Free Radical Damage to Protein and DNA: Mechanisms Involved and Relevant Observations on Brain Undergoing &dative Stress Robert A. Floyd, PhD," and John M. Carney, PhDf
Iron mediates damage to proteins and DNA. The mechanisms of damage not only involve iron but also oxygen free radical intermediates. Oxidative damage to DNA causes not only strand breaks, but also formation of specific base adducts, such as 8-hydroxy-2'-deoxyguanosine.Oxidative damage also inactivates certain enzymes such as glutamine synthetase. Novel methods of assessing oxidative damage to tissue, including quantitation of salicylate hydroxylation as an index of hydroxyl free radical flux as well as specific lesions to proteins and DNA, have yielded results that clearly show that ischemialreperfusion injury to mongolian gerbil brain involves oxidatively damaging events. Aging in gerbil as well as human brain is also associated with increased oxidative damage. Recent novel observations have shown that the spin-trapping agent phenyl a-tert-butylnitrone (PBN)offers protection in gerbil brain during ischemia/ reperfusion injury. We also show that oxidative damage to brain during aging is decreased by chronic administration of PBN. The mechanism of action of PBN may be related to its trapping of specific free radicals, which triggers a cascade of oxidative events that eventually lead to tissue injury. Floyd RA, Carney JM. Free radicaI damage to protein and DNA: mechanisms involved and relevant observations o n brain undergoing oxidative stress. Ann Neurol 1992;32:S22-S27
All aerobic organisms experience an imposed oxidative stress constantly C 11. This statement is becoming virtually axiomatic as knowledge in this area continues to accumulate. The mere fact that oxygen is consumed by an organism results in an imposed oxidative damage potential (PJ. Oxidative damage potential is caused by a small fraction of the total oxygen consumed forming activated oxygen by-products, which, unless quenched, cause oxidative damage to biological molecules. Oxidative damage causes lesions to many biological molecules, including lipids, proteins, and nucleic acids. The activated oxygen species involved include lipid hydroperoxides, hydrogen peroxide, superoxide, hydroxyl free radicals, and singlet oxygen. The antioxidant defense capacity (A,) of the system acts protectively to oppose the oxidative damage potential. The antioxidant defense capacity is composed of the enzymes superoxide dismutase (SOD), catalase, glutathione peroxidase (GSHPx), as well as vitamin E and glutathione. Natural systems appear to operate in a state such that most but not all of the oxidative damage potential imposed is quenched by the antioxidant de-
From the "Molecular Toxicology Research Program, Oklahoma Medical Research Foundation, and the Department of Biochemistry and Molecular Biology, University of Oklahoma Health Sciences Center, Oklahoma City, OK; and the ?Department of Pharmacology, University of Kentucky Medical School, Lexington, KY.
s22
fense capacity (i.e., a small amount of oxidative damage occurs at all times due to the net flux of damaging activated oxygen species, Po', leaking through the antioxidant defense capacity). The oxidative damage potential and the antioxidant defense capacity are in a dynamic equilibrium at all times (Fig 1). Oxidative damage to biological molecules involves the combined action of oxygen free radicals and the trace elements Fe, Cu, or both. Oxidatively damaging events occur in Fenton-type reactions where Fe acts catalytically. Thus, if a very small amount of available Fe is present, then large amounts of oxidative damage may occur if the appropriate flux of oxygen free radicals is present in the immediate vicinity and can react with the available Fe, Cu, or both. The site specificity observed in oxidative damage to DNA and proteins is due to the site-specific iron-binding properties of these macromolecules. Thus, damage to specific amino acids in proteins is dictated by their proximity to the binding site of Fe or their location proximal to the active site of an enzyme that contains Fe or Cu essential for activity. There is compelling evidence that oxidative damage
Address correspondence to Dr Floyd, Department of Biochemistry and Molecular Biology, University of Oklahoma Health Sciences Center, Oklahoma City, OK 7 3 104-5046,
~~~
~
Lipid Hydroperoxides
'9,OH /&
SOD, CATAUSE
GSHPx GSH, Vitamin E
Po-Ac
=
PA
Fig 1. Dynamic equilibrium between oxidative damage potential (Po) and the antioxidant defense capacity (Ad. A small amount of oxidatively damaging species (Po')leaks throughout the defense, thus causing oxidative damage. SOD = superoxide dismutare; GSHPx = glutathione peroxidase; GSH = reduced glutathione.
to brain proteins does occur during brain injury caused by an ischemia/reperfusion insult (IRI) as well as during aging [2, 31. In fact, loss of specific enzymes due to oxidative damage may be a major factor causing brain injury [2}.These areas are reviewed using mostly our own data, which have been collected in collaboration with other laboratories. With regard to oxidative damage to DNA and how this contributes to neurological damage, this area of research is at an early stage and therefore is addressed only from a summary standpoint arrived at from observations conducted on in vitro model systems.
Oxidative Damage to DNA The combination of oxygen free radicals and the trace metals Fe or Cu causes damaging lesions to nucleic acids. The predominant lesion observed is strand breakage, but DNA base modification products are also formed. Many studies have been conducted utilizing supercoiled DNA suspended in phosphate buffer and exposed to ascorbate. Phosphate buffer contains trace amounts of Fe, and thus when exposed to ascorbate, H202 and hydroxyl free radicals (*OH)are formed. Utilizing this system, it can be clearly demonstrated that DNA strand breakage is mediated by Fe and active oxygen intermediates [4}. We demonstrated that one of the essential active oxygen intermediates formed in the Felascorbate system is H202[4, 51. The fact that catalase prevents supercoiled DNA strand breakage supports this notion 14, 5 1. In addition, it has been shown on the basis of inhibitor studies as well as salicylate trapping studies that .OH is produced in the ascorbateiFe system 141. Thus, it is likely that autooxidation of ascorbate produces H202,which then reacts with ferrous iron, which is bound to the DNA strand.
