2-Thioxanthines Are Mechanism-based Inactivators of Myeloperoxidase That Block Oxidative Stress during Inflammation*

Background: The enzyme myeloperoxidase produces chlorine bleach at sites of inflammation. Results: 2-Thioxanthines are potent mechanism-based inactivators of myeloperoxidase. Conclusion: 2-Thioxanthines block production of chlorine bleach during inflammation. Significance: Mechanism-based inactivators of myeloperoxidase should limit oxidative stress at sites of inflammation. Myeloperoxidase (MPO) is a prime candidate for promoting oxidative stress during inflammation. This abundant enzyme of neutrophils uses hydrogen peroxide to oxidize chloride to highly reactive and toxic chlorine bleach. We have identified 2-thioxanthines as potent mechanism-based inactivators of MPO. Mass spectrometry and x-ray crystal structures revealed that these inhibitors become covalently attached to the heme prosthetic groups of the enzyme. We propose a mechanism whereby 2-thioxanthines are oxidized, and their incipient free radicals react with the heme groups of the enzyme before they can exit the active site. 2-Thioxanthines inhibited MPO in plasma and decreased protein chlorination in a mouse model of peritonitis. They slowed but did not prevent neutrophils from killing bacteria and were poor inhibitors of thyroid peroxidase. Our study shows that MPO is susceptible to the free radicals it generates, and this Achilles' heel of the enzyme can be exploited to block oxidative stress during inflammation.

Myeloperoxidase (MPO) is a prime candidate for promoting oxidative stress during inflammation. This abundant enzyme of neutrophils uses hydrogen peroxide to oxidize chloride to highly reactive and toxic chlorine bleach. We have identified 2-thioxanthines as potent mechanism-based inactivators of MPO. Mass spectrometry and x-ray crystal structures revealed that these inhibitors become covalently attached to the heme prosthetic groups of the enzyme. We propose a mechanism whereby 2-thioxanthines are oxidized, and their incipient free radicals react with the heme groups of the enzyme before they can exit the active site. 2-Thioxanthines inhibited MPO in plasma and decreased protein chlorination in a mouse model of peritonitis. They slowed but did not prevent neutrophils from killing bacteria and were poor inhibitors of thyroid peroxidase. Our study shows that MPO is susceptible to the free radicals it generates, and this Achilles' heel of the enzyme can be exploited to block oxidative stress during inflammation.
Oxidative stress is invariably associated with inflammation. This is particularly so when the inflammatory infiltrate is dominated by neutrophils, which have an enormous capacity to generate reactive oxygen species (1). Myeloperoxidase (MPO), 2 a heme enzyme of neutrophils, plays a pivotal role in producing oxidants and has a growing reputation for promoting oxidative stress in numerous inflammatory diseases (2,3). It uses hydrogen peroxide to catalyze the production of hypohalous acids as well as a plethora of free radicals (4). These reactive intermediates readily oxidize lipids, proteins, and DNA (3,5,6). Hypochlorous acid and hypobromous acid are kinetically the most reactive two-electron oxidants produced in vivo (7). Their reactivity with biomolecules is orders of magnitude greater than that of peroxynitrite and hydrogen peroxide. Hypochlorous acid is a potent toxin, and at low levels it activates stress response pathways within cells (8). Identification of specific biomarkers of hypochlorous acid at sites of inflammation has confirmed that MPO contributes to protein damage in cystic fibrosis (9), atherosclerosis (10), atrial fibrillation (11), lung disease of prematurity (12), and sepsis (13). MPO has also been implicated in oxidative stress associated with chronic obstructive pulmonary disease, rheumatoid arthritis, atherogenesis, Parkinson disease, and Alzheimer disease (3).
The compelling evidence that MPO produces damaging oxidants at sites of inflammation has focused attention on it as a pharmacological target. Currently, there is no effective inhibitor of the enzyme and limited appreciation of the best routes to block its activity. When hydrogen peroxide reacts with the ferric MPO (Reaction 1), it produces the redox intermediate compound I in which the heme iron has a formal oxidation state of 5 ϩ (4,14). Compound I of MPO is unique among mammalian peroxidases because its high two-electron reduction potential of 1.16 V (15) enables it to oxidize the halides (X Ϫ ), chloride, bromide, and iodide as well as thiocyanate, to their respective hypohalous acids (HOX; Reaction 2) (16). Its even higher oneelectron reduction potential of 1.36 V allows it to remove a single electron from myriad substrates (RH) to produce free radical intermediates (R ⅐ ) (17). In these reactions, compound I is reduced to compound II (Reaction 3), in which the heme iron has a formal oxidation state of 4 ϩ . Free radicals are also produced when substrates reduce compound II and recycle the enzyme back to its native state (Reaction 4). Physiological oneelectron reducing substrates for compound I and compound II include urate, ascorbate, nitric oxide, nitrite, serotonin, superoxide, and tyrosine (3,4,18).
