Oxidative DNA Damage and Cardiovascular Disease

Oxidative DNA Damage and Cardiovascular Disease Seon Hwa Lee and Ian A. Blair*

Reactive oxygen species can directly cause covalent modifications to DNA. Alternatively, they can initiate the formation of lipid hydroperoxides, which undergo homolytic decomposition to the a,b-unsaturated aldehyde genotoxins, 4-oxo-2-nonenal, 4,5-epoxy-2(E)-decenal, and 4hydroxy-2-nonenal through two quite separate pathways. One pathway involves a complex rearrangement of the alkoxy radical derived from the lipid hydroperoxide. The other pathway involves the intermediate formation of 4-hydroperoxy-2-nonenal. Lipid hydroperoxides can also be derived from the action of lipoxygenases and cyclooxygenases on polyunsaturated fatty acids. 4,5-Epoxy-2(E)-decenal forms etheno-29-deoxyadenosine adduct with DNA, a mutagenic lesion observed in human tissue DNA samples. Several new ethano- and etheno-DNA adducts have been identified from the reaction of 4-oxo2-nonenal with DNA. Malondialdehyde, another genotoxic bifunctional electrophile, forms a propano adduct with 29-deoxyguanosine (M1G-dR) rather than an etheno adduct. Very little is known about the consequences of lipid hydroperoxide-mediated DNA damage in cardiovascular diseases. This should prove to be an important area for future research. (Trends Cardiovasc Med 2001;11:148–155). © 2001, Elsevier Science Inc.

Oxygen in its ground triplet state has two unpaired electrons with parallel spins and so it can be regarded as a radical even though this is its most stable form. The parallel spins mean that oxygen can only accept one electron at a time. Singlet oxygen, which is produced without the addition of an electron, has

Seon Hwa Lee and Ian A. Blair are from the Center for Cancer Pharmacology, University of Pennsylvania, Philadelphia, Pennsylvania. * Address correspondence to: Ian A. Blair, Ph.D., Director, Center for Cancer Pharmacology, 1254 BRB II/III, 421 Curie Bvd., University of Pennsylvania, Philadelphia, PA 19104-6160, USA. Tel.: 215-573-9880; fax: 215-573-9889; e-mail: [email protected]. © 2001, Elsevier Science Inc. All rights reserved. 1050-1738/01/$-see front matter

148

antiparallel spins. This very reactive form of oxygen is produced in vivo during phagocytosis by peroxidases (Kiryu et al. 1999) and during photodynamic therapy with photosensitizers (Hsi et al. 1999). The reactive oxygen species . (ROS) superoxide (O22 ), peroxide . 22 (O2 ), and hydroxyl radical (HO ), are generated constantly in vivo from ground state triplet oxygen by a variety of endogenous processes (Figure 1) (Ames et al. 1993). ROS are normally detoxified by antioxidant defense systems such as superoxide dismutase, catalase, reduced glutathione (GSH)dependent peroxidases (Ames et al. 1993), and thioredoxin (Mustacich and Powis 2000). There are many other endogenous processes that protect against ROS-mediated damage including the sequestration of hydrogen peroxide

generating enzymes and the chelation of free transition metal ions by transferrin, ferritin, and ceruloplasmin. Some of the ROS are able to escape these defenses in order to perform important metabolic roles. This means that there is always a potential for ROS-mediated damage to macromolecules such as DNA and proteins. Increased ROS production occurs in inflammation, during radiation, or during metabolism of hormones, drugs, and environmental toxins. This can overwhelm endogenous protective mechanisms and increase ROS-mediated damage to cellular macromolecules. It has been suggested that ROS generation is a major contributor to the degenerative diseases of aging including cardiovascular disease, cancer, immune-system decline, brain dysfunction, and cataracts (Ames et al. 1993). DNA damage can occur by direct reaction with ROS (Ames et al. 1993) or from ROS-derived lipid-hydroperoxide breakdown to endogenous genotoxins (Marnett 2000). Lipid hydroperoxides can also be formed by the action of lipoxygenases (LOXs) (Brash 1999) and cyclooxygenases (COXs) (Laneuville et al. 1995) on polyunsaturated fatty acids (PUFAs). Linoleic acid, the major v-6 PUFA in human plasma, is an excellent substrate for 15-LOX and COX-2. Therefore, LOX- and COX-mediated pathways of PUFA metabolism can potentially provide a rich source of lipid hydroperoxide-derived genotoxins. The lesions that result from ROS and lipid hydroperoxide-derived genotoxins are normally repaired in order to maintain the fidelity of the DNA. If the lesions are not repaired, subsequent DNA replication can lead to mutations (Burcham 1998, Marnett 2000) or apoptosis (Johnson et al. 1996). Endothelial cell apoptosis appears to play a major regulatory role in neovascularization. Excessive apoptosis may limit angiogenesis which in turn would lead to vessel regression (Dimmeler and Zeiher 2000). ROS-derived covalent modifications to DNA have been identified in human atherosclerotic plaques (De Flora et al. 1997) but the technique that was employed most likely resulted in an over-estimation of their content (Cadet et al. 1998). In spite of the methodological problems, this study serves to highlight the potential for the formaTCM Vol. 9, No. 3/4, 2001

