dismutase (corrected or not corrected for citrate synthase), and catalase were ... Brain superoxide disrnutase, catalase, and glutathione peroxidase activities in.
Brain Superoxide Dismutase, Cadase, and Glutahone Peroxidase Activities in Amyotrophc Lateral Sclerosis Serge Przedborski, MD, PhD,* David Donaldson, BS,* Michael Jakowec, PhD,* Stephen J. Kish, PhD,$ M. Guttman, MD,$ Gorazd Rosoklija, M D , PhD,? and Arthur P. Hays, MDT
Amyotrophic lateral sclerosis is a fatal paralytic disorder of unknown cause. Recent evidence implicated the role of free radicals in the death of motor neurons in this disease. To investigate this hypothesis further, we measured the activity of the main free radical scavenging enzymes copperlzinc superoxide dismutase, manganese superoxide dismutase, catalase, and glutathione peroxidase in postmortem brain samples from 9 patients with sporadic amyotrophic lateral sclerosis and from 9 control subjects. We examined samples from the precentral gyrus of the cerebral cortex, a region affected in amyotrophic lateral sclerosis, and from the cerebellar cortex, a region not affected. The two groups did not differ in age or postmortem delay. In the precentral gyrus from amyotrophic lateral sclerosis samples, glutathione peroxidase activity as measured by spectrophotometric assay (13.8 & 2.6 nmol/min/mg protein [mean & standard error of mean]) was reduced significantly compared to the activity in the precentral gyrus from control samples (22.7 & 0.5 nmol/min/ mg protein). In contrast, glutathione peroxidase activity was not significantly altered in the cerebellar cortex from amyotrophic lateral sclerosis patients compared to controls. Copper/zinc superoxide dismutase, manganese superoxide dismutase (corrected or not corrected for citrate synthase), and catalase were not significantly altered in the precentral gyrus or cerebellar cortex in the patient samples. This study indicated that glutathione peroxidase activity is reduced in a brain region affected in amyotrophic lateral sclerosis, thus suggesting that free radicals may be implicated in the pathogenesis of the disease. Przedhorski S, Donaldson D, Jakowec M, Kish SJ, Guttman M, Rosoklija G, Hays AP. Brain superoxide disrnutase, catalase, and glutathione peroxidase activities in amyotrophic lateral sclerosis. Ann Neurol 1996;39:158- 165
Amyotrophic lateral sclerosis (ALS) is a progressive paralytic neurodegenerative disorder of unknown etiology characterized mainly by the loss of upper and lower motor neurons [l]. Mutations of the copper/zinc superoxide dismutase (Cu/Zn-SOD, EC 220.127.116.11) gene are found in approximately 20% of patients affected with familial ALS [2, 31. Most studies of these patients reported a reduction in erythrocyte superoxide dismurase (SOD) activity .Since SOD is a key enzyme in the protective mechanisms against free radical-induced toxicity [ 5 ] , it is suggested that oxidative stress may play a physiopathogenic role in familial ALS [4, 61. Sporadic ALS accounts for more than 90% of all ALS [ I ] . The clinical syndromes of familial and sporadic ALS are similar [l], and it is believed that they share a common mechanism of neuronal death [4, 61. Consistent with this view, clues support that oxidative stress may be implicated in the physiopathogenesis of both
familial and sporadic ALS [4, 61. Although the oxidative stress hypothesis is supported by identified alterations in SOD in familial ALS, the abnormality responsible for free radical-induced damage in sporadic ALS is unknown. Superoxide radicals, hydrogen peroxide (H2O2),and hydroxyl radicals are oxygen-centered reactive species  that have been implicated in several neurotoxic disorders [S-lo]. They are produced by many normal biochemical reactions, but their concentrations are kept in a harmless range by potent protective mechanisms [I I]. Increased free radical concentrations, resulting from either increased production or decreased detoxification, can cause oxidative damage to various cellular components, ultimately leading to cell death (111. In the present study, we hypothesized that the mechanism underlying oxidative stress in sporadic ALS is an alteration in one (or more) of the main free radical scaveng-
From the Department of "Neurology and tDivision of Neuropathology, Department of Pathology, Columbia University, New York, NY, and the $Human Brain Laboratory, Clark Institute of Psychiatry, Toronto, Ontario, Canada.
Received Jun 19, 1995, and in revised form Aug 11 and Sep 25. Accepted for publication Oct 13, 1995. Address correspondence to D r Przedborski, RR-307, Department of Neurology, College of Physicians and Surgeons, Columbia University, 650 West 168th Street, New York, NY 10032.
158 Copyright 0 1996 by the American Neurological Asmciation
as deprenyl, antispastics, trophic factors, or levodopa for at least 6 weeks prior to death. For each brain region, control and ALS samples were randomly matched and assayed simultaneously.
