiae. The yeast system recapitulates many of the findings made with higher eukaryotes. For example, yeast cells expressing α-syn accumulate lipid droplets, have vacuolar/lysosomal defects, and exhibit markers of apoptosis, including the externalization of phosphatidylserine, the release of cytochrome c, and the accumulation of reactive oxygen species. This MiniReview focuses on the mechanisms by which α-syn induces oxidative stress and the mechanisms by which yeast cells respond to this stress. Three classes of therapeutics are discussed.
Parkinson's disease (PD) is characterized by the selective loss of dopamine-producing neurons that comprise the substantia nigra pars compacta and the presence of proteinaceous inclusion bodies termed Lewy bodies and Lewy neurites in the affected neurons (Lucking & Brice, 2000; Dawson & Dawson, 2003). The discovery that the protein α-synuclein (α-syn) is the main component of Lewy bodies (Baba et al., 1998) has fueled the hypothesis that α-syn is causally linked to PD. This hypothesis was subsequently reinforced by discoveries that mutations in α-syn (A30P or A53T) produce rare, autosomal dominant forms of PD (Polymeropoulos et al., 1997; Kruger et al., 1998). These various findings have sparked an intense search for the mechanism of α-syn toxicity. We now know that α-syn is a 140-amino acid intrinsically unfolded protein that is thought to regulate cell differentiation, synaptic plasticity, and dopaminergic neurotransmission. α-Syn binds to lipids (Narayanan & Scarlata, 2001), inhibits phospholipase D (Jenco et al., 1998), forms a myriad of differently sized protofibrils and fibers (Caughey & Lansbury, 2003), induces the accumulation of reactive oxygen species (ROS) (Xu et al., 2002; Flower et al., 2005), and causes proteasome dysfunction (Lindersson et al., 2004; Betarbet et al., 2005). We believe that the lack of knowledge concerning how α-syn induces oxidative stress in cells, which can lead to dysregulated apoptosis, constitutes a major gap in our understanding of this disease. This MiniReview focuses on using yeast to understand the mechanism of α-syn-induced ROS accumulation and apoptosis and possible therapeutic approaches to the inhibition of these two processes.
Intense efforts are underway to uncover the molecular details that underlie the slow degeneration of dopamine-producing neurons in patients with PD. A possible breakthrough has come from recent studies showing that wild-type α-syn (WT) induces apoptosis (Saha et al., 2000; Stefanova et al., 2001; Tanaka et al., 2001), and that the toxicity of α-syn is related to dopamine production (Xu et al., 2002). In one of these studies, Xu et al. (2002) discovered that overexpressing WT α-syn in human fetal dopaminergic neurons causes ROS accumulation and apoptosis, whereas expressing α-syn in nondopaminergic human cortical neurons actually protects the cells and significantly increases neuronal survival. It was also discovered by this group that WT α-syn exists in stable complexes with the antiapoptotic 14-3-3 protein in dopaminergic cells. Titrating away the antiapoptotic 14-3-3 protein via complexation with α-syn could explain the increased susceptibility to apoptosis. The Xu study showed that α-syn triggers ROS and apoptosis in dopamine-producing neurons, and that vitamin E, an antioxidant, effectively inhibits ROS accumulation.
Several groups are using S. cerevisiae to study the toxicity of WT α-syn and the two mutants associated with early onset PD (A30P and A53T). Although S. cerevisiae does not have an α-syn ortholog, a recent study revealed that one integrated copy of α-syn had no appreciable effect on cell growth, whereas two integrated copies severely inhibited cell growth (Outeiro & Lindquist, 2003). Thus, α-syn toxicity increases with increasing concentration. Additionally, it was shown that the various α-syns induce accumulation of lipid droplets and vacuolar dysfunction. A companion study that assayed 4850 yeast deletion strains for synthetic lethality with WT α-syn revealed that genes of lipid metabolism and vesicle trafficking are the most prevalent enhancers of α-syn toxicity (Willingham et al., 2003; Scherzer & Feany, 2004). Other studies showed that WT and A53T, but not A30P, are delivered to the plasma membrane via the classic secretory pathway (Dixon et al., 2005), and that α-syn-expressing cells exhibit an apoptotic-phenotype (Flower et al., 2005), as discussed below.
