1521-0103/353/2/288–298$25.00 http://dx.doi.org/10.1124/jpet.114.221788THE JOURNAL OF PHARMACOLOGY AND EXPERIMENTAL THERAPEUTICS J Pharmacol Exp Ther 353:288–298, May 2015Copyright ª 2015 by The American Society for Pharmacology and Experimental Therapeutics

PF-1355, a Mechanism-Based Myeloperoxidase Inhibitor,Prevents Immune Complex Vasculitis and Anti–GlomerularBasement Membrane Glomerulonephritis s

Wei Zheng, Roscoe Warner, Roger Ruggeri, Chunyan Su, Christian Cortes,Athanasia Skoura, Jessica Ward, Kay Ahn, Amit Kalgutkar, Dexue Sun, Tristan S. Maurer,Paul D. Bonin, Carlin Okerberg, Walter Bobrowski, Thomas Kawabe, Yanwei Zhang,Timothy Coskran, Sammy Bell, Bhupesh Kapoor, Kent Johnson, and Leonard BuckbinderPfizer World Wide Research and Development, Cambridge, Massachusetts (W.Z., R.R., C.S., C.C., A.S., J.W., K.A., A.K., D.S.,T.S.M., Y.Z., B.K., L.B.); Department of Pathology, University of Michigan, Ann Arbor, Michigan (R.W., K.J.); and Pfizer WorldWide Research and Development, Groton, Connecticut (P.D.B., C.O., W.B., T.K., T.C., S.B.)

Received December 4, 2014; accepted February 18, 2015

ABSTRACTSmall vessel vasculitis is a life-threatening condition and patientstypically present with renal and pulmonary injury. Disease path-ogenesis is associated with neutrophil accumulation, activation,and oxidative damage, the latter being driven in large part bymyeloperoxidase (MPO), which generates hypochlorous acidamong other oxidants. MPO has been associated with vasculitis,disseminated vascular inflammation typically involving pulmonaryand renal microvasculature and often resulting in critical con-sequences. MPO contributes to vascular injury by 1) catabolizingnitric oxide, impairing vasomotor function; 2) causing oxidativedamage to lipoproteins and endothelial cells, leading to athero-sclerosis; and 3) stimulating formation of neutrophil extracel-lular traps, resulting in vessel occlusion and thrombosis. Herewe report a selective 2-thiouracil mechanism-based MPOinhibitor (PF-1355 [2-(6-(2,5-dimethoxyphenyl)-4-oxo-2-thioxo-

3,4-dihydropyrimidin-1(2H)-yl)acetamide) and demonstrate thatMPO is a critical mediator of vasculitis in mouse disease models.A pharmacokinetic/pharmacodynamic responsemodel of PF-1355exposure in relation with MPO activity was derived from mouseperitonitis. The contribution of MPO activity to vasculitis was thenexamined in an immune complexmodel of pulmonary disease. Oraladministration of PF-1355 reduced plasma MPO activity, vascularedema, neutrophil recruitment, and elevated circulating cytokines.In a model of anti–glomerular basement membrane disease,formerly known asGoodpasture disease, albuminuria and chronicrenal dysfunction were completely suppressed by PF-1355 treat-ment. This study shows that MPO activity is critical in drivingimmune complex vasculitis and provides confidence in testing thehypothesis that MPO inhibition will provide benefit in treatinghuman vasculitic diseases.

IntroductionMyeloperoxidase (MPO) is a heme-containing peroxidase

produced in bone marrow and stored in the azurophilicgranules of neutrophils, where it constitutes up to 5% of thecellular protein; MPO is also found, albeit to a lesser extent, insome humanmonocytes andmacrophages (Hansson et al., 2006).Upon phagocyte activation, MPO activity appears extracellu-larly and within the phagolysosome. MPO catalyses theproduction of hypochlorous acid (HOCl) by utilizing chloride(Cl2) and hydrogen peroxide (H2O2), predominantly gener-ated by NADPH oxidase. MPO also directly oxidizes a varietyof electron-rich aromatic substrates to generate aryl radicals

(Goldman et al., 1999). MPO is required for neutrophil extra-cellular trap (NET) formation in response to phorbol ester(Metzler et al., 2011; Parker et al., 2012). NETs constitute theterminal act of neutrophil activation, whereby DNA, histones,and granule proteins are expelled (Branzk and Papayannopoulos,2013), capturing and colocalizing pathogens with the antimi-crobial constituents including MPO but also promoting endo-thelial cell damage and platelet/leukocyte aggregation. Themainfunction of MPO is considered to be microbicidal; however, MPOdeficiency occurs in approximately 1 in 2000 individuals (Kutter,1998) and is not usually associated with increased infection risk(Lekstrom-Himes and Gallin, 2000).Compelling human data support a causal role for MPO in

vasculitis, characterized by the influx of leukocytes into the ves-sel wall resulting in endothelial cell damage, impaired endothe-lial barrier function, and increased risk of thromboembolism

dx.doi.org/10.1124/jpet.114.221788.s This article has supplemental material available at jpet.aspetjournals.org.

