Myeloperoxidase-generated oxidants and atherosclerosis

Abstract

Atherosclerosis is a chronic inflammatory process where oxidative damage within the artery wall is implicated in the pathogenesis of the disease. Mononuclear phagocytes, an inflammatory cell capable of generating a variety of oxidizing species, are early components of arterial lesions. Their normal functions include host defense and surveillance through regulated generation of diffusible radical species, reactive oxygen or nitrogen species, and HOCl (hypochlorous acid). However, under certain circumstances an excess of these oxidizing species can overwhelm local antioxidant defenses and lead to oxidant stress and oxidative tissue injury, processes implicated in the pathogenesis of atherosclerosis. This review focuses on oxidation reactions catalyzed by myeloperoxidase (MPO), an abundant heme protein secreted from activated phagocytes which is present in human atherosclerotic lesions. Over the past several years, significant evidence has accrued demonstrating that MPO is one pathway for protein and lipoprotein oxidation during the evolution of cardiovascular disease. Multiple distinct products of MPO are enriched in human atherosclerotic lesions and LDL recovered from human atheroma. However, the biological consequences of these MPO-catalyzed reactions in vivo are still unclear. Here we discuss evidence for the occurrence of MPO-catalyzed oxidation reactions in vivo and the potential role MPO plays in both normal host defenses and inflammatory diseases like atherosclerosis.

Introduction

An elevated level of low-density lipoprotein (LDL), the major carrier of cholesterol in plasma, is a primary risk factor for the development of atherosclerosis. However, our current understanding of the atherosclerotic process suggests that LDL must first be oxidized or chemically modified before it is rendered atherogenic [1], [2], [3], [4], [5], [6]. In vivo studies support the notion that protein and lipid components of the arterial wall undergo oxidative damage early in atherogenesis. Moreover, a wealth of in vitro studies demonstrate that oxidation of LDL can lead to the acquisition of biological activities considered to be “atherogenic,” such as the capacity to promote cholesterol accumulation and foam cell formation, the induction of leukocyte chemoattractant molecules and adhesion proteins, and the elaboration of mitogenic or cytotoxic factors [1], [2], [3], [4], [5], [6]. Numerous chemical models have been used for promoting LDL oxidative damage in vitro, including exposure to free transition metal ions, ionizing radiation, both heme and nonheme proteins, and even “aging” in a refrigerator. However, elaboration of the precise pathways and reactive intermediates involved in oxidative modification of LDL, proteins, and lipids in vivo has remained elusive.

Recent studies have begun to unravel some of the specific oxidation pathways that contribute to (lipo)protein and lipid oxidation during vascular disease. One pathway for which significant data are now available involves the leukocyte protein myeloperoxidase (MPO). MPO is present in human atherosclerotic lesions [7]. Both immunohistochemical and mass spectrometric studies support a role for this heme protein in catalyzing oxidative modification of targets in the human artery wall. In this review we examine evidence supporting a role for MPO in promoting oxidation reactions in atherosclerotic tissues. We discuss many of the reactive oxidants and diffusible radical species formed by MPO, structural characterization of the oxidation products they form, and evidence that these pathways and products occur in the human artery wall. We also discuss results from recent studies identifying novel mechanisms for regulating MPO activity and some potential biological roles of MPO-generated oxidants in vivo.

MPO (donor: hydrogen peroxide, oxidoreductase, EC 1.11.1.7) is a tetrameric, heavily glycosylated, basic (PI > 10) heme protein of ∼150 kDa. It is comprised of two identical disulfide-linked protomers, each of which possesses a protoporphyrin-containing 59–64 kDa heavy subunit and a 14 kDa light subunit [8]. MPO is abundant in neutrophils and monocytes, accounting for 5%, and 1 to 2%, respectively, of the dry weight of these cells [8], [9]. The heme protein is stored in primary azurophilic granules of leukocytes and secreted into both the extracellular milieu and the phagolysosomal compartment following phagocyte activation by a variety of agonists [10]. Typically, phagocyte activation and MPO secretion are accompanied by an oxidative burst where superoxide (O₂^•−) and its dismutation product, hydrogen peroxide (H₂O₂), are formed by the NADPH oxidase complex [11]. MPO amplifies the oxidizing potential of H₂O₂ by using it as co-substrate to generate a host of reactive oxidants and diffusible radical species through a classic peroxidase cycle [9], [10], [12], [13], [14], [15], [16], [17], [18].

