Address correspondence and reprint requests to Dr Rugao Liu, Kosair Children's Hospital Research Institute, Department of Pediatrics, 570 S. Preston St., Suite 321, University of Louisville, Louisville, KY 40202, USA. E-mail: r0Liu002@gwise.louisville.edu
The molecular mechanisms of selective motor neuron degeneration in human amyotrophic lateral sclerosis (ALS) disease remain largely unknown and effective therapies are not currently available. Mitochondrial dysfunction is an early event of motor neuron degeneration in transgenic mice overexpressing mutant superoxide dismutase (SOD)1 gene and mitochondrial abnormality is observed in human ALS patients. In an in vitro cell culture system, we demonstrated that infection of mouse NSC-34 motor neuron-like cells with adenovirus containing mutant G93A-SOD1 gene increased cellular oxidative stress, mitochondrial dysfunction, cytochrome c release and motor neuron cell death. Cells pretreated with highly oxidizable polyunsaturated fatty acid elevated lipid peroxidation and synergistically exacerbated motor neuron-like cell death with mutant G93A-SOD1 but not with wild-type SOD1. Similarly, overexpression of mitochondrial antioxidative genes, MnSOD and GPX4 by stable transfection significantly increased NSC-34 motor neuron-like cell resistance to mutant SOD1. Pre-incubation of cells with␣spin trapping molecule, 5′,5′-dimethylpryrroline-N-oxide (DMPO), prevented mutant SOD1-mediated mitochondrial dysfunction and cell death. Furthermore, treatment of mutant G93A-SOD1 transgenic mice with DMPO significantly delayed paralysis and increased survival. These findings suggest a causal relationship between enhanced oxidative stress and mutant SOD1-mediated motor neuron degeneration, considering that enhanced oxygen free radical production results from the SOD1 structural alterations. Molecular approaches aimed at increasing mitochondrial antioxidative activity or effectively blocking oxidative stress propagation can be potentially useful in the clinical management of human ALS disease.
sodium dodecyl sulfate–polyacrylamide gel electrophoresis
Human amyotrophic lateral sclerosis (ALS) is a progressive and fatal neurodegenerative disease that mainly affects motor neurons in brainstem and spinal cord (Williams and Windebank 1991). The landmark discovery by Rosen et al. (1993) has unveiled that mutations of SOD1 are responsible for at least some forms of ALS disease in humans. An initial study by Gurney et al. (1994) has further demonstrated that overexpression of G93A-SOD1 gene, not the wild-type superoxide dismutase (SOD)1 gene, led to the clinical and pathological phenotype of ALS-like disease in a transgenic mouse model. Additional transgenic mice expressing different mutations of SOD1 gene, either with elevated or unchanged SOD1 activity, also developed an ALS-like disease (Borchelt et al. 1994; Ripps et al. 1995; Wong et al. 1995; Bruijn et al. 1997). In contrast, SOD1 knockout mice had normal phenotype and did not develop ALS-like disease under normal living conditions (Reaume et al. 1996). Taken together, these studies strongly suggest that a gained function or gained functions by mutant SOD1 contribute to the characteristic motor neuron degeneration of ALS disease.