ABBREVIATIONS: ANCA, antineutrophil cytoplasmic antibody; BSA, bovine serum albumin; DAB, diaminobenzidine; GBM, glomerular basementmembrane; HOCl, hypochlorous acid; HRP, horseradish peroxidase; MPO, myeloperoxidase; NET, neutrophil extracellular trap; PBS, phosphate-buffered saline; PF-1355, 2-(6-(2,5-dimethoxyphenyl)-4-oxo-2-thioxo-3,4-dihydropyrimidin-1(2H)-yl)acetamide; TBST, Tris-buffered saline/Tween20; TEM, transmission electron microscopy; THF, tetrahydrofuran; TNF, tumor necrosis factor; TPO, thyroid peroxidase; TX-1, 3-isobutyl-2-thioxo-7H-purin-6-one; WT, wild-type.

288

http://jpet.aspetjournals.org/content/suppl/2015/02/19/jpet.114.221788.DC1Supplemental material to this article can be found at:

at ASPE

T Journals on O

ctober 12, 2020jpet.aspetjournals.org

Dow

nloaded from

(Halbwachs and Lesavre, 2012; Gaffo, 2013). Small vesselvasculitis classification includes microscopic polyangiitis,eosinophilic granulomatosis with polyangiitis, and granulo-matosis with polyangiitis. Renal, respiratory, and/or cutane-ous injury is associated with disease morbidity and mortality.Antineutrophil cytoplasmic antibodies (ANCAs) are oftenassociated with small vessel vasculitis and are directed againsttwo abundant neutrophil proteins: MPO and proteinase 3.Although ANCA is defined by lack of immune deposits(pauci-immune) in target organs, ANCA antibodies mayalso copresent in immune complex diseases includinganti–glomerular basement membrane (GBM) disease, lupus,drug-induced vasculitis, and rheumatoid arthritis (Shortet al., 1995; Bart�unková et al., 2003; Yang et al., 2007; Dammaccoet al., 2013).MPO is only present at high levels on the cell surface after

neutrophil activation. MPO-ANCA disease flares are oftenprecipitated by an infection, with primed neutrophils beingfurther activated by circulating anti-MPO antibody interactingwith MPO on the cell surface (Kettritz, 2012). Thus, MPO canbe both an antigen and a source of oxidative damage (Johnsonet al., 1987). Consistent with this, neutrophil-derived oxidantswere associatedwith vasculitis (Warren et al., 1990) and leukocytedepletion as well asMPO deficiency–blocked albuminuria in anti-GBM models (Feith et al., 1996; Odobasic et al., 2007). However,since nonenzymatic activities of MPO including leukocyte re-cruitment have been described (Klinke et al., 2011), the preciserole of MPO activity in vasculitic diseases remained incompletelydefined.Here we describe and characterize the activity of PF-1355

[2-(6-(2,5-dimethoxyphenyl)-4-oxo-2-thioxo-3,4-dihydropyrimidin-1(2H)-yl)acetamide], a selective, mechanism-based inhibitor ofMPO (Fig. 1A). PF-1355 blocked HOCl formation in biochemi-cal and cellular assays and was used to test the hypothesis thatMPO activity is a critical mediator of disease activity in im-mune complex vasculitis mouse models and to support theconfidence in rationale for this mechanism as a therapeuticapproach to treat related human conditions. Disease severitywas attenuated by the prophylactic administration of PF-1355,reducing perivessel edema, leukocyte infiltration, and pro-duction of proinflammatory cytokines. MPO recovered fromtreated animals was found to be irreversibly inhibited byPF-1355 treatment. In anti-GBM glomerulonephritis, PF-1355completely blocked kidney injury acutely and prevented anyprogressive disease. As in the pulmonary model of immunecomplex vasculitis,MPO inhibition also reduced the infiltrationof neutrophils into glomeruli in experimental anti-GBM disease.These findings indicate that MPO activity is critical in drivingimmune complex vasculitis, neutrophil recruitment, and proin-flammatory signaling.

Materials and MethodsPF-1355 Synthesis. A 20-l reaction vessel was charged with

magnesium ethoxide (3.61 mol, 413.52 g) and tetrahydrofuran (THF)(6.6 l). The resulting mixture was stirred as ethyl hydrogen malonate(7.23 mol, 888.89 ml, 994.67 g; diluted with 20 ml THF) and wasadded. The mixture was heated at 45°C for 4 hours. Meanwhile, a20-l reactor was charged with 2,5-dimethoxybenzoic acid (3.29 mol,600.00 g) and THF (3.6 l). We added 1,19-carbonyldiimidazole (3.61 mol;585.98 g) to this mixture in portions to avoid excess foaming, stirringat room temperature. After stirring for 3 hours at room temperature,the second solution was added gradually to the first solution. After the

addition, the reaction mixture was heated to 45°C. After 20 hours, thereaction mixture was concentrated under reduced pressure beforeadding ethyl acetate (6 l) followed by 2 N HCl (3 l). After mixing, thelayers were separated and the organic phase was washed sequentiallywith 2 N HCl (3 l), saturated sodium bicarbonate (3 l), and water (3 l).The organic phase was concentrated under reduced pressure, and theresidue was taken up in ethyl acetate (6 l) and concentrated again toafford an oil, which was transferred to a 20-l reaction vessel with 5 l ofethyl acetate and treated with sodiummethoxide (3.45 mol, 793.00 mlof a 4.35 M solution in methanol). After stirring at room temperaturefor 3 hours, an additional 6 l of ethyl acetate was added and the solidwas collected by vacuum filtration and dried overnight in a vacuumoven at 40°C to give 661 g sodium 1-(2,5-dimethoxyphenyl)-3-ethoxy-3-oxoprop-1-en-1-olate.