α-Synuclein, apoptosis and yeast
Saccharomyces cerevisiae undergoes apoptosis in response to a variety of stimuli, including acetic acid, hydrogen peroxide, sugar and salts, pheromone, and Bax (a human protein that promotes apoptosis) (Madeo et al., 2004). When treated with these various compounds, or upon the expression of Bax, yeast exhibits the classic markers of apoptosis such as nuclear fragmentation, chromatin condensation, DNA cleavage, phosphatidylserine externalization, and the accumulation of ROS. Although many of the initial findings in this field were treated with skepticism, the discovery that S. cerevisiae harbors several orthologs of genes that have been implicated in apoptosis in multicellular organisms suggests that the apoptosis machinery is ancient. For example, yeast contain two antiapoptotic 14-3-3 genes (BMH1 and BMH2) (van Heusden & Steensma, 2006), a pro-apoptotic metacaspase (YCA1) (Madeo et al., 2002), an apoptosis-inducing factor (AIF1) (Wissing et al., 2004), and two genes involved in mitochondrial fission–fusion, which is a process linked to cell death (FIS1, DNM1) (Fannjiang et al., 2004). One possibility is that yeast, particularly aging yeasts, undergo apoptosis when nutrients are scarce, and this makes nutrients available to the younger, fitter members of the population (Herker et al., 2004).
Strikingly, α-syn induces S. cerevisiae to undergo apoptosis, e.g. cells expressing α-syn externalize phosphatidlyserine, accumulate ROS, release cytochrome c from the mitochondria into the cytosol, have fragmented nuclei (T. Flower and S.N. Witt, unpublished results), and become supersensitive to killing by hydrogen peroxide (H2O2) (Flower et al., 2005). The same study showed that a heat shock, geldanamycin (a chemical activator of the heat shock response), overexpression of the Hsp70 chaperone Ssa3p, deletion of the yeast caspase gene YCA1, and glutathione protect α-syn-expressing yeast cells from α-syn-induced ROS accumulation. Notably, several of these findings with yeast are in parallel with those from Xu and colleagues (Xu et al., 2002): in yeast and human cells, α-syn induces ROS, apoptotic markers are evident, and chemical reductants (vitamin E or glutathione) abolish α-syn-induced ROS. Protection by geldanamycin in the yeast system also agrees with recent studies using transgenic Drosophila showing that geldanamycin protects fly neurons from a degenerative death that is otherwise caused by human α-syn (Auluck & Bonini, 2002). That α-syn triggers ROS accumulation in yeast and mammalian cells is exciting, and perhaps this indicates a common mechanism.
We are interested in the mechanisms by which α-syn triggers ROS accumulation in cells. In this section, we first discuss mitochondrial complex I, and how its inhibition by pesticides and 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) causes ROS. Second, we discuss how α-syn itself can damage mitochondria and cause ROS.
Respiration and oxidative stress
Oxidative phosphorylation is the process whereby living cells utilize free energy from the oxidation of NADH and FADH2 to synthesize ATP. Mammalian mitochondria have four protein complexes (Complexes I–IV) embedded in the inner mitochondrial membrane that shuttle electrons from NADH and FADH2, in a step-wise fashion, to molecular oxygen. Complex I (NADH:Coenzyme Q oxidoreductase) catalyzes the oxidation of NADH and electron transfer to coenzyme Q, which is a small-molecule electron carrier located within the inner mitochondrial membrane (reaction 1).