More than 70 mutations in human SOD1 gene have thus far been identified for ALS disease (Deng et al. 1993; Orrell 2000; Julien 2001), and most occur within regions of the SOD1 subunit folding that leads to the delocalization of copper in the active site. The aberrant properties of mutant SOD1 result in the production of peroxynitrite (Beckman et al. 1993; Beal et al. 1997) and Fenton-like reactions (Wiedau-Pazos et al. 1996; Yim et al. 1997). However, the biochemistry and free radical chemistry leading to the production of oxygen free radicals remain controversial in the presence of mutant SOD1 (Singh et al. 1998). Using a G93A-SOD1 transgenic mouse model, we and others have demonstrated that enhanced production of oxygen free radicals in the spinal cords is associated with ALS-like disease onset and progression (Bogdanov et al. 1998; Liu et al. 1998). In support of these findings, the oxidation markers, protein carbonyls and protein nitration, have been found elevated in both human ALS patients (Bowling et al. 1993) and in ALS-like transgenic mice (Beal et al. 1997; Ferrante et al. 1997). More recently, it has been shown that mutant SOD1 forms protein aggregates in human ALS patients (Kato et al. 1996; Shibata et al. 1996), transgenic mice (Bruijn et al. 1998; Johnston et al. 2000) and in cell culture (Johnston et al. 2000). However, the contribution of oxidative stress to motor neuron degeneration has not been clearly defined, and the causal relationship between increased oxidative stress and protein aggregation by mutant SOD1 has not been conclusively established. Mitochondrial defects have been demonstrated to associate ALS-like disease onset and progression in a transgenic mouse model (Dal Canto and Gurney 1995; Kong and Xu 1998). More significantly, partially decreased mitochondrial MnSOD activity increased early ALS-disease onset and progression and decreased survival (Andreassen et al. 2000). In this paper, we employed in vitro and in vivo models to study the molecular mechanisms and potential rescue of mutant SOD1-mediated motor neuron degeneration. We demonstrated that mutant SOD1-mediated lipid peroxidation and cytochrome c release contributed to the motor neuron-like cell death in vitro, while increased mitochondrial antioxidative activity prevents mutant SOD1-mediated motor neuron-like cell death. Furthermore, we showed that treatment of cells and mutant G93A-SOD1 transgenic mice with the spin trapping compound, DMPO, significantly reduced mutant SOD1-mediated motor neuron death, delayed paralysis and prolonged survival.
Materials and methods
Chemicals and reagents
N-Acetyl-cysteine (N-AC), 5′,5′-dimethylpryrroline-N-oxide (DMPO), steric acid (SA) [18 : 0], linolenic acid (LA) [18 : 3n − 6],␣docosahexaenoic acid (DHA) [22 : 6n − 3], cumene hydroperoxide, xanthine, xanthine oxidase and lactate dehydrogenease (LDH) assay kits were purchased from Sigma Chemical Co. (St Louis, MO, USA). Cell culture medium and transfection reagents were obtained from Life Technology Inc. (Galesburg, MD, USA). Polyclonal antibody against SOD1 was purchased from Chemicon International (Temecula, CA, USA). LPO-586 and monoclonal antibody against cytochrome c were purchased from R & D Systems (Minneapolis, MN, USA).
NSC-34 motor neurons cells (Passages 4–21) (kindly provided by Dr Neil Cashman, University of Toronto) and Neuro-2a cells (Passages 163–172) (purchased from ATCC, Manassas, VA, USA) were cultured in a humidified atmosphere of 95% air −5% CO2 at a 37°C incubator in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% heat inactivated fetal bovine serum (FBS), 100 units/mL penicillin and 100 µg/mL streptomycin (DMEM complete medium) as previously described (Cashman et al. 1992; Bishop et al. 1999). Cell culture medium was usually replaced every 3–4 days. Once 90% cell confluence was reached, cells were disrupted with 0.15% trypsin and 1 mm EDTA for 5 min. Dissociated cells were spun down and replated in new culture dishes or flasks (Corning Inc., Acton, MA, USA).
Enrichment of cells with PUFA was carried out as previously described (North et al. 1992; North et al. 1994). Briefly, 70–80% confluence cells were incubated with 20 µm PUFA in culture medium for 24 h, and infected with either adenovirus alone or adenovirus containing WT-SOD1 or mutant SOD1 for 24 h. After washing twice with sterile 1 × phosphate-buffered saline (PBS), cells were cultured for various lengths of time and processed for analyses.
Cell viability of NSC-34 or Neuro-2a cells upon treatments was determined by trypan blue exclusion assay as described (Liu et al. 1999). Briefly, cells were incubated with 0.1% trypan blue dye for 5 min at room temperature (22–25°C) and were counted on a hemocytometer. Cell viability was expressed as the number of viable cells (dye-excluding) divided by total number of cells.
Adenovirus preparation and infection
The construction of adenoviral vectors containing human WT-SOD1, G93A-SOD1 gene or devoid of any transgene (vector control) was based on previously published procedures (Wang and Finer 1996; Craig et al. 1997). Recombinant adenoviruses were propagated in 293 cells and purified by cesium chloride density centrifugation. After titration, adenovirus was stored in − 80°C as described (Craig et al. 1997).
NSC-34 cells were infected with 2 × 106 PFU/mL of adenovirus alone or adenovirus containing WT-SOD1 or G93A-SOD1 gene for 24 h. The infected cells were then washed twice with 5 mL sterile 1 × PBS twice and incubated with complete medium for various lengths of time before analysis.