A 5-l reaction vessel was charged with methanol (3.3 l), sodiummethoxide (102.4 g, 1.8 mol), and glycinamide hydrochloride (202 g,1.8 mol). The mixture was heated at 65°C for 1 hour before cooling to50°C and adding acetic acid (514.25 mmol, 30.88 g, 29.47 ml) and300 g or the preceding product. After heating at reflux for 16 hours,the reaction mixture was stirred as it was cooled to 10°C. After30 minutes, the resulting solid was collected by vacuum filtration andpulled dry to form a cake that was dried in a vacuum oven (20 mmHg,65°C) for 14 hours to afford (Z)-ethyl 3-((2-amino-2-oxoethyl)amino)-3-(2,5-dimethoxyphenyl)acrylate (339.4 g).

A 5-l reaction vessel equipped with an efficient stirrer was chargedwith (Z)-ethyl 3-((2-amino-2-oxoethyl)amino)-3-(2,5-dimethoxyphenyl)acrylate (1.30 mol, 400.00 g), butyl acetate (3.4 l) and trimethylsilylisothiocyanate (4.15 mol, 585.67 ml, 544.96 g) and the mixture washeated to reflux. After 16 hours, the mixture was cooled to 40°C andtreated with 2 N aqueous sodium hydroxide (1.95 l). The organic layerwas separated and extracted with another portion of 2 N sodiumhydroxide (0.325 l). The combined aqueous phases were filtered,extracted twice with dichloromethane (2 � 1.6 l), and added slowly toa well stirred 3 N aqueous HCl solution (1.3 l) at room temperature.After stirring for 30 minutes, the resulting solid was isolated byvacuum filtration, rinsed with water, and pulled dry to afford a waterwet cake (640 g). The cake was dissolved in dimethylformamide (2.4 l)at 90°C and stirred as water (2 l) was added slowly to the solution. Themixture was cooled gradually to room temperature and the resultingsolid was isolated by vacuum filtration, rinsed with water, and pullingdry to afford 245 g of solid. This solid was then suspended in 1.25 l ofmethanol and was stirred as 1.25 l of water was added. The mixturewas heated with stirring at 50°C for 2 hours. It was then cooled to10°C for 2 hours before collecting the solid by vacuum filtration andwas pulled dry before drying in a vacuum oven (20 mm Hg, 60°C) toafford the desired product PF-1355.

MS (ES1) 322.2[M11]1. 1H NMR (500 MHz, dimethylsulfoxide-d6)d ppm 12.80 (s, 1 H) 7.32 (broad singlet, 1 H) 7.06–7.11 (m, 2 H) 7.06(broad singlet, 1 H) 6.74–6.77 (m, 1 H) 5.82 (d, J 5 2.20 Hz, 1 H)5.37 (broad singlet, 1 H) 3.88 (broad singlet, 1 H) 3.78 (s, 3 H) 3.70(s, 3 H).

Mice. C57BL6 wild-type (WT) and genetically deficient (C57BL6MPO2/2) mice were obtained from The Jackson Laboratory (BarHarbor, ME). Mice aged 8–10 weeks and weighing approximately 25 gbody weight were used in all experiments. All experiments wereapproved by the Pfizer Institutional Animal Care and Use Commit-tees within Massachusetts or by the animal care committee at theUniversity of Michigan, and were performed according to institutionaland national guidelines.

Peritonitis Model for Pharmacokinetics/Pharmacodynamicsof PF-1355. Animals received an intraperitoneal injection of 4%thioglycollate broth in phosphate-buffered saline (PBS) (Sigma-Aldrich,St. Louis, MO) for neutrophil recruitment (Zhang et al., 2002). Twentyhours later, PF-1355 or vehicle (1% hydroxypropyl methylcellulose, 0.5%2-amino-2-hydroxymethyl-propane-1,3-diol, 0.5% hypromellose acetatesuccinate, pH 9.5) was administered p.o., followed by intraperitonealadministration of opsonized zymosan (Sigma-Aldrich) or saline. After3 hours, the mice were euthanized and received intraperitoneal injection

MPO Inhibition Prevents Vasculitis and Anti-GBM Diseases 289

at ASPE

T Journals on O

ctober 12, 2020jpet.aspetjournals.org

Dow

nloaded from

of 2ml cold PBS. Bloodwas collected and animalswere shaken vigorouslybefore the collection of peritoneal lavage.