A recently proposed working kinetic model for MPO is shown in Fig. 1. MPO is a complex heme protein which possesses multiple intermediate states, each of which are influenced by the availability of reduced oxygen species such as O₂^•− and H₂O₂, and nitric oxide (NO, nitrogen monoxide) [19]. At ground state, MPO exists in the ferric (Fe(III)) form. Upon addition of H₂O₂, the heme group of MPO is oxidized two e⁻ equivalents forming a reactive ferryl π cation radical intermediate termed Compound I. In the presence of halides such as Cl⁻, Br⁻, and I⁻, and the psuedohalide thiocyanate (SCN⁻), Compound I is readily reduced in a single two e⁻ step, regenerating MPO-Fe(III) and the corresponding hypohalous acid (HOX). At plasma levels of halides and thiocyanate (100 mM Cl⁻, 100 μM Br⁻; 50μM SCN⁻, 100 nM I⁻), chloride is a preferred substrate and hypochlorous acid (HOCl), a potent chlorinating oxidant, is formed [20], [21]. The ability of MPO to generate chlorinating oxidants

$$\text{Cl}{}^{\mathrm{-}}\text{+H}{}_2\text{O}{}_2\text{+H}{}^{\mathrm{+}}\text{MPO}\text{→}\text{HOCl+H}{}_2\text{O}$$

under physiological conditions is a unique and defining activity for the mammalian enzyme.

Compound I can also oxidize numerous organic substrates while the heme undergoes two sequential one e⁻ reduction steps, generating compound II and MPO-Fe(III), respectively (Fig. 1). The active site of MPO is located at the base of a narrow and deep hydrophobic pocket which restricts accessibility of substrates [22], [23]. Thus, low molecular weight compounds primarily serve as substrates for MPO, generating diffusible oxidants and free radical species which can then convey the oxidizing potential of the heme to distant targets. In addition to halides and SCN⁻, some of the naturally occurring substrates for MPO include nitrite (NO₂⁻) [24], tyrosine [25], ascorbate [26], urate [27], catecholamines [28], estrogens [29], and serotonin [30]. The relative contribution of peroxidation of these substrates to the overall activity of MPO in vivo is unknown. The products they form, and the reactions they initiate may play a significant biological function. For example, the diffusible oxidant species they generate can lead to formation of potent signaling molecules, such as through initiation of lipid peroxidation [18], [31], [32], [33], or the activation of xenobiotic intermediates [34], [35], [36], [37]. MPO-Fe(III) can also be reduced to an inactive ferrous form, MPO-Fe(II) [9], [16]. MPO-Fe(III) and MPO-Fe(II) bind to O₂^•−, and O₂, respectively, forming a ferrous dioxy intermediate, compound III (MPO-Fe(II)-O₂) (Fig. 1). Spectral studies demonstrate that addition of H₂O₂ to Compound III ultimately forms compound II. Thus, compound III may indirectly promote one e⁻ peroxidation reactions.

The regulation of MPO activity is typically thought to primarily rely upon the rate of O₂^•− production, the availability of H₂O₂ and other cosubstrates [9], [10], [16], or the abundance of antioxidant species such as ascorbate or methionine [16], [26]. However, recent studies identify a role for NO, a relatively long-lived free radical generated by nitric oxide synthase (NOS), in modulating MPO peroxidase activity [19]. NO plays important bioregulatory roles in many physiological processes [38], [39], and is implicated in the regulation of many hemoproteins [38], [39], [40]. MPO and the inducible isoform of NOS are colocalized in the primary granule of leukocytes. During phagocyte activation, such as during ingestion of bacteria, MPO and NOS are secreted into the phagolysosome and extracellular compartments, and nitration of bacterial proteins is observed [41]. Rapid kinetics studies demonstrate that at low levels of NO, the initial rate of MPO-catalyzed peroxidation of substrates is enhanced. The mechanism is through acceleration of the rate-limiting step in MPO catalysis, reduction of compound II to MPO-Fe(III) (Fig. 1) [19], [42]. At higher levels of NO, reversible inhibition of MPO occurs through formation of a spectroscopically distinguishable nitrosyl complex, MPO-Fe(III)-NO [19]. NO also can serve as a substrate for MPO compound I, resulting in its reduction to Compound II [42]. Furthermore, in the presence of NO, the overall turnover rate of MPO through the peroxidase cycle is enhanced nearly 1000-fold [42]. The net result is that MPO can serve as a catalytic sink for NO, limiting its bioavailability [42]. Finally, NO also reversibly binds to MPO-Fe(II) forming the corresponding MPO-Fe(II)-NO intermediate, which is in equilibrium with MPO-Fe(II) and MPO-Fe(III)-NO (Fig. 1) [19], [42]. The regulation of MPO in vivo is thus complex since the enzyme performs its functions in a wide variety of environments with differing pH and levels of NO, H₂O₂, O₂^•−, O₂, inorganic and organic substrates, and reducing agents.