Table I . Clinical Dataa
No. of patients Age (yr) at the time of death Sex Postmortem delay (hr) Length of conservationb
9 56.7 ? 4.5
5 women, 4 men
3 women, 6 men 13.9 Z 1.2
11.9 ? 1.8 3.4 2 3.2 yr
61.1 ? 3.0
Chemicals and Equipment
2.5 ? 1.8 yr
“Values represent mean f standard deviation. Differences between amyotrophic lateral sclerosis (ALS) and control brains were tested by Student’s t test; none of the tests showed any significant difference
( p > 0.05).
Except for the bathocuprioinedisulfonic acid, which was from ICN (Costa Mesa, CA), all other compounds were purchased at highest-grade purity from Sigma (St. Louis, MO). All spectrophotometric assays were performed on Shimatzu U V 160 and Beckman D U 7400 instruments. Multiwell plate readings for the enzyme-linked immunosorbent assay (ELISA) were performed on a Bio-Rad 3550 computerized microplate reader (Microplate Manager for Windows, version 2.01, Bio-Rad, Richmond, CA).
Glutathione Peroxidase Activiq ing enzymes. To test this hypothesis, we measured in sporadic ALS and control postmortem brains the activities of Cu/Zn-SOD and manganese SOD (Mn-SOD), the two intracellular SOD isoenzymes present in eukaryotes [ 5 ] , as well as the activities of catalase (EC 1.1 1.1.6) and glutathione peroxidase (EC 18.104.22.168).
Materials and Methods Brain Collection and Dissection Nine brains from patients with ALS and 9 from control patients with nonneurological diseases were obtained from the Brain Bank, Department of Pathology at Columbia University (n = 7 per group) and from the Human Brain Laboratory, Clarke Institute of Psychiatry (n = 2 per group). In the ALS group, the cause of death was pneumonia (n = 6), respiratory failure (n = 2), or starvation (n = 1); in the control group, it was pulmonary embolism (n = 3), pneumonia (n = 2), acute pancreatitis (n = I), heart failure (n = I ) , acute myocardial infarction (n = l ) , or septicemia (n = 1). At autopsy, brains were removed and divided sagittally. One half of the brain was immediately frozen on dry ice and stored at -80°C until dissection. O n the day of the dissection, small frozen blocks of cerebellar cortex and of precentral gyrus were cut from each frozen half-brain. The other half-brain was placed in 10% buffered formalin and was subjected to neuropathological examination. There was no significant difference between the time from death to autopsy or between the time of autopsy to assay for the ALS and control groups (Table 1). The clinical diagnosis of ALS was confirmed pathologically in all 9 patients. In these brains, neuronal loss was observed in the precentral gyrus, while no remarkable pathological changes were noted in the cerebellar cortex. In all control brain, no remarkable pathological changes were noted either in frontal cerebral cortex or in cerebellum. For the ALS patients, the mean age (? standard deviation [SD]) of onset was 57.3 ? 2.9 years, with the mean duration of disease being 3.4 -+ 0.8 years and ranging from 2 to 5 years. None of the ALS patients had a family history for the illness. None of the ALS patients or control subjects were treated by chronic administration of drugs such
Glutathione peroxidase activity was determined according to the method described by Sinet and coauthors  based on NADPH oxidation followed at 340 nm at 37°C as described previously , with some modifications . O n the day of the assay, for each patient and for each region, approximately 250 mg of frozen brain tissue was homogenized by hand with a glass tissue grinder in 2.5 vol (wt/vol) of 10 mM Tris-hydrochloric acid (HCI) (pH 7.4) buffer containing 0.25 M sucrose. Homogenates were centrifuged at 15,000 g for 10 minutes at 4°C. Supernatants were mixed with an equal volume of a 4 mM potassium ferricyanide/20 mM potassium cyanide solution to prevent hemoglobin (Hb) interference with the assay [ 151. Then 100 pl of this solution was added to the reaction mixture (total volume of 3 ml) containing 1 m M reduced glutathione, 2 units of glutathione reductase, and 0.2 mM NADPH. After 10 minutes of preincubation, the absorbance was recorded for 4 minutes to determine the hydroperoxide-independent oxidation of NADPH. Thereafter, 1 mM t-butyl hydroperoxide was added to the mixture, and the rate of NADPH oxidation was monitored for 4 minutes. For a blank, the sample was replaced by an equal volume of buffer. Glutathione peroxidase activity was calculated as the overall rate of NADPH oxidation minus the rate of hydroperoxide-independent NADPH oxidation minus the blank value.