tion of ROS-derived DNA adducts in atherosclerosis. • Reactive Oxygen Species Molecular oxygen undergoes three successive one-electron reductions (Linn 1998) to form ROS (Figure 1A). The initially formed superoxide radical anion dismutates to hydrogen peroxide and oxygen either non-enzymatically, or through the action of superoxide dismutases (Figure 1B). Alternatively, it can undergo the Haber-Weiss reaction to give hydroxyl radicals (Kehrer 2000). The Haber–Weiss reaction between superoxide radical anion and hydrogen peroxide has a second-order rate constant of zero in biological systems (Figure 1C). Therefore, it requires catalysis by a transition metal ion such as Fe31 (Figure 1D and E). The initial reaction involves Fe31-mediated oxidation of superoxide radical anion to oxygen (Figure 1D). The resulting Fe21 then initiates the Fenton reaction (Figure 1E), which involves the reduction hydrogen peroxide to hydroxyl radical and hydroxyl anion with concomitant re-generation of Fe31 (Kehrer 2000). It has been proposed recently that hypochlorous acid (generated from myeloperoxidase) reacts with superoxide radical anion to give hydroxyl radicals (Herdener et al. 2000). Hydrogen peroxide can be converted directly to hydroxyl radicals through the Fenton reaction (Figure 1E) (Linn 1998). Mixtures of ascorbic acid and transition metal ion also induce ROS formation from molecular oxygen (Buettner and Jurkiewicz 1996). The combination of Cu21 and ascorbic acid is a particularly rich source of ROS (Buettner and Jurkiewicz 1996). ROS are also formed during the conversion of catechols to quinones (Bolton et al. 2000) (Figure 2). The first oneelectron oxidation of the catechol to a semiquinone radical anion, which is initiated by superoxide radical anion, results in the formation of hydrogen peroxide (McCoull et al. 1999). The second one-electron oxidation of the semiquinone radical anion to the quinone is mediated by molecular oxygen and results in the formation of more superoxide radical anion. NADPH-dependent two-electron reduction of the quinone results in regeneration of the original catechol. This sets up a futile redox TCM Vol. 11, No. 3/4, 2001

Figure 1. (A) Sequential one-electron reductions of molecular oxygen to superoxide radical anion, hydrogen peroxide, and hydroxyl radical. (B) Dismutation of superoxide radical anion. (C) Haber–Weiss reaction. (D) Oxidation of superoxide radical anion by Fe31. (E) Fenton reaction.

cycle, which amplifies the formation of ROS and leads to direct DNA damage or initiates the formation of lipid hydroperoxides (Figure 2). Catechols thought to be involved in the generation of ROS through redox cycling include polycyclic aromatic hydrocarbon catechols (McCoull et al. 1999), catechol estrogens (Bolton et al. 2000), and the catechol metabolite of the topoisomerase II inhibitor etoposide (Lovett et al. 2001). Additional pathways for ROS generation include normal mitochondrial aerobic respiration, phagocytosis of bacteria or virus-containing

cells, peroxisomal-mediated degradation of fatty acids, and cytochromeP450-mediated metabolism of xenobiotics DNA (Ames et al. 1993). • Direct ROS-Mediated Oxidative DNA Damage The direct reaction of ROS with DNA bases is complex and dependent upon the system used (Breen and Murphy 1995). The most widely accepted index of ROS-mediated DNA damage involves quantitation of 7,8-dihydro-8-oxo-29deoxyguanosine (8-oxo-dGuo; Figure 3)

Figure 2. Cytochrome P-450 (CYP)-mediated oxidation followed by epoxide hydrolase-(EH) mediated hydrolysis of the polycyclic aromatic hydrocarbon benzo[a]pyrene (BP) to BP-7,8-diol. This is followed by aldo-keto reductase (AKR)-mediated oxidation to BP-7,8-catechol, which then redox cycles to BP-7,8-quinone with concomitant formation of reactive oxygen species.

149

(Cadet et al. 1999). This is most likely because of the high sensitivity with which it can be analyzed using electrochemical detection (ECD) (Helbock et al. 1998). A comprehensive study of oxidative damage to 29-deoxyguanosine (dGuo) was conducted using hydrogen peroxide and FeIII (Henle et al. 1996). In this system, 8-oxo-dGuo was a minor product and at least 10 other covalent modifications to dGuo were characterized. Mixtures of ascorbic acid, hydrogen peroxide, and CuII have also been used to generate oxidized DNA bases (Kennedy et al. 1997). Under carefully controlled conditions, yields in excess of 25% of 8-oxo-dGuo can be obtained (Lee and Blair, unpublished observation). The currently accepted mechanism for the formation of 8-oxo-dGuo from hydroxyl radicals involves initial attack at C-8 with concomitant formation of a nitrogen-centered radical at N7 and a carbon-centered radical located primarily at C-4 (Breen and Murphy 1995). A one-electron reduction gives the formamidopyrimidine (FAPY) derivative; whereas, a one-electron oxidation yields 8-oxo-dGuo (Figure 3). Singlet oxygen generated from N,N9-di(2,3-dihydroxypropyl)-1, 4-naphthalenedipropanamide (1,4-endoperoxide of DHPNO2) also results in the formation of 8-oxo-dGuo (Ravanat et al. 2000). ROS-derived modifications to other DNA bases include: 5,6-dihydroxy5,6-dihydro-thymidine, 5-hydroxy-29deoxyuridine, 5-hydroxymethyl-29-deoxyuridine, 8-oxo-7,8-dihydro-29-deoxyadenosine

(8-oxo-dAdo), 4,6-diamino-5-formadopyrimidine, and 2,6-diamino-4-hydroxy5-formadopyrimidine (Frelon et al. 2000). • Lipid Hydroperoxides Lipid hydroperoxides are formed nonenzymatically by hydroxyl radicalmediated abstraction of a bis-allylic hydrogen atom of the PUFA followed by attachment of molecular oxygen (Porter et al. 1995). Linoleic acid, the major v-6 PUFA present in plasma lipids is converted to a complex mixture of 9- and 13-hydroperoxy-octadecadienoic acid (HPODE) isomers. Recently, 11-HPODE was identified as a product of vitamin E-controlled autoxidation of linoleate (Brash 2000). Lipoxygenases (LOXs) (Brash 1999) and cyclooxygenases (COXs) (Hamberg 1998) can also convert linoleic acid into HPODEs but with much greater stereoselectivity than is observed in free radical reactions. Human 15-LOX produces mainly 13(S)-hydroperoxy(Z,E)-9,11-octadecadienoic acid (13HPODE) (Kamitani et al. 1998). COX-1 and COX-2 produce mainly 13-HPODE and 9(R)-hydroperoxy-(E,Z)-10,12-octadecadienoic acid (9-HPODE) by a mechanism similar to that described for dihomo-g-linolenic acid (Hamberg and Samuelsson 1967). The HPODEs are subsequently reduced to the corresponding 13(S)- and 9(R)-hydroxyoctadecadienoic acids (HODEs) through the peroxidase activity of the COXs (Hamberg 1998, Schneider and Brash