Complex II catalyzes the oxidation of FADH2 and electron transfer to coenzyme Q, which thereby increases the pool of reduced coenzyme Q. Complex III catalyzes the oxidation of reduced coenzyme Q and electron transfer to cytochrome c, the highly conserved electron carrier protein. Reduced cytochrome c binds to and injects electrons into complex IV, which passes the electron to dioxygen, generating two molecules of water (4e−+ 4H++ O2→ 2H2O). In concert with these various redox reactions, complexes I, II and IV pump protons from the matrix into the interstitial space between the inner and outer mitochondrial membranes. This results in the formation of an electrochemical hydrogen ion gradient that is used by the F1F0-ATPase to synthesize ATP. (Note that S. cerevisiae do not have a homolog of mammalian complex I, instead yeast expresses three NADH dehydrogenases that oxidize NADH, and that do not pump protons; Schagger, 2002.) In a perfectly efficient process, each electron derived from NADH and FADH2 would transfer to dioxygen, and each molecule of dioxygen would produce two molecules of water. Unfortunately, respiration is inefficient: highly reactive oxygen species such as superoxide (), hydrogen peroxide (H2O2), or the hydroxyl radical (HO·) escape enzyme active sites and diffuse into solution, where they oxidize lipids, proteins, and DNA. To prevent irreversible oxidative damage to cellular components, organisms have evolved a quality control system consisting of enzymes (Wheeler & Grant, 2004), such as catalases, glutaredoxins, peroxidases, superoxide dismutases and thioredoxins, as well as small-molecule reductants, such as ascorbic acid, glutathione, ubiquinol, and tocoferol (vitamin E) that react with toxic reactive oxygen species and render them nontoxic. This multicomponent quality control system permits cells to thrive in oxygen atmospheres.
Complex I inhibition by pesticides and MPTP and oxidative stress
Numerous compounds inhibit the respiratory enzymes, and inhibition impedes the flow of electrons to molecular oxygen, which in turn alters membrane potential, decreases ATP output, and increases ROS output. Rotenone, paraquat, and MPTP are complex I inhibitors that are of particular interest to the field of PD (Dawson & Dawson, 2003). Rotenone and paraquat are pesticides, whereas MPTP is an impurity in synthetic heroin (Langston et al., 1983). By inhibiting complex I, each of these three compounds impedes the flow of electrons to oxidized coenzyme Q; instead, electrons flow to molecular oxygen and generate powerful oxidants such as superoxide and hydrogen peroxide (Li et al., 2003), which causes chronic oxidative stress. Infusing rotenone into a rat results in a movement disorder strikingly similar to PD (Sherer et al., 2003), and studies have shown that cells treated with rotenone accumulate ROS (Li et al., 2003). Ingestion of synthetic heroin by addicts results in a movement disorder strikingly similar to PD (Langston et al., 1983; Vila et al., 2000). Years of detective work have established a causal link between decrements in complex I activity, chronic oxidative stress (ROS) and PD, but without exposure to a complex I poison, how is mitochondrial dysfunction triggered? Because α-syn triggers ROS accumulation in yeast cells, and because of its relative ease of genetic manipulation, yeast is an ideal organism to test several hypotheses regarding how α-syn triggers mitochondrial dysfunction.
How does α-syn cause oxidative stress?
α-Synuclein probably causes mitochondrial dysfunction by several pathways, and perhaps two or more of these pathways occur in parallel. To complicate matters, the actual α-syn culprit that affects mitochondria, whether directly or indirectly, may be nitrosated (Giasson et al., 2000), phosphorylated (Fujiwara et al., 2002), or a cleaved form of α-syn (Liu et al., 2005), rather than full-length unmodified α-syn. We propose that α-syn causes mitochondrial dysfunction, and attendant oxidative stress, in three ways:
(i) Inhibition of the respiratory-chain enzymes. In the absence of exposure to complex I poisons, does α-syn itself inhibit complex I? α-Syn would have to enter the mitochondrial matrix or the interstitial space between the inner and outer membranes in order for the inhibition of complex I to occur. Since α-syn has no obvious mitochondrial localization sequence, it is hard to imagine how α-syn could enter the mitochondrial matrix. On the other hand, a recent quantitative mass spectroscopy proteomics study using a cultured dopaminergic cell line treated with rotenone found α-syn in association with two mitochondrial precursor and channel proteins (TOM34 and TIM 13 A) (Zhou et al., 2004); thus, perhaps a route exists by which α-syn enters mitochondria. This same proteomics study, as well as an earlier study (Elkon et al., 2002), also found that α-syn associates with a subunit of the mitochondrial complex IV enzyme (cytochrome c oxidase). Inhibition of the complex IV enzyme, whether by α-syn or any other molecule, would have a profound deleterious effect on mitochondrial function. If α-syn can enter mitochondria – either through a channel or by poking holes in the outer membrane – then perhaps it can bind to complex I or IV and this association leads to the generation of ROS.