Many nonsteroidal anti-inflammatory drugs and related phenols are good inhibitors of the chlorination activity of purified MPO (19). These inhibitors act by converting the enzyme to compound II, which is incapable of oxidizing chloride (19,20). It is unlikely, however, that they will affect the activity of MPO in vivo. This is because physiological substrates reduce compound II back to the ferric enzyme and limit inhibition of the enzyme (21). Paracetamol is a substrate for MPO and is oxidized via its peroxidation activity (22). At pharmacological concentrations, it competes with chloride for oxidation by compound I and inhibits production of hypochlorous acid. However, it is oxidized to toxic radicals and a quinone imine by peroxidases (23). Benzoic acid hydrazides are effective suicide substrates of MPO (24). Their action is also limited because they inactivate the enzyme only at high concentrations of hydrogen peroxide, which are unlikely to be attained in vivo.
In this study, we have identified 2-thioxanthines as suicide substrates of MPO. These inhibitors are converted to strongly oxidizing radicals that react avidly with the heme prosthetic groups of MPO before they can exit the active site of the enzyme. In this way, they prevent production of hypochlorous acid without concomitant release of free radicals. Our finding that they are active in vivo raises the prospect that mechanismbased inhibitors may prove useful in elucidating the role MPO plays in inflammatory tissue damage. They may also have potential as pharmacological agents in diseases in which MPO is shown to be a catalyst of oxidative stress.

MATERIALS AND METHODS
MPO used in in vitro activity assays was purchased from Planta Natural Products (Vienna, Austria). Lactoperoxidase were purchased from the Sigma, and human thyroid peroxidase was supplied by RSR Ltd., UK. For characterization of complexes between MPO and the 2-thioxanthines, MPO was purified from HL-60 cells, which were obtained from American Type Culture Collection (Manassas, VA). Cells were grown in DMEM/F-12 (Invitrogen) plus 5% fetal calf serum and 5 mM glutamine in a 50-liter reactor to a cell density of 1.7 ϫ 10 6 cells/ml. The purification is a modification of the protocol described previously (25). In the modified protocol, the ammonium sulfate precipitation steps were excluded, and the final purification was achieved using Superdex 200 (GE Healthcare) size exclusion chromatography. Purity and identity of MPO were determined by 10% SDS-PAGE and N-terminal sequencing.
Activity Assays for MPO-The chlorination activity of MPO was determined by measuring the production of hypochlorous acid. Reaction mixtures contained 10 nM MPO, 140 mM sodium chloride, 5 mM taurine, and various concentrations of 2-thioxanthine in 10 mM phosphate buffer, pH 7.4. Reactions were carried out at 21°C, started by adding 50 M hydrogen peroxide and stopped after 5 min by adding 20 g/ml of catalase. The concentration of accumulated taurine chloramine was assayed using iodide to catalyze the oxidation of 3,3Ј,5,5Ј-tetramethylbenzidine (27). The reaction time was chosen so that no more than 50% of the hydrogen peroxide was converted to hypochlorous acid.
The activities of MPO and thyroid peroxidase were compared by monitoring the oxidation of ascorbate at 266.5 nm (28). With MPO, the reactions were started by adding hydrogen peroxide (25 M) to MPO (5 nM) in 10 mM phosphate buffer, pH 7.4, at 21°C containing 100 M ascorbate, 140 mM sodium chloride, and 100 M diethylenetriaminepentaacetic acid, in the absence or presence of 2-thioxanthine. Conditions were the same when using thyroid peroxidase (20 nM) except reactions also contained 50 M sodium iodide. The activities of MPO and lactoperoxidase were compared by measuring their ability to consume hydrogen peroxide using a hydrogen peroxide electrode (29).
To assess whether MPO was inactivated by 2-thioxanthines, it was incubated at 100 nM in 100 mM phosphate buffer, pH 7.4, at 21°C with varying concentrations of 2-thioxanthine and 10 M hydrogen peroxide in the absence or presence of 100 mM sodium chloride. Five minutes after adding hydrogen peroxide to start the reaction, the enzyme was diluted 200-fold, and its residual peroxidation activity was measured using 3,3Ј,5,5Јtetramethylbenzidine as the reducing substrate (30). To determine the partition ratio for the thioxanthines, the residual activity of MPO was determined 10 min after incubating the enzyme with various concentrations of the inhibitors and 10 M hydrogen peroxide (30).