Ascorbate mediates the reduction of ferric ion to ferrous iron. The reaction of H20, with ferrous iron yields .OH, which oxidatively lesions the ribose moiety, thus causing DNA strand breakage. It has been demonstrated that oxygen is an absolute requirement for the ascorbate/Fe system to cause DNA breaks [4}. We also found that the reactions proceed faster as the temperature increases within the 4" to 37°C range IS}. Oxidative damage to DNA mediated by the trace metals Fe or Cu produces mutagenic lesions [G, 71, probably due to formation of oxidatively modified bases such as thymine glycol, 5-hydroxymethyluracil, and 8-hydroxyguanine. The amount of 8-hydroxyguanine in DNA can be determined very sensitively using high-performance liquid chromatographic (HPLC)electrochemical detection [8},which allows assessment of the amount of this modified base in as little as 10 Fg DNA. This sensitivity is very useful when using very small amounts of tissue. The amount of oxidized bases present in DNA is not just a reflection of oxidative damage to the macromolecule per se, but represents a steady-state level of the repair processes that occur.
Oxidative Damage to Proteins Proteins are oxidatively damaged by the combined action of activated oxygen species and the trace metals Fe, Cu, or both [9}. The amino acids lysine, proline, histidine, and arginine have been found to be the most sensitive to oxidative damage [ 101. Oxidative damage to these amino acids yields an increased level of carbony1 groups on proteins. Methods have been developed to assess the carbonyl content of proteins as an estimate of the amount of oxidative damage [ll}. Thus, as oxidative damage to tissue occurs, there is an increased amount of oxidized protein. Oxidized proteins are degraded by a novel protease, a neutral protease, discovered by Rwett [ 12). The relationship between the formation of oxidized protein by Fe and H202and its degradation by neutral protease is shown in Figure 2. Oxidative damage causes rapid loss of enzymatic activity in many proteins [lo}. Glutamine synthetase (GS) is rapidly inactivated by oxidative damage [lo}. The histidine located near the Fe at the active site of GS is lesioned by oxidative damage, thus causing loss in activity [lo}.
High Vulnerability of Brain to Oxidative Damage There are several factors that predispose brain tissue to be highly susceptible to oxidative damage: (1) brain is highly enriched in the easily peroxidizable 22 :6 and 20:4 unsaturated fatty acids; (2) certain regions of brain, particularly human brain, are highly enriched in iron; (3) brain is not particularly well endowed with
Floyd and Carney: Free Radical Damage to Protein and D N A S23
c!7
p Natural Protein
F)
_I%o,,
JR
-
OH
6 1+Hob OH
I
-
Neutral
&
Amino Acids
Protease
I
OH
I
Oxidized Protein
Fig 2. Relationship showing that Fe bound to protein will react with H202 to cause oxidatively damaged protein, resulting in increased carbonyl groups on the protein. The oxidized protein is degraded to individual amino acids by a neutral protease.
antioxidant defense systems; ( 4 ) brain utilizes proportionately a large amount of oxygen in relation to its weight; (5) brain is a highly organized information processing center; and (6) neurons are postmitotic, thus death of a neuron represents a permanent lesion. Research work with in vitro incubation systems using rat brain homogenate 1131 has demonstrated the following: ( 1 ) brain will readily peroxidize if it is organizationally disrupted; (2) it peroxidizes rapidly in phosphate buffer if the incubation temperature is 37"C, but not at 4°C.; (3) the peroxidation rate of each brain region is proportional to the total iron content of the brain region, with the exception of striatum; where ( 4 ) the increased dopamine content acts to inhibit peroxidation, as can be demonstrated by the addition of dopamine to brain homogenate; and ( 5 ) chelators of iron inhibit brain peroxidation, but active oxygen scavengers are not very effective { 131. The consequences of oxidative damage to brain in terms of carrying out essential neurochemical activities, such as dopamine synthesis, are illustrated by the studies of Zaleska and colleagues 1141. We demonstrated that dopamine synthesis from tyrosine in striatal synaptosomes was dramatically curtailed by the addition of very small levels of adenosine diphosphate (ADP)-chelated Fe in the presence of ascorbate. Peroxidation of the synaptosomes was induced by the ADP-Fe/ascorbate treatment 114). Peroxidation of synaptosomes and the decrease in dopamine synthesis was inversely linked, showing a dramatic decrease in dopamine synthesis, with one-half the activity lost with 2.5 pmol/L Fe {14). Determination of Oxygen Free Radical Flux I n Vivo It is a difficult task to determine the amount of oxygen free radicals present in brain. The difficulties arise from the fact that oxygen free radicals react rapidly (i.e., as soon as they are formed); thus, the amount present at any specific time is very small, perhaps lo-" mol/L or less. Thus, it has been necessary to use indirect methods to assess oxygen free radical flux in vivo. Two general methods have been utilized: (1) determination S24 Annals of Neurology
0
CH,
0ec-LC " ,+ I
R
- Of-F'-+.