(ii) Hole punching. Hole punching into the outer mitochondrial membrane is another route by which α-syn could produce mitochondrial dysfunction. This would lead to the leakage of ROS and cytochrome c into the cytosol and ultimately the collapse of the mitochondrial membrane potential. This hypothesis was proposed based on observations that α-syn binds to isolated rat mitochondria in vitro (Ding et al., 2002). It was suggested that protofibrillar forms of α-syn target and punch holes in membranes like bacterial toxins (Ding et al., 2002). Given that WT and A53T α-syn bind to the plasma membrane of yeast cells (Outeiro & Lindquist, 2003), the binding of α-syn to mitochondria membranes in vivo certainly seems possible. However, to date, there is no evidence that α-syn itself disrupts the outer mitochondrial membrane in vivo to cause cytochrome c release, and since acetic acid treatment even induces cytochrome c release from yeast mitochondria (Ludovico et al., 2002), a route to mitochondrial dysfunction and the release of cytochrome c exists independent of α-syn expression.
(iii) A caspase substrate, notα-syn, causes mitochondrial dysfunction. Another possibility is that a processed caspase substrate, rather than α-syn, triggers mitochondrial dysfunction. This idea can be developed by recalling studies that used hydrogen peroxide to induce cell death in yeast. Hydrogen peroxide induces yeast to exhibit classic markers of apoptosis such as chromatin condensation, DNA fragmentation, PS flipping, and ROS (Ludovico et al., 2002; Madeo et al., 2002). Two features of this triggered cell death are intriguing. First, that hydrogen peroxide fails to induce an apoptotic phenotype in a caspase delete strain (yca1Δ) indicates that hydrogen peroxide-induced cell death is caspase-dependent, thus it is not simply oxidative damage that induces cell death. Second, that hydrogen peroxide fails to induce an apoptotic phenotype in cells pre-treated with cycloheximide, a protein synthesis inhibitor, indicates that hydrogen peroxide-induced apoptosis requires active protein synthesis. Consider a scheme where hydrogen peroxide induces transcription of certain genes (one of which codes for the caspase substrate), and the mRNAs are rapidly translated by the protein synthesis machinery. Yca1p then processes a newly synthesized substrate, and the processed substrate triggers the various metabolic changes associated with apoptosis. In this model, cycloheximide would block the expression of the apoptosis-inducing caspase substrate, thus inhibiting apoptosis. This is exactly what was observed experimentally.
In yeast α-syn also induces an apoptotic phenotype defined by caspase-dependence and super-sensitivity to hydrogen peroxide (Flower et al., 2005). Perhaps α-syn induces the formation of hydrogen peroxide in situ, which would explain why α-syn-induced cell death is caspase-dependent. In this sense, α-syn-induced apoptosis may be a variant of hydrogen peroxide-induced cell death. It is intriguing that α-syn or an α-syn fragment, referred to as NAC (residues 61–95), when incubated in phosphate-buffered saline containing Fe (III), generates detectable amounts of hydrogen peroxide, whereas identical concentrations of β- and γ-synucleins fail to generate comparable amounts of hydrogen peroxide (Turnbull et al., 2001). If α-syn induces hydrogen peroxide formation in the cytosol, this constitutes a fourth (iv) way in which α-syn induces oxidative stress in cells.
Dopamine and α-syn
Because α-syn is causally linked to the selective degeneration of dopamine-producing neurons, dopamine itself has come under intense scrutiny as a co-conspirator of α-syn. Dopamine can affect α-syn in two ways. First, because dopamine is chemically unstable and oxidizes to dopamine quinone, which results in the generation of superoxide and hydrogen peroxide (Maguire-Zeiss et al., 2005), dopamine metabolism causes oxidative stress. One possibility is that these two ROS oxidize α-syn, which makes α-syn more susceptible to fiber formation (Souza et al., 2000). α-Syn fibers and aggregates could lead to decrements in proteasome function (Lindersson et al., 2004). Second, catecholamines, such as dopamine quinone, react with proteins, including α-syn, to form covalent adducts (Rochet et al., 2004). This covalent modification of α-syn impedes the protofibril-to-fiber transition, thus the potentially more cytotoxic protofibrils accumulate in cells that synthesize dopamine (Conway et al., 2001). We conclude that dopamine, directly or indirectly, can alter α-syn structure, and perhaps this altered structure is more cytotoxic. Dopamine-modified α-syn may be more active than unmodified α-syn in any of the four activities (i–iv) described above.