Stopped Flow Spectrophotometry-Kinetic studies were carried out on an Applied Photophysics SX-20MV stopped flow (Leatherhead, UK) with a xenon arc lamp. Experiments were completed in the double mixing mode. Reactions were carried out in 50 mM sodium phosphate buffer, pH 7.0. An excess of 10ϫ hydrogen peroxide was the minimum concentration required for complete conversion of MPO into compound I. All double mixing experiments used a delay time of 40 ms. For reduction of compound I by TX1, 1 M ferric MPO was premixed in the aging loop with 10 M hydrogen peroxide for 40 ms. Then compound I was then mixed with TX1. Formation of compound II was monitored at 454 nm (31). The reaction of TX1 with compound II was studied using the same conditions described above, and the conversion of compound II back to ferric MPO was monitored in the spectroscopic range 380 -700 nm from 0.5 to 200 s (see supplemental material for complete details).
Oxidant Production by Human Neutrophils-Blood was obtained from healthy donors after obtaining written informed consent, which was approved by the Southern A Regional Ethics Committee. Neutrophils were isolated from whole blood as described previously (32). The isolated cells (2 ϫ 10 6 /ml) were incubated at 37°C in 10 mM phosphate buffer, pH 7.4, containing 140 mM chloride, 5 mM taurine, 1 mM calcium chloride, 0.5 mM magnesium chloride, 1 mg/ml glucose, and various concentrations of 2-thioxanthine. Cells were stimulated by adding 100 ng/ml phorbol myristate acetate, and after 30 min the reactions were stopped by adding 20 g/ml catalase. Cells were pelleted, and the concentration of accumulated taurine chloramine was measured using iodide catalyzed oxidation of 3,3Ј,5,5Ј-tetramethylbenzidine (27). Under these conditions, cells produced ϳ50 M hypochlorous acid.
Effects of 2-Thioxanthines on Absorption Spectrum of MPO-MPO (1.6 M) was incubated at 21°C in 50 mM phosphate buffer, pH 7.4, containing 50 M 2-thioxanthine. Spectral changes were monitored after addition of 50 M hydrogen peroxide. Spectra between 200 and 700 nm were recorded every minute using an Agilent 7500 diode array spectrophotometer. Each spectrum was the average of 10 readings taken over 1 s. After 10 min, 5 g/ml catalase was added to remove residual hydrogen peroxide and the spectral change was recorded again.
Analysis of Heme Groups after Inactivation of MPO-MPO (5 M) in a volume of 200 l was incubated at room temperature (ϳ20°C) in 20 mM phosphate buffer, pH 7.4, containing 20 M TX5 and 1 mM methionine. Reactions were started by adding 20 M hydrogen peroxide and stopped after 30 min by adding 5 g/ml catalase. Samples were vacuum-dehydrated to ϳ30 l to which an equal volume of 6 M guanidine hydrochloride in 50 mM ammonium bicarbonate was added. The MPO was then denatured by heating samples at 57°C for 1 h. The denatured protein was diluted with 300 l of 50 mM ammonium bicarbonate prior to the addition of trypsin (protease/protein, 1:50) and incubated overnight at 37°C. The digestion was terminated by the addition of formic acid to a final concentration of 0.1%. The digested peptides were vacuum-dehydrated and subsequently dissolved in 100 mM Tris, pH 7.5, containing 0.5% SDS and 10 mM CaCl 2 , and then further digested with Pronase at 40°C for 115 h. The protease to protein ratio was 1:5. The digestion was stopped by the addition of formic acid (0.1% final). The resulting samples were vacuum-dehydrated and subsequently made up to 25 l with water. They were analyzed by liquid chromatography with mass spectrometry (LC/MS) and diode array detection to characterize the liberated heme. For LC/MS conditions see the supplemental material.
Crystallographic Analysis of Complexes between MPO and 2-Thioxanthines-TX2 or TX5 was added to MPO dissolved in 50 mM HEPES, pH 7.4, to yield a molar ratio of ϳ10:1. The mixture of MPO and compound was titrated with hydrogen peroxide until no further spectroscopic changes in the 400 -700-nm region could be observed. Excess compound and hydrogen were removed prior to crystallization. For details on crystallization and data collection see the supplemental material. Data were collected at beam line 711 at MAXLab, Lund, Sweden, at a wavelength of 1.085 Å (MPO-TX2) or on an FR-Eϩ SuperBright Microfocus rotating anode generator with VariMax-VHF optic at a wavelength of 1.542 Å (MPO-TX5). These data were processed using MOSFLM (33), scaled, and further reduced using the CCP4 suite of programs (34). Initial phasing was done by molecular replacement using a high resolution ligand-free structure of MPO (Protein Data Bank code 1cxp (35)) as a starting model. Model rebuilding was performed within O (36), and refinement was performed using REFMAC5 (34).