Transfection and selection of cell clones
For stable transfection, 10 × 106 cells were incubated with 20 µg MnSOD and/or GPX4 expression plasmid DNA in 0.5 mL Opti-MEM (Life Technology, Galesburg, MD, USA) at room temperature for 10 min, were electroporated with a BTX electroporator using a low voltage mode at 160 V, 1 pulse, and 25 ms/V of pulse length, kept at room temperature for 30 min, and then cultured in complete medium in the presence of 0.8 mg/mL G418. After 3 weeks, cell clones resistant to G418 were isolated, analyzed by RT-PCR and further confirmed by activity assay. For convenience, one MnSOD, one GPX4, and one control vector stable expression clones were selected for further studies. Transfected cells were routinely cultured in the presence of 0.5 mg/mL of G418. However, 2 days prior to and during the experiments, transfected cells were cultured in complete DMEM medium without G418.
Mitochondrial function assay
The effect of WT-SOD1 and mutant G93A-SOD1 on mitochondrial function was carried out as described by Oddis and Finkel (1995), and Madesh et al. (1997). Briefly, cells infected with vector control adenovirus, or adenovirus containing WT-SOD1 cDNA or G93A-SOD1 cDNA were incubated with 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) for 4 h. The medium was then removed and cells were washed twice gently with 1 × PBS. The production of formazan by the cells in the wells was dissolved with 200 µL of dimethyl sulfoxide. Absorbency was measured at 540 nm using an automatic microplate reader. In addition, a lactate dehydrogenase (LDH) release assay was used to confirm the MTT method for mitochondria function assay. Briefly, the culture medium from vector control, WT-SOD1 or mutant G93A-SOD1 infected cells was collected after certain period time of incubation. After centrifugation at 10 000 g for 20 min at 4°C, the cell-free medium was allowed to react with the LDH detection reagents, and measurement was carried out with a microplate reader at 340 nm according to the manufacturer's instructions (Sigma).
Reactive oxygen species production assay and lipid peroxidation assay
A dichlorofluorescein (DCF) assay was used to determine cellular oxygen free radical generation in adenovirus containing WT-SOD1 gene or G93A-SOD1 gene infected cells (LeBel et al. 1992; Wang and Joseph 1999). Briefly, 2 × 105 cells/well cultured in a 96-well plate were incubated with 100 µm of 6-carboxy-2′,7′-dicholorfluorescin diacetate (DCFH-DA) (Molecular Probes, Eugene, OR, USA) for 30 min after washing with sterile PBS twice, DCFH-DA-loaded cells were placed in a multiwell fluorescence plate reader with temperature maintained at 37°C. The excitation and emission wavelengths were set at 485 nm and 530 nm, respectively. Data analysis was performed as described by Wang and Joseph (1999).
An LPO-586 kit was used to measure lipid peroxidation in the NSC-34 cell culture model according to manufacturer's instructions. Specifically, 20 × 106 vector control, WT-SOD1 and mutant G93A-SOD1 infected cells were homogenized in 10 mm phosphate buffer (pH 7.0). Equal amounts of protein in each sample were allowed to react with a chromogenic reagent at 45°C for 40 min. The samples were centrifuged and supernatant were measured at 586 nm. The level of lipid peroxidation was calculated with the standard curve according to the manufacturer's instructions (R & D Systems).
MnSOD activity assay and GPX4 activity assay
Stable transfected cells overexpressing MnSOD, GPX4, or both (about 20 × 106 cells), were vigorously washed with PBS and scraped off the dishes in cold PBS. Cell pellets were resuspended in 500 µL of 10 mm Tris-HCl activity buffer (pH 7.4) and sonicated on ice for 10 s three times. MnSOD activity was assayed as described by Spitz and Oberley (1989). Briefly, SOD activity was determined spectrophotometricaly at 560 nm by measuring reduction of nitroblue tetrazolium (NBT) in a superoxide generating system. The superoxide generated from the xanthine and xanthine oxidase system (Sigma Chemical Company) reduces NBT.␣However, the reduction of NBT is competitively inhibited in the presence of SOD that catalyzes superoxide to hydrogen peroxide. By measuring the amount of protein necessary for half maximal inhibition, the SOD activity in the sample was determined. Sodium cyanide was used to inhibit Cu,Zn-SOD activity and therefore the SOD activity in the presence of 5 mm sodium cyanide represented MnSOD activity. SOD activity values were expressed in units per milligram of protein.