Complicating matters, α-syn has also been linked to the degeneration of cells that do not produce dopamine. Multiple system atrophy (Wenning & Jellinger, 2005) and dementia with Lewy bodies (Rampello et al., 2004) are two other devastating α-synucleinopathies. In these two diseases, Lewy bodies enriched with α-syn develop in a variety of brain cells, some of which do not produce dopamine or other catecholamines. Affected cells slowly die, and there are no effective treatments. Reinforcing the view that α-syn is cytotoxic to nondopamine-producing cells, α-syn induces apoptosis in cultured human peripheral lymphocytes (Kim et al., 2004) and in S. cerevisiae (Flower et al., 2005), neither of which synthesize dopamine. Faced with this knowledge that α-syn is toxic to dopaminergic neurons as well as to cells that fail to synthesize dopamine, one conclusion, which we favor, is that α-syn kills cells via multiple mechanisms, and some of these mechanisms were outlined above. An alternative conclusion is that α-syn kills different eukaryotic cells by the same mechanism; the only difference is that in dopamine-producing cells dopamine enhances α-syn toxicity, whereas in nondopamine-producing cells nitrosation, phosphorylation, or proteolytic processing enhances α-syn toxicity. In this view, whether nitrosated, phosphorylated, cleaved, or a dopamine adduct, α-syn kills cells by the same mechanism. That α-syn kills a variety of eukaryotic cells reinforces the importance of using yeast, which are so amenable to genetic analysis, to study PD.
In the following sections we discuss mechanisms of protection against α-syn-induced ROS and apoptosis in the yeast model. Three therapeutic approaches for the treatment of PD are discussed.
Glutathione (GSH) is the intracellular redox buffer in most cells (Wheeler & Grant, 2004). Reduced glutathione exists in cells in the concentration range 1–10 mM, and it equilibrates with oxidized glutathione (GSSG) to give a GSH : GSSG ratio of approximately 50 : 1. The GSH : GSSG redox pair is found not only in the cytosol of cells but also in the mitochondrial matrix. Given its abundance, glutathione is the principal small-molecule antioxidant that eliminates ROS from cells. Moreover, since ROS is a primary trigger of apoptosis in yeast and human cells, by its ability to eliminate ROS, glutathione is an antiapoptotic molecule.
The antiapoptotic nature of glutathione and the enzymes involved in glutathione metabolism is illustrated by the following examples. First, consider the effect of knocking out glutathione synthetase (gsh1Δ) on yeasts viability: yeast that fail to synthesize glutathione exhibit the classic markers of apoptosis, including chromatin condensation, DNA fragmentation, and phosphatidylserine flipping from the inner leaflet to the outer leaflet of the plasma membrane (Madeo et al., 1999). This result best illustrates the antiapoptogenic nature of glutathione. Second, a screen for inhibitors of Bax-induced cell death in yeast uncovered a novel glutathione S-transferase from tomato that suppresses Bax lethality (Kampranis et al., 2000). The expression of Bax in yeast depletes glutathione, and co-expression of this novel tomato GST transferase increases glutathione back to normal concentrations. Third, a recent screen of 4850 yeast deletion strains discovered that WT α-syn is synthetically lethal in mutants that lack genes involved in glutathione metabolism. One mutant (gtt1Δ) lacks glutathione transferase, whereas the other mutant (glo4Δ) lacks mitochondrial glyoxylase II, which is an enzyme that uses glutathione to detoxify cells of methylglyoxal. GLO4 has a human homolog but GTT1 does not. Fourth, the various α-syns (WT, A30P, or A53T) induce ROS in yeast cells, and the percentage of cells exhibiting ROS in these α-syn-expressing cells is similar to that observed for cells treated with the apoptosis-inducing agent acetic acid (Flower et al., 2005). Notably, glutathione, when added exogenously, abolishes ROS formation in cells expressing α-syn. These examples show the complex interplay between glutathione metabolism, oxidative stress, and cell death.