Bacterial Killing by Neutrophils-Neutrophils were incubated at 37°C with Staphylococcus aureus at a ratio of 1:10 in 10 mM phosphate buffer, pH 7.4, containing 140 mM chloride, 5 mM taurine, 1 mM calcium chloride, 0.5 mM magnesium chloride, and 10% serum in the presence or absence of 10 M TX1. Rate constants for phagocytosis and bacterial killing were determined by assessing the viability of extracellular and intracellular bacteria over 30 min as described in detail elsewhere (37).
Inhibition of MPO Activity in Plasma-Isolated neutrophils were added to autologous plasma (75%) at a concentration of 4 ϫ 10 6 /ml and incubated in the presence or absence of 2-thioxanthines at 37°C. They were stimulated with formyl-methionyl-leucyl-phenylalanine (100 nM) and cytochalasin B (10 g/ml) to induce the production of hydrogen peroxide and release of MPO, respectively. Neutrophils were maintained at 37°C with repeated mixing, and after 15 min the reactions were stopped by pelleting the cells and removing the supernatant. The concentration of MPO in supernatants was determined by adapting an ELISA (38). The presence of allantoin in supernatants, due to the MPO-dependent oxidation of urate, was measured by LC/MS/MS as described previously (18).
Establishing the Efficacy of Thioxanthines in a Mouse Model of Inflammation-Leukocytes were recruited into the peritoneal cavity of 9 -10-week-old C57BL/6 mice (n ϭ 5) by intraperitoneal injection with 1 ml of 4% thioglycollate broth. Twenty h after recruitment, leukocytes were activated in situ by intraperitoneal injection with 0.5 ml of 10 mg/ml Zymosan A (from Saccharomyces cerevisiae) in saline. The mice received a single oral dose of TX3 at 20, 60, or 180 mol/kg in 30% hydroxypropyl-␤-cyclodextrin in saline (w/v) 1 h before phagocyte activation. Five hours after administration of 2-thioxanthine, mice were killed by CO 2 asphyxiation, and the peritoneal cavity was lavaged with 1 ml of ice-cold phosphate-buffered saline. Peritoneal exudates cells were pelleted by centrifugation, and supernatants were analyzed for 3-chlorotyrosine in proteins and glutathione sulfonamide. 3-Chlorotyrosine (9) and glutathione sulfonamide (39) were measured by stable isotope dilution mass spectrometry. All animal experiments were performed in accordance with relevant guidelines and regulations provided by the Swedish Board of Agriculture. The ethical permission was provided by an ethical board specialized in animal experimentation.

Measurement of Thyroid Hormone Thyroxine (T 4 ) in Plasma
Samples-Nine-to 11-week-old C57BL/6 mice (n ϭ 30) received oral doses of 2-thioxanthine (TX3) at 20, 60, or 180 mol/kg in 30% hydroxypropyl-␤-cyclodextrin in saline (w/v) twice daily during 20 days. Plasma was isolated from blood collected by retro-orbital bleeding into microtainer tubes containing heparin. The amount of thyroid hormone thyroxine (T 4 ) in plasma samples was analyzed with an assay kit, coat-a-coat Total T 4 from Diagnostic Products Corp. All samples were made in duplicate and counted in a gamma counter (Wallac Wizard 1470). The amount of total T 4 was estimated from a standard curve.

RESULTS
Inhibition of MPO by 2-Thioxanthines-N-Isobutylthioxanthine (TX1) was identified in a high throughput screen as a potent inhibitor of the peroxidation activity of MPO (supplemental Fig. S1). It inhibited the chlorination activity of MPO obtaining 50% inhibition (IC 50 ) at 0.8 M (Fig. 1A). TX1 also inhibited hypochlorous acid production by neutrophils stimulated with phorbol myristate acetate with a similar IC 50 value to the purified enzyme (Fig. 1A). Neither tyrosine nor superoxide affected the ability of TX1 to inhibit hypochlorous acid production by MPO or neutrophils (Fig. 1B). Thus, it is unlikely that TX1 acts by trapping MPO as inactive compound II because these substrates readily reduce this redox intermediate back to the native enzyme (40,41). TX1 did not inhibit superoxide production or release of granule enzymes by neutrophils (results not shown). These results suggest that TX1 acted on neutrophils by inhibiting the enzymatic activity of MPO. Other related 2-thioxanthines were also potent inhibitors of hypochlorous acid production by MPO ( Table 1).