Figure 3. Mechanism for the formation of 8-oxo-dGuo and the formamidopyrimidine (FAPY) derivative.

150

2000). The other C-18 PUFAs, linolenic acid (v-3) and dihomo-g-linolenic acid (v-6) (minor constituents of plasma lipids) are also metabolized by 15-LOX. Linolenic acid is a poor substrate for COX-1 and COX-2 (Laneuville et al. 1995) but dihomo-g-linolenic acid has not been studied in detail. C-20 PUFAs all undergo 15-LOX-mediated conversion to hydroperoxides. The major 5-LOX and COX-derived products from C-20 PUFAs (apart from eicosadienoic acid) are prostaglandins, thromboxanes, and leukotrienes rather than lipid hydroperoxides. Transition metal ions induce the decomposition of v-6 lipid hydroperoxides to the a,b-unsaturated aldehyde genotoxins, 4-oxo-2-nonenal and 4-hydroxy2-nonenal (Lee and Blair 2000). We have recently demonstrated that the unstable epoxides, trans-4,5-epoxy-2(E)decenal and cis-4,5-epoxy-2(E)decenal are also formed during the reaction. These genotoxins appears to arise from two quite different pathways. The first pathway, based on that described previously by Pryor and Porter (1990), is initiated by alkoxy radical formation as shown in Figure 4. Alkoxy radical I rearranges to a C-12,13epoxide with concomitant addition of molecular oxygen to give the 9-hydroperoxy-12,13-epoxy intermediate II. This hydroperoxide then undergoes a second one-electron reduction to alkoxy radical intermediate III, which rearranges to the symmetrical intermediate IV or undergoes an a-cleavage to give primarily trans-4,5-epoxy-2(E)-decenal and a small amount of the cis-isomer. Intermediate IV adds molecular oxygen to give hydroperoxide V, which then either undergoes a concerted rearrangement to 4-oxo-2nonenal or forms the unstable epoxide VI, which rearranges to 4-hydroxy-2nonenal (Figure 4). Decomposition of 13-HPODE to 4-hydroperoxy-2-nonenal cannot be rationalized by any previously proposed mechanism. It is particularly difficult to understand the oxygenation at C-10 of the 13-HPODE and so we are currently conducting experiments to determine how this occurs. We have recently established that 4-hydroperoxy2-nonenal undergoes transition metal ion-mediated breakdown to 4-oxo-2nonenal and 4-hydroxy-2-nonenal (Lee and Blair, unpublished observation). This provides an additional pathway by which these bifunctional electrophiles TCM Vol. 11, No. 3/4, 2001

Figure 4. Mechanism for the formation of the bifunctional electrophiles, 4-hydroperoxy-2nonenal (4-HPNE), 4-oxo-2-nonenal (4-ONE), 4-hydroxy-2-nonenal (4-HNE), and 4,5-epoxy2(E)-decenal (4,5-EDE) by homolytic decomposition of 13-HPODE.

can be formed (Figure 4). Lipid hydroperoxides derived from arachidonic acid (a C-20 v-6 PUFA), undergo homolytic decomposition (Pryor and Stanley 1975) to malondialdehyde (MDA; b-hydroxyacrolein), a genotoxic bifunctional electrophile (Marnett 1999). MDA is also formed by hydroxyl radical attack of the 29-deoxyribose DNA backbone (Rashid et al. 1999) and during the biosynthesis of thromboxane A2 (Hecker and Ullrich 1989).

lently modify dGuo to generate propano or ethano adducts, respectively (Chung et al. 1996). Therefore, it was tempting to attribute the formation of the observed products to either one or both of these reactive intermediates. However, it was difficult to reconcile the anticipated adduct structures with the mass spectral data that were obtained. Thus, it appeared that novel adduction products

were formed in the reaction between 13HPODE and dGuo. We subsequently showed that the initially formed adducts all dehydrated to give a single product (Rindgen et al. 1999). Based on considerations of the collision-induced dissociation (CID) tandem mass spectrometry (MS/MS) spectra we hypothesized that the novel dehydration product was derived from 4-oxo-2-nonenal. Two-dimensional nuclear magnetic resonance data supported a heptanone-etheno-dGuo structure for the adduct (Figure 5). Based on the known reactivity of dGuo, the etheno group was thought to be attached through N1 and N2 of the purine ring rather than N2, N3 (Kusmierek and Singer 1992). Furthermore, it was shown previously that an analogous reaction occurred between 4-oxo-2-pentenal and dGuo (Hecht et al. 1992). The substituted-etheno dGuo adduct identified in this reaction also arose from a-acetoxy-N-nitrosopiperidine, indicating that 4-oxo-2-pentenal was the reactive bifunctional electrophile responsible for the covalent modification of dGuo. This is directly analogous to our findings, which suggested that 13HPODE-derived 4-oxo-2-nonenal was responsible for the covalent modification to dGuo. We recently demonstrated unequivocally that 4-oxo-2-nonenal is indeed a major product of homolytic 13HPODE decomposition (Lee and Blair 2000). Three major products were detected in the reaction between dAdo and 13HPODE (Rindgen et al. 2000). The rate of formation of these adducts was in-