Although glutathione protects against transient bursts or low-level sustained ROS, chronic oxidative stress depletes glutathione concentrations (Sherer et al., 2002). The depletion of glutathione impairs the ability of cells to catabolize a toxic byproduct of glycolysis, namely methylglyoxal. Methylglyoxal is a highly reactive α-oxoaldehyde that rapidly reacts with proteins to yield advanced glycation end products (AGEs). Glycation of a protein may lead to a loss of function or a toxic gain of function. Interestingly, AGEs are found in Lewy bodies from patients with PD (Munch et al., 2000), which indicates that methylglyoxal concentrations may increase in the neurons of patients with PD.
To understand how glutathione depletion increases the concentration of methylglyoxal, one must know how cells catabolize methylglyoxal. Two enzymes rid cells of methylglyoxal, glyoxalase I (GLO2) and glyoxalase II (GLO2/GLO4) (Bito et al., 1997). Glo2p is a cytosolic glyoxalase II, whereas Glo4p is a mitochondrial glyoxalase II. Recall that the study by Willingham found that WT α-syn is synthetically lethal in cells in which GLO4 is deleted. Methylglyoxal (MG) spontaneously condenses with a molecule of glutathione (GSH) to form a hemithioacetal (HTA) (reaction 2). Glyoxalase I catalyzes the conversion of the hemithioacetal to S-d-lactoylglutathione (SLG), and glyoxalase II catalyzes the conversion of this intermediate to d-lactate and GSH. By rapidly driving reaction (2) to completion, the glyoxalase enzymes keep MG concentrations at negligible levels, and this minimizes harmful competing side reactions.
However, if GSH concentration decreases, the MG to d-lactate conversion is inhibited and the MG concentration increases, resulting in unwanted side reactions. For example, MG attacks arginine residues of proteins, resulting in irreversible modification. Thus, in eukaryotic cells, whenever cellular conditions allow α-syn to induce chronic oxidative stress a downward spiral commences, which is characterized by decreasing glutathione, increasing methylglyoxal, and increasing AGEs, which tends to amplify ROS (Fig. 1). To make matters worse, chronic oxidative stress itself, as well as the damage done to mitochondria by chronic oxidative stress, biases cells towards apoptosis. With the availability of large libraries of bioactive compounds, yeast can be used to discover novel compounds that suppress α-syn-induced ROS. Given that α-syn induces ROS in yeast and human cells (Xu et al., 2002; Flower et al., 2005), it is likely that a common mechanism exists, which means that compounds discovered using yeast may also work to suppress α-syn-induced ROS in mammalian cells. If chronic ROS could be suppressed, then the characteristic neuronal loss in Parkinson's disease might be delayed or prevented.
Programmed cell death is, by necessity, a highly regulated process. In mammalian cells, caspases are one set of programmed cell death regulators. Once an external or internal cue triggers a cell's apoptotic program, caspases start to selectively degrade key cellular proteins, and this leads to cell death. Until recently, there was no indication that unicellular organisms such as S. cerevisiae expressed caspases. However, it turns out that S. cerevisiae indeed contains a metacaspase, YOR197w (YCA1), which has structural homology to mammalian caspases. The importance of YCA1 as a regulator of hydrogen peroxide-induced apoptosis in yeast was revealed by experiments in which YCA1 was deleted: Wild-type cells treated with hydrogen peroxide exhibit the classic markers of apoptosis, as described above, whereas yca1Δ cells treated with the same concentrations of hydrogen peroxide fail to undergo apoptosis (Madeo et al., 2002). YCA1 also appears to regulate the toxicity of the human protein α-syn: wild-type yeast cells expressing the various α-syns (WT, A30P, or A53T) accumulate ROS, whereas yca1Δ cells expressing any of these α-syns fail to accumulate ROS (Flower et al., 2005). Together, these results suggest that caspase inhibitors would block the ability of α-syn to trigger apoptosis in yeast (Fig. 1). The significance of the yeast system is that it can be used to screen large libraries of compounds to find compounds that inhibit cell death (Griffioen et al., 2006), and perhaps some of these compounds will be found to be caspase inhibitors.
