Myeloperoxidase functions as a major enzymatic catalyst for initiation of lipid peroxidation at sites of inflammation.

Initiation of lipid peroxidation and the formation of bioactive eicosanoids are pivotal processes in inflammation and atherosclerosis. Currently, lipoxygenases, cyclooxygenases, and cytochrome P450 monooxygenases are considered the primary enzymatic participants in these events. Myeloperoxidase (MPO), a heme protein secreted by activated leukocytes, generates reactive intermediates that promote lipid peroxidation in vitro. For example, MPO catalyzes oxidation of tyrosine and nitrite to form tyrosyl radical and nitrogen dioxide ((.)NO(2)), respectively, reactive intermediates capable of initiating oxidation of lipids in plasma. Neither the ability of MPO to initiate lipid peroxidation in vivo nor its role in generating bioactive eicosanoids during inflammation has been reported. Using a model of inflammation (peritonitis) with MPO knockout mice (MPO(-/-)), we examined the role for MPO in the formation of bioactive lipid oxidation products and promoting oxidant stress in vivo. Electrospray ionization tandem mass spectrometry was used to simultaneously quantify individual molecular species of hydroxy- and hydroperoxy-eicosatetraenoic acids (H(P)ETEs), F(2)-isoprostanes, hydroxy- and hydroperoxy-octadecadienoic acids (H(P)ODEs), and their precursors, arachidonic acid and linoleic acid. Peritonitis-triggered formation of F(2)-isoprostanes, a marker of oxidant stress in vivo, was reduced by 85% in the MPO(-/-) mice. Similarly, formation of all molecular species of H(P)ETEs and H(P)ODEs monitored were significantly reduced (by at least 50%) in the MPO(-/-) group during inflammation. Parallel analyses of peritoneal lavage proteins for protein dityrosine and nitrotyrosine, molecular markers for oxidative modification by tyrosyl radical and (.)NO(2), respectively, revealed marked reductions in the content of nitrotyrosine, but not dityrosine, in MPO(-/-) samples. Thus, MPO serves as a major enzymatic catalyst of lipid peroxidation at sites of inflammation. Moreover, MPO-dependent formation of (.)NO-derived oxidants, and not tyrosyl radical, appears to serve as a preferred pathway for initiating lipid peroxidation and promoting oxidant stress in vivo.

A characteristic feature of inflammation is the concomitant peroxidation of lipids and formation of bioactive lipid peroxidation products (1)(2)(3). Often considered a deleterious process, peroxidation of membrane and lipoprotein phospholipids has been linked to disruption of biomembranes and cellular dysfunction. When localized to specific cellular compartments and under strict regulatory control, generation of lipid oxidation products of polyunsaturated fatty acids, both in their free and esterified forms, is also linked to generation of potent chemoattractants, pro-inflammatory eicosanoids, and various oxidized species that serve as signaling molecules (4 -11). Thus, lipid peroxidation is in some settings a normal physiological process, whereas in others is a potential participant in the pathophysiological sequelae of acute and chronic inflammatory diseases (6 -13).
The primary enzymatic participants involved in lipid peroxidation in vivo are not established. Lipoxygenases, cyclooxygenases, and cytochrome P450 mono-oxygenases are widely thought to serve as the major enzymatic pathways involved (2, 6, 14 -17). These enzymes are expressed in leukocytes and catalyze the direct insertion of molecular oxygen into polyenoic fatty acids, forming hydroperoxides, which are both rapidly reduced to their corresponding alcohols and converted into more advanced oxidation products such as prostaglandins and leukotrienes. For example, hydroperoxy-eicosatetraenoic acids (HPETEs) 1 and their corresponding reduced forms, hydroxyeicosatetraenoic acids (HETEs), are initial products of arachidonic acid metabolism (extensively reviewed in Ref. 18). These species have potent pro-inflammatory actions and serve as precursors for numerous eicosanoids (18). Their formation has predominantly been attributed to lipoxygenase pathways, and to a lesser extent, cyclooxygenase and cytochrome P450 monooxygenase pathways (2, 14 -16). Three isomers, 5-, 12-, and 15-, constitute the main isomers of HETEs and HPETEs (i.e. H(P)ETEs) detected in media from cultured cells and are synthesized by corresponding 5-, 12-, and 15-lipoxygenases (18). The contribution of pathway(s) alternative to lipoxygenases, cyclooxygenases, and cytochrome P450s to the initiation of lipid oxidation in vivo has not yet been fully defined.