GPX4 activity assay was carried out according to Lawrence and Burk (1976) and Maiorino et al. (1990) using cumene hydroperoxide as a substrate. Briefly, 300 µg/100 µL cellular extracts were incubated with 700 µL of a cocktail mixture containing 50 mm potassium phosphate buffer (pH 7.8), 1 mm EDTA, 1 mm NaN3, 1 mm glutathione and 1 unit glutathione reductase for 10 min NADPH consumption was measured at 340 nm for 3 min after addition of 100 µL of 1.5 mm NADPH. NADPH oxidation was measured by addition of 100 µL of 2 mm cumene hydroperoxide at 340 nm for 3 min at 20-s intervals. GPX4 activity was calculated according to Lawrence and Burk (1976) and Maiorino et al. (1990).
Cytochrome c release analysis
Cytochrome c release to cytoplasmic fraction was analyzed according to Fujimura et al. (2000) with minor modifications. Briefly, cells infected with adenovirus alone, adenovirus containing WT-SOD1 or mutant SOD1 were harvested and homogenized with ice cold suspension buffer [20 mm HEPES–KOH, pH 7.5, 250 mm sucrose, 10 mm KCl, 1.5 mm MgCl2, 1 mm EDTA, 1 mm EGTA, 1 mm dithiothreitol (DTT), 0.1 mm phenylmethylsulfonyl fluoride (PMSF), 2 µg/mL aprotinin, 10 µg/mL leupeptin]. The homogenates were centrifuged at 100 000 g for 60 min at 4°C, and protein concentration was measured in the supernatants. Western blots were performed to determine the cytochrome c release to cytosol.
Western blot analysis was carried out in a similar way to those described previously (Liu et al. 1999). Briefly, parental, vector control cells and adenovirus infected cells were washed with 1 × PBS twice and scraped off the dishes in the same solution. After centrifugation at 1200 gfor 5 min, cell pellets were lysed vigorously in NP-40 lysis buffer (50 mm Tris-HCl, pH 7.5, 150 mm, 0.5% NP-40, 1 mm EDTA, 1 mm PMSF, 10 µg/mL leupeptin and 10 µg/mL aprotinin). Lysates were spun down at 10 000 g for 5 min at 4°C. The protein concentrations in the supernatants were measured by Bradford method of Bio-Rad (Hercules, CA, USA). A total of 10 µg to 20 µg of proteins from each sample were separated by 15% sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE), followed by transferring of the separated proteins to a polyvinylidene difluoride (PVDF) membrane and immunoreacted with specific antibody. The immunoreactive protein was detected by enhanced chemiluminescence (ECL).
Analysis of ALS-like disease progression and survival in transgenic mouse overexpressing mutant G93A-SOD1 gene
The transgenic mouse line overexpressing mutant G93A-SOD1 gene [TgN(SOD1-G93A)1Gur, a high expression mouse line] and the mating pairs were purchased from Jackson laboratory (Bar Harbor, ME, USA). Mice were housed and bred as described previously (Gurney et al. 1996) in accordance with the Institutional Animal Care and Use Committee guidelines. Seventy day-old mice (16 mice per group) were injected intraperitoneally (i.p.) with either 0.2 mL of vehicle alone (0.9% saline buffer) or 0.2 mL of␣25 mg/mL DMPO three times a week, until animals reached the terminal stage of ALS-like disease. Mice were examined daily for paralysis, disease progression and survival analysis. The initial symptoms of hind leg paralysis were considered as the disease progression threshold. Mice were killed at terminal stage, i.e. severely paralysis and inability to seek food and water. Survival and paralysis curves were plotted using Sigmaplot software (Chicago, IL, USA).
Data are reported as means ± SD. Statistical comparisons with specific controls were made by two-tailed Student's t-tests. A␣p-value␣< 0.05 was considered as reaching statistical significance.