CH,
Spin-Trap
Free Radical
H
0 CH,
R
CH,
Spin- Adduct
la-phenyl-tert-butylNitrone1
Fig 3. Reactions showing that salicylate reacts with bydroxyl free radicals (OH) t o form the 2,3- and 2,5-dibydroxybenzoic acid products and the spin-trap phenyl a-tert-butylnitrone re'acts with a free radical to form a spin-adduct.
of the amount of unique products formed when oxygen free radicals react with biological molecules, and (2) addition of exogenous trapping compounds that react with oxygen free radicals to form unique and relatively stable products, which can be quantitated. Measurement of oxygen free radical-mediated products of DNA bases and proteins are examples of the first general method, whereas use of spin-traps to react with free radicals and salicylate to trap hydroxyl free radicals are examples of the second general method. Figure 3 illustrates the reactions involved with two exogenous traps. The spin-trap, PBN (phenyl a-tevt-butylnitrone), reacts with free radicals to form spin adducts. Salicylate reacts with hydroxyl free radicals at a diffusion-limited rate to produce the 2,3- and 2,5-dihydroxybenzoic acid (DHBA) products, which can be quantitated very sensitively using HPLC-electrochemical detection { 151. We have used the salicylate hydroxylation method to estimate hydroxyl free radical flux in several biological systems, including the IRI-lesioned gerbil brain [ 161. Oxidative Damage i n the Ischemia/Reperfused Gerbil Brain The Mongolian gerbil is an ideal animal model to study ischemia/reperfusion in jury in brain because in almost all gerbils, the forebrain is entirely perfused by the two common carotids. This anatomical peculiarity allows examination of the effect of cessation of blood flow and its subsequent reintroduction into the forebrain by ligation and then release of the common carotids { 161. Utilizing this model, it is also possible to compare other regions of the brain, especially the cerebellum and the brainstem, and use these regions as internal controls to ascertain if the effect is region-specific. The cerebellum and the brainstem are not perfused by the carotid arteries and thus do not suffer oxygen deprivation when the carotids are ligated. Earlier concerns that gerbils are prone to seizures have been minimized by use of seizure-resistant lines.
Supplement to Volume 32, 1992
Figure 4 summarizes pertinent data from a series of several different studies examining oxidative damage in gerbil brain undergoing an IRI. Several methods of assessing oxidative damage in gerbil brain given a 10-minute period of ischemia followed by 60 minutes of reperfusion are compared for cortex and brainstem. Salicylate hydroxylation in cortex is significantly increased by an IRI, but not in brainstem [17}. Protein oxidation as assessed by carbonyl content of protein is markedly increased in cortex, but not in the brainstem of the same animals 131, whereas GS is markedly decreased in cortex of IRI-lesioned gerbils, but not in the brainstem of the same animals 131. The remarkable correlation of all three parameters of oxidative damage in the two separate regions of brain convincingly underscore the notion that oxygen deprivation of brain and its subsequent reintroduction leads to a series of oxidative events that lead to much damage, including oxidative lesioning of proteins and specific enzymes.
Cortex
Brain Stem
175
-.2
IRI
150 125
P
E" loo . -5
Control IRI
75
350
8
25
0 14-7I
PBN Protects Brain from IRI-induced Damage We have made observations that clearly indicate the spin-trapping compound, PBN, protects gerbils from IRI-induced brain damage. A combined summary of the results of several studies 11-31 are shown in Figure 5. The data clearly show that administration of PBN prior to a brain IRI helped to prevent (1) loss in GS activity, ( 2 ) increase in oxidative damage to brain protein, and ( 3 ) lethality brought on by a very damaging IRI. There is an age effect in the severity of the damage brought on by brain ischemia as manifested by lethality, assessed at 7 days. In young gerbils ( 3 to 4 months old) 15 minutes of ischemia caused 50% lethality; in older gerbils (18 months old), 10 minutes of ischemia caused death in all animals. PBN-mediated Reversal of Age-associated Oxidative Damage There is an increased (nearly two-fold) amount of oxidized proteins in the brains of normal older gerbils (18 months old) as compared with normal young gerbils (3-4 months old). In addition, the activity of neutral protease in older gerbil brain is only 33% of that in younger gerbil brain, and the activity of GS in older gerbils is only 65% of that observed in younger gerbils {27. Chronic intraperitoneal administration of PBN (32 mg/kg twice a day) for 14 days caused the amount of oxidized proteins to decrease in the brains of older gerbils nearly to that observed in younger gerbils 12). In addition, PBN administration to the older gerbils also caused an increase in GS and neutral protease activity nearly to that observed in the younger animals 121. A summary of these data 1181 is presented in Figure 6. In addition to the changes noted in protein oxidation and enzyme activity, we also found that the in-
Fig 4. Data from our recent work summarizing salicylate hydroxykztion 117) and protein oxidation, as well as loss in glutamine synthetase (GS) activity {3) in gerbils with ischemialrepe$usion insult and comparable sham-operated control gerbils. The cerebral cortex region is compared to the brainstem. The gerbils were given a 10-minute ischemic insult, and a 60-minute repetfusion insult.
Floyd and Carney: Free Radical Damage to Protein and DNA
S25
300-
200-
5
'6
s
IRI
100-
OJ
IRI + PBN
NORMALS
PBN TREATED
OFF PBN
Fig 6. Data summarizing the protective effort of phenyl a-tertbutylnitrone (PBN) administered chronically (14 days. twice daily at 32 mglkg, intraperitoneally) to older gerbils {2}. Data show that PBN administration caused a decrease in oxidized protein and concomitant increases in glutamine synthetase (GS) and neutral protease activity. In addition, errors in short-term spatial memory, assessed with a radial a m maze test, was decreased by PBN administration. (From {I 8). Used with permission; copyright 0 1991 by the AAAS.)
control
= i =
_-
~TI
control
IRI
I
Fig 5 . Summary data from our research results { I 2) demonstrating that phenyl a-tert-butylnitrone iPBN) administration (300 mglkg) prior t o an ischemialreperfuJion insult ilR1) protects gerbils from death ( 1 ) and prevents increase in brain protein oxidation as well as loss of glutamine synthetase IGS) activity {2). I
S26 Annals of Neurology Supplement
to
creased amount of errors created in a radial arm maze by older gerbils is decreased significantly, nearly to that observed in the younger animals, after 14 days of PBN administration [ 2 ] (see Fig 6). Naive animals (n = 18) were used in each group; the test, which measures short-term/spatial memory, was conducted 24 hours after cessation of PBN administration. In contrast to the older gerbils, PBN administration to the younger animals caused no changes in the amount of errors committed [2].