We have suggested that myeloperoxidase (MPO), a heme protein secreted by activated neutrophils, monocytes, and some macrophages, may serve as an alternative enzymatic participant in the initiation of lipid peroxidation in vivo (3, 19 -21, 23). Although no studies to date have directly examined the role of MPO in lipid peroxidation in vivo, numerous in vitro studies support this hypothesis. Lipid and lipoprotein oxidation have been documented using model systems containing purified MPO (10,19,24) and both isolated human neutrophils (19,23,25) and monocytes (20,21,26), suggesting that MPO-generated diffusible radical species may similarly initiate lipid peroxidation in vivo.
One reactive intermediate formed by MPO that may serve as a potential initiator of lipid peroxidation is tyrosyl radical (24), a resonance-stabilized intermediate capable of both initiating lipid peroxidation and forming stable phenolic cross-links on proteins (27)(28)(29). Protein-bound dityrosine, a characteristic molecular fingerprint, is enriched in human atheroma and at other sites of inflammation (29 -31), suggesting that MPO-generated tyrosyl radical may participate in lipid peroxidation in vivo.
An alternative diffusible radical species formed by MPO that may participate in lipid peroxidation is nitrogen dioxide ( ⅐ NO 2 ), a product of peroxidase-H 2 O 2 -nitrite (NO 2 Ϫ ) systems (32). ⅐ NO 2 is a lipophilic oxidant that efficiently abstracts a bis-allylic hydrogen atom initiating lipid peroxidation in model systems (19,(33)(34)(35) and following inhalation exposures (36). ⅐ NO 2 may also participate in protein oxidation forming nitrotyrosine, a post-translational modification specific for reactive nitrogen species (32,(37)(38)(39). Nitrotyrosine, like dityrosine, is enriched in human atheroma and other sites of inflammation (32, 40 -42). A role for MPO in the generation of ⅐ NO-derived oxidants in vivo has recently been shown by demonstrating suppression in nitrotyrosine formation in tissues of MPO knockout (MPO Ϫ/Ϫ ) mice in several acute inflammatory models (32,43,44). A role for MPO-generated ⅐ NO 2 in initiation of lipid peroxidation in vivo has yet to be examined.
Although the above data are consistent with the hypothesis that MPO may participate in lipid and lipoprotein peroxidation in vivo, numerous studies have questioned this role for the hemoprotein. For example, only modest differences in lipid hydroperoxide formation were reported in a recent study (45) examining the capacity of leukocytes isolated from wild type and MPO Ϫ/Ϫ mice to promote oxidation of low density lipoprotein in model systems ex vivo. Further, an inhibitory effect of NO 2 Ϫ has been reported when monitoring MPO-catalyzed oxidation of low density lipoprotein (46). Antioxidant rather than pro-oxidant effects of MPO-generated tyrosine oxidation products during low density lipoprotein oxidation have similarly been proposed (47,48). It has also been proposed that MPO functions not to generate biologically active lipids but rather to destroy them through oxidative mechanisms, thereby altering inflammatory responses (49). Thus, the role of MPO in initiation of lipid peroxidation in vivo is uncertain, and only through direct experimental investigation will the contribution of MPO to this process be enumerated.