Plasma-Glutathione Peroxidase Protein Determination Quantification of plasma-glutathione peroxidase protein was performed using a commercial “sandwich” ELISA kit (Bioxytech, Bonneuil sur Marne, France). For each patient and each region, approximately 100 mg of frozen brain tissue was homogenized in 2.5 vol (wtivol) of Trisisucrose buffer and centrifuged as described above. Supernatants were collected and twofold serial dilutions were prepared using 50 mM Tris-HC1 (pH 7.8) buffer containing 1 mglml of bovine casein, 150 m M sodium chloride (NaCI), 0.1% Tween 20, and 0.2% sodium azide. From each dilution, an aliquot of 100 pl was incubated in a microwell that had been coated with rabbit anti-human plasma-glutathione peroxidase polyclonal antibody. As per the manufacturer’s protocol, each microwell was incubated sequentially with the same anti-
Przedborski et al: Free Radical Scavenging Enzymes in ALS
human plasma-glutathione peroxidase antibody conjugated to biotin, streptavidin conjugated to alkaline phosphatase, and para-nitrophenol-phosphate. Based on information supplied by the manufacturer, these antibodies were obtained using a synthetic plasma-glutathione peroxidase antigen and purified by affinity chromatography. The limit of detection was approximately 2.5 ng/ml. No crossreactivity was found with cellular-glutathione peroxidase from human erythrocyres. Microplates were read at 405 nm. Optical densities were converted to nanograms of plasma-glutathione peroxidase using a standard curve generated with purified human plasma-glutathione peroxidase protein (Kioxytech).
Supevoxide Dismutase Isoenzymes, Catalase, and Citrate Synthase Determination For SOD isoenzymes, catalase, and citrate synthase, approximately 250 mg of frozen brain samples was homogenized in 5 vol (wtlvol) of Tridsucrose buffer as described above. Homogenates were centrifuged at 2,000 g for 5 minutes at 4°C. Pellets were discarded and supernatants were centrifuged at 15,000 g for 10 minutes at 4°C. Supernatants of the second centrifugation were stored at - 80°C until assayed for cytosolic S O D and catalase activity. Pellets were gently resuspended and used for mitochondria1 purification using a Ficoll gradient as previously described . At the end of the purification procedure, mitochondria were subjected to three cycles of freezelthaw prior to centrifugation at 100,000 g for 1 hour at 4°C. Supernatants were stored at -80°C until they were assayed for mitochondria1 SOD and citrate synthase activity. Total and Mn-SOD activities were determined according to the method described by Spitz and Oherley  based on nitroblue tetrazolium (NBT) reduction by superoxide radicals followed at 560 nm at room temperature; xanthinexanthine oxidase was utilized to generate a superoxide radical flux. The reaction mixture (total volume, 1 id) contained 0.15 mM xanthine, 0.6 tnM NBT, 1 mM diethylenetriaminepentaacetic acid, 1 unit of catalase, 0.13 mg of defatted bovine serum albumin (BSA), 0.25 m M hathocuprioinedisulfonic acid, and 100 p1 of sample. After 3 minutes of preincubation, the reaction was initiated by the addition of xanthine oxidase (amount necessary to achieve a reference rate [Le., rate of NBT reduction in the absence of tissue] of 0.02 Aabsorbancelmin). One unit of SOD activity is defined as the amount of enzyme that inhibits the reaction rate by 50%. The Mn-SOD was distinguished from the cyanide-sensitive CulZn-SOD by the inclusion of 5 mM NaCN ; the samples were incubated i n the reaction mixture at room temperature for 30 to 60 minutes before the reaction was started, to ensure complete inhibition of the CulZn-SOD . C u l Zn-SOD activity was calculated by subtracting the cyanideresistant SOD activity from the total SOD activity. Catalase activity was determined according to the method decomposition foldescribed by Aebi  based on HIOz lowed at 240 nm at room temperature as previously described . Each sample was diluted to 1 : 500 in 50 mM phosphate buffer (pH 7.0) immediately prior to being assayed. The reaction mixture (total volume, 3 ml) contained 2 ml of the diluted sample and 60 mM H 2 0 2 ,which was added to initiate the reaction. The blank contained only sam-
160 Annals of Neurology
ple and buffer. Catalase activity was determined by calculating the rate constant of a first-order reaction (k) . Citrate synthase activity was assayed according to the method of Srere  based on formation of mercaptide ions followed at 412 nm at 30°C as previously described [IG]. The reaction mixture (total volume, 1 ml) contained 0.1 m M 5,5’-dithiohis-(2-nitrobenzoicacid), 0.3 mM acetyl coenzyme A, and 20 pl of sample. After 5 minutes of preincubation, 0.5 mM oxaloacetate was added to initiate the reaction. The increase in absorbance was recorded for 2 minutes.