Conclusions and Implications for Future Studies The data summarized herein clearly implicate that oxidative damage does occur not only in normal aging brain, but also in brain that has experienced a large IRI. The lesions that occur to the CA, neurons during an IRI in gerbil brain may be due in part to the loss of GS activity, which would be expected to result in an increase in glutamate (i.e., GS enzymatically converts glutamate [using NH, and adenosine triphosphate as substrates) to glutamine; thus, a decrease in its activity may allow glutamate, which is neurotoxic, to accumulate). The simplest interpretation of the ability of PBN to cause reversal of the increased oxidation state in older gerbil brain is by implicating its ability to trap free radicals. Thus, if it traps crucial free radicals very early in the cascade of events leading to oxidative damage, then its protective activity may be much greater than
Volume 32, 1992
its tissue concentration may implicate. It is known that PBN effectively and rapidly (within 20 min) permeates all tissues and remains there with a half-life of a few hours before being eliminated in the urine 1191. The lessons learned from the PBN experiments certainly show that the oxidative damage set-point is altered with age but is fairly rigidly regulated at an advanced age {lS}. The experiments also show that interference with the equilibrium processes that control the oxidative damage set-point will reestablish the original oxidative damage set-point if the perturbing influence is removed (i.e., PBN administration ceases). It is clear that the results provide leads that may yield therapeutic interventions with age-associated brain lesions. Research with these goals in mind is now actively being pursued. Our recent results, which show that oxidized protein increases in a logarithmic fashion as a function of age in human brain [20), clearly indicate that the data obtained in gerbil may also be of pertinence in human aging. This research was supported in part by National Institutes of Health grants AG 09690 and NS 23307. We thank Drs Earl R. Stadtman, Pamela Stark-Reed, and Cynthia Oliver, whose collaborative effort made these studies possible.
References Floyd RA. Role of oxygen free radicals in carcinogenesis and brain ischemia. FASEB J 1990;4:2587-2597 Carney JM, Stake-Reed PE, Oliver CN, et al. Reversal of agerelated increase in brain protein oxidation, decrease in enzyme activity, and loss in temporal and spatial memory by chronic administration of the spin-trapping compound N-tert-butyl-aphenylnitrone. Proc Natl Acad Sci USA 1991;88:3633-3636 Oliver CN, Starke-Reed PE, Stadtman ER, et al. Oxidative damage to brain proteins, loss of glutamine synthetase acrivity, and production of free radicals during ischemia/reperfusioninduced injury to gerbil brain. Proc Natl Acad Sci USA 1990; 87:5144-5147 Schneider JE, Browning MM, Floyd RA. Ascorbate/iron mediation of hydroxyl free radical damage to pBR322 plasmid DNA. J Free Radic Biol Med 1988;5:287-295 5. Schneider JE, Browning MM, Zhu X, et al. Characterization of
hydroxyl free radical mediated damage to plasmid pBR322 DNA. Mutat Res 1989;214:23-31 6. Loeb LA, James EA, Waltersdorph AM, et al. Mutagenesis by the autoxidation of iron with isolated DNA. Proc Natl Acad Sci USA 1988;85:3918-3922 7. Tkeshelashvili LK, McBride T, Spence K. Mutation spectrum of copper-induced DNA damage. J Biol Chem 1991;266: 6401-6406 8. Floyd RA, West MS, Eneff KL, et al. Conditions influencing yield and analysis of 8-hydroxy-2'-deoxyguanosinein oxidatively damaged DNA. Anal Biochem 1990;188:155-158 9. Stadtman ER, Oliver CN. Metal-catalyzed oxidation of proteins. J Biol Chem 1991;266:2005-2008 10. Stadtman ER. Metal ion catalyzed oxidation of proteins: biochemical mechanism and biological consequences. J Free Radic Biol Med 1990;9:315-325 11. Levine RL, Garland D, Oliver CN, et al. Determination of carbony1 content in oxidatively modified proteins. Methods Enzymol 1970;186:464-4 78 12. Rivett AJ. Preferential degradation of the oxidatively modified form of glutamine synthetase by intracellular mammalian proteases. J Biol Chem 1985;260:300-305 13. Zaleska MM, Floyd RA. Regional lipid peroxidation in rat brain in vitro: possible role of endogenous iron. Neurochem Res 1985;10:397-410 14. Zaleska MM, Nagy K, Floyd RA. Iron-induced lipid peroxidation and inhibition of dopamine synthesis in striatum synaptosomes. Neurochem Res 1989;14:597-605 15. Floyd RA, Henderson R, Watson JJ, et al. Use of salicylare with high pressure liquid chromatography and electrochemical detection (LCED) as a sensitive measure of hydroxyl free radicals in adriamycin treated rats. J Free Radic Biol Med 1986; 2: 13-18 16. Cao W, Carney JM, Duchon A, et al. Oxygen free radical involvement in ischemia and reperfusion injury to brain. Neurosci Lett 1988;88:233-238 17. Floyd RA, Carney JM. Age influence on oxidative events during brain ischemialreperfusion. Arch Gerontol Geriatr 1991;12: 155-177 18. Floyd RA. Oxidative damage to behavior during aging. Science 1991;254:1597 19. Chen G, Bray TM, Janzen EG, et al. Excretion, metabolism and tissue distribution of a spin trapping agenr, a-phenyl-N-butylnitrone (PBN) in rats. Free Radic Res Commun 1990;9:317323 20. Smith CD, Carney JM, Starke-Reed PE, et al. Excess brain protein oxidation and enzyme dysfunction in normal aging in Alzheimer's disease. Proc Natl Acad Sci USA 1991;88:1054010543
Floyd and Carney: Free Radical Damage to Protein and DNA
S27