It should be pointed out that the therapeutic potential of caspase inhibition for PD was recently demonstrated in a study employing cultured rat neurons. The A53T mutant, when expressed in PC12 cells, induces endoplasmic reticulum stress, triggers ROS, and activates the mitochondrial cell death pathway. Notably, the mitochondrial cell death pathway is partially blocked by a pan caspase inhibitor (z-VAD), or inhibitors of caspase-9 and -3 (Smith et al., 2005). That α-syn triggers oxidative stress and caspase-dependent cell death (apoptosis) in organisms as different as yeast, rat and human suggests that a common mechanism underlies α-syn toxicity. Caspase inhibitors, with their potential to halt unwanted neuronal apoptosis, are a promising class of therapeutic agents against PD.
Accelerated intracellular vesicle trafficking suppresses apoptosis downstream of ROS
When S. cerevisiae is grown in water with glucose but no added nitrogen source, the cells accumulate ROS and undergo apoptosis; this phenomenon is now called sugar-induced apoptosis (Granot et al., 2003). In a previous section, we described the results from screening a tomato cDNA library to find suppressors of Bax lethality. An analogous genetic screen was conducted to find suppressors of sugar-induced lethality; however, in this case, an Arabidopsis cDNA library was used (Levine et al., 2001). The screen uncovered an Arabidopsis gene, referred to as AtVAMP, that blocks sugar-induced killing as well as hydrogen peroxide-induced killing. AtVAMP showed the highest homology to the rat synaptobrevin gene VAMP7 (41% identity and 64% similarity). VAMPS are vesicle-associated membrane proteins that belong to the family of v-SNARES. VAMPs and v-SNARES, together with their counterpart t-SNARES, mediate vesicle docking events throughout the secretory pathway, including the endoplasmic reticulum, the Golgi apparatus, early endosomes, and the plasma membrane. Membrane proteins transit through the ER and Golgi apparatus in vesicles, and these vesicles ultimately dock and fuse with the plasma membrane or different organelles. The transport of newly synthesized proteins to the plasma membrane is balanced by retrograde transport of damaged proteins from the plasma membrane to the vacuole. Since oxidative stress damages membranes and membrane proteins, efficient membrane recycling is essential to maintain the integrity of the plasma membrane. Failure to repair the plasma membrane leads to leaky membranes and cell death. Thus, it was suggested by Granot and colleagues that AtVAMP accelerates membrane recycling, and this suppresses apoptosis downstream from the oxidative stress. Perhaps novel compounds could be discovered that stimulate vesicle trafficking.
Hsp70 chaperones suppress neurodegeneration
A Drosophila model of PD has revealed the importance of molecular chaperones, specifically Hsp70s, in protecting against neurodegeneration induced by the directed expression of human α-syn (Auluck et al., 2002). For example, flies expressing WT α-syn in dopaminergic neurons had 50% fewer neurons after 20 days compared to the control group which did not contain the transgene. In contrast, when human Hsp70 was coexpressed with WT α-syn, no loss of fly neurons occurred after 20 days. Microscopic examination of dopaminergic neurons from flies expressing both human Hsp70 and WT α-syn revealed that Lewy bodies were still present in neurons. These experiments showed that Hsp70 protects dopaminergic neurons from a degenerative cell death induced by α-syn, and that Lewy bodies, rather than being toxic and causing disease, may be inert inclusions. Later work by this group showed that flies with directed expression of α-syn in their dopaminergic neurons are protected from a degenerative cell death if the flies are fed a diet enriched in geldanamycin (Auluck & Bonini, 2002; Auluck et al., 2005). Geldanamycin is a chemical activator of the heat shock response: it induces the expression of Hsp70s and other chaperones. Bonini and coworkers hypothesized that chaperones are part of a cellular arsenal that protect against α-syn toxicity. Alternatively, perhaps α-syn interferes with normal chaperone function, and this loss of function causes neurodegeneration. Adding Hsp70, whether by a transgene or by a chemical stimulator such as geldanamycin, would boost Hsp70 and prevent neurodegeneration.