Here we employ a combination of murine models of inflammation and quantification of specific lipid oxidation products to define the role of MPO in the initiation of lipid peroxidation in vivo. Further, through parallel analyses of protein molecular markers of distinct oxidation pathways, the relative contributions of MPO-generated tyrosyl radical and reactive nitrogen species pathways to lipid peroxidation are assessed.

EXPERIMENTAL PROCEDURES
Chemicals and Solvents-Free fatty acids were purchased from Cayman Chemical Company (Ann Arbor, MI). Organic solvents were obtained from Fisher Scientific Co. (Pittsburgh, PA). All other reagents were purchased from Sigma unless otherwise indicated.
General Procedures-All buffers were passed over a Chelex-100 resin column (Bio-Rad) and supplemented with 0.1 mM diethylenetriamine pentaacetic acid (DTPA) to remove potential contaminant transition metal ions that might catalyze lipid oxidation. Protein content was determined using Bradford-based Bio-Rad protein assay using IgG as protein standard. Wild type (MPO ϩ/ϩ ) and MPO Ϫ/Ϫ mice were on a C57BL/6J background (Ͼ98% genetic homogeneity). Age-and sexmatched MPO ϩ/ϩ and MPO Ϫ/Ϫ mice were used for all studies. Superoxide and HOCl generation in recovered peritoneal lavage cells from MPO ϩ/ϩ and MPO Ϫ/Ϫ mice were performed as recently described (50). Nitrotyrosine, chlorotyrosine, and dityrosine were quantified by stable isotope dilution mass spectrometry-based methods as previously described (32). All animal studies were performed using approved protocols from the Animal Research Committee of the Cleveland Clinic Foundation.
Thioglycollate (Tg)/zymosan (Z)-induced Peritonitis Model-MPO ϩ/ϩ and MPO Ϫ/Ϫ mice were injected intraperitoneally with 1 ml of 4% thioglycollate broth. Twenty hours after recruitment, mice were injected with zymosan (250 mg/kg). Peritoneal lavage was performed in control animals (20 h post intraperitoneal injection of normal saline), 20 h following thioglycollate injection, or 24 h following the combination of both thioglycollate injection (20 h) followed by zymosan injection (4 h), as indicated (Tg/Z). Peritoneal lavages were performed with phosphate-buffered saline containing antioxidant (0.1 mM butylated hydroxytoluene (BHT)) and metal chelator (2 mM DTPA), transferred to screw-capped tubes covered with argon atmosphere, and then immediately centrifuged at 1000 rpm for 10 min at 4°C. All analyses of protein and lipid peroxidation products were performed on cell-free lavage supernatants. Cell pellets were resuspended in phosphate-buffered saline containing 0.1 mM BHT and 0.1 mM DTPA, and differentials were performed on cytospin preparations of isolated cells. Cell pellets were also used for Western analyses (see below). Both cells and cell-free lavage fluid not immediately analyzed were overlaid with argon, snap frozen in liquid nitrogen, and stored at Ϫ80°C until analysis.
Western Analysis of Expression of Cyclooxygenase-1(COX-1), Cyclooxygenase-2(COX-2), and 12-Lipoxygenase (12-LO)-Cell pellets recovered from peritoneal lavages were lysed, concentration measured, and 25 g total protein loaded per lane in 10% Tris-HCl gel. Following SDS-PAGE analysis, proteins were transferred to polyvinylidene difluoride membrane and probed with rabbit antibodies against mouse COX-1, COX-2, and 12-LO (all from Cayman Chemical) and affinity purified rabbit antibody generated against a peptide fragment of actin (Sigma). Detection was performed using secondary anti-rabbit antibodies conjugated with horseradish peroxidase for chemiluminescent detection. Quantification of bands was performed by densitometric analysis. Band intensities were normalized to actin to account for loading variations between lanes. The mean relative band intensity (normalized to actin) from MPO ϩ/ϩ mice were assigned a relative value of 1.0 for comparisons between MPO ϩ/ϩ and MPO Ϫ/Ϫ mice.