Iron mediates damage to proteins and DNA. The mechanisms of damage not only involve iron but also oxygen free radical intermediates. Oxidative damage to DNA causes not only strand breaks, but also formation of specific base adducts, such as 8-hydroxy-2'-deoxyguanosine.Oxidative damage also inactivates certain enzymes such as glutamine synthetase. Novel methods of assessing oxidative damage to tissue, including quantitation of salicylate hydroxylation as an index of hydroxyl free radical flux as well as specific lesions to proteins and DNA, have yielded results that clearly show that ischemialreperfusion injury to mongolian gerbil brain involves oxidatively damaging events. Aging in gerbil as well as human brain is also associated with increased oxidative damage. Recent novel observations have shown that the spin-trapping agent phenyl a-tert-butylnitrone (PBN)offers protection in gerbil brain during ischemia/ reperfusion injury. We also show that oxidative damage to brain during aging is decreased by chronic administration of PBN. The mechanism of action of PBN may be related to its trapping of specific free radicals, which triggers a cascade of oxidative events that eventually lead to tissue injury. Floyd RA, Carney JM. Free radicaI damage to protein and DNA: mechanisms involved and relevant observations o n brain undergoing oxidative stress. Ann Neurol 1992;32:S22-S27
All aerobic organisms experience an imposed oxidative stress constantly C 11. This statement is becoming virtually axiomatic as knowledge in this area continues to accumulate. The mere fact that oxygen is consumed by an organism results in an imposed oxidative damage potential (PJ. Oxidative damage potential is caused by a small fraction of the total oxygen consumed forming activated oxygen by-products, which, unless quenched, cause oxidative damage to biological molecules. Oxidative damage causes lesions to many biological molecules, including lipids, proteins, and nucleic acids. The activated oxygen species involved include lipid hydroperoxides, hydrogen peroxide, superoxide, hydroxyl free radicals, and singlet oxygen. The antioxidant defense capacity (A,) of the system acts protectively to oppose the oxidative damage potential. The antioxidant defense capacity is composed of the enzymes superoxide dismutase (SOD), catalase, glutathione peroxidase (GSHPx), as well as vitamin E and glutathione. Natural systems appear to operate in a state such that most but not all of the oxidative damage potential imposed is quenched by the antioxidant de-
From the "Molecular Toxicology Research Program, Oklahoma Medical Research Foundation, and the Department of Biochemistry and Molecular Biology, University of Oklahoma Health Sciences Center, Oklahoma City, OK; and the ?Department of Pharmacology, University of Kentucky Medical School, Lexington, KY.
s22
fense capacity (i.e., a small amount of oxidative damage occurs at all times due to the net flux of damaging activated oxygen species, Po', leaking through the antioxidant defense capacity). The oxidative damage potential and the antioxidant defense capacity are in a dynamic equilibrium at all times (Fig 1). Oxidative damage to biological molecules involves the combined action of oxygen free radicals and the trace elements Fe, Cu, or both. Oxidatively damaging events occur in Fenton-type reactions where Fe acts catalytically. Thus, if a very small amount of available Fe is present, then large amounts of oxidative damage may occur if the appropriate flux of oxygen free radicals is present in the immediate vicinity and can react with the available Fe, Cu, or both. The site specificity observed in oxidative damage to DNA and proteins is due to the site-specific iron-binding properties of these macromolecules. Thus, damage to specific amino acids in proteins is dictated by their proximity to the binding site of Fe or their location proximal to the active site of an enzyme that contains Fe or Cu essential for activity. There is compelling evidence that oxidative damage
Address correspondence to Dr Floyd, Department of Biochemistry and Molecular Biology, University of Oklahoma Health Sciences Center, Oklahoma City, OK 7 3 104-5046,
~~~
~
Lipid Hydroperoxides
'9,OH /&
SOD, CATAUSE
GSHPx GSH, Vitamin E
Po-Ac
=
PA
Fig 1. Dynamic equilibrium between oxidative damage potential (Po) and the antioxidant defense capacity (Ad. A small amount of oxidatively damaging species (Po')leaks throughout the defense, thus causing oxidative damage. SOD = superoxide dismutare; GSHPx = glutathione peroxidase; GSH = reduced glutathione.
to brain proteins does occur during brain injury caused by an ischemia/reperfusion insult (IRI) as well as during aging [2, 31. In fact, loss of specific enzymes due to oxidative damage may be a major factor causing brain injury [2}.These areas are reviewed using mostly our own data, which have been collected in collaboration with other laboratories. With regard to oxidative damage to DNA and how this contributes to neurological damage, this area of research is at an early stage and therefore is addressed only from a summary standpoint arrived at from observations conducted on in vitro model systems.
Oxidative Damage to DNA The combination of oxygen free radicals and the trace metals Fe or Cu causes damaging lesions to nucleic acids. The predominant lesion observed is strand breakage, but DNA base modification products are also formed. Many studies have been conducted utilizing supercoiled DNA suspended in phosphate buffer and exposed to ascorbate. Phosphate buffer contains trace amounts of Fe, and thus when exposed to ascorbate, H202 and hydroxyl free radicals (*OH)are formed. Utilizing this system, it can be clearly demonstrated that DNA strand breakage is mediated by Fe and active oxygen intermediates [4}. We demonstrated that one of the essential active oxygen intermediates formed in the Felascorbate system is H202[4, 51. The fact that catalase prevents supercoiled DNA strand breakage supports this notion 14, 5 1. In addition, it has been shown on the basis of inhibitor studies as well as salicylate trapping studies that .OH is produced in the ascorbateiFe system 141. Thus, it is likely that autooxidation of ascorbate produces H202,which then reacts with ferrous iron, which is bound to the DNA strand.