As described above, MPO can utilize a variety of cosubstrates with H₂O₂ to generate reactive oxidants as intermediates. Many stable end-products generated by these species have been characterized (Fig. 2) and shown to be enriched in proteins, lipids, and LDL recovered from human atherosclerotic lesions [43], [44], [45], [46], [47]. The breadth of products formed by MPO, and their identification in vascular tissues by direct mass spectrometric methods, represents some of the most compelling evidence that MPO is a catalyst for oxidative modification of biological targets like LDL in vivo [45], [48], [49], [50]. Fig. 2 summarizes some of the reactive intermediates and products formed by MPO, many of which are known to be enriched in vascular lesions. For many, pathways alternative to those involving MPO also exist for their formation. However, MPO is the only known pathway for generating reactive chlorinating species at physiological levels of halides in humans [12], [13], [45], [51]. Detection of specific stable chlorinated oxidation products has therefore been exploited as a powerful tool to identify tissues where MPO promotes oxidative damage in vivo [45], [52].

The reactivity of chlorinating oxidants like HOCl against biological targets is well established and includes chlorination of amines [21], [53], [54] and unsaturated lipids [55], [56], oxidation of thiols and thiol esters [57], and oxidative bleaching of heme groups and iron sulfur centers [58]. In addition, interaction of HOCl with H₂O₂ and O₂^•− is reported to form singlet oxygen and hydroxyl radical, respectively [59], [60]. The vast majority of oxidation products formed by MPO-generated chlorinating oxidants are not ideally suited to serve as molecular markers for identifying sites of MPO-catalyzed oxidation in vivo. For example, many are labile chlorinated intermediates which readily lose their halide group forming noninformative products (e.g., chloramines decompose to amines [21], [54] and chlorohydrins decompose into epoxides [55], [61], [62]). Alternatively, many of the oxidation products formed are not uniquely generated by MPO (e.g., dityrosine, nitrotyrosine, lipid peroxidation products, oxysterols, methionine sulfoxide, disulfides). However, one stable chlorinated product formed by MPO is 3-chlorotyrosine [45], [48], [52], [63].

3-Chorotyrosine is a nonphysiologic oxidation product of tyrosine which is both acid stable and not readily formed by artificial mechanisms—characteristics which make it well suited to serve as a specific molecular marker for MPO-catalyzed oxidation [52]. Gas chromatography/mass spectrometry (GS/MS) studies demonstrated that the 3-chlorotyrosine content of LDL recovered from human atherosclerotic aorta was 30-fold higher than that present in LDL recovered from plasma of healthy donors [45]. Significant increases in the content of 3-chlorotyrosine were also observed in proteins in atherosclerotic vs. normal aortic intima [45]. Independent immunohistological confirmation of a role for chlorinating oxidants in modification of human atherosclerotic tissues was reported using antibodies specific for proteins modified by HOCl [44]. The epitope recognized by these antibodies is not known, but does not appear to be chlorotyrosine since molar excesses of either free chlorotyrosine or synthetic peptides containing chlorotyrosine fail to block antibody recognition (Podrez and Hazen, unpublished data).