The potency of TX1 in the chlorination assay decreased with increasing concentrations of chloride (Fig. 1C), suggesting that chloride and TX1 compete in a reaction with MPO. Therefore, we used stopped flow spectrophotometry to determine whether TX1 reduces compound I. Compound I was formed by mixing MPO with a 10-fold excess of hydrogen peroxide (42). TX1 was added 40 ms later, and it promoted a rapid increase in the absorbance of MPO at 454 nm. This is indicative of the reduction of compound I to compound II ( Fig. 2A, inset) (14). The rate of the reaction was dependent on the concentration of TX1, and the rate constant was calculated to be 6.8 ϫ 10 5 M Ϫ1 s Ϫ1 (Fig. 2A). TX1 also reacted with compound II (Fig. 2B). The spectral changes, however, did not resemble clean conversion to ferric enzyme. Rather, they suggested that the enzyme was converted to an additional species because the Soret maximum of the final spectrum was 434 nm rather than 430 nm (Fig. 2B) (14). Time-resolved absorbance changes at 428 nm revealed an initial fast formation of ferric enzyme followed by a slower decline (Fig. 2C). Similarly, at 456 nm there was an initial fast loss followed by a slower decline. From these results, we conclude that TX1 reduces compound I and compound II, but the enzyme is modified during oxidation of the substrate.
2-Thioxanthines Are Mechanism-based Inactivators of MPO-To investigate the mechanism by which TX1 inhibits MPO, it was added to the enzyme in the absence and presence of hydrogen peroxide. After a set time, the enzyme was diluted 200-fold, and its residual activity was measured. TX1 alone caused about 30% inactivation of MPO, whereas hydrogen peroxide alone minimally affected activity (Fig. 3A). In combination, however, TX1 and hydrogen peroxide caused almost complete loss in  OCTOBER 28, 2011 • VOLUME 286 • NUMBER 43 enzyme activity. Chloride did not prevent enzyme inactivation. From these results, it is apparent that MPO uses hydrogen peroxide to oxidize TX1 to a product that irreversibly inactivates the enzyme.

Suicide Inactivation of Myeloperoxidase
The partition ratio (r) is a measure of an enzyme's commitment to turnover compared with reactions with its substrate that promote inactivation. The enzyme is fully inactivated after r ϩ 1 turnovers (43). This value is the intercept on the x axis when the percentage of residual enzyme activity is plotted against the mole ratio of inhibitor to enzyme (43). To establish the partition ratio for TX1, it was incubated at varying concentrations with MPO and hydrogen peroxide. After the reactions had gone to completion, the residual activity of MPO was measured. MPO lost activity with increasing concentrations of TX1 but was not fully inactivated (Fig. 3B). A similar pattern of inactivation has been reported for horseradish peroxidase and alkylhydrazines (44,45). Correction for incomplete inactivation gave a partition ratio for TX1 of 4. TX2 was more potent than TX1 and completely inactivated MPO with a partition ratio of 0.2 (Fig. 3B). At room temperature, inactivation of MPO by TX1 occurred over a few minutes (Fig. 3C). It was markedly slowed at 4°C, and no more than 40% of the enzyme was inactivated. In contrast, TX2 caused rapid inactivation (Fig. 3C), which was hardly affected by dropping the temperature of the reaction to 4°C.

2-Thioxanthines Modify the Heme Prosthetic Groups of MPO-
The distinct absorption spectra of the redox intermediates of MPO can be monitored during oxidation of substrates by the enzyme. When TX1 was added to MPO, it did not appreciably affect the absorption spectrum of the native enzyme (Fig. 4A). However, upon addition of hydrogen peroxide, the absorption spectrum underwent marked changes. The Soret band shifted to 456 nm, and a new peak appeared with a maximum at 630 nm. The wavelengths of these peaks and their relative absorbances resemble those for compound II (14). Adding catalase to scavenge residual hydrogen peroxide resulted in decay of the spectrum for compound II. It was converted to a spectrum with a Soret maximum at 433 nm having an absorption coefficient less than that of the native enzyme. It also had a small absorbance peak at 630 nm. This spectrum, which is similar to that observed in the stopped flow experiments (Fig. 2B), has previously not been observed for MPO. It is definitely not that of the native enzyme or compounds I-III (14). After oxidation of TX2, the absorption spectrum of MPO was also altered with the Soret maximum appearing at 450 nm and a smaller visible peak at 622 nm (data not shown).
To confirm that the prosthetic groups of MPO were modified when it oxidized the 2-thioxanthines, the enzyme was extensively digested by proteases, and the liberated hemes were separated by HPLC. The enzyme alone (Fig. 4B), with hydrogen peroxide or inhibitors, gave a major heme-containing peak that eluted at ϳ39.4 min. Incubation of TX5 with MPO and hydrogen peroxide caused a decrease in the peak at 39.4 min and gave rise to a new species that eluted at 37.4 min (Fig. 4B). Mass spectral analysis of the heme peak in unmodified MPO showed that its positive ion had an m/z value of 694 (supplemental  Fig. S2B). These ions are consistent with a previous mass spectral analysis of the heme of MPO (46). It was not possible to directly observe the ion for the modified heme due to co-elution of ions with greater intensity. Therefore, we monitored ions with an m/z value of 918, which would be the mass of a covalent adduct between the heme (694) and TX5 (226) minus two protons lost due to oxidation. Such an ion co-eluted with the modified heme and fragmented to give a major ion with an m/z value of 692. This is the expected positive ion if TX5 fragments off the covalent heme adduct (Fig. 4C, inset). Single reaction monitoring of the fragmentation of 918 to 692 m/z gave one peak that coeluted with the modified heme (Fig. 4C). We conclude that when the 2-thioxanthines are oxidized by MPO, they become covalently attached to the heme groups of the enzyme.