IgG Immune Complex–Mediated Acute Alveolitis. WT C57BL6and MPO2/2 mice were anesthetized by intraperitoneal injection of amixture of Ketaset (Fort Dodge Animal Health, Fort Dodge, IA) and

Rompun (Bayer Corporation, Shawnee Mission, KS) at doses of 1.66and 0.033 mg/g body weight, respectively. IgG immune complex lunginjury was induced by intratracheal instillation of 0.5 mg rabbitanti–bovine serum albumin (BSA) antibody (MP Biomedicals, Solon,OH), andBSAantigenwas instilled intravenously (Warner et al., 2001).

Fig. 1. Determination of potency and reversibility of PF-1355 forMPO inhibition. (A) Structure of PF-1355. TheMPO reactions, rapid dilution experiments,and all data fitting and analysis were performed as previously described (Ward et al., 2013). (B) MPO inhibition progress curves at varying concentrations ofPF-1355 (0.12–30 mM) display curvatures indicating time-dependent inhibition. (C) The kobs values obtained by fitting the progress curves in (B) wereplotted as a function of [PF-1355] to determine values of the kinetic inhibition constants kinact and KI. Data are averages, and error bars represent the S.D.from six separate experiments. (D) The percent inhibition at each inhibitor concentration was plotted as a function of [PF-1355] to determine the IC50 valueforMPO inhibition (solid circles). The concentrations of PF-1355were varied from 0.050 to 100 mM for TPO inhibition. The percent inhibitions for TPO (solidtriangles) are averages and error bars represent the S.D. from two separate experiments. (E) MPO was incubated with 5 mM PF-1355 or DMSO with orwithout 2 mM H2O2. After 15 minutes, an aliquot of the preincubation mixture was diluted 300-fold into MPO buffer containing H2O2 and Amplex Redsubstrates andMPO activity was monitored. (F) Isolated human neutrophils were treated with PF-1355 or vehicle and then stimulated with PMA inmediacontaining taurine. The trapped taurine chloramines were then reacted with TMB and sulfuric acid, producing a chromogenic product. The percentinhibition from duplicates of five individual donors (solid circles) with 95% confidence intervals (shaded area) are shown. (G) PF-1355 was added to freshlyisolated human blood followed by LPS addition. After 4 hours, MPO was captured using anti-MPO–coated plates and residual MPO activity was measuredin a reaction containing Amplex Red and H2O2. The fluorescence value was compared relative to a standard curve using purified human MPO. The percentinhibition from duplicates of 17 individual donors (solid circles) with 95% confidence intervals (shaded area) are shown. DMSO, dimethylsulfoxide; LPS,lipopolysaccharide; PMA, phorbol 12-myristate 13-acetate; RFU, relative fluorescence unit; TMB, 3,39,5,59-tetramethylbenzidine.

290 Zheng et al.

at ASPE

T Journals on O

ctober 12, 2020jpet.aspetjournals.org

Dow

nloaded from

Animals were euthanized after 4 hours and the degree of lung injurywas quantified by histopathology, which correlates with lung perme-ability changes.

Anti-GBM Glomerulonephritis Model. Mice were adminis-tered intravenous 200 ml sheep anti-rat GBM serum (PTX-001;Probetex Inc., San Antonio, TX), nonimmune sheep serum (PTX-000S;Probetex Inc.), or PBS controls. The MPO inhibitor PF-1355 wasdosed p.o. 1 hour before injection of anti-GBM serum. Urine wascollected in metabolic cages. Mice were euthanized after 2 hours or 21days and plasma and kidney tissues were collected. Kidneys wereembedded in optimal cutting temperature compound and frozen forimmunohistochemical analysis. A piece of kidney tissue was alsoplaced in 0.1 M phosphate-buffered 4% paraformaldehyde plus 1%glutaraldehyde for transmission electron microscopy (TEM) study.

Morphologic Studies and Quantitative Evaluation. At thetime of euthanasia, the lungs were removed en bloc and inflated with10% buffered formalin, embedded in paraffin, and processed formicroscope analysis using hematoxylin and eosin–stained sections(5 mm). Sections were examined (40�) on a light microscope (CarlZeiss Inc., Thornwood, NY) for the intensity of neutrophil influx andedema as an assessment of the injury. Seven or more animals pergroup were evaluated.

The extent of lung injury was analyzed by quantitative morphom-etry. A minimum of 20 random 40� fields were analyzed from eachanimal using an Olympus BX40 microscope (Olympus Corporation,Center Valley, PA) with a video camera attached to a digital videoimaging analysis system. The system uses software obtained from IPLaboratories Spectrum (Signal Analytics, Vienna, VA) and NationalInstitutes of Health ImageJ to precisely quantify parameters, suchas intra-alveolar hemorrhage, leukocyte influx, and surface area ofedema around blood vessels. This method of analysis has been usedpreviously (Warner et al., 2001).