Ascorbate mediates the reduction of ferric ion to ferrous iron. The reaction of H20, with ferrous iron yields .OH, which oxidatively lesions the ribose moiety, thus causing DNA strand breakage. It has been demonstrated that oxygen is an absolute requirement for the ascorbate/Fe system to cause DNA breaks [4}. We also found that the reactions proceed faster as the temperature increases within the 4" to 37°C range IS}. Oxidative damage to DNA mediated by the trace metals Fe or Cu produces mutagenic lesions [G, 71, probably due to formation of oxidatively modified bases such as thymine glycol, 5-hydroxymethyluracil, and 8-hydroxyguanine. The amount of 8-hydroxyguanine in DNA can be determined very sensitively using high-performance liquid chromatographic (HPLC)electrochemical detection [8},which allows assessment of the amount of this modified base in as little as 10 Fg DNA. This sensitivity is very useful when using very small amounts of tissue. The amount of oxidized bases present in DNA is not just a reflection of oxidative damage to the macromolecule per se, but represents a steady-state level of the repair processes that occur.
Oxidative Damage to Proteins Proteins are oxidatively damaged by the combined action of activated oxygen species and the trace metals Fe, Cu, or both [9}. The amino acids lysine, proline, histidine, and arginine have been found to be the most sensitive to oxidative damage [ 101. Oxidative damage to these amino acids yields an increased level of carbony1 groups on proteins. Methods have been developed to assess the carbonyl content of proteins as an estimate of the amount of oxidative damage [ll}. Thus, as oxidative damage to tissue occurs, there is an increased amount of oxidized protein. Oxidized proteins are degraded by a novel protease, a neutral protease, discovered by Rwett [ 12). The relationship between the formation of oxidized protein by Fe and H202and its degradation by neutral protease is shown in Figure 2. Oxidative damage causes rapid loss of enzymatic activity in many proteins [lo}. Glutamine synthetase (GS) is rapidly inactivated by oxidative damage [lo}. The histidine located near the Fe at the active site of GS is lesioned by oxidative damage, thus causing loss in activity [lo}.
High Vulnerability of Brain to Oxidative Damage There are several factors that predispose brain tissue to be highly susceptible to oxidative damage: (1) brain is highly enriched in the easily peroxidizable 22 :6 and 20:4 unsaturated fatty acids; (2) certain regions of brain, particularly human brain, are highly enriched in iron; (3) brain is not particularly well endowed with
Floyd and Carney: Free Radical Damage to Protein and D N A S23
c!7
p Natural Protein
F)
_I%o,,
JR
-
OH
6 1+Hob OH
I
-
Neutral
&
Amino Acids
Protease
I
OH
I
Oxidized Protein
Fig 2. Relationship showing that Fe bound to protein will react with H202 to cause oxidatively damaged protein, resulting in increased carbonyl groups on the protein. The oxidized protein is degraded to individual amino acids by a neutral protease.
antioxidant defense systems; ( 4 ) brain utilizes proportionately a large amount of oxygen in relation to its weight; (5) brain is a highly organized information processing center; and (6) neurons are postmitotic, thus death of a neuron represents a permanent lesion. Research work with in vitro incubation systems using rat brain homogenate 1131 has demonstrated the following: ( 1 ) brain will readily peroxidize if it is organizationally disrupted; (2) it peroxidizes rapidly in phosphate buffer if the incubation temperature is 37"C, but not at 4°C.; (3) the peroxidation rate of each brain region is proportional to the total iron content of the brain region, with the exception of striatum; where ( 4 ) the increased dopamine content acts to inhibit peroxidation, as can be demonstrated by the addition of dopamine to brain homogenate; and ( 5 ) chelators of iron inhibit brain peroxidation, but active oxygen scavengers are not very effective { 131. The consequences of oxidative damage to brain in terms of carrying out essential neurochemical activities, such as dopamine synthesis, are illustrated by the studies of Zaleska and colleagues 1141. We demonstrated that dopamine synthesis from tyrosine in striatal synaptosomes was dramatically curtailed by the addition of very small levels of adenosine diphosphate (ADP)-chelated Fe in the presence of ascorbate. Peroxidation of the synaptosomes was induced by the ADP-Fe/ascorbate treatment 114). Peroxidation of synaptosomes and the decrease in dopamine synthesis was inversely linked, showing a dramatic decrease in dopamine synthesis, with one-half the activity lost with 2.5 pmol/L Fe {14). Determination of Oxygen Free Radical Flux I n Vivo It is a difficult task to determine the amount of oxygen free radicals present in brain. The difficulties arise from the fact that oxygen free radicals react rapidly (i.e., as soon as they are formed); thus, the amount present at any specific time is very small, perhaps lo-" mol/L or less. Thus, it has been necessary to use indirect methods to assess oxygen free radical flux in vivo. Two general methods have been utilized: (1) determination S24 Annals of Neurology
0
CH,
0ec-LC " ,+ I
R
- Of-F'-+.