Protein and Hemoglobin Determination Protein content was determined by the method of Lowry and colleagues , using BSA as the standdrd. Hb content was determined by the method of Drabkin and Austin , using cyanmethemoglobin as the standard.
Statistics All values are expressed as means +- standard error of the mean (SEM). Differences between the ALS and control groups were tested by two-tailed Student‘s t test or the Mann-Whitney rank sum test. For each analysis, data sets were subjected to normality test and equal variance test to determine whether Student’s t test or the Mann-Whitney rank sum test should be used. Pearson’s product correlation coefficient, and subsequent linear regression, was performed using a least-squares curve fit model. In all analyses, the null hypothesis was rejected at the 0.05 level. All statistical analyses were performed using SigmaStat for Windows (version 1.0, Jandel Corporation, San Rafael, CA).
Results In the present study, the calculated rate of liydroperoxide-dependent oxidation of NADPH was used as a specific measure of glutathione peroxidase activity and represented more than 85% of the total rate of NADPH oxidation. In addition, the nonspecific hydroperoxide-independent oxidation of NADPH was low and similar in the two groups in both the precentral gyrus (control = 2.73 2 0.66 nmol/min/nig protein, ALS = 2.88 F 0.36 nmollminlmg protein, t = -0.220, df = 16, p = 0.844) and the cerebellar cortex (control = 2.79 t 0.45 nmol/min/mg protein, ALS = 3.18 f 0.93 nmol/min/mg protein, t = -0.377, df = 16, p = 0.711). Glutathione peroxidase activity in the precentral gyrus was significantly lower in ALS compared to control brains (Fig 1A; t = -2.49, df = 16, p = 0.0240). In contrast, glutathione peroxidase activity in the cerebellar cortex was similar in ALS and control brains (see Fig 1A; t = 0.207, df = 16, p = 0.839). To account for individual variations, we normalized the activities of glutathione petoxidase in the precentral gyrus (the affected region) by the activities in the cerebellar cortex (the nonaffected control region). The mean precentral gyrus-cerebellar cortex ratio was 5196 smaller in ALS (0.59 ? 0.1 1) compared to control brains (1.20 f 0.19; t = -2.82, d f = 16, p = 0.0123). In the ALS group, precentral
gyrus glutathione peroxidase activities and precentral gyrus-cerebellar cortex ratios correlated positively with duration of the disease ( r 2 > 0.45, n = 9, p < 0.048) (Fig 2). Glutathione peroxidase activities in the cerebellar cortex did not correlate with the duration of the disease ( 2 = 0.16, n = 9, p = 0.295). Neither precentral gyrus glutathione peroxidase activities nor precentral gyrus-cerebellar cortex ratios correlated significantly with postmortem delay (v' < 0.10, n = 9, p > 0.25) or age at onset of the disease (Y' < 0.15, n
9, p > 0.20).
To determine whether the reduction in glutathione peroxidase activity corresponds to a decrease in the glu-
Fig 1. Glutathione peroxiduse (GPx) enzymatic activity (A) and plasma-glutathione peroxidase (PI-GPx) protein content (B) measured in sporadic amyotrophic lateral sclerosis (ALS) and control postmortem brain samples. The activigi of glutathione peroxidase was significantly reduced in the precentral g r u s (PCG) of sporadic ALS compared to control brains; in precentral gyrus, plasma-glutathione peroxidase protein content was similar in ALS and in control groups. In the cerebellar cortex (CC) none of the measurements was dzfferent in the ALS compared to the control group. Bars represent means and the error bars represent standmd error of mean f o r 9 samples per group and per brain region. The asterisk indicates p = 0.024 (Student? t test).