• Lipid Hydroperoxide-Mediated DNA Damage Our investigation of the reaction between 13-HPODE and dGuo revealed the presence of several adduction products that contained nine carbon units of the starting fatty acid (Rindgen et al. 1999). The number of hydroperoxide carbons included in the products was supported by the data obtained from enzymatic reactions with soybean LOX conducted in the presence of [13C18]linoleic acid. Previous studies had shown that the C9 compounds derived from fatty acid hydroperoxides, 4-hydroxy-2-nonenal and 2,3-epoxy-4-hydroxynonanal, could covaTCM Vol. 11, No. 3/4, 2001

Figure 5. DNA adducts from 4-oxo-2-nonenal (4-ONE) and 4,5-epoxy-2(E)-decenal (4,5-EDE).

151

creased in the presence of FeII indicating that they were formed through a homolytic process. The adducts arose through reaction with 13-HPODE-derived 4-oxo2-nonenal. Initial nucleophilic addition of N6 to the C-1 aldehyde of 4-oxo-2nonenal was followed by reaction of N1 at C-2 of the resulting a,b-unsaturated ketone to generate a mixture of two ethano adducts that could inter-convert with each other. The ethano adducts subsequently dehydrated to give a single heptanone etheno-dAdo adduct (Figure 5). Unsubstituted etheno-dAdo adducts have been suggested to arise from the reaction of the epoxide of 4-hydroxy-2nonenal with DNA (Chung et al. 1996). It seems unlikely that epoxidation of 4-hydroxy-2-nonenal occurs in vivo because it is such a good substrate for glutathione-S-transferases and aldo-keto reductases (Burczynski et al. 2001). We have now demonstrated that ethenodAdo is formed by the reaction of 4,5epoxy-2(E)-decenal with dAdo, which provides an alternative pathway for generation of the etheno-dAdo adduct in vivo. Subsequent loss of the 29-deoxyribose moiety results in formation of the corresponding adenine (Ade) adduct (Figure 5). Etheno-dAdo is also formed in the reaction between the environmental carcinogen vinyl chloride and DNA (Doerge et al. 2000). Reactions between the DNA-bases dGuo or dAdo and the a,b-unsaturated aldehydes, 4-hydroxy-2-nonenal and MDA (no electron-withdrawing substituent at C-4) results in the formation of exocyclic propano adducts (Burcham 1998). Michael addition occurs initially at the b-carbon from N2 or N1 of dGuo followed by nucleophilic addition of N1 or N2 of dGuo at the carbonyl carbon. When the a,b-unsaturated aldehyde has a substituent at the b-carbon, the resulting steric hindrance inhibits nucleophilic attack from N1. Kinetic control of the reaction favors the regioisomer in which N2 is attached to the b-carbon atom and N1 is attached to the carbonyl carbon. This results in the formation of two pairs of diastereomeric 1,N2-propano-dGuo adducts from 4-hydroxy-2-nonenal (Yi et al. 1997) and pyrimido[1,2-a]purin-10(3H)-one (M1G) from MDA (Marnett 2000). DNA adducts have not yet been identified from reaction of DNA bases with the bifunctional electrophile 4-hydroperpoxy-

152

2-nonenal. Interestingly, 4-hydroxy-2nonenal can react with model amino acids through a pathway that has been proposed to involve the intermediate formation of 4-oxo-2-nonenal (Xu and Sayre 1998). 4,5-Epoxy-2(E)-heptenal (the v-3 analog of 4,5-epoxy-2(E)-decenal), is known to modify histidine residues in human serum albumin (Hidalgo and Zamora 2000). Therefore, there is also a potential role for lipid hydroperoxide-derived bifunctional electrophiles in modifying protein structure and function. • Covalent Modifications to DNA in vivo from Reactive Oxygen Species and Lipid Hydroperoxides Substantial efforts have been made over the last decade to develop methodology for the quantification of DNA lesions formed in vivo from direct reaction with ROS and from lipid hydroperoxidederived bifunctional electrophiles. The measurement of oxidative damage has relied heavily on the analysis of 8-oxodGuo in DNA or urine (Cadet et al. 1999). These measurements have been extremely controversial because of the ease with which adventitious oxidation can occur during DNA-hydrolysis and derivatization of the resulting DNAbases (Cadet et al. 1999, Helbock et al. 1998). It has been estimated that biological buffers normally contain at least 1 mM of transition metal ions (Buettner and Jurkiewicz 1996). Molecular oxygen undergoes a transition metal ionmediated one-electron reduction to superoxide radical anion (Figure 1A), which dismutates into hydrogen peroxide and oxygen (Figure 1B). Further transition metal ion-mediated reduction of the hydrogen peroxide to hydroxyl radicals can then occur by Fenton chemistry (Figure 1E). High concentrations of unmodified bases are present in DNA hydrolysates and so a significant amount of hydroxyl radical-mediated oxidation can then occur. This provides artificially increased concentrations of oxidized bases, which has led to erroneous quantitative results in a number of studies (Cadet et al. 1998). Oxidation of DNA bases during derivatization reactions for GC/MS analyses can be minimized by the use of antioxidants (Jenner et al. 1998). Careful purification of buffers by Chelex treatment coupled with