Lipid Extraction-All steps were performed under either argon or nitrogen atmosphere. Peritoneal lavage volume was reduced under vacuum centrifugation prior to extraction and preparation for mass spectrometry analysis. Immediately prior to extraction, 10 ng each of two deuterated internal standards, 12(S)-hydroxy-5,8,10,14-eicosatetraenoic-5,6,8,9,11,12,14,15-d8 acid (12-HETE-d 8 ) and prostaglandin F 2␣ (PGF 2␣ -d 4 ) (Cayman Chemical Company) were added to each sample. Hydroperoxides in samples were then reduced to their corresponding stable alcohols by adding 1 mM SnCl 2 (23). Lipids were then extracted by adding a solvent mixture (1 M acetic acid/2-isopropanol/ hexane (2:20:30, v/v/v) to the sample at a ratio of 2.5 ml of solvent mixture/1 ml of sample, vortexing, and then adding 2.5 ml hexane. Following vortex and centrifugation, lipids were recovered in the hexane layer. Peritoneal lavages were re-extracted by addition of an equal volume of hexane, followed by vortex and centrifugation. The combined hexane layers were dried under N 2 flow and then analyzed following saponification to release all esterified fatty acid oxidation products. For saponification, N 2 -dried lipids were resuspended in 1.5 ml of 2-isopropanol, and then fatty acids were released by base hydrolysis with 1.5 ml of 0.2 M NaOH at 60°C for 30 min under argon atmosphere. Hydrolyzed samples were first acidified to pH 3.0 with 0.5 M HCl, and then fatty acids were extracted twice with 4 ml of hexane. The combined hexane layers were dried under N 2 flow, resuspended in 100 l of methanol, and stored under argon at Ϫ80°C until analysis by reverse phase high performance liquid chromatography (HPLC) with on-line electrospray ionization tandem mass spectrometry (LC/ESI/MS/MS), as described below.
Lipid Analyses by LC/ESI/MS/MS-LC/ESI/MS/MS was employed to quantify the multiple distinct oxidation products of arachidonic acid and linoleic acid, including individual HETEs, F 2 -isoprostanes, and hydroxy-octadecadienoic acids (HODEs). Immediately prior to analysis, one volume of H 2 O was added to 20 volumes methanol-suspended sample, which was then passed through a 0.22-m filter (Millipore Corporation, Bedford, MA). Sample (50 l) was injected onto a C-18 column (2 ϫ 250 mm, 5 m ODS) (Phenomenex, Rancho Palos Verdes, CA) at a flow rate of 0.2 ml/min. The separation was performed using a gradient starting from 85% methanol over 15 min, then to 100% methanol over 1 min, followed by 100% methanol for 15 min. HPLC column effluent was split so that only 50% was introduced into a Quattro II triple quadrupole mass spectrometer (Micromass, Inc.). Analyses were performed using electrospray ionization in negativeion mode with multiple reaction monitoring of parent and characteristic daughter ions specific for each isomer monitored. The transitions monitored were mass-to-charge ratio (m The nitrogen curtain gas-flow rate was 5 liters/min and nebulizer gas-flow rate was held at 0.2 liters/min. Collision-induced dissociation was obtained using argon gas. The internal standard 12-HETE-d 8 was used to for quantification of HETEs as well as to calculate extraction efficiencies of HODEs and HETEs (which were Ͼ 85%). The internal standard PGF 2␣ -d 4 was used to for quantification of F 2 -isoprostanes.
Statistical Analysis-Data are presented as mean Ϯ S.D. Comparisons between groups were done using unpaired Student's t test. p Ͻ 0.05 was considered significant between groups.