To confirm that an irreversible complex was formed between the 2-thioxanthines and MPO, the structure of the enzyme after it was inactivated with hydrogen peroxide and TX2 was studied using x-ray crystallography. The F o Ϫ F c difference map showed positive residual density in the distal heme cavities in each half of the protein corresponding to bound TX2. The ligand was positioned at the entrance of the narrow distal heme cavity and was oriented so that the plane of the 2-thioxanthine was approximately perpendicular to the heme plane (Fig. 5A). The only interaction between MPO and TX2 was between the exocyclic sulfur of the thioxanthine ring and one of the methyl groups of the heme. The distance was ϳ1.7 Å, which is consistent with a covalent thioether bond. Although there are several hydrogen bond donors and acceptors in the ligand, none of these were involved in direct polar interactions with the protein. To appreciate how the substituent on the 2-thioxanthines affects its orientation in the active site, the structure of MPO complexed to TX5 was determined. TX5 was found to be covalently bound to the heme the same way as TX2 with a strong difference density for the sulfur. However, the electron density for the rest of the ligand was weaker indicating flexibility. The thioxanthine plane was tilted relative to the position seen for TX2 (Fig. 5B). Again, no polar interactions between the ligand and MPO could be observed.
TX1 Inhibits MPO Selectively-To ascertain whether TX1 is selective for MPO rather than a general peroxidase inhibitor, we also determined how effectively it inhibited thyroid peroxidase and lactoperoxidase. The effects of TX1 on the halogenation activities of MPO and thyroid peroxidase were compared by following the oxidation of ascorbate (Fig. 6). In this assay, oxidation of ascorbate was reliant on the presence of chloride or iodide when using MPO or thyroid peroxidase, respectively (data not shown). At 2 M, TX1 almost completely blocked   OCTOBER 28, 2011 • VOLUME 286 • NUMBER 43 oxidation of ascorbate by MPO (Fig. 6A), although it had a minimal impact on the activity of thyroid peroxidase, inhibiting the initial rate by less than 10% (Fig. 6B). At 10 M, TX1 inhibited thyroid peroxidase by ϳ50%. TX1 was also a poor inhibitor of lactoperoxidase (supplemental Fig. S3).

Suicide Inactivation of Myeloperoxidase
We determined whether 2-thioxanthines affect thyroid hormone synthesis by measuring the amount of T 4 in plasma from mice treated with 2-thioxanthines for 20 days. The reference interval of total T 4 in untreated C57BL/6 mice was 3-3.5 g/dl of plasma. Mice treated with 20, 60, or 180 mol/kg TX3 twice daily had mean plasma concentrations of total T 4 between 2.6 and 4.5 g/dl (Fig. 6C). The corresponding average maximal plasma concentration (C max ) values at these three doses 30 min after oral administration were 12, 53, and 178 M. We conclude that 2-thioxanthines have only a minor affect on thyroid hormone synthesis.
TX1 Slows the Rate of Bacterial Killing by Neutrophils-MPO is largely responsible for the oxidative killing of S. aureus by neutrophils (47). Therefore, we investigated whether TX1 could affect the kinetics by which neutrophils phagocytose and kill these bacteria. Phagocytosis was determined by measuring the loss of extracellular bacteria when they were incubated with neutrophils. Killing was assessed from the degree to which live bacteria accumulated within neutrophils (Fig. 7A) (37). At 10 M, TX1 had no effect on phagocytosis but caused an increase in the number of live bacteria within neutrophils (Fig. 7A). From these data, we calculated the rate constants for phagocytosis and killing of bacteria (Fig. 7B). TX1 decreased the rate constant for bacterial killing by ϳ50%, but it did not affect the rate constant for phagocytosis (Fig. 7B).