the use of deferoxamine to chelete transition metal ions and chaotropic sodium iodide DNA extraction methods can prevent most of the artifactual oxidation during DNA adduct isolation (Helbock et al. 1998). The current state of the art involves careful elimination of transition metal ions from buffer solutions followed by liquid chromatography (LC)/ECD analysis. Using this methodology, levels of 8-oxo-dGuo of 4.0/107 dGuo have been found in liver DNA of young rats (4 months old) and 1.1/106 dGuo in liver DNA of old rats (26 months old). These results are 1–2 orders of magnitude lower than measurements made using older methodology (Helbock et al. 1998). The interpretation of 8-oxo-dGuo excretion into the urine is complicated by the 8-oxo-dGuo formed from mitochondrial DNA turnover and repair, DNAdegradation from apoptotic cell death, and oxidation of nucleobase pools. Furthermore, 8-oxo-dGuo may be further oxidized or catabolized prior to excretion. With these caveats in mind, the urinary excretion of 8-oxo-dGuo was determined as 370 6 63 pmol/kg/day in rats (n 5 30) and 172 6 79 pmol/kg/day (n 5 63) in humans (Park et al. 1992). This difference in rate of 8-oxo-dGuo excretion appears to reflect the reduced relative rate of oxygen consumption by humans compared with rats. Methodology based on stable isotope dilution LC/electrospray (ESI)/MS/MS has been developed for 8-oxo-dGuo but the sensitivity is somewhat poorer than that observed with LC/ECD (Frelon et al. 2000). GC/ MS methodology has shown that 8-oxodGuo is the major oxidized DNA base present in human urine (Ravanat et al. 1999). Several studies have attempted to correlate oxidative DNA damage with cardiovascular disease (Collins et al. 1998, De Flora et al. 1997). Unfortunately, appropriate precautions to prevent artifactual oxidation of the dGuo were not taken, so these studies will need to be repeated under more stringent conditions (Helbock et al. 1998). The urinary excretion of 8-oxo-dGuo was measured in the Antioxidant Supplementation in Athersclerosis Prevention Trial (ASAP). Surprisingly, no significant effects were observed with vitamin C, vitamin E, or a combination of the two (Porkkala-Sarataho et al. 2000). MDA is probably the most intensively studied lipid-derived endogenous genoTCM Vol. 11, No. 3/4, 2001

toxin. The cyclic DNA adduct it forms with dGuo (M1G) can also be formed from base propenals by treatment of DNA with calicheamicin or bleomycin (Dedon et al. 1998). M1G has been detected in both animal (Chaudhary et al. 1994a) and human (Chaudhary et al. 1994b) liver DNA and in circulating human leukocytes (Rouzer et al. 1997). Using sensitive and specific gas chromatography (GC)/electron capture negative chemical ionization (ECNCI)/MS, levels in the range of 4.8 M1G/108 to 2.0 M1G/107 normal bases were observed (Chaudhary et al. 1994a and 1994b, Rouzer et al. 1997). The levels of M1G in human leukocyte were less than onetenth the level found in human liver DNA (Rouzer et al. 1997). This was ascribed to the fact that leukocytes are relatively short-lived cells, and may not be expected to accumulate high levels of adduct during their lifetime. On the other hand, leukocytes have active COXand LOX-enzymes, which could have given rise to significant MDA production. Therefore, it was expected that higher M1G levels would have been observed. The leukocytes from several smokers were examined and no striking differences were observed with age, or between smokers and nonsmokers in the study. A small difference was observed between the M1G levels of male and female donors, although there were insufficient samples to draw any firm conclusions from this study (Rouzer et al. 1997). A study by Chen et al. (1999) reported the etheno-Ade adduct at a level of 2–4 adducts/106 Ade bases in human placental DNA. This was considerably higher than M1G levels in liver and leukocyte DNA. Stable isotope dilution GC/ECNCI/MS methodology was used for quantification and LC/MS/MS was used to confirm the structural assignment. A more recent study was able to improve the sensitivity of the LC/MS/MS technique considerably, so that etheno-dAdo adducts could be quantified in human placental DNA with high specificity. Levels of 1.1 etheno dAdo/108 normal bases were much closer to those found for M1G in human leukocytes (Doerge et al. 2000). The discrepancy between the two studies was suggested to result from artifactual formation of etheno dAdo during the isolation procedure, and highlights the difficulties associated with TCM Vol. 11, No. 3/4, 2001

high sensitivity determinations of lipid hydroperoxide-derived DNA adducts in human tissue samples. • Summary and Future Directions A large body of evidence has accumulated, which suggests that ROS can cause direct covalent modifications to DNA (Ames et al. 1993) and that lipid hydroperoxide-derived bifunctional electrophiles can covalently modify DNA (Burcham 1998, Marnett 2000). Lipid hydroperoxides are derived from ROS, or enzymatically by the action of LOXs and COXs on PUFAs. The recent discovery of 4-oxo-2-nonenal as a homolytic breakdown product of lipid hydroperoxides (Lee and Blair 2000) has led to the characterization of several new DNA adducts (Lee et al. 2000, Rindgen et al. 1999). However, nothing is known about how these lesions affect cellular proliferation or apoptosis. A role for 4-oxo-2nonenal in modifying protein structure is suggested by the finding that it can covalently bind to model amino acids (Xu and Sayre 1998). Therefore, this a,bunsaturated aldehydic bifunctional electrophile, which reacts with DNA bases with remarkable regioselectivity, should provide a fertile ground for further mechanistic investigations. 4-Hydroxy-2-nonenal, another bifunctional electrophile derived from homolytic decomposition of lipid hydroperoxides, was shown recently to up-regulate COX-2 expression (Kumagai et al. 2000). As COX-2 can convert linoleic acid into lipid hydroperoxides (Hamberg 1998), this potentially provides a mechanism for increased production of genotoxic bifunctional electrophiles. The identification of 4-hydroperoxy-2-nonenal as another homolytic breakdown product of lipid hydroperoxides (Lee and Blair 2000) suggests that there may be other, as yet unidentified, lesions present in DNA arising from this bifunctional electrophile. Little attention has been given in the past to 4,5-epoxy-2(E)-decenal as a bifunctional electrophile derived from v-6 lipid hydroperoxides. We have recently demonstrated that 4,5-epoxy2(E)-decenal reacts with DNA to give unsubstituted etheno-dAdo adducts (Figure 5), which provides an important link between lipid peroxidation and a DNA adduct that has been detected in human tissue samples (Doerge et al. 2000). Sub-