Bronchoalveolar Fluid and Cells. Animals were euthanized4 hours after initiation of injury by lethal injection of ketamine.Animals were exsanguinated and their thoracic cavities were openedto reveal lungs and trachea. Lungs were lavaged three times with0.7 ml cold PBS. The cells were pelleted after each lavage. Supernatantfrom the first lavage was saved for MPO measurements and thesubsequent two lavages were discarded. The cell pellets were combinedand resuspended in 200 ml PBS, and the total cell number wasdetermined with a hemocytometer. Cell numbers were determinedusing a hemocytometer, and morphologic determination of neutro-phils was made using modified Wright’s stain.

MPO Peroxidation Activity. Plates were coated with the MPOcapture antibody (1:200) overnight at 4°C, washed with PBS, and thennonspecific binding blocked with PBS/1% BSA. Plasma or peritonealexudate samples were diluted 1:4 in PBS and 50 ml was added totriplicate wells for 1 hour at room temperature. Plates were washedthree times with PBS containing 0.05% Tween, followed by PBSwashes. Assay buffer (50 ml containing 50 nM phosphate buffer,pH 7.4, containing 140 mMNaCl, 10 mMNa2NO2, 40 mMAmplex Red,and 10 mM H2O2) was added with kinetic reads and an excitation/emission wavelength of 530/580 nm on a fluorescence plate reader(Molecular Devices, Sunnyvale, CA). Assay linearity was typicallymaintained for .300 seconds (R2 . 0.99) and Vmax represented by thechange in relative fluorescence units divided by time to yield MPOactivity (units per second). Active MPO was back-calculated usingpurified human MPO standard (R&D Systems, Minneapolis, MN).

MPO Enzyme-Linked Immunosorbent Assay. Antibody coat-ing was as described above for the MPO activity assay. Plasma orexudate was diluted 1:4 with Biolegend diluent. Plates were in-cubated 2 hours at 25°C and then washed three times with Tris-buffered saline/Tween 20 (TBST). The detection antibody (HM1051BT;Hycult, Plymouth Meeting, PA) was diluted 1:1000 with BiolegendDiluent and the plate was incubated 1 hour at 25°C. The plate waswashed three times with TBST and developed with 50 ml horseradishperoxidase (HRP)–streptavidin (DY998; R&D Systems) and Picoenzyme-linked immunosorbent assay substrate (Pierce, Rockford, IL).

Plates were read on the Wallac TriLux plate reader (PerkinElmer,Waltham, MA) and MPO back-calculated using an MPO standard(R&D Systems).

MPO Activity in Human Whole Blood (Residual CaptureActivity). MPO activity was determined using modifications toa previously described method (Franck et al., 2009), using humanwhole blood from healthy volunteers, collected in heparinized tubes.The test compound was incubated with human whole blood stimulatedwith bacterial lipopolysaccharide for 4 hours, followed by capture ofMPO on immobilized anti-MPO antibody–coated plates. The capturedMPO was washed and residual MPO activity was determined usingAmplex Red and H2O2 as described above.

MPO Chlorination Activity in Isolated Human Neutrophils.A modification of the manufacturer’s (Northwest Life Science Spe-cialties, LLC, Vancouver, WA) protocol and published procedures(Tidén et al., 2011) were used. Venous blood was collected in heparintubes from fasting healthy donors and neutrophils isolated usingLympholyte-poly media (Cedarlane Laboratories, Burlington, ON,Canada). PF-1355 was added to the purified neutrophils over a di-lution series of 31.5 nM to 31.5 mM or vehicle and then cells werestimulated with 0.4 mM phorbol 12-myristate 13-acetate in mediacontaining taurine, which traps unstable HOCl. Taurine chloramineswere quantified after reaction with 3,39,5,59-tetramethylbenzidineand sulfuric acid, producing a chromogenic product with absorbanceat 450 nm.

Preparation of Human Thyroid Peroxidase. The cDNAencoding the full-length human thyroid peroxidase (TPO) (accessionno. NM_000547, amino acids 1–933) was cloned into the inducibleexpression vector pcDNA5/frt/to (Life Technologies, Norwalk, CT).After transfection of Flp-In-T-Rex-293 cells with the TPO expressionvector, stable clones (HEK293-hTPO) were selected using selectionmedia (Dulbecco’s modified Eagle’s medium containing 10% fetalbovine serum, 100 mg/ml hygromycin, and 15 mg/ml blasticidine). Whencells reached 50%–60% confluence, TPO expression was induced inselection media containing 10 mg/ml doxycycline hyclate and 5 mg/mlhemin (Sigma-Aldrich). All of the operations below were at 4°C unlessotherwise noted. The HEK293-hTPO cells were washed in PBS andcollected by centrifugation at 1000g for 5 minutes, resuspended in lysisbuffer (1 mM sodium bicarbonate, pH 7.4) containing EDTA-freeprotease inhibitor cocktail, incubated on ice for 10 minutes, and lysedby Dounce homogenization. Nuclei and cell debris were removed bycentrifugation at 1000g for 10 minutes. The supernatant was thencentrifuged at 25,000g for 20 minutes. The resulting pellet wasresuspended in lysis buffer and centrifuged again at 25,000g for20 minutes. The final pellet was resuspended in storage buffer (50 mMTris-HCl, pH 7, 150mMNaCl) containingEDTA-free protease inhibitorcocktail, aliquotted, flash frozen in liquid N2, and stored at280°C untiluse. Protein concentration was determined using the BCA ProteinAssay (Pierce).