Mutant G93A-SOD1 increased oxidative stress-mediated NSC-34 motor neuron cell death
The expression of WT-SOD1 or G93A-SOD1 protein was dramatically increased in adenovirus containing WT- or G93A-SOD1 infected NSC-34 cells compared with parental and vector control cells as determined by western analysis, in which the upper band represents the human SOD1 expressed from adenovirus infection and the lower band represents the endogenous mouse SOD1 (Fig. 1a). The increased G93A-SOD1 expression significantly enhanced time-dependent oxidative stress (Fig. 1b) as measured by 2′,7′-DCF, and was associated with increased cell death when assessed by trypan blue exclusion (Fig. 1c) or the MTT assay (data not shown). Thus, the parallel increases in ROS production and motor neuron cell death suggest that increased oxidative stress by mutant SOD1 significantly contributes to motor neuron-like cell death.
Mutant G93A-SOD1 increased lipid peroxidation and mitochondrial dysfunction in NSC-34 cells
Analysis of transgenic mice overexpressing G93A-SOD1 gene has suggested that lipid peroxidation may participate in selective motor neuron degeneration in human ALS disease (Hall et al. 1998). NSC-34 cells infected with G93A-SOD1 displayed dramatic increases in lipid peroxidation in a time- and dose-dependent fashion, while infection with adenovirus containing WT-SOD1 did not have a significant effect on lipid peroxidation compared with the adenovirus control vector (Figs 2a and b). The increased oxidative stress by mutant G93A-SOD1 led to mitochondrial dysfunction as determined by MTT assay (Fig. 2c) and by increased LDH release (data not shown). Pre-incubation of NSC-34 cells with 20 mm of the antioxidant, N-AC improved mitochondrial function (Fig. 2c) and cell viability, indicating that oxidative stress-mediated mitochondrial dysfunction contributes, at least partially, to the motor neuron-like cell death.
Because mutations of SOD1 selectively target to motor neuron cells in ALS disease, we hypothesized that mutant SOD1-mediated lipid peroxidation would be specifically increased in motor neuron cells. To examine this issue, we infected both mouse Neuro-2a neuroblastoma cells and NSC-34 motor neuron-like cells with the same dose of adenovirus containing G93A-SOD1 gene. As shown in Fig. 3(a), the expression of G93A-SOD1 protein in Neuro-2a and NSC-34 cells was increased at similar levels. However, mutant G93A-SOD1-mediated lipid peroxidation and cell death were significantly higher in NSC-34 motor neuron cells than in Neuro-2a neuroblastoma cells (Figs 3b and c). Thus, increased lipid peroxidation by mutant G93A-SOD1 is a key oxidative reaction leading to motor neuron cell degeneration.
Since fatty acid contents in cell and organelle membrane may affect lipid peroxidation status, modification of cellular contents by oxidizable PUFA can potentially increase lipid peroxidation (Alexander-North et al. 1994; North et al. 1994; Wagner et al. 1994). We therefore incubated NSC-34 cells with 20 µm of SA [18 : 0], LA [18 : 3n-6], and docosahexaenoic acid (DHA) [22 : 6n-3]. Cells were then infected with adenovirus vector alone, adenovirus containing WT-SOD1 or G93A-SOD1. PUFA did not significantly affect SOD1 expression (data not shown). However, the highly oxidizable PUFA synergistically enhanced G93A-SOD1-mediated lipid peroxidation (Fig. 4a) and cell death (Fig. 4b). Furthermore, expression of WT-SOD1 had only a modest protective effect on motor neuron death in the presence of PUFA. The enhanced toxicity by highly oxidizable PUFA in G93A-SOD1 infected cells further suggests that lipid peroxidation and cellular PUFA composition in motor neuron cells may play an important role in selective motor neuron degeneration in ALS disease.
Mutant G93A-SOD1 increased cytochrome c release in motor neuron-like cells
Oxidative stress-mediated mitochondrial dysfunction may result in cytochrome c release and subsequent cell death. We therefore examined cytochrome c release to cytosol from mitochondria by western analysis in adenovirus infected NSC-34 cells. Figure 5a shows that cytochrome c release into the cytoplasmic fraction was dramatically increased in NSC-34 cells infected with adenovirus containing G93A-SOD1 gene, while infection with adenovirus containing WT-SOD1 gene elicited cytochrome c release that were similar to those found with control adenovirus vector. The elevated cytochrome c release was associated with increased cell death as determined by trypan blue exclusion (Fig. 5b) and MTT assays (data not shown). Pre-incubation of cells with 20 mmN-AC reduced cytochrome c release and increased cell survival (Fig. 5b), suggesting that oxidative stress by mutant SOD1 contributes, at least partially, to motor neuron-like cell death.