tathione peroxidase isoenzyme, plasma-glutathione peroxidase, we used an immunoassay to quantify the amount of plasma-glutathione peroxidase protein in the precentral gyrus and cerebellar cortex in both ALS and control brains. We found detectable amounts of plasma-glutathione peroxidase in all brain samples (see Fig 1B). In addition, we found that plasma-glutathione peroxidase protein content was similar in the precentral gyrus and cerebellar cortex, and no significant difference in plasma-glutathione peroxidase protein content between ALS and control groups in either brain region (see Fig 1B). Inclusion of 5 mM NaCN with the cytosolic fraction from the precentral gyrus and cerebellar cortex caused reduction in SOD activity greater than 75%, indicating that most SOD activity in this fraction corresponds to Cu/Zn-SOD. We found no significant difference in Cu/Zn-SOD activity in either brain region between ALS and control groups (Table 2). The tissue samples used were frozen; therefore, the integrity of peroxisomes was not preserved and all of the catalase activity was found in the cytosolic fraction. Catalase activity did not differ between ALS and control brains (see Table 2). Unlike previous experiments , we measured Mn-SOD in purified mitochondria, the organelle that contains most of this SOD isoenzyme [ 5 ] . As indicated by the activity in citrate synthase, mitochondrial preparations from ALS and control brains were comparable (see Table 2). As expected, more than 90% of SOD activity measured in the mitochondrial fraction was resistant to 5 mM NaCN, indicating that most SOD activity in this fraction corresponds to MnSOD. No significant difference was observed in MnSOD activity in mitochondrial fractions between ALS and control brains (see Table 2). Likewise, Mn-SOD activity normalized by the mitochondrial enzyme citrate synthase was not different between ALS and control brains (data not shown). Since all samples used for these assays contained some blood, we measured Hb content in tissue homogenates prior to centrifugation. Crude homogenates contained comparable amounts of Hb in the precentral gyrus (ALS = 4.05 t 0.75 mg/ml, control = 3.94 &
Przedborski et al: Free Radical Scavenging Enzymes in ALS
Discussion It has been demonstrated that the carbonyl content, a marker of oxidative damage in proteins, is increased by 85% in the frontal cortex of sporadic ALS postmortem
tification of the mechanism by which this oxidative damage occurs in sporadic ALS will provide important insight into the cause of the disease. In the present study, we found that the activity of glutathione peroxidase, a key enzyme in the protective mechanism against free radicals, is significantly lower in ALS compared to control precentral gyrus (see Fig 1). A similar change was not observed in the cerebellum (see Fig l), suggesting that reductions in glutathione peroxidase activity have a specific association with the pathological processes in sporadic ALS. We also found that glutathione peroxidase activity in the precentral gyrus, but not in the cerebellar cortex, positively correlates with the disease duration (see Fig
._. Fig 2. Scatterplot and linear regression for the disease duration and the activity o f glutathione pesoxidase (GPx) measured in the precentral gyrus (PCG) (A) and the ratio of glutathione pesoxidase activiq measured in the precentval gyms to glutathione peroxidase activity measured in the cerebellar cortex (CC) (B). The solid lines show the best leastsquares fit (in A: Y = -7.16 f 6.07X, r' = 0.45,p = 0.048; in B: Y = -0.256 f 0.246X,r 2 = 0.57, p = 0.018), and die dotted lines represent the 9 5 % conjidence interval.
162 Annals of Neurology Vol 39
and control brains. Thus, approximately 4 mg of Hb is found per milliliter of homogenate, which in turn contains 285 mg of wet-weight brain tissue. Since normal blood contains between 120 to 160 mg of H b l ml, we have 25 to 33 p1 of blood in 285 mg of brain tissue or 0.09 to 0.12 1.11 of blood per milligram of brain tissue. Based on this calculation we conclude that blood contamination of the brain samples accounted for less than 12% (vol/wt).
following disease of the longest duration. Our review of the pathological data revealed that the degree of spinal and cortical neurodegeneration was comparable among these patients; this positive correlation (see Fig 2) suggests that alterations in brain-glutathione peroxidase activity can influence the rate of disease progression. Our study revealed reductions in glutathione peroxidase activity in the precentral gyrus from 9 ALS patients. In another affected region, the spinal cord, Ince and coworkers found a marked increase in activity in 10 ALS patients. The reason for the discrepancy in the postmortem precentral gyrus and spinal cord in ALS is unclear. Technical differences are, however, unlikely as the methods used in the two studies were fairly similar. In our study, we measured free radical scavenging enzyme activities in only a single affected region
Table 2. Superoxide Dismutase (SOD) Isoenzymes, Catalase, and Citrate Synthase Activity in Brain Extracts fFom Amyotrophic Lateral Sclerosis (ALS) Patients and Control Subjects"
Cyrosolic Fraction Catalase (klmg protein)
Mn-SOD (unitslmg protein)
Citrate Synthase (nmollminlmg protein)
96.4 2 19.7 98.0 t 22.8
0.185 t 0.057 0.194 t 0.022
92.5 t 20.5 127.1 t 22.5
1.71 i- 0.46 1.88 i- 0.25
110.2 t 22.5 95.8 2 17.1
0.163 t 0.047 0.167 ? 0.038
117.9 2 14.0 137.7 t 28.7
2.12 i- 0.46 2.01 f 0.27
CulZn-SOD (unitslmg protein)
Precentral gyrus Controls ALS Cerebellar cortex Controls
'Values represent mean i standard error of mean of 9 samples per patient group and per brain region. Differences between ALS and control brains were tested by Student's t test or the Mann-Whitney rank sum test; none of the tests showed any significant difference ( p > 0.05).