MPO Serves as a Major Source of Oxidant Stress During
Inflammation-F 2 -isoprostanes are a family of prostaglandinlike oxidation products of arachidonic acid that are predominantly generated by free radical-mediated processes (51,52). Elevated in a variety of inflammatory disorders, they are recognized as markers of oxidant stress in vivo (52,53). In initial studies we sought to determine the role of MPO in promoting enhanced oxidant stress during inflammation. A peritonitis model was selected for these investigations because, in prior studies (32,50), we have shown that MPO is catalytically active in this model using mass spectrometry to monitor protein bound chlorotyrosine formation, a specific marker of protein modification by MPO-generated chlorinating oxidants (54). Moreover, we have shown that this model results in comparable leukocyte recruitment into the peritoneum in mice from the two groups and comparable superoxide generation in elicited leukocytes in MPO ϩ/ϩ and MPO Ϫ/Ϫ mice (32,50). Consistent with these findings, analyses of leukocytes recovered from lavages analyzed for lipid oxidation products in this study (as described below) at baseline, following thioglycollate recruitment, and following subsequent zymosan challenge, demonstrated no significant differences in total cell count, differentials, or superoxide generation between MPO ϩ/ϩ and MPO Ϫ/Ϫ mice, whereas only those harvested from MPO ϩ/ϩ mice produced HOCl (data not shown).
LC/ESI/MS/MS was employed to quantify F 2 -isoprostanes levels at baseline, following thioglycollate-elicited recruitment of leukocytes, and following subsequent leukocyte activation by intraperitoneal injection of yeast spore coat (zymosan) (Fig. 1). Levels of F 2 -isoprostanes were similar in MPO ϩ/ϩ and MPO Ϫ/Ϫ mice at baseline and following leukocyte recruitment with thioglycolate. However, subsequent activation of leukocytes by intraperitoneal injection of zymosan resulted in marked production of F 2 -isoprostanes in MPO ϩ/ϩ , but not MPO Ϫ/Ϫ , mice. Inflammation-triggered formation of F 2 -isoprostanes was reduced by ϳ85% in the MPO Ϫ/Ϫ mice compared with that observed in MPO ϩ/ϩ mice (Fig. 1). Comparable results (i.e. 75-85% reduction in MPO Ϫ/Ϫ mice) were observed when F 2 -isoprostanes measured were expressed in units of total mass, normalized to either animal size or elicited leukocyte number in lavage (data shown in Fig. 1A is expressed per gram of body weight to account for any differences in mouse sizes), or expressed as a product/precursor ratio (i.e. F 2 -isoprostanes/ arachidonic acid, mmol/mol; Fig. 1B).

Myeloperoxidase Functions as a Major Enzymatic Catalyst for Initiation of Lipid Peroxidation at Sites of Inflammation-
The above results strongly suggest that MPO serves as a primary catalyst for initiating lipid peroxidation in vivo during acute inflammatory insults. To confirm this hypothesis we sought to monitor multiple distinct species of H(P)ETEs and H(P)ODEs as indices of lipid peroxidation in vivo. HPLC with LC/ESI/MS/MS methods were developed for the simultaneous quantification of specific isomers based upon monitoring unique transitions between the mass-to-charge ratio of the parent ion [M-H]Ϫ and characteristic daughter ions for each species and their appropriate retention time ( Fig. 2A). Addition of heavy isotope-labeled internal standard (12-HETE-d 8 ) to samples prior to sample preparation permitted adjustment for lipid extraction efficiency, and standard curves (Fig. 2B) were constructed with each authentic isomer to account for differences in extraction efficiency and ionization efficiencies for each unique parent 3 daughter ion transition monitored.
Similar to results observed with F 2 -isoprostanes, comparable levels of each HETE monitored were noted in lavage fluid recovered at baseline and following leukocyte recruitment with

FIG. 1. MPO plays a dominant role in formation of F 2 -isoprostanes at sites of inflammation.