The chlorinating intermediate generated by MPO-catalyzed oxidation of Cl⁻ is believed to be HOCl or its conjugate base, hypochlorite (ClO⁻; pK_a 7.4 [12], [56]). However, recent studies suggest that a chlorine gas (Cl₂)-like oxidant, and not HOCl/OCl⁻, serves as the proximate chlorinating intermediate in chlorotyrosine formation from the free amino acid [15]. HOCl is in equilibrium with Cl₂ via a reaction that requires Cl⁻ and H⁺ (Eqn. 2). The formation of Cl₂ is thus favored by acidic pH and the

$$\text{HOCl+Cl}{}^{\mathrm{-}}\text{+H}{}^{\mathrm{+}}\text{=Cl}{}_2\text{+ H}{}_2\text{O(pK}{}_{\mathrm{a}}\text{3.3;[56])}$$

presence of Cl⁻, suggesting that under acidic conditions such as in the phagolysosome, Cl₂ could potentially execute oxidation/halogenation reactions normally ascribed to HOCl/ClO⁻ [15]. Consonant with this notion, the HOCl-dependent oxidation of a variety of biological compounds including free tyrosine has a requirement for Cl⁻ and is augmented by acidic pH [15], [58], [62]. Moreover, GC/MS analyses of condensates of head space gas over acidic reaction mixtures containing MPO, H₂O₂ and Cl⁻ demonstrate that MPO generates Cl₂ [15]. The proximate chlorinating oxidant formed during MPO-mediated chlorination of protein tyrosine residues in vivo is not established. The acidic pH optima and requirement for Cl⁻ in HOCl/OCl⁻-dependent chlorination of protein tyrosine residues suggests that a Cl₂-like oxidant is involved. In addition to Cl₂, a potential role for sequence-specific monochloramines as intermediates has been suggested [63].

The biological activities of LDL modified by MPO-generated chlorinating oxidants is unknown. Exposure of LDL to bolus additions of reagent HOCl promotes lipoprotein aggregation [64] and conversion into a high uptake particle for macrophages via phagocytosis [65]. However, recent studies demonstrated that exposure of LDL to a comparable amount of HOCl generated by MPO with a low continuous flux of H₂O₂, as might occur in vivo, failed to induce lipoprotein aggregation and conversion into a high uptake form [32]. Similarly, exposure of LDL to bolus additions of reagent HOCl in the presence of NO₂⁻ [66] promotes lipoprotein aggregation, but not when a continuous low steady flux of oxidant is used [32].

GC/MS and multinuclear NMR studies examining the structures of tyrosine oxidation products formed by the MPO-H₂O₂-Cl⁻ system of neutrophils demonstrated that at neutral pH very little (< 1% yield) chlorotyrosine was formed. Rather, the major product formed was an amphipathic aldehyde, p-hydroxyphenylacetaldehyde (pHA), that accounted for over 90% of the H₂O₂ consumed by the enzyme [67]. Similar reactions (i.e., aldehyde formation) were subsequently demonstrated for virtually all other common α-amino acids [68]. The mechanism for aldehyde formation has been elaborated and involves initial formation of a labile α-monochloramine intermediate that subsequently decomposes into an aldehyde during the concerted loss of CO₂ and NH₄⁺ [69]. Multiple covalent adducts formed following reaction of some of these MPO-generated aldehydes with nucleophilic moieties on proteins and lipids have been identified in human vascular tissues. For example, immunohistochemical and mass spectrometric studies demonstrate that acrolein, an aldehyde which can be generated by exposure of threonine to either HOCl or the MPO-H₂O₂-Cl⁻ system of neutrophils [68], [70], covalently modifies proteins lysine residues in human atheroma [71]. Likewise, glycolaldehdye, the product of serine oxidation by MPO-generated HOCl/OCl⁻ [68], [70], serves as an intermediate in formation of advanced glycosylation end products like carboxymethyl lysine [70], a product observed in human atheroma [72]. GC/MS studies have also confirmed that Schiff base adducts between pHA and both the ε-amino moiety of protein lysine residues and amino lipids are present in human atherosclerotic tissues and other inflammatory sites [73], [74], [75], [76]. The biological consequences of MPO-dependent generation of amino acid-derived aldehydes is unknown. It is interesting to note that many of the aldehydes are formed by alternative pathways and are known cytotoxins (e.g., acrolein, acetaldehyde) [77], [78], while others have been used as vaccination adjuvants because of their ability to modulate T cell differentiation and phenotype (e.g., pHA [79]).