2-Thioxanthines Inactivate MPO in Human Plasma-
The potential of 2-thioxanthines to inactivate MPO in vivo was tested by determining whether they could inactivate MPO in plasma as well as prevent MPO-dependent oxidation of urate to allantoin (18). Isolated neutrophils were added to plasma and stimulated with formyl-methionyl-leucyl-phenylalanine in the presence of cytochalasin B to effect a respiratory burst and release of MPO. When stimulated, neutrophils released substantial amounts of MPO (supplemental Fig. S4), and it was largely active (Fig. 8A). However, increasing concentrations of TX1, TX2, or TX5 caused progressive inactivation of the enzyme (Fig. 8A) but did not affect its release from neutrophils (supplemental Fig. S4). Stimulation of neutrophils also caused the oxidation of urate in plasma to allantoin (Fig. 8B), which is reliant on MPO (18). Concomitant with the loss in MPO activity, the 2-thioxanthines also decreased the production of allantoin in a concentration-dependent manner.
2-Thioxanthines Inhibit MPO during Inflammation-To establish whether 2-thioxanthines could inhibit MPO in vivo, inflammation was induced in the peritoneal cavity of mice using thioglycollate and zymosan. Intraperitoneal injection of thioglycollate into mice elicits an influx of leukocytes into the peritoneum 20 h after the challenge. Subsequent intraperitoneal injection with the yeast cell wall extract zymosan triggers activation and degranulation of these leukocytes, which are predominantly neutrophils. Mice were treated with 20, 60, or 180 mol/kg TX3 1 h before zymosan injection, resulting in average plasma exposures 5 h after dosing of 0.7, 2.0, and 7.7 M, respectively. Treatment with TX3 had no effect on the total number of cells in the peritoneal exudates as determined by live cell counting under a light microscope, excluding trypan blue-stained cells (data not shown). The inflammatory peritoneal exudates were analyzed for 3-chlorotyrosine (Fig. 9A) in proteins and glutathione sulfonamide in supernatants (Fig. 9B). These biomarkers of hypochlorous acid (48) were present in the inflammatory exudates, and their levels decreased with increasing doses of TX3. These observations support the conclusion that 2-thioxanthines can inhibit the activity of MPO during inflammation.

DISCUSSION
We have shown that 2-thioxanthines are mechanism-based inactivators of MPO that modify the heme prosthetic groups of the enzyme. Our data demonstrate that they have good specificity and are oxidized to a limited extent before the enzyme is   fully inactivated. In addition, they had only a moderate effect on bacterial killing by neutrophils even at concentrations that completely blocked extracellular production of hypochlorous acid. Importantly, they were able to inactivate MPO in plasma and inhibit its ability to produce chlorine bleach in vivo. Consequently, we have demonstrated that it is possible to design suicide substrates for MPO that do not affect thyroid hormone production, promote adverse reactions, or prevent neutrophils from killing bacteria. These inhibitors should be extremely useful for probing the role MPO plays in inflammatory tissue damage and have promise as therapeutic agents in numerous diseases associated with inflammation.
The 2-thioxanthines were good inhibitors of hypochlorous acid production by both the purified enzyme and isolated neutrophils. The IC 50 values with neutrophils were better than those for other inhibitors, including numerous nonsteroidal anti-inflammatory drugs (19,22,49). Indeed, TX2 and TX4 are the best known inhibitors of the chlorination activity of MPO (IC 50 0.2 M) that do not act by converting the enzyme to compound II. 2-Thioxanthines are mechanism-based inactivators of MPO because they promoted inactivation of the enzyme to a far greater extent in the presence of added hydrogen peroxide than in its absence. Inactivation resulted from modification of the heme through covalent attachment of the 2-thioxanthine. This was evident from the increase in mass of the modified heme and analysis of the x-ray crystal structure of the inactivated enzyme.
We showed that TX1 reacted with compound I of MPO with a rate constant sufficiently large (6.8 ϫ 10 5 M Ϫ1 s Ϫ1 ) for it to compete with chloride (2.5 ϫ 10 4 M Ϫ1 s Ϫ1 ) (42) for oxidation. Reduction of compound I resulted in formation of compound II, which indicates that the 2-thioxanthines must undergo oneelectron oxidation to produce a free radical intermediate. We propose that these free radicals then react with the enzyme and inactivate it. This reaction is calculated to be exothermic using density functional methods (supplemental material). Oxidation of the 2-thioxanthine would involve an abstraction of a hydrogen atom on the 2-thioxanthine by the oxygen bound to the iron atom of compound I (structure 1 in Scheme 1). This produces a radical on the 2-thioxanthine and simultaneously reduces the heme iron to an iron IV species (compound II; structure 2 in Scheme 1). This 2-thioxanthine radical is located predominantly on the sulfur (0.8 electrons) as assessed by density functional methods.