TPOAssay. TPO activity wasmeasured bymonitoring the formationof resorufin from the oxidation of Amplex Red using conditions similar tothose in the previously described MPO assay (Ward et al., 2013). Assaymixtures (100 ml) contained 50 mM NaPi, pH 7.4, 150 mM NaCl, 2 mMH2O2, 30 mM Amplex Red, 1 mM diethylene triamine pentaacetic acid,and 2% dimethylsulfoxide. The reactions were initiated by the addition ofTPO. Reaction mixtures to determine the background reaction rateconsisted of all assay components and 4 ml of 500 U/ml bovine catalase in50 mM KPi, pH 7.0. The background rate was subtracted from eachreaction progress curve. All data were analyzed using nonlinearregression analysis in Microsoft Excel (Microsoft Corporation, Redmond,WA) and KaleidaGraph software (version 3.5; Synergy Software, Dubai,United Arab Emirates).

Cytokine Measurements. Cytokine levels were measured witha Milliplex Map–based assay (Millipore, Billerica, MA) using themanufacturer’s recommendations for cytokine quantitation usinga Bio-Plex 200 system (Bio-Rad Laboratories, Hercules, CA).

Urine Albumin/Creatinine Ratio. Micewere individually placed inmetabolic cages for 24 hours. Urine was centrifuged at 1500 rpm for

MPO Inhibition Prevents Vasculitis and Anti-GBM Diseases 291

at ASPE

T Journals on O

ctober 12, 2020jpet.aspetjournals.org

Dow

nloaded from

10 minutes. Urine albumin concentration was measured using a compet-itive enzyme-linked immunosorbent assay method (Albuwell M kit;Exocell Inc., Philadelphia, PA) and was normalized to creatinine. Urinecreatinine levels were determined using a Hitachi clinical analyzer(Roche, Indianapolis, IN).

TEM. Minced tissue samples from two animals per treatment groupwere fixed in 0.1 M phosphate-buffered 4% paraformaldehyde plus 1%glutaraldehyde, postfixed in 0.1 M phosphate-buffered 1% osmiumtetroxide, dehydrated, and processed to epoxy resin block for con-ventional TEM. Toluidine blue–stained microscopy sections (0.6 mm)were prepared from select blocks and examined via light microscopy.Blocks were further trimmed to include regions with glomeruli, andelectron microscopy sections (approximately 90 nm) were prepared andcounterstained with uranyl acetate and lead citrate. Electron micros-copy examination was performed on an Hitachi H-7100 and digitalimages (AMT Inc., Woburn, MA) of representative areas were recorded.

Fluorescence Microscopy. Frozen tissue samples (n 5 5 pergroup) were sectioned at 8-mm thickness, fixed in either 4% para-formaldehyde or with amixture of 3:1 acetone/ethanol. Sections fixed inthe acetone/ethanol mixture were air dried before immunohistochem-ical staining. Sections were treated with Dako Protein Block (Dako,Carpinteria, CA) and then incubated with a primary antibody cocktailconsisting of donkey anti-sheep IgG conjugated to Alexa Fluor 594 anddonkey anti-mouse IgG conjugated to Alexa Fluor 488 (JacksonImmunoResearch, West Grove, PA) at 4°C in a humidified chamberovernight. Slides were then rinsed with TBST and Hoechst 33342 wasapplied. Specimens were examined on a Zeiss Axioplan2 Imagingfluorescence microscope (Carl Zeiss Corp., Jena, Germany) with a40�/0.95NAKorr Plan-Apochromat objective lens. Digital images wereacquired with a Hamamatsu OrcaR2 camera (Hamamatsu Corp.,Bridgewater, NJ) using Metamorph software (Molecular Devices).Microscopy evaluations were completed at room temperature. Cap-tured images were assembled into figures using Adobe Photoshopsoftware (Adobe Systems, San Jose, CA).

Sections fixed in 4% paraformaldehyde were rinsed and then blockedfor 1 hour. Slides were incubated with primary anti-MPO (ThermoFisher Scientific, Waltham, MA) and anti-Ly6G (BD Biosciences, SanJose, CA) in a light-protected humidified chamber at 4°C overnight. Thenext day, sections were rinsed with PBS containing 0.05% Tween andincubated with secondary antibodies (conjugated with Alexa Fluor 488,Alex Fluor 594, or Alex Fluor 647; Invitrogen, Carlsbad, CA) at roomtemperature for 1 hour. Alexa Fluor 647–conjugated anti-sheep IgGwas used to analyze glomeruli in animals treated with an anti-GBMantibody. Hoechst 33342 was used to stain nuclei. Confocal microscopeimages were acquired with a Zeiss LSM710 Axio Examiner with63�/NA 1.4 oil differential interference contrast and ZEN lite software0.13 mm � 0.13 mm � 1.0 mm/pixel.