Neuro-2a cells were again used to study the differential effects of mutant SOD1 on neuron death. Similar increases in mutant SOD1 protein expression occurred in Neuro-2a and NSC-34 cells after infection with adenovirus containing G93A-SOD1 gene. However, release of cytochrome c and cell death in NSC-34 cells were significantly increased compared with Neuro-2a cells (Fig. 5c). Thus, cytochrome c release may represent another important pathway involved in the selective motor neuron degeneration of human ALS disease.
Overexpression of MnSOD gene and GPX4 gene reduced G93A-SOD1-mediated NSC-34 motor neuron-like cell death
Because oxidative stress-mediated mitochondrial dysfunction is associated with motor neuron degeneration in ALS transgenic mice, we reasoned that increased expression of MnSOD and GPX4 could reduce G93A-SOD1-mediated cell toxicity. To examine this possibility, NSC-34 cells were initially transfected with a mammalian expression vector containing MnSOD gene and/or GPX4 gene. After selection for 3 weeks in the presence of 0.8 mg/mL of G418, resistant clones were analyzed by RT-PCR for MnSOD and/or GPX4 expression (data not shown) and activity (Fig. 6a). Individual clones from control vector, MnSOD, GPX4, or MnSOD and GPX4, were chosen to study mitochondrial antioxidative/antiperoxidative activity and its effect on motor neuron survival. MnSOD and GPX4 activity was significantly increased in cells transfected with MnSOD and/or GPX4 genes compared with that of control vector-transfected cells (Fig. 6a). As shown above, infection with adenovirus containing the G93A-SOD1 gene elicited increased cytochrome c release (Fig. 6b), lipid peroxidation and cell death as compared with that of control vector-transfected cells (Figs 6c and d). However, overexpression of MnSOD and/or GPX4 genes prevented mutant SOD1-mediated cytochrome c release and significantly increased cell survival (Figs 6b–d), suggesting that increased antioxidative activity in mitochondria may prevent motor neuron degeneration.
Spin trapping molecule DMPO reduced G93A-SOD1-mediated cell death in NSC-34 cells, delayed paralysis and increased survival in G93A-SOD1 transgenic mice
Increased oxygen free radical production by mutant SOD1 contributes significantly to the motor neuron death in ALS disease (Wiedau-Pazos et al. 1996; Yim et al. 1997; Bogdanov et al. 1998; Liu et al. 1998). Spin trapping molecules could therefore prevent mutant SOD1-mediated motor neuron-like cell death. Chemically, the spin trapping molecules can effectively react with oxygen free radicals and prevent free radical propagation (Buettner 1987; Buettner 1993). We therefore tested whether DMPO, a highly effective spin trapping compound, could prevent mutant G93A-SOD1-mediated cell death in the NSC-34 cell culture system. After preincubation of cells with 25 mm DMPO, mutant G93A-SOD1-mediated lipid peroxidation, cytochrome c release and cell death were dramatically reduced, while control and WT-SOD1 overexpressed cells were not significantly affected (Fig. 7).
Because DMPO can cross the brain–blood barrier and little if any cell toxicity, we next investigated the therapeutic efficacy of DMPO on delaying paralysis and increasing survival in G93A-SOD1 transgenic mice. Treatment of G93A-SOD1 transgenic mice with 5 mg/mouse of DMPO dissolved in sterile 0.9% saline buffer by i.p. injection three times a week significantly delayed mouse paralysis and increased survival compared with vehicle-treated transgenic mice (Fig. 8).
Two major hypotheses have thus far been proposed to explain mutant SOD1-mediated motor neuron toxicity in ALS disease, namely aberrant oxidative damage and protein aggregation (Gurney et al. 1998; Cleveland 1999; Cleveland and Liu 2000; Julien 2001). While these two hypotheses appear to be vastly different and non-overlapping, substantial evidence has suggested that these two models are closely related (Wiedau-Pazos et al. 1996; Yim et al. 1997; Bruijn et al. 1998; Liu et al. 1999; Johnston et al. 2000). The presence of mutant SOD1 initiates a cascade of oxidative reactions ultimately resulting in motor neuron degeneration. The enhanced production of oxygen radicals could also modulate and facilitate the formation of mutant protein aggregation and cause cell toxicity (Fig. 9). The present study provides direct evidence supporting the role of mutant SOD1-induced oxidative stress in the initiation of motor neuron-like cell degeneration, and also demonstrates that increasing antioxidative activity enhances motor neuron survival in both a cell line and a transgenic mouse model of ALS-like disease.