In steps 2 and 3, a hydrogen atom from the methyl group on the D-pyrrole ring is transferred to the radical of the 2-thioxanthine. The radical formed in this step is mainly located on the iron atom, but ϳ20% is still located on the CH 2 group on the D-pyrrole ring. The character of the CH 2 group is more sp2 than sp3, and the C-C bond length is 1.36 Å. The following steps could follow a radical path (path A as shown in Scheme 1) or via a deprotonation of the thioxanthine followed by a nucleophilic attack of the sulfur on the CH 2 carbon (path B as shown in Scheme 1) forming structure 5. An analogous mechanism has been shown to operate when alkylhydrazines inactivate horseradish peroxidase (44,50). Inactivation of horseradish peroxidase by alkylhydrazines, however, was much less efficient as judged by partition ratios that varied from 11 to 80 compared  with 0.2 for TX2 and TX4 (44). Based on the proposed mechanism, it is likely that substrates of MPO will inactivate the enzyme only when their radical has a high enough reduction potential to abstract an electron from the terminal methyl group. The structural information presented here suggests that the 2-thioxanthines, while tethered to the heme via the thioether bond, can adopt several positions and thus accommodate a range of different substituents.
The different effects of temperature on inactivation of MPO by TX1 and TX2 can be explained by the mechanism outlined in reaction Scheme 1. Oxidation of the 2-thioxanthines to reac-tive free radicals will depend on the rate of turnover of compound II because all substrates react more slowly with this redox intermediate than with compound I (14). Turnover of compound II will decline with temperature. For TX1, which had a partition ratio of ϳ 4, several turnovers of the enzyme are required for complete inactivation. Hence, lower temperatures should slow inactivation. For TX2, however, inactivation can occur without enzyme turnover because its partition ratio is less than 1. Therefore, inactivation should be independent of temperature. These results suggest that the TX2 radical must react rapidly with the heme rather than exit the active site. Consequently, with this type of inhibitor where reaction with the heme is maximized, potential adverse reactions due to free radical metabolites will be minimized.
Tyrosine and superoxide are both good substrates for compound II (40,41). Yet they did not prevent TX1 from inhibiting hypochlorous acid production by MPO. These results support the proposal that the incipient free radicals react rapidly with compound II. They also demonstrate that simple conversion of the enzyme to compound II does not explain how 2-thioxanthines inhibit MPO. Otherwise these physiological substrates of the enzyme would have prevented inhibition of hypochlorous acid production.
There is considerable homology among the mammalian heme peroxidases (14). Thus, mechanism-based inhibitors have the potential to affect all these enzymes. Our results demonstrate, however, that it is possible to target the activity of MPO while having a modest effect on thyroid peroxidase and lactoperoxidase. This is clearly important because of the essential role the former plays in thyroid hormone synthesis and the recent realization that oxidants generated by lactoperoxidase contribute to host defense in the airways (51). A unique feature of MPO is the high reduction potential of compound I (14). Thus, it should be possible to exploit its redox chemistry to fine-tune the specificity of mechanism-based inhibitors toward MPO.
The demonstration that TX1 slowed the killing of S. aureus by neutrophils reinforces several earlier studies that MPO plays a substantive role in oxidant-dependent killing by these inflammatory cells (47,52). It is unlikely, however, that 2-thioxanthines and related inhibitors will affect the ability of neutrophils to clear infections. This is because the immune system is not reliant on MPO to effectively combat pathogenic micro-organisms. It is now well established that although neutrophils that lack MPO have impaired killing of numerous bacteria (47,52), MPO deficiency is essentially benign (53). Its prevalence is relatively high in populations of European descent (ϳ1 in 3000), but among these individuals rates of infectious, inflammatory, or malignant complications are mostly normal. This phenomenon highlights the redundancy in the mechanisms used by the immune system to clear pathogens. We found that TX1 slowed killing of S. aureus to a lesser extent than that found previously for neutrophils lacking MPO (47). Consequently, inhibitors of this type should be expected to have no more impact on host defense than what is observed in MPO deficiency.
Collectively, our data indicate that 2-thioxanthines should attenuate oxidative stress promoted by MPO in vivo. These suicide substrates of MPO inactivated MPO in plasma and in the process decreased the oxidation of endogenous urate to allantoin. Furthermore, TX3 decreased production of two specific biomarkers of hypochlorous acid, namely 3-chlorotyrosine and glutathione sulfonamide (48), in an inflammatory model of peritonitis, Thus, it is reasonable to conclude that the reaction mechanism outlined in Scheme 1 is likely to operate in vivo.
In conclusion, we have shown that suicide substrates that block the chlorination activity of MPO can be specifically targeted at this enzyme to dampen oxidative stress without promoting excessive formation of reactive free radicals or compromising host defense. This type of inhibitor has considerable promise in illuminating the contribution MPO makes to inflammatory diseases, and it may be useful as an anti-inflammatory agent. Potentially, they may prove to be one of the first efficacious antioxidants because they will prevent formation of the strongest two-electron oxidant that is generated in humans.