Lipids in peritoneal lavage of wild type (WT) and MPO knockout (MPO Ϫ/Ϫ ) mice were extracted and analyzed as described in "Experimental Procedures." Intraperitoneal injection of thioglycollate (Tg) was used to elicit leukocytes to the peritoneal cavity, and yeast spore coat (zymosan, Z) was used to activate recruited peritoneal leukocytes. Animals were injected intraperitoneally with normal saline served as controls (Ctrl). A, data represent total F 2isoprostanes recovered in peritoneal lavage per gram mouse weight. intraperitoneal injection of thioglycollate, in MPO ϩ/ϩ and MPO Ϫ/Ϫ mice (Fig. 3). Leukocyte activation with intraperitoneal injection of yeast spore coat resulted in marked increases in formation of each HETE in MPO ϩ/ϩ , but not MPO Ϫ/Ϫ mice. Presentation of HETE levels, either as total produced (normalized per gram of body weight to account for any differences in mouse sizes, Fig. 3A) or as a product/precursor ratio (i.e. HETE/ arachidonic acid, mol/mol; Fig. 3B), yielded comparable results, supporting a major role for MPO in initiation of lipid peroxidation in vivo. Analyses of oxidation products (i.e. H(P)ODEs) derived from linoleic acid yielded comparable results, with marked reductions noted in every HODE isomer monitored in samples recovered from the MPO Ϫ/Ϫ mice (Fig. 4).

Alterations in Expression of Cyclooxygenases and Lipoxygenases Does Not Account for the Marked Reduction in Lipid Peroxidation Noted in MPO
Ϫ/Ϫ Mice-The preceding results strongly support a role for MPO as the major catalyst for initiation of lipid peroxidation at sites of inflammation. The near uniform distribution of various isomers of HETEs and HODEs formed during inflammation in itself is also consistent with a free radical, rather than regioselective (i.e. like lipoxygenases and cyclooxygenases), oxidation process as the mechanism involved. None the less, we thought it important to confirm that levels of cyclooxygenases and lipoxygenases present in elicited and activated peritoneal leukocytes in MPO ϩ/ϩ and MPO Ϫ/Ϫ mice are comparable, given the widely held belief that cyclooxygenases and lipoxygenases serve as the primary enzymatic pathways for initiation of lipid peroxidation and formation of eicosanoids at sites of inflammation. To examine whether the levels of these enzymes are changed in mice with targeted deletion for the MPO gene, the expression of COX-1, COX-2, and 12-LO in peritoneal lavage were analyzed by Western blot. Illustrative Western blot analyses for MPO ϩ/ϩ and MPO Ϫ/Ϫ mice (Fig. 5A) and quantification of levels for large groups of animals (n Ն 8 each genotype) are shown (Fig. 5B). No significant differences were noted in expression levels of COX-1, COX-2, and 12-LO, further confirming that lipid peroxidation and formation of bioactive lipids at the site of inflammation is predominantly mediated by MPO.

MPO-generated Reactive Nitrogen Species, but Not Tyrosyl Radical, Serves as a Pathway for MPO-catalyzed Initiation of Lipid Peroxidation in a Peritonitis Model of Inflammation-In a final series of studies, we sought to define which reactive intermediate(s) serves as the relevant participant(s) in MPOinitiated lipid peroxidation in vivo.