Geldanamycin, as well as the overexpression of Hsp70 (Ssa3p or Ssa4p), was shown to protect yeast expressing the various α-syns (WT, A30P, or A53T) from ROS accumulation (Flower et al., 2005). The findings in yeast and cultured human or rat neurons (Xu et al., 2002; Flower et al., 2005; Smith et al., 2005) – that α-syn induces ROS – may shed light on the nature of α-syn toxicity in Drosophila. We suggest that the protection conferred by human Hsp70 in the Drosophila model is due to suppression of α-syn-induced ROS and α-syn-induced apoptosis. Such a hypothesis could be tested by feeding flies with α-syn expressed in their dopaminergic neurons a free radical spin trap, such as TEMPO, or glutathione. If a diet supplemented with TEMPO or glutathione suppresses neurodegeneration at day 20 compared to untreated controls, it would indicate that α-syn induces ROS in dopamine-producing neurons, and this plays a role in their neurodegeneration. Alternatively, perhaps TEMPO and glutathione induce Hsp70 expression; this could be tested by Western blotting.
How does Hsp70 protect cells from α-syn-induced ROS? Hsp70 chaperones reversibly bind unfolded substrate proteins, such as α-syn, in an ATP-dependent reaction cycle, where free energy from ATP binding and hydrolysis are probably used by Hsp70 to unfold or disentangle its substrates (Slepenkov & Witt, 2002). We propose that via its association with Hsp70, Hsp70 suppresses α-syn toxicity: (i) by preventing α-syn membrane hole punching activity; (ii) by preventing protofibril and aggregate formation (Dedmon et al., 2005); (iii) by converting α-syn from a toxic to nontoxic conformation; or (iv) by protecting α-syn from oxidation. Overexpression of Hsp70 increases the rates of reactions (i–iv) and this suppresses α-syn toxicity (ROS) (Fig. 1).
Hsp70 and apoptosis
Hsp70 is a potent inhibitor of apoptosis in mammalian cells. Hsp70 binds to a protein called Apaf-1, and this interaction inhibits formation of the apoptosome, which is a multimeric structure required for the activation of caspases and subsequent selective proteolysis of key cellular proteins (Beere et al., 2000; Saleh et al., 2000). Failure to form the apoptosome prevents apoptosis. Within this framework, increasing Hsp70 concentrations in mammalian cells via overexpression inhibits apoptosis; conversely, decreasing Hsp70 concentrations should bias cells towards apoptosis. Two processes, one subtle and one obvious, can decrease Hsp70 concentration: first, given the propensity of α-syn to bind to Hsp70s, and given its abundance (α-syn comprises 1% of the cytosolic protein in neurons), α-syn saturates Hsp70 binding sites, which dramatically decreases the concentration of unbound Hsp70. Second, the deposition of Hsp70 molecules into Lewy bodies decreases the concentration of Hsp70. Decrements in Hsp70 concentrations due to either of these two processes could bias cells towards apoptosis (Fig. 1). In the face of these processes that decrease the concentration of Hsp70, an obvious therapeutic approach, pioneered by Bonini and colleagues, is to use chemical activators of the heat shock response, such as geldanamycin or other natural products, to boost Hsp70 expression. One caveat about this therapy is that, because Hsp70 inhibits apoptosis, Hsp70 can be considered a proto-oncogene (Garrido et al., 2003). Chemical activators of the heat shock response, like caspase inhibitors, may be a double-edged sword: they might protect neurons by inhibiting unwanted apoptosis, but by inhibiting apoptosis in other cells, they might promote tumor formation.
α-Syn induces chronic oxidative stress and apoptosis in a variety of eukaryotic cells, including yeast, rat and human cells, but the mechanistic details have not been elucidated. Nevertheless, even with an incomplete understanding of the mechanism of α-syn-induced oxidative stress, cells can be protected from α-syn toxicity in various ways. We have outlined how Hsp70 overexpression, caspase inhibitors, and antioxidants may be used to thwart oxidative stress and apoptosis in cells that express α-syn. Perhaps in the future, individuals with Parkinson's disease will take a cocktail of compounds that serve to inhibit different pathobiological processes activated by α-syn. It is likely that the yeast system will contribute to the discovery of these various therapeutic compounds.
This work was supported in part by grants from the Parkinson's Resource of Northwest Louisiana and from the National Institutes of Health (R21NS053678) to S.N.W.