stantial changes in gene expression occur as a result of ROS generation during chemically induced atherogenesis (Ramos 1999). Furthermore, DNA adduct formation is known to induce apoptosis, which inhibits angiogenesis and promotes vessel regression. Therefore, it will be important in the future to determine whether endogenous DNA adducts derived from ROS or lipid hydroperoxides are associated with cardiovascular disease. Our recent discovery of LC/ electron capture atmospheric pressure ionization/MS/MS for the high sensitivity analysis of intact DNA adducts should greatly facilitate such studies (Singh et al. 2000). • Acknowledgments We acknowledge the support of NIH grant CA65878.

References Ames BN, Shigenaga MK, Hagen TM: 1993. Oxidants, antioxidants, and the degenerative diseases of aging. Proc Natl Acad Sci USA 90:7915–7922. Bolton JL, Trush MA, Penning TM, et al.: 2000. Role of quinones in toxicology. Chem Res Toxicol 13:136–160. Brash AR: 1999. Lipoxygenases: occurrence, functions, catalysis, and acquisition of substrate. J Biol Chem 274:23,679–23,682. Brash AR: 2000. Autoxidation of methyl linoleate: identification of the bis-allylic 11-hydroperoxide. Lipids 35:947–952. Breen AP, Murphy JA: 1995. Reactions of oxyl radicals with DNA. Free Rad Biol Med 18:1033–1077. Buettner GR, Jurkiewicz BA: 1996. Catalytic metals, ascorbate and free radicals: combinations to avoid. Radiat Res 145:532– 541. Burcham PC: 1998. Genotoxic lipid peroxidation products: their DNA damaging properties and role in formation of endogenous DNA adducts. Mutagenesis 13:287–305. Burczynski ME, Sridhar GR, Palackal NT, et al.: 2001. The reactive oxygen species- and Michael acceptor-inducible human aldoketo reductase AKR1C1 reduces the a,bunsaturated aldehyde 4-hydroxy-2-nonenal to1,4-dihydroxy-2-nonene. J Biol Chem 276:2890–2897. Cadet J, Delatour T, Douki T, et al.: 1999. Hydroxyl radicals and DNA base damage. Mutat Res 424:9–21.

153

Cadet J, D’Ham C, Douki T, et al.: 1998. Facts and artifacts in the measurement of oxidative base damage to DNA. Free Rad Res 29:541–550. Chaudhary AK, Nokubo M, Marnett LJ, et al.: 1994a. Analysis of malondialdehyde-29deoxyguanosine adduct in rat liver DNA by gas chromatography/electron capture negative chemical ionization mass spectrometry. Biol Mass Spectrom 23:457–464. Chaudhary AK, Nokubo M, Reddy GR, et al.: 1994b. Detection of endogenous malondialdehyde-deoxyguanosine adducts in human liver. Science 265:1580–1582. Chen H-JC, Chiang L-C, Tseng MC, et al.: 1999. Detection and quantification of 1,N6ethenoadenine in human placental DNA by mass spectrometry. Chem Res Toxicol 12:1119–1126. Chung F-L, Chen H-JC, Nath RG: 1996. Lipid peroxidation as a potential endogenous source for the formation of exocyclic DNA adducts. Carcinogenesis 17:2105–2111. Collins AR, Gedik CM, Olmedilla B, et al.: 1998. Oxidative DNA damage measured in human lymphocytes: large differences between sexes and between countries, and correlations with heart disease mortality rates. FASEB J 13:1397–1400. Dedon PC, Plastaras JP, Rouzer CA, et al.: 1998. Indirect mutagenesis by oxidative DNA damage: formation of the pyrimidopurinone adduct of deoxyguanosine by base propenal. Proc Natl Acad Sci USA 95:11,113–11,116. De Flora S, Izzotti A, Walsh D, et al.: 1997. Molecular epidemiology of atherosclerosis. FASEB J 11:1021–1031. Dimmeler S, Zeiher AM: 2000. Endothelial cell apoptosis in angiogenesis and vessel regression. Circ Res 87:434–439. Doerge DR, Churchwell MI, Fang JL, et al.: 2000. Quantification of etheno DNA adducts using liquid chromatography, online sample processing, and electrospray tandem mass spectrometry. Chem Res Toxicol 13:1259–1264. Frelon S, Douki T, Ravanat J-L, et al.: 2000. High performance liquid chromatographytandem mass spectrometry measurement of radiation-induced base damage to isolated and cellular DNA. Chem Res Toxicol 13:1002–1010. Hamberg M: 1998. Stereochemistry of oxygenation of linoleic acid catalyzed by prostaglandin-endoperoxide H synthase-2. Arch Biochem Biophys 349:376–380. Hamberg M, Samuelsson B: 1967. On the mechanism of the biosynthesis of prostaglandins E1 and F1a. J Biol Chem 242:5336–5343. Hecht SS, Young-Sciame R, Chung F-L: 1992. Reaction of a-acetoxy-N-nitrosopiperidine with deoxyguanosine: oxygen-dependent