CH,
Spin-Trap
Free Radical
H
0 CH,
R
CH,
Spin- Adduct
la-phenyl-tert-butylNitrone1
Fig 3. Reactions showing that salicylate reacts with bydroxyl free radicals (OH) t o form the 2,3- and 2,5-dibydroxybenzoic acid products and the spin-trap phenyl a-tert-butylnitrone re'acts with a free radical to form a spin-adduct.
of the amount of unique products formed when oxygen free radicals react with biological molecules, and (2) addition of exogenous trapping compounds that react with oxygen free radicals to form unique and relatively stable products, which can be quantitated. Measurement of oxygen free radical-mediated products of DNA bases and proteins are examples of the first general method, whereas use of spin-traps to react with free radicals and salicylate to trap hydroxyl free radicals are examples of the second general method. Figure 3 illustrates the reactions involved with two exogenous traps. The spin-trap, PBN (phenyl a-tevt-butylnitrone), reacts with free radicals to form spin adducts. Salicylate reacts with hydroxyl free radicals at a diffusion-limited rate to produce the 2,3- and 2,5-dihydroxybenzoic acid (DHBA) products, which can be quantitated very sensitively using HPLC-electrochemical detection { 151. We have used the salicylate hydroxylation method to estimate hydroxyl free radical flux in several biological systems, including the IRI-lesioned gerbil brain [ 161. Oxidative Damage i n the Ischemia/Reperfused Gerbil Brain The Mongolian gerbil is an ideal animal model to study ischemia/reperfusion in jury in brain because in almost all gerbils, the forebrain is entirely perfused by the two common carotids. This anatomical peculiarity allows examination of the effect of cessation of blood flow and its subsequent reintroduction into the forebrain by ligation and then release of the common carotids { 161. Utilizing this model, it is also possible to compare other regions of the brain, especially the cerebellum and the brainstem, and use these regions as internal controls to ascertain if the effect is region-specific. The cerebellum and the brainstem are not perfused by the carotid arteries and thus do not suffer oxygen deprivation when the carotids are ligated. Earlier concerns that gerbils are prone to seizures have been minimized by use of seizure-resistant lines.
Supplement to Volume 32, 1992
Figure 4 summarizes pertinent data from a series of several different studies examining oxidative damage in gerbil brain undergoing an IRI. Several methods of assessing oxidative damage in gerbil brain given a 10-minute period of ischemia followed by 60 minutes of reperfusion are compared for cortex and brainstem. Salicylate hydroxylation in cortex is significantly increased by an IRI, but not in brainstem [17}. Protein oxidation as assessed by carbonyl content of protein is markedly increased in cortex, but not in the brainstem of the same animals 131, whereas GS is markedly decreased in cortex of IRI-lesioned gerbils, but not in the brainstem of the same animals 131. The remarkable correlation of all three parameters of oxidative damage in the two separate regions of brain convincingly underscore the notion that oxygen deprivation of brain and its subsequent reintroduction leads to a series of oxidative events that lead to much damage, including oxidative lesioning of proteins and specific enzymes.
Cortex
Brain Stem
175
-.2
IRI
150 125
P
E" loo . -5
Control IRI
75
350
8
25
0 14-7I
PBN Protects Brain from IRI-induced Damage We have made observations that clearly indicate the spin-trapping compound, PBN, protects gerbils from IRI-induced brain damage. A combined summary of the results of several studies 11-31 are shown in Figure 5. The data clearly show that administration of PBN prior to a brain IRI helped to prevent (1) loss in GS activity, ( 2 ) increase in oxidative damage to brain protein, and ( 3 ) lethality brought on by a very damaging IRI. There is an age effect in the severity of the damage brought on by brain ischemia as manifested by lethality, assessed at 7 days. In young gerbils ( 3 to 4 months old) 15 minutes of ischemia caused 50% lethality; in older gerbils (18 months old), 10 minutes of ischemia caused death in all animals. PBN-mediated Reversal of Age-associated Oxidative Damage There is an increased (nearly two-fold) amount of oxidized proteins in the brains of normal older gerbils (18 months old) as compared with normal young gerbils (3-4 months old). In addition, the activity of neutral protease in older gerbil brain is only 33% of that in younger gerbil brain, and the activity of GS in older gerbils is only 65% of that observed in younger gerbils {27. Chronic intraperitoneal administration of PBN (32 mg/kg twice a day) for 14 days caused the amount of oxidized proteins to decrease in the brains of older gerbils nearly to that observed in younger gerbils 12). In addition, PBN administration to the older gerbils also caused an increase in GS and neutral protease activity nearly to that observed in the younger animals 121. A summary of these data 1181 is presented in Figure 6. In addition to the changes noted in protein oxidation and enzyme activity, we also found that the in-
Fig 4. Data from our recent work summarizing salicylate hydroxykztion 117) and protein oxidation, as well as loss in glutamine synthetase (GS) activity {3) in gerbils with ischemialrepe$usion insult and comparable sham-operated control gerbils. The cerebral cortex region is compared to the brainstem. The gerbils were given a 10-minute ischemic insult, and a 60-minute repetfusion insult.
Floyd and Carney: Free Radical Damage to Protein and DNA
S25
300-
200-
5
'6
s
IRI
100-
OJ
IRI + PBN
NORMALS
PBN TREATED
OFF PBN
Fig 6. Data summarizing the protective effort of phenyl a-tertbutylnitrone (PBN) administered chronically (14 days. twice daily at 32 mglkg, intraperitoneally) to older gerbils {2}. Data show that PBN administration caused a decrease in oxidized protein and concomitant increases in glutamine synthetase (GS) and neutral protease activity. In addition, errors in short-term spatial memory, assessed with a radial a m maze test, was decreased by PBN administration. (From {I 8). Used with permission; copyright 0 1991 by the AAAS.)
control
= i =
_-
~TI
control
IRI
I
Fig 5 . Summary data from our research results { I 2) demonstrating that phenyl a-tert-butylnitrone iPBN) administration (300 mglkg) prior t o an ischemialreperfuJion insult ilR1) protects gerbils from death ( 1 ) and prevents increase in brain protein oxidation as well as loss of glutamine synthetase IGS) activity {2). I
S26 Annals of Neurology Supplement
to
creased amount of errors created in a radial arm maze by older gerbils is decreased significantly, nearly to that observed in the younger animals, after 14 days of PBN administration [ 2 ] (see Fig 6). Naive animals (n = 18) were used in each group; the test, which measures short-term/spatial memory, was conducted 24 hours after cessation of PBN administration. In contrast to the older gerbils, PBN administration to the younger animals caused no changes in the amount of errors committed [2].