Enhanced oxidative stress by mutant SOD1 contributes to motor neuron degeneration
It has been previously shown that mutant SOD1 enhanced production of hydroxyl radicals in vitro (Wiedau-Pazos et al. 1996; Yim et al. 1997; Liu et al. 1999) and in vivo (Liu et al. 1998; Bogdanov et al. 1998). We further employed plasmid transfection and adenovirus infection methods to study the molecular mechanisms underlying mutant SOD1-mediated neuron degeneration. We have now shown that expression of the mutant G93A-SOD1 gene increased cellular production of ROS in mouse NSC-34 motor neuron-like cells, and that increased lipid peroxidation contributed to mitochondrial dysfunction and motor neuron death. Furthermore, addition of the highly oxidizable LA or DHA to the media synergistically enhanced mutant SOD1-mediated lipid peroxidation and cell death in NSC-34 cells, while the lesser oxidizable SA modestly decreased mutant SOD1-mediated lipid peroxidation and cell death. Taken together, these findings provide direct evidence to support the causal relationships between oxidative stress and motor neuron death in the presence of mutant SOD1.
More recently, mutant SOD1 protein aggregation has been implicated in motor neuron death in ALS transgenic mice expressing the mutant SOD1 gene (Bruijn et al. 1998; Johnston et al. 2000). However, the toxic effect of mutant SOD1 to motor neurons is independent of wild-type SOD1 activity, because neither increasing nor eliminating wild-type SOD1 affected ALS disease onset and progression in G85R-SOD1 transgenic mice (Bruijn et al. 1998). In a cell culture model, Johnston et al. (2000) demonstrated that expression of mutant, but not wild-type SOD1 protein was associated with the formation of insoluble protein complexes, the latter being further correlated with cell toxicity. Notwithstanding these findings, it still remains unclear as to how the mutant SOD1 protein forms aggregates and how these aggregates target to the motor neuron death. One possibility entails that enhanced oxygen radical production by mutant SOD1 may participate in the formation of protein aggregates, although the causal relationship between oxidative stress and protein aggregation has not been established. Because increased oxygen free radical production is resulted from the structural alterations of mutant SOD1, we propose that enhanced oxidative stress is an early event followed by other molecular processes, such as protein aggregation, leading to motor neuron degeneration (Fig. 9).
Mitochondrial dysfunction is associated with motor neuron degeneration
It is well recognized that mitochondria are a vulnerable target to oxidative stress (Beal 2000). Our current findings demonstrate that increased lipid peroxidative damage correlates with mitochondrial dysfunction in NSC-34 motor neuron-like cells overexpressing mutant G93A-SOD1 gene, but not the WT-SOD1 gene (Fig. 1). In addition, we showed that NSC-34 cells pretreated with a spin trapping molecule that reduced oxidative stress propagation were resistant to mutant SOD1-mediated lipid peroxidation, mitochondrial dysfunction and cell death (Figs 2 and 7 and data not shown), confirming the role of mitochondrial damage in motor neuron death (Dal Canto and Gurney 1995; Kong and Xu 1998).
In a transgenic mouse model, Kostic et al. (1997) showed that overexpression of Bcl2 gene delayed the ALS-like disease onset and progression and increased survival. More recently, Li et al. (2000) demonstrated that intracerebroventricular administration of zVAD-fmk, a broad caspase inhibitor, also significantly increased motor function, delayed disease onset, and increased survival in mutant G93A-SOD1 transgenic mice. Using the NSC-34 motor neuron cell culture model, we now showed that overexpression of mutant G93A-SOD1 gene increased cytochrome c release to cytosol and caused motor cell death, while overexpression of WT-SOD1 did not have significant effect. Increased cytochrome c release was also observed in the G93A-SOD1 transgenic mouse spinal cord in association with ALS disease onset and progression (data not shown). Thus, we conclude that oxidative stress-mediated mitochondrial dysfunction leading to cytochrome c release plays an important role in the initiation of motor neuron death by mutant SOD1 effects.