= rate constant of the first-order reaction (for catalase activity) .
of the brain. Thus, it would be important in linking glutathione peroxidase alterations to the pathogenesis of ALS to extend our work to other affected and unaffected regions. This future endeavor could also be important in understanding the reasons for the opposite changes in glutathione peroxidase activity found in ALS postmortem samples by us and by Ince and collaborators 1261. Glutathione peroxidase purified from human plasma is a glycosylated selenoprotein distinct from cellularglutathione peroxidase . Although the two isoenzymes cannot be distinguished by the enzymatic assay, they can be differentiated by immunoassay [27, 281. Plasma-glutathione peroxidase has been identified in several tissues  and cell lines , but its precise tissue distribution remains incomplete. Therefore, the apparent reduction in glutathione peroxidase activity can be due to changes in either cellular- or plasmaglutathione peroxidase, or both. T o address this issue, we quantified the amount of plasma-glutathione peroxidase in the ALS and control brain samples using an immunoassay specific for this isoenzyme. We found detectable amounts of plasma-glutathione peroxidase in all samples. This is the first demonstration of measurable amounts of plasma-glutathione peroxidase in the human brain. In an earlier study , no plasmaglutathione peroxidase was detected in the rat brain. Whether this discrepancy represents interspecies variations or differences in the technique or the antibodies used remains to be determined. We found that neither in the precentral gyrus nor in the cerebellar cortex did the content of plasma-glutathione peroxidase differ between the ALS and control samples (see Fig 1). This suggests that in the precentral gyrus of ALS patients the reduction in glutathione peroxidase is related to a deficit in the cellular-glutathione peroxidase. We also measured the activities of other major free radical scavenging enzymes including CuiZn-SOD, Mn-SOD, and catalase. Consistent with the study by
Bowling and coworkers , we found that brain Cu/ Zn-SOD activity in sporadic ALS was comparable to that in controls. In addition, we found that brain MnSOD activity as well as brain catalase activity were not significantly changed in sporadic ALS. Free radical scavenging enzymes are present in high concentrations in the blood [ 111; hence, contamination of tissue samples with blood can lead to spurious determination of brain enzyme activities. We measured blood-glutathione peroxidase (9.59 -+ 0.49 pmol/min/ gm Hb), S O D (28,920 2 1,218 units/gm Hb), and catalase (309.8 ? 15.3 k/gm Hb) activities in 24 normal volunteers; these values did not differ from those measured in 17 age-matched sporadic ALS patients . Based on these values and on normal blood Hb concentrations, we estimated that glutathione peroxidase activity is 1.34 t 0.07 pmollminigm brain tissue. Since the calculated blood contamination of brain samples was 12%, brain-glutathione peroxidase activity originating from blood is 0.161 pmolfminlml blood (1.34 ? 0.07 pmol/min/ml blood X 0.12). Our brain value was 22.7 nmollminlmg protein in control brains and 13.8 nmol/min/mg protein in ALS brains. Since we had 0.08 to 0.10 gm of protein per gram of brain tissue, we calculated values of 1.1 to 1.35 pmol/min/ gm brain tissue for ALS and 1.8 to 2.2 Fmoliminigm brain tissue for controls. Therefore, glutathione peroxidase originating from blood represented 12 to 15% of measured glutathione peroxidase activity in ALS and 7 to 9% in control brains. Using the same calculation, we found that the contribution of blood-derived free radical scavenging enzyme activities may account for less than 5% in SOD (4,049 2 171 unitdm1 blood) and less than 28% in catalase (43.4 ? 2.14 k/ml blood). This suggests that free radical scavenging enzyme activities measured in the present work originated predominantly from brain tissue. Based on published values of plasma-glutathione peroxidase concentration in normal human volunteers (i.e., 3.3 2 0.1 pgirnl
Przedborski et al: Free Radical Scavenging Enzymes in ALS