Immunohistochemistry. Kidneys frozen in optimal cutting tem-perature compound were sectioned at 10 mm and then stored at280°Cuntil use. Slides were warmed to room temperature and then fixed in3:1 acetone/ethanol before staining. Briefly, endogenous peroxidasewas blocked using 0.3%H2O2 and nonspecific sites were blocked usingBiocare Rodent Block M (Biocare Medical, Concord, CA). Rat anti-mouse Ly6G, clone RB6-8C5 (BD Biosciences), was applied at a 1:500dilution for 60 minutes at room temperature. Detection of positiveneutrophil staining was accomplished by using the Biocare Rat Probefollowed by the Biocare Rat-on-Mouse HRP-Polymer, and was visualizedwith Dako Liquid diaminobenzidine (DAB1). Slides were counterstainedwith Mayer’s hematoxylin and mounted.

Neutrophil Image Analysis. Slides were scanned on the Leica/Aperio AT2 Whole Slide Digital Scanner (Leica, Buffalo Grove, IL)using the bright-field and 20� objective settings. Images were stored inLeica eSlideManager in .svs format and were analyzed using TissueStudio Software (Definiens, Munich, Germany). A rule set was createdthat allowed for a manual outline of the kidney tissue area. Within thekidney tissue area, thresholds were set to identify and quantifyhematoxylin staining (nuclear areas) and DAB staining (neutrophil-stained cells). A measurement of DAB stain area per kidney and

a measurement of DAB stain count per kidney tissue were completedfor each animal. Ratios were calculated per group in Microsoft Excel.

Statistical Methods. All numerical values presented in graphsare means6 S.E.M. unless otherwise stated. The effect of PF-1355 onactiveMPO (Fig. 2, A and B), activeMPO by time point (Fig. 3D), totalMPO (Supplemental Fig. 1), total MPO by time point (Fig. 3E), andneutrophil counts by time point (Fig. 3C) were analyzed using one-way analysis of variance, and statistical significance was based onDunnett’s test. The effect of PF-1355 on relative area (Fig. 3B), activeMPO (Fig. 4K), total MPO (Fig. 4L), cytokines by time point (Fig. 3, Fand G; Supplemental Fig. 3), and %DAB marker area (Fig. 4J) wereanalyzed using one-way analysis of variance, and statistical signifi-cance was based on multivariate t test via simulation. The effect ofPF-1355 on the urine albumin/creatinine ratio (Fig. 4I) was analyzedusing a linear mixed-effects model for repeated measures andstatistical significance was based on the multivariate t test viasimulation. All statistical analyses were implemented using SAS 9.2software (SAS Institute Inc., Cary, NC) unless noted otherwise.

ResultsPF-1355 Is a Selective, Mechanism-Based MPO In-

hibitor, Efficacious in Cell Assays and In Vivo. A novel,2-thiouracil MPO inhibitor PF-1355 was developed and thestructure is shown in Fig. 1A. The mechanism of MPOinhibition by PF-1355 was evaluated using purified humanMPO and monitoring enzymatic activity using H2O2 andAmplex Red as substrates. In the presence of PF-1355,the reaction progress curves exhibited curvature and time-dependent inhibition consistent with an irreversible mecha-nism of inactivation (Fig. 1B). Plotting the kobs values asa function of [PF-1355] revealed saturation indicating a two-step mechanism for inactivation (Fig. 1C). To assess whetherMPO inhibition by PF-1355 is mechanism based and requiresMPO catalysis, we performed rapid dilution experimentsin the presence or absence of H2O2 during preincubation.As shown in Fig. 1D, no recovery of MPO activity was observedwith PF-1355 when H2O2 was included in the preincubationmixture. However, in the absence of H2O2, little or no MPOinhibition by PF-1355 was observed, indicating that PF-1355is an irreversible mechanism-based inhibitor. PF-1355 hashigh selectivity over the closely related and clinically relevantthyroid peroxidase enzyme (Fig. 1E) and is selective againsta broad panel of more than 50 targets including receptors,enzymes, transporters, and ion channels (SupplementalTable 1). In a dose-responsive fashion, PF-1355 inhibitedMPO activity in phorbol ester–stimulated human neutro-phils as measured by taurine chlorination (EC50 5 1.47 mM;Fig. 1F) as well as lipopolysaccharide-treated human bloodmeasuring residual MPO activity (EC50 5 2.03 mM; Fig. 1G).The pharmacokinetic/pharmacodynamic relationship of

PF-1355 and MPO activity was characterized in a murineperitonitis model. Neutrophils were recruited to the peritoneumovernight by intraperitoneal administration of thioglycollatebroth. The next day, PF-1355 or vehicle was administered p.o.,followed by intraperitoneal administration of opsonized zymo-san. Peritoneal lavage and plasma were collected after 3 hoursand residual MPO activity was determined by antibody captureof MPO and measurement of MPO activity as well as plasmalevels of PF-1355 (Supplemental Methods and SupplementalTable 2). MPO activity in peritoneal lavage and plasma wasreduced by PF-1355 in a dose- and exposure-responsive manner(Fig. 2, A and B, respectively), without affecting total MPO