Conclusions and Implications for Future Studies The data summarized herein clearly implicate that oxidative damage does occur not only in normal aging brain, but also in brain that has experienced a large IRI. The lesions that occur to the CA, neurons during an IRI in gerbil brain may be due in part to the loss of GS activity, which would be expected to result in an increase in glutamate (i.e., GS enzymatically converts glutamate [using NH, and adenosine triphosphate as substrates) to glutamine; thus, a decrease in its activity may allow glutamate, which is neurotoxic, to accumulate). The simplest interpretation of the ability of PBN to cause reversal of the increased oxidation state in older gerbil brain is by implicating its ability to trap free radicals. Thus, if it traps crucial free radicals very early in the cascade of events leading to oxidative damage, then its protective activity may be much greater than
Volume 32, 1992
its tissue concentration may implicate. It is known that PBN effectively and rapidly (within 20 min) permeates all tissues and remains there with a half-life of a few hours before being eliminated in the urine 1191. The lessons learned from the PBN experiments certainly show that the oxidative damage set-point is altered with age but is fairly rigidly regulated at an advanced age {lS}. The experiments also show that interference with the equilibrium processes that control the oxidative damage set-point will reestablish the original oxidative damage set-point if the perturbing influence is removed (i.e., PBN administration ceases). It is clear that the results provide leads that may yield therapeutic interventions with age-associated brain lesions. Research with these goals in mind is now actively being pursued. Our recent results, which show that oxidized protein increases in a logarithmic fashion as a function of age in human brain [20), clearly indicate that the data obtained in gerbil may also be of pertinence in human aging. This research was supported in part by National Institutes of Health grants AG 09690 and NS 23307. We thank Drs Earl R. Stadtman, Pamela Stark-Reed, and Cynthia Oliver, whose collaborative effort made these studies possible.
References Floyd RA. Role of oxygen free radicals in carcinogenesis and brain ischemia. FASEB J 1990;4:2587-2597 Carney JM, Stake-Reed PE, Oliver CN, et al. Reversal of agerelated increase in brain protein oxidation, decrease in enzyme activity, and loss in temporal and spatial memory by chronic administration of the spin-trapping compound N-tert-butyl-aphenylnitrone. Proc Natl Acad Sci USA 1991;88:3633-3636 Oliver CN, Starke-Reed PE, Stadtman ER, et al. Oxidative damage to brain proteins, loss of glutamine synthetase acrivity, and production of free radicals during ischemia/reperfusioninduced injury to gerbil brain. Proc Natl Acad Sci USA 1990; 87:5144-5147 Schneider JE, Browning MM, Floyd RA. Ascorbate/iron mediation of hydroxyl free radical damage to pBR322 plasmid DNA. J Free Radic Biol Med 1988;5:287-295 5. Schneider JE, Browning MM, Zhu X, et al. Characterization of
hydroxyl free radical mediated damage to plasmid pBR322 DNA. Mutat Res 1989;214:23-31 6. Loeb LA, James EA, Waltersdorph AM, et al. Mutagenesis by the autoxidation of iron with isolated DNA. Proc Natl Acad Sci USA 1988;85:3918-3922 7. Tkeshelashvili LK, McBride T, Spence K. Mutation spectrum of copper-induced DNA damage. J Biol Chem 1991;266: 6401-6406 8. Floyd RA, West MS, Eneff KL, et al. Conditions influencing yield and analysis of 8-hydroxy-2'-deoxyguanosinein oxidatively damaged DNA. Anal Biochem 1990;188:155-158 9. Stadtman ER, Oliver CN. Metal-catalyzed oxidation of proteins. J Biol Chem 1991;266:2005-2008 10. Stadtman ER. Metal ion catalyzed oxidation of proteins: biochemical mechanism and biological consequences. J Free Radic Biol Med 1990;9:315-325 11. Levine RL, Garland D, Oliver CN, et al. Determination of carbony1 content in oxidatively modified proteins. Methods Enzymol 1970;186:464-4 78 12. Rivett AJ. Preferential degradation of the oxidatively modified form of glutamine synthetase by intracellular mammalian proteases. J Biol Chem 1985;260:300-305 13. Zaleska MM, Floyd RA. Regional lipid peroxidation in rat brain in vitro: possible role of endogenous iron. Neurochem Res 1985;10:397-410 14. Zaleska MM, Nagy K, Floyd RA. Iron-induced lipid peroxidation and inhibition of dopamine synthesis in striatum synaptosomes. Neurochem Res 1989;14:597-605 15. Floyd RA, Henderson R, Watson JJ, et al. Use of salicylare with high pressure liquid chromatography and electrochemical detection (LCED) as a sensitive measure of hydroxyl free radicals in adriamycin treated rats. J Free Radic Biol Med 1986; 2: 13-18 16. Cao W, Carney JM, Duchon A, et al. Oxygen free radical involvement in ischemia and reperfusion injury to brain. Neurosci Lett 1988;88:233-238 17. Floyd RA, Carney JM. Age influence on oxidative events during brain ischemialreperfusion. Arch Gerontol Geriatr 1991;12: 155-177 18. Floyd RA. Oxidative damage to behavior during aging. Science 1991;254:1597 19. Chen G, Bray TM, Janzen EG, et al. Excretion, metabolism and tissue distribution of a spin trapping agenr, a-phenyl-N-butylnitrone (PBN) in rats. Free Radic Res Commun 1990;9:317323 20. Smith CD, Carney JM, Starke-Reed PE, et al. Excess brain protein oxidation and enzyme dysfunction in normal aging in Alzheimer's disease. Proc Natl Acad Sci USA 1991;88:1054010543
Floyd and Carney: Free Radical Damage to Protein and DNA
S27