serum) , we estimated that the contribution of blood-derived plasma-glutathione peroxidase may account for less than 10% of its brain measurement. The main function of glutathione peroxidase is to detoxify H20,.Although the reactivity of H 2 0 Lis limited and its direct implication in toxic processes is doubtful, it has been recognized as an important intermediate compound in reactions producing highly reactive damaging species [ 1 I]. Glutathione peroxidase also catalyzes the reduction of lipid peroxides , preventing lipid peroxidation and thus protecting biological membranes. Several lines of evidence indicate that a reduction in glutathione peroxidase activity can be damaging [l 1, 321. For example, a 21 ?h inhibition in glutathione peroxidase activity in cultured human fibroblasts decreased cell number by 50% .This in vitro model supports the hypothesis that moderate changes in glutathione peroxidase activity may be noxious. Therefore, the 40% reduction in glutathione peroxidase activity in the ALS precentral gyrus is potentially damaging, and may be implicated in the pathogenesis of sporadic ALS. Alternatively, the reduction in glutathione peroxidase activity in ALS might be secondary to the neurodegenerative process, namely the loss of motor neurons. However, in the human brain, glutathione peroxidase is mainly if not exclusively localized in glial cells , which are not lost in ALS. In conjunction with this, there was n o change in the activity of the other free radical scavenging enzymes, which would be expected if the reduction of glutathione peroxidase was only consecutive to motor neuron loss. O n the other hand, many factors have been recognized to modulate glutathione peroxidase activity in vivo . Thus, if one of these factors is altered in ALS, a deficit in glutathione peroxidase activity may result. For instance, a deficiency in selenium can cause a dramatic reduction in glutathione peroxidase activity 1351. However, no study has documented a deficit in selenium either in the brain or in the spinal cord of ALS patients . Glutathione peroxidase and selenium deficiency has been suggested to be involved in the pathogenesis of different muscular and neurological disorders . In addition, some evidence exists for a small (1 !)(yo) reduction in glutathione peroxidase activity in the substantia nigra of patients with Parkinson’s disease . Moreover, selenium deficiency commorily affects skeletal muscles in various species including humans  and some studies associated a decrease in glutathione peroxidase activity with muscle cramps and weakness . In conclusion, we speculate that the observed deficit in glutathione peroxidase activity is implicated in the pathogenesis of sporadic ALS. Furthermore, we believe that this deficit could be involved in the oxidative damage observed in sporadic ALS .Although our study needs to be extended to a greater number of postmor-
Annals of Neurology
No 2 February 1996
tem samples and a greater number of regions of the nervous system, it provides strong support to the hypothesis that free radicals are implicated in the neurodegenerative process in sporadic ALS.
This work was supported by The Muscular Dystrophy Association, the National Institute of Neurological Disorders and Stroke grant 1K08-NSO 1724-0 1, the Lowenstein Foundation, and the Parkinson’s Disease Foundation. Dr Przedborski is recipient of the Colonel Berger Junior Investigator Award and the Irving A. Hansen Memorial Foundation Award.
We are grateful to Dr J. Goldman for providing the brain samples from the Columbia Brain Bank, Dr Yan Liu for her help with brain sample dissection, Drs N. Latov and S. A. Sadiq for their advice on ELISA experiments, and Drs I. Tayarani and A. Lemainque from Bioxytech for their advice on the plasma-glurathione peroxidase ELISA. We are also grateful to Drs L. 1’. Rowland and R. H. Brown, Jr., for their constant support and to Drs Timothy Lynch and Daniel Togasaki for their insightful comments on the manuscript.
References 1. Rowland LP. Hereditary and acquired motor neuron diseases. In: Rowland LP, ed. Merritt’s textbook of neurology. 9th ed. Philadelphia: Williams & Wilkins, 1995:742-749 2. Rosen DR, Siddique T, Patterson D, et al. Mutations in Cu/ superoxide dismutase gene are associated with familial amyorrophic lateral sclerosis. Nature 1993;362:59-62 3. Deng H-X, Hentati A, Tainer JA, et al. Amyotrophic lateral sclerosis and structural defects in Cu,& superoxide dismutase. Science 1993;261:1047-105 1 4. Rowland LP. Amyotrophic lateral sclerosis: human challenge for neuroscience. Proc Natl Acad Sci USA 1995;92:1251-1253 5. Bannister JV, Bannister W, Rotilio G. Aspects of the structure, function, and applications of superoxide dismurase. Crit Rev Biochem 1987;22:111-179 6. Smith RG, Appd SH. Molecular approaches to amyotrophic lateral scierosis. Annu Rev Med 1995;46:133-145 7. Freeman BA, Cr;ipo JD. Hiology of disease. Free radicals and tissue injury. Lab Invest 1982;47:412-426 8. Cohen G, Werner P. Free radicals, oxidative stress, and neurodegeneration. In: C a h e DB, ed. Neurodegeiierative diseases. Philadelphia: WB Saunders, i 994: 139- 16 1 9. Fahn S, Cohen C,. The oxidant stress hypothesis in Parkinson’s disease: evidence supporting it. Ann Neurol 1992;32:804-8 12 10. Coyle IT, Puttfdrcken I). Oxidative stress, glutamate, and neurodegenerative disorders. Science 1993;262:689-694 11. Halliwell B, Gutteridge JM. Free radicals in biology and medicine. 