GC/MS studies demonstrate that dityrosine, a radical-radical addition product formed between two tyrosyl radical species, is enriched in both LDL recovered from human vascular lesions and intima from early atherosclerotic lesions [47]. MPO-dependent formation of tyrosyl radical in vitro has been demonstrated through electron spin resonance spectroscopy [80] and structural identification of dityrosine and other tyrosyl radical addition products [81]. Exposure of (lipo)proteins to MPO-generated tyrosyl radical promotes protein cross-linking [14] and initiation of lipid peroxidation [31], features characteristic of (lipo)proteins and lipids recovered from human atheroma. Modification of high density lipoprotein with an MPO-generated tyrosyl radical system enhances the ability of the lipoprotein to support reverse cholesterol transport [82], [83]. Understanding the biological consequences of tyrosyl radical-dependent oxidation of cellular targets remains essentially unexplored.

Both immunohistochemical [84], [85] and GC/MS studies [46] demonstrate that nitotyrosine is enriched in human atherosclerotic intima and LDL recovered from human atheroma. Detection of nitrotyrosine in tissues had previously been used as evidence for the in vivo generation of peroxynitrite (ONOO⁻), a potent nitrating intermediate formed by interaction of NO and O₂^•− (Fig. 3, pathway C) [86], [87]. However, recent studies demonstrate the existence of at least two alternative pathways for nitrotyrosine formation, both through MPO dependent generation of reactive nitrogen species. The first (Fig. 3, pathway A) involves MPO-dependent oxidation of nitrite (NO₂⁻), a stable end-product of NO metabolism, forming a reactive nitrogen species, presumably nitrogen dioxide (NO₂) [17], [18], [24]. The second (Fig. 3, pathway B) involves secondary oxidation of NO₂⁻ by MPO-generated HOCl [17], [88]. Studies with both isolated MPO and human neutrophils demonstrate that free and protein-bound tyrosine residues can be converted into nitrotyrosine by these pathways [17], [18], [24], [32], [88]. However, the role of secondary oxidation of NO₂⁻ by HOCl (Fig. 3, pathway B) in promoting aromatic nitration reactions in the neutrophil phagolysosome [89] and by monocytes [18] has been questioned. Mass spectrometric studies demonstrate that the MPO-H₂O₂-NO₂⁻ pathway effectively initiates lipid peroxidation in complex biological mixtures like human serum [33]. Finally, recent studies examining the pathways available to monocytes for generating NO-derived oxidants identify MPO-dependent oxidation of NO₂⁻ as a primary mechanism for promoting LDL protein nitration and initiating LDL lipid peroxidation, particularly at low rates of NO flux [18] (Fig. 3).

The relative contribution of MPO to nitrotyrosine formation on LDL in the human artery wall is not known. Likewise, the biological consequences of LDL nitration are not fully established. LDL modified by millimolar levels of reagent ONOO⁻ is aggregated and converted into a high uptake particle for macropages via phagocytosis [32], [90]. LDL modified by reactive nitrogen species generated by the MPO-H₂O₂-NO₂⁻ system of activated human monocytes is converted into a form (NO₂-LDL) that is avidly taken up and degraded by macrophages, resulting in cholesterol deposition and foam cell formation, essential steps in lesion development [32]. The macrophage scavenger receptor CD36 is the major receptor responsible for high affinity and saturable cellular recognition of NO₂-LDL by murine and human macrophages [91]. The potential physiological significance of this pathway was further illustrated by demonstrating that modification of LDL by the MPO-H₂O₂-NO₂⁻ system in the presence of up to 80% lipoprotein deficient serum (LPDS) only partially inhibited conversion of the lipoprotein into a ligand for macrophages via CD36, whereas addition of < 5% lipoprotein deficient serum to reaction mixtures totally blocked Cu²⁺-catalyzed LDL oxidation and conversion into a ligand for CD36 [91]. Whether this represents a beneficial activity, such as removal of excess LDL from the subendothelium, a harmful one, such as leading to lesion development if the process occurs in excess, or both, remains to be established.