In a recent study (23) we employed a systematic approach to isolate and chemically define the low molecular weight components in plasma capable of enabling MPO to initiate peroxidation of plasma lipids. Tyrosine and nitrite, and to a lesser degree thiocyanite, were the major plasma components used by MPO as substrate to support lipid oxidation. Dityrosine and nitrotyrosine are stable markers of protein oxidative modification by MPO-generated tyrosyl radical and ⅐ NO 2 , the reactive intermediates formed during the 1 electron oxidation of tyrosine and nitrite, respectively. We therefore quantified alterations in levels of proteinbound dityrosine and nitrotyrosine in the same peritoneal lavage supernatants monitored for alterations in lipid peroxidation products. Dityrosine levels increased substantially following inflammatory challenge in MPO ϩ/ϩ mice. Remarkably, marked increases in dityrosine production were also noted in MPO Ϫ/Ϫ mice. Importantly, no significant differences were noted in all comparisons for dityrosine levels observed in samples recovered from MPO ϩ/ϩ versus MPO Ϫ/Ϫ mice (Fig. 6A). In marked contrast, nitrotyrosine levels were significantly reduced (p ϭ 0.046) following leukocyte activation with intraperitoneal zymosan in the MPO Ϫ/Ϫ mice (Fig. 6B), consistent with prior findings (32). To confirm that MPO was catalytically active in the model, protein-bound chlorotyrosine levels were simultaneously monitored by stable isotope dilution gas chromatography mass spectrometry analyses. Activation of MPO in MPO ϩ/ϩ mice was confirmed by dramatic increases in chlorotyrosine formation (Fig. 6C). As expected, no detectable chlorotyrosine was observed in samples recovered from all MPO Ϫ/Ϫ mice. DISCUSSION The present study provides unambiguous evidence that MPO plays a dominant role in the initiation of lipid peroxidation in vivo. Marked reductions in formation of multiple distinct isomers mice (i.e. following intraperitoneal injection with thioglycollate to elicit leukocytes to the peritoneal cavity, and then yeast spore coat (zymosan) to activate recruited peritoneal leukocytes). Pellets were then analyzed by SDS-PAGE and Western blot for the indicated proteins as described under "Experimental Procedures." A, typical Western analyses using antibodies against COX-1, COX-2, 12-LO and actin. B, quantification of COX-1, COX-2, and 12-LO in peritoneal lavage pellets. For each enzyme examined by Western, the band intensity was normalized to that of actin intensity to account for sample loading variation, and mean levels observed in WT mice were normalized to a value of 1.0 for relative comparisons with MPO Ϫ/Ϫ mice. Results represent mean Ϯ S.D. of at least 8 animals in each group.
of H(P)ETEs and H(P)ODEs, lipid peroxidation products of arachidonic acid and linoleic acid, respectively, were noted in inflammatory fluids recovered from MPO Ϫ/Ϫ versus MPO ϩ/ϩ mice. The LC/ESI/MS/MS methods employed are both sensitive and specific and confirmed parallel alterations in numerous independent lipid peroxidation products in inflammatory tissues. Further, extensive characterization of the model employed demonstrated comparable cellular and enzymatic components that might impact upon lipid peroxidation in peritoneal lavage fluids analyzed from both MPO Ϫ/Ϫ and MPO ϩ/ϩ mice. These included similar leukocyte recruitment (total cell count and differential), capacity for cell-mediated O 2 . formation, and levels of alternative potential enzymatic participants in lipid peroxidation, including COX-1, COX-2, and 12-LO. The present studies also demonstrate a principal role for MPO in promotion of oxidant stress at sites of inflammation, as monitored by F 2 -isoprostane formation. F 2 -isoprostanes are widely used as a systemic marker of oxidant stress and have been shown to be enhanced in numerous chronic inflammatory and degenerative diseases (52,53). F 2 -isoprostane formation above baseline levels in response to inflammatory challenge was almost exclusively MPO-mediated, as monitored by stable isotope dilution LC/ESI/MS/MS analyses in peritoneal lavages recovered from MPO Ϫ/Ϫ versus MPO ϩ/ϩ mice. The present studies thus strongly suggest a major role for MPO in promoting oxidant stress in inflammatory diseases and suggest a closer inspection of the potential role of MPO in inflammatory disorders.