We previously reported on mitochondrial abnormalities in PC12 cells transfected with mutant SOD1 gene (Liu et al. 1999), and similar findings occur in transgenic mice overexpressing mutant G93A-SOD1 (Dal Canto and Gurney 1995; Mourelatos et al. 1996; Kong and Xu 1998), as well as in the human ALS disease-affected spinal cord (Bowling et al. 1993). The most prominent feature was the vacuolation of mitochondria at an early stage preceding and during the early stages of disease in mutant G93A-SOD1 transgenic mice (Dal Canto and Gurney 1995; Wong et al. 1995; Mourelatos et al. 1996; Kong and Xu 1998). Onset of ALS-like disease was accompanied by an exponential increase in the number of vacuoles in mitochondria and this was followed by rapid motor neuron death (Dal Canto and Gurney 1995; Kong and Xu 1998). The relatively prolonged period during which mitochondrial degeneration occurs and which preceded the onset of disease progression clearly offers a therapeutic window for ALS disease prevention and therapy through attempts to increase mitochondrial antioxidative/antiperoxidative activity.
Mitochondrial antioxidative/antiperoxidative enzymes and spin trapping molecules prevent mutant SOD1-mediated cell death
Mutant SOD1-mediated mitochondrial dysfunction prompted us to study the functional role of the mitochondrial antioxidative/antiperoxidative enzymes, MnSOD and GPX4 in the prevention of motor neuron degeneration. Transfection and expression of MnSOD and/or GPX4 reduced mutant SOD1-mediated lipid peroxidation and cell death (Fig. 6). More importantly, the cooperation between MnSOD and GPX4 prevented mutant SOD1-mediated oxidative mitochondrial damage and increased motor neuron-like cell survival (Fig. 6). In a bi-transgenic mouse model, Andreassen et al. (2000) have recently shown that partial reduction of MnSOD activity increased the early onset and progression of ALS disease, suggesting that increased mitochondrial antioxidative activity may prevent mutant SOD1-mediated motor neuron degeneration. Thus, targeting of mitochondrial antioxidant/antiperoxidative activity may provide an effective approach to prevent ALS disease onset and progression (Fig. 9).
Because enhanced oxygen free radical production contributes to the motor neuron death, it is reasonable to assume that molecules that can effectively block oxidative stress propagation may prevent motor neuron death and delay disease onset and progression. To this end, we screened a variety of spin trapping compounds to assess their efficacy in preventing oxidative stress-mediated motor neuron death. DMPO was shown highly effective reductions in mutant SOD1-mediated lipid peroxidation, cytochrome c release and cell death in the NSC-34 cell culture system. However, DMPO did not significantly affect the lipid peroxidation status in vector control and WT-SOD1 overexpressed cells. We reasoned that the basal level of oxygen free radical production in control and WT-SOD1 overexpressed cells was relatively low and preincubation of DMPO did not had additional effects on the inhibition of basal level of lipid peroxidation. Because of the effectiveness of DMPO on mutant SOD1-mediated NSC-34 motor neuron like cell death in vitro, we selected DMPO to treat mutant G93A-SOD1 transgenic mice at an age that precedes the onset of ALS disease, and found that this intervention delayed paralysis and increased survival when compared with mutant mice treated with vehicle alone (Fig. 8). DMPO has been shown to react with oxygen radicals at a high rate constant (Buettner 1987; Buettner 1993), we postulate that such protective function is specifically related to the attenuation of oxidative stress propagation initiated by mutant SOD1 gain of function. This effect of DMPO is clearly different from that of vitamin E, since the latter delayed disease onset without improving survival (Gurney et al. 1996). Thus, our in vivo and in vitro experiments suggest that spin trapping molecules can potentially be used to treat human ALS disease caused by mutant SOD1 since it readily crosses the blood brain barrier and displays minimal toxicity.
Oxidative stress-mediated mitochondrial degeneration and protein aggregation play major roles in the progressive motor neuron death by mutant SOD1 effects (Bogdanov et al. 1998; Bruijn et al. 1998; Liu et al. 1998; Johnston et al. 2000; Warita et al. 2001) (Fig. 9). Our molecular and pharmacological approaches aiming to increase mitochondrial antioxidative activity demonstrate that such approach is worthwhile of consideration as a viable strategy for the clinical management of human ALS disease.
We thank Dr Neil Cashman of University of Toronto for the NSC-34 cell line. This work was partially supported by EPSCORE/NSF, EPS-9874764.