154

formation of 4-oxo-2-pentenal and a 1,N2ethenodeoxyguanosine adduct. Chem Res Toxicol 5:706–712. Hecker M, Ullrich V: 1989. On the mechanism of prostacyclin and thromboxane A2 biosynthesis. J Biol Chem 264:141–150. Helbock HJ, Beckman KB, Shigenaga MK, et al.: 1998. DNA oxidation matters: the HPLC-electrochemical detection assay of 8-oxo-deoxyguanosine and 8-oxo-guanine. Proc Natl Acad Sci USA 95:288–293. Henle ES, Luo Y, Gassman W, et al.: 1996. Oxidative damage to DNA constituents by iron-mediated Fenton reactions. The deoxyguanosine family. J Biol Chem 271: 35,21177–21186. Herdener M, Heigold S, Saran M, et al.: 2000. Target cell-derived superoxide anions cause efficiency and selectivity of intercellular induction of apoptosis. Free Rad Biol Med 29:1260–1271. Hidalgo FJ, Zamora R: 2000. Modification of bovine serum albumin structure following reaction with 4,5(E)-epoxy-2(E)-heptenal. Chem Res Toxicol 13:501–508. Hsi RA, Rosenthal DI, Glatstein E: 1999. Photodynamic therapy in the treatment of cancer: current state of the art. Drugs 57:725– 734. Jenner A, England TG, Aruoma O, et al.: 1998. Measurement of oxidative DNA damage by gas chromatography-mass spectrometry: ethanethiol prevents artifactual generation of oxidized DNA bases. Biochem J 331:365–369. Johnson TM, Yu ZX, Ferrans VJ, et al.: 1996. Reactive oxygen species are downstream mediators of p53-dependent apoptosis. Proc Natl Acad Sci USA 93:11,848–11,852. Kamitani H, Geller M, Eling T: 1998. Expression of 15-lipoxygenase by human colorectal carcinoma Caco-2 cells during apoptosis and cell differentiation. J Biol Chem 273:21,569–21,577. Kehrer JP: 2000. The Haber-Weiss reaction and mechanisms of toxicity. Toxicology 149:43–50. Kennedy LJ, Moore K, Caulfield JF, et al.: 1997. Quantitation of 8-oxoguanine and strand breaks produced by four oxidizing agents. Chem Res Toxicol 10:386–392. Kiryu C, Makiuchi M, Miyazaki J, et al.: 1999. Physiological production of singlet molecular oxygen in the myeloperoxidase-H2O2chloride system. FEBS Lett 443:154–158. Kumagai T, Kawamoto Y, Nakamura Y, et al.: 2000. 4-Hydroxy-2-nonenal, the end product of lipid peroxidation, is a specific inducer of cyclooxygenase-2 expression. Biochem Biophys Res Commun 273:437–441. Kusmierek JT, Singer B: 1992. 1,N2-Ethenodeoxyguanosine: properties and formation in

chloroacetaldehyde-treated polynucleotides and DNA. Chem Res Toxicol 5:634–638. Laneuville O, Breuer DK, Xu N, et al.: 1995. Fatty acid substrate specificities of human prostaglandin-endoperoxide H synthase-1 and -2. J Biol Chem 270:19,330–19,336. Lee SH, Blair IA: 2000. Characterization of 4-oxo-2-nonenal as a novel product of lipid peroxidation. Chem Res Toxicol 13:698–702. Lee SH, Rindgen D, Bible RA, et al.: 2000. Characterization of 29-deoxyadenosine adducts derived from 4-oxo-2-nonenal, a novel product of lipid peroxidation. Chem Res Toxicol 13:565–574. Linn S: 1998. DNA damage by iron and hydrogen peroxide in vitro and in vivo. Drug Metab Rev 30:313–326. Lovett BD, Strumberg D, Blair IA, et al.: 2001. Etoposide metabolites enhance DNA topoisomerase II cleavage proximal to leukemia-associated MLL translocation breakpoints. Biochemistry 40:1159–1170. Marnett LJ. 1999. Occurrence and methods of detection. Chemistry and biology of DNA damage by malondialdehyde. In Singer B, Bartsch H (eds.). Exocyclic DNA Adducts in Mutagenesis and Carcinogenesis. Lyon, IARC Scientific Publications 150, pp 17–27. Marnett LJ: 2000. Oxyradicals and DNA damage. Carcinogenesis 21:361–370. McCoull KD, Rindgen D, Blair IA, et al.: 1999. Synthesis and characterization of polycyclic aromatic hydrocarbon orthoquinone depurinating N7-guanine adducts. Chem Res Toxicol 12:237–246. Mustacich D, Powis G: 2000. Thioredoxin reductase. Biochem J 346:1–8. Park EM, Shigenaga MK, Degan P, et al.: 1992. Assay of excised oxidative DNA lesions: isolation of 8-oxoguanine and its nucleoside derivatives from biological fluids with a monoclonal antibody column. Proc Natl Acad Sci USA 89:3375–3379. Porkkala-Sarataho E, Salonen JT, Nyyssonen K, et al.: 2000. Long-term effects of vitamin E, vitamin C, and combined supplementation on urinary 7-hydro-8-oxo-29-deoxyguanosine, serum cholesterol, and oxidation resistance of lipids in nondepleted men. Arterioscler Thromb Vasc Biol 20:2087–2093. Porter NA, Caldwell SE, Mills KA: 1995. Mechanisms of free radical oxidation of unsaturated lipids. Lipids 30:277–290. Pryor WA, Porter NA: 1990. Suggested mechanisms for the production of 4-hydroxy-2nonenal from the autoxidation of polyunsaturated fatty acids. Free Rad Biol Med 8:541–543. Pryor WA, Stanley JP: 1975. A suggested mechanism for the production of malonaldehyde during the autoxidation of polyunsaturated