2nd ed. Oxford: Clarendon Press, 1991 12. Sinet PM, Michelson AM, Bazin A, et al. Increase in glutathione peroxidase activity in eryrhrocytes from trisomy 21 subjects. Biochem Biophys RKSCommun 1975;67:910-915 13. Przedborski S, Jackson-Lewis V, Kostic V, et al. Superoxide dismutase, catalase, and glutathione peroxidase activities in copper/zinc-superoxide dismutase transgenic mice. 1. Neurochem 1992;58: 1760- 1767 14. Flohe I,, Gunzler WA. Assays of glurathione peroxidase. Methods Enzymol 1984;105:114-121 15. Gunzler WA, Kremers H, Flohe L. An improved coupled test procedure for glutathione peroxidase (EC 1.1 I .1.9) in blood. Z Klin Chem Klin Biochem 1974;12:444-448 16. Przedborski S, lackson Lewis V, Muthane U, et al. Chronic
levodopa administration alters cerebral mitochondrial respiratory chain activity. Ann Neurol 1993;34:715-723 Spitz DR, Oberley LW. An assay for superoxide dismutase acrivity in mammalian tissue homogenates. Anal Biochem 1989; 1798-18 Ledig M, Fried R, Ziessel M, er al. Regional distribution of superoxide dismutase in rat brain during postnatal development. Dev Brain Res 1982;4:333-337 Bize IB, Oberley LW, Morris HI’. Superoxide dismurase and superoxide radical in Morris hepatomas. Cancer Res 1980;40: 3686-3693 Aebi H. Catalase. In: Bergmeyer HU, Bergmeyer J, Grab1 M, eds. Methods of enzymatic analysis. 3rd ed. FL: Chemie, 1983: 273-296 Srere PA. Citrate synthase. Merhods Enzyrnol 1969;13:3-11 Lowry OH, Rosenbrough NJ, Farr AL, et al. Protein measurement with the folin phenol reagent. J Biol Chem 1951;193: 265-275 Drabkin DL, Ausrin JH. Spectrophotomerric studies. 11. Preparations from washed blood cells; nitric oxide hemoglobin and sulfhemoglobin. J Biol Chem 1935;112:51-65 Bowling AC, Schulz JB, Brown RH Jr, et al. Superoxide dismutase activity, oxidative damage, and mitochondrial energy metabolism in familial and sporadic amyotrophic lateral sclerosis. J Neurochem 1993;61:2322-2325 Tandan R, Bradley WG. Amyotrophic lateral sclerosis: part 2. Eriopathogenesis. Ann Neurol 1985;18:419-431 Ince PG, Shaw PJ, Candy JM, et al. Iron, selenium and glutarhione peroxidase activity are elevated in sporadic motor neuron disease. Neurosci Lett 1994;182:87-90 Takahashi K, Avissar N, Whitin JC, et al. Purification and characrerizarion of human plasma glutathione peroxidase: a
selenoglycoprorein disrinct from rhe known cellular enzyme. Arch Biochem Biophys 1987;256:677-686 Huang W, Akesson B. Radioimmunoassay of glutathione peroxidase in human serum. Clin C h i n Acta 1993;219:139148 Yoshimura S , Watanabe K, Suemizu H , et al. Tissue specific expression of the plasma glutathione peroxidase gene in rat kidney. J Biochem 1991;109:918-923 Avissar N, Whitin JC, Allen PZ, et al. Plasma selenium-dependent glutathione peroxidase. Cell of origin and secretion. J Biol Chem 1989;264:15850-15855 Przedborski S, Donaldson DM, Murphy PL, et al. Blood stiperoxide dismutase, catalase, and glutathione peroxidase activiries in familial and sporadic amyotrophic lateral sclerosis. Neurodegen 1996 (in press) Michiels C, Raes M, Toussaint 0 , et al. Importance of Seglutathione peroxidase, caralase, and Cu/Zn-SOD for cell survival against oxidarive stress. Free Radic Biol Med 1994;17: 235-248 Michiels C, Remacle J. Use of the inhibition of enzymatic antioxidant systems in order to evaluate their physiological importance. Eur J Biochem 1988;177:435-441 Damier P, Hirsch EC, Zhang P, et al. Clutarhione peroxidase, glial cells and Parkinson’s disease. Neuroscience 1993;52:1-6 Flohe L. The selenoprotein glurathione peroxidase. In: Dolphin D, Poulson R, Avramovic 0, eds. Glutathione. New York: Wiley, 1989:643-731 Kasarkis EJ. Neurotoxicology: heavy metals. In: Smith RA, ed. Handbook of amyotrophic lateral sclerosis. New York: Marcel Dekker, 1992:559-574 Kish SJ, Moriro CL, Hornykiewicz 0. Glutathione peroxidase activity in Parkinson’s disease. Neurosci Lett 1985;58:343-346
Przedborski et al: Free Radical Scavenging Enzymes in ALS