By simultaneously monitoring production of a combination of both protein and lipid oxidation products, we were able to provide insights into the potential mechanisms through which MPO promotes lipid oxidation in vivo. In contrast to anticipated results, no changes in protein dityrosine could be detected in extracellular proteins recovered in inflammatory fluid lavages recovered from MPO Ϫ/Ϫ versus MPO ϩ/ϩ mice. It is thus hard to implicate MPO-generated tyrosyl radical as a mechanism contributing to initiation of lipid peroxidation in extracellular targets in vivo, at least for the peritonitis model examined in the present study. In a recent study (55) that employed a comparable murine peritonitis model, dityrosine levels in mice with functional deficiency in the NADPH oxidase complex were noted to have marked reductions in leukocyte and urinary levels of dityrosine formed. It was concluded that MPO-catalyzed formation of dityrosine contributes significantly to oxidative cross-linking and lipid peroxidation of extracellular targets in vivo (55,56). The present studies, however, indicate that the proposed role of MPO in dityrosine cross-linking, or for generating tyrosyl radicals that initiate lipid peroxidation, though logical inferences based upon results of in vitro studies (23,24), were erroneous. Rather, the present results, combined with the recent studies reported using NADPH oxidase-null mice (55) . , is capable of forming dityrosine crosslinks (30,58), and might also participate in the rise in dityrosine observed following leukocyte activation with intraperitoneal inoculum of zymosan. However, the significant reductions in nitrotyrosine observed in MPO Ϫ/Ϫ versus MPO ϩ/ϩ mice suggest that MPO is a major source of ⅐ NO-derived oxidants in this model. Thus, a primary role for MPO-generated ⅐ NO 2 as the reactive nitrogen species that promotes lipid peroxidation and oxidant stress in vivo is likely. ⅐ NO 2 is lipophilic and readily abstracts a bisallylic hydrogen atom from membrane lipids (19,33). Formation of ⅐ NO-derived oxidants in vivo by MPO may thus play a significant role in both acute and chronic inflammatory processes where lipid peroxidation and enhanced oxidant stress have been noted. Two years ago we reported that MPO and other members of the mammalian heme peroxidase superfamily possess the remarkable ability to utilize ⅐ NO as a physiological substrate (59). Stopped-flow kinetics studies demonstrated multiple ⅐ NOperoxidase interactions (59,60), and subsequent organ chamber studies with preconstricted vascular (61) and tracheal rings (62) revealed that catalytic amounts of peroxidase were capable of intercepting ⅐ NO, preventing smooth muscle relaxation, and NO-mediated ring dilation. In more recent studies (63), these earlier findings were repeated and confirmed. ⅐ NO is a potent terminator of lipid peroxidation (22,64). Thus, an additional mechanisms through which MPO may participate in promotion FIG. 6. MPO ؊/؊ mice generate significantly less nitrotyrosine, but not dityrosine, than WT mice in a peritonitis model of acute inflammation. Intraperitoneal injection of thioglycollate (Tg) was used to elicit leukocytes to the peritoneal cavity, and yeast spore coat (zymosan, Z) was used to activate recruited peritoneal leukocytes. Peritoneal lavage was then recovered from WT and MPO Ϫ/Ϫ mice as described in "Experimental Procedures." Animals injected intraperitoneally with normal saline served as controls (Ctrl). Protein-bound dityrosine (A), nitrotyrosine (B), and chlorotyrosine (C) in peritoneal lavage were prepared and analyzed in supernatants as described in "Experimental Procedures." Data are represented as values proportional to the total amount of each amino oxidation product recovered, as monitored by the ratio of each amino acid oxidation product to its precursor tyrosine ϫ the protein concentration recovered from peritoneal lavage. Results represent mean Ϯ S.D. of at least 6 animals in each group. of both lipid peroxidation and generation of ⅐ NO-derived oxidants is via direct scavenging of ⅐ NO by active peroxidase at the site of inflammation.
The present studies invite a reappraisal of the enzymatic participants in lipid peroxidation in vivo, with MPO added to (and even superceding in some settings) the ranks of enzymes like lipoxygenases, cyclooxygenases, and cytochrome P450 complexes. They also show that complex and unanticipated patterns of protein oxidation products from distinct pathways can be observed in otherwise "simple" inflammatory models. Finally, the present studies again point both to the importance of ⅐ NO-MPO interactions in inflammation as well as to the need for development of inhibitors for MPO as a novel anti-inflammatory therapeutic.