TCM Vol. 11, No. 3/4, 2001

fatty acids. Nonenzymatic production of prostaglandin endoperoxides during autoxidation. J Org Chem 40:3615–3617. Ramos KS: 1999. Redox regulation of c-Haras and osteopontin signaling in vascular smooth muscle cells: implications in chemical atherogenesis. Annu Rev Pharmacol Toxicol 39:243–265. Rashid R, Langfinger D, Wagner R, et al.: 1999. Bleomycin versus OH-radical-induced malonaldehydic-product formation in DNA. Int J Radiat Biol 75:101–109. Ravanat J-L, Di Mascio P, Medeiros MHG, et al.: 2000. Singlet oxygen induces oxidation of cellular DNA. J Biol Chem 275:40601–40604. Ravanat J-L, Guicherd P, Tuce Z, et al.: 1999. Simultaneous determination of five oxidative DNA lesions in human urine. Chem Res Toxicol 12:802–808. Rindgen D, Lee SH, Nakajima M, et al.: 2000. Formation of a substituted 1,N6-etheno-29deoxyadenosine adduct by lipid hydroperoxide-mediated generation of 4-oxo-2-nonenal. Chem Res Toxicol 13:846–852. Rindgen D, Nakajima M, Wehrli S, et al.: 1999. Covalent modifications to 29-deoxyguanosine by 4-oxo-2-nonenal a novel product of lipid peroxidation. Chem Res Toxicol 12:1195–1204. Rouzer CA, Chaudhary AK, Nokubo M, et al.: 1997. Analysis of the malondialdehyde-29deoxyguanosine adduct pyrimidopurinone in human leukocyte DNA by gas chromatography/electron capture/negative chemical ionization/mass spectrometry. Chem Res Toxicol 10:181–188. Schneider C, Brash AR: 2000. Stereospecificity of hydrogen abstraction in the conversion of arachidonic acid to 15R-HETE by aspirin-treated cyclooxygenase-2. J Biol Chem 275:4743–4746. Singh G, Xu K, Gutierrez A, et al.: 2000. Liquid chromatography/electron capture atmospheric pressure chemical ionization/mass spectrometry: analysis of pentafluorobenzyl derivatives of biomolecules and drugs in the attomole range. Anal Chem 72: 3007–3013. Xu G, Sayre LM: 1998. Structural characterization of a 4-hydroxy-2-alkenal-derived fluorophore that contributes to lipoperoxidation-dependent protein cross-linking in aging and degenerative disease. Chem Res Toxicol 11:247–251. Yi P, Zhan DJ, Samokyszyn VM, et al.: 1997. Synthesis and 32P-postlabeling/highperformance liquid chromatography separation of diastereomeric 1,N2 -(1,3-propano)29-deoxyguanosine 39-phosphate adducts formed from 4-hydroxy-2-nonenal. Chem Res Toxicol 10:1259–1265. PII S1050-1738(01)00094-9

TCM Vol. 11, No. 3/4, 2001

TCM

The Role of High-Density Lipoproteins in Oxidation and Inflammation Brian J. Van Lenten*, Mohamad Navab, Diana Shih, Alan M. Fogelman, and Aldons J. Lusis

High-density lipoproteins (HDL) in the basal state are anti-inflammatory, capable of destroying oxidized lipids that generate an inflammatory response. However, HDL during acute inflammation are altered and become pro-inflammatory. This “chameleon-like” nature of HDL is considered to be due to the complex composition of HDL. The data reviewed here demonstrate the key role of HDL in modulating inflammation and its implications for atherogenesis. (Trends Cardiovasc Med 2001;11:155–161). © 2001, Elsevier Science Inc.

High-density lipoproteins (HDL) are a heterogeneous collection of particles consisting of about 50% protein and 50% lipid. Mature HDL have a hydrophobic core of cholesteryl esters and triglycerides, surrounded by a monolayer of phospholipids and unesterified cholesterol in which are embedded various proteins. The major proteins of HDL are apolipoproteins (apo) AI and AII. In addition, HDL contain a large number of less abundant proteins including apoCI, apoCII, apoCIII, apoE, apoJ, apoL, lecithin:cholesterol acyltransferase (LCAT), serum paraoxonase

Brian J. Van Lenten, Mohamad Navab, Diana Shih, and Alan M. Fogelman are from the Department of Medicine, University of California, Los Angeles, California. Aldons J. Lusis is from the Department of Microbiology, Immunology, and Molecular Genetics, Department of Human Genetics, and the Molecular Biology Institute, University of California, Los Angeles, California. * Address correspondence to: Brian J. Van Lenten, Ph.D., UCLA School of Medicine, Atherosclerosis Research Unit, Division of Cardiology, Department of Medicine 47-123 CHS, Los Angeles, CA 90095-1679, USA. Tel.: 310-206-1150; fax: 310-206-3537; e-mail: [email protected]. © 2001, Elsevier Science Inc. All rights reserved. 1050-1738/01/$-see front matter

(PON1), and platelet-activating factor acetylhydrolase (PAF-AH). During inflammation, a new set of proteins, including serum amyloid A (SAA) and ceruloplasmin, bind to HDL. The metabolism of HDL is pictured in Figure 1. A major function of HDL is to promote the efflux of free cholesterol from cells and to transport it to the liver, a process termed “reverse cholesterol transport.” HDL also act as a carrier vehicle for small molecules, including bacterial lipopolysaccharide and vitamins such as beta-carotene. Finally, HDL exhibit antioxidant and anti-inflammatory properties, the subject of this review. HDL probably protect against atherosclerosis by several mechanisms, and under certain conditions such as infection, HDL may even promote atherogenesis. The relationship between HDL levels and atherosclerosis has been observed in many epidemiologic studies. Also, most HDL deficiencies result in increased coronary heart disease (CHD), and apoAI transgenic mice, with unusually high levels of HDL, exhibit resistance to atherosclerosis. On the other hand, some disorders with dramatically reduced HDL levels, such as LCAT deficiencies, are not associated with greatly increased CHD. This has raised the possibility that HDL may not directly protect against atherosclerosis and that the inverse relationship between HDL levels and CHD may

155