292 Zheng et al.

at ASPE

T Journals on O

ctober 12, 2020jpet.aspetjournals.org

Dow

nloaded from

mass levels (Supplemental Fig. 1). In this robust inflamma-tory model, the inhibitory effect of PF-1355 on MPO in thelocal exudate was similar to that in circulation (compareFig. 2, A and B). The relationship between [PF-1355] andMPO activity in plasma (Fig. 2C) was described by aninhibitory Imax model yielding an IC50 of 437 ng/ml (1.4 mM),similar to that observed in neutrophils and human blood(IC50 5 1.7 and 2.0 mM; Fig. 1, F and G, respectively).Additional characterization of PF-1355 pharmacokinetics inrat, mouse, and dog is provided in Supplemental Table 3.MPO Is a Key Mediator of Pulmonary Immune

Complex Vasculitis. The lung is a common vasculitistarget. It is highly perfused, having ready exchange withthe systemic circulation, and resulting injury can have acuteand critical consequences. MPO has been implicated in lungvasculitis associated with ANCA and anti-GBM mentionedpreviously, as well as in other diseases such as hemolyticuremic syndrome and transfusion-related acute lung injury(Fuchs et al., 2012; Thomas et al., 2012). Mice were treatedwith PF-1355 or vehicle and subjected to immune complexvasculitis by intravenous administration of BSA and lunginstillation of anti-BSA antibodies. Immune complex injurycaused perivessel lung edema (Fig. 3, Ai and Aii, comparecontrol with injury, respectively), which was diminished byPF-1355 in a dose- and exposure-dependent manner (Fig. 3,Aiii and Aiv; Supplemental Table 4), approximately 50% atthe high dose (Fig. 3B). To further characterize the role ofMPO in this model, we also examined MPO2/2 animals.Studies showed that disease severity was suppressed relativeto WT animals (Supplemental Fig. 2A), and edema scoreswere similar to that of WT animals treated with 100 mg/kgPF-1355 (Supplemental Fig. 2B). Furthermore, PF-1355 didnot significantly affect the already reduced edema scores inMPO-deficient mice (Supplemental Fig. 2B). Lung injury wasaccompanied by increased MPO plasma levels (Fig. 3E),which were partially suppressed by PF-1355 treatment at2 hours. However, despite the similar total MPO levels at4 hours (Fig. 3E), residual plasma MPO activity wasalmost completely inhibited by PF-1355 treatment in bothdose groups (Fig. 3D) and was consistent with PF-1355

plasma concentrations (Supplemental Table 4), which exceededthe estimated human whole blood IC50 at all times (Fig. 1G).Injury-induced lung edema appeared to be associated with

increased infiltrating cells (Fig. 3Aii). The role of MPO onleukocyte trafficking in the lung was studied by collectingbronchoalveolar lavage at the times indicated and neutrophilswere quantified (Fig. 3C). Neutrophil accumulation was ob-served by 30 minutes and increased with time. Greaterneutrophil numbers were found at 30 minutes in the high-dose PF-1355 group compared with vehicle controls, but thisdifference was gone at 2 hours. Recent studies showed thatH2O2 is an important neutrophil chemoattractant and MPOactually consumes substantial amounts of H2O2 in generatingHOCl and other oxidants (Pase et al., 2012). Therefore, theinitial MPO inhibitor–associated increase in neutrophilscould be related to elevated H2O2 bioavailability. Neutrophilaccumulation in bronchoalveolar lavage was suppressed byPF-1355 at later times, but continued to increase in vehiclecontrols. Thus, although MPO activity may augment neutro-phil recruitment at early times after immune complex injury,possibly involving H2O2 consumption, MPO activity clearlypotentiates neutrophil recruitment at later times.MPO promotes the production of inflammatory cytokines

through activation of endothelial cells and macrophages(Lefkowitz et al., 1992), a function stimulated by enzymaticallyinactive MPO under some conditions (Lefkowitz et al., 2000).We considered that reduced edema and leukocyte infiltrationassociated with MPO inhibition may involve such mediators.Immune complex lung injury was accompanied by markedelevation of tumor necrosis factor (TNF)-a, monocyte chemo-attractant protein-1, KC, and macrophage inflammatoryprotein-2; and this induction was inhibited by both dose levelsof PF-1355 (Fig. 3, F and G; Supplemental Fig. 3). Since totalplasma MPO levels were similar with or without PF-1355treatment at 0.5 and 4 hours (Fig. 3E), this suggests that MPOactivity is necessary for elevated cytokine levels. However,plasmaMPOmass levels were reduced by PF-1355 treatment at2 hours (Fig. 3E); therefore, a local or temporal effect of inactiveMPO on cytokine release cannot be excluded. Nonetheless,inhibition of MPO activity by PF-1355 significantly reduced

Fig. 2. Establishing a pharmacokinetic/pharmacodynamic relationship between plasma [PF-1355] and MPO inhibition in a mouse peritonitis model.Thioglycollate was administered by intraperitoneal injection; the next day, vehicle (0) or PF-1355 was dosed p.o. as indicated, followed by Zymosanadministration intraperitoneally (n = 4). (A and B) MPO activity from peritoneal lavage (A) and plasma (B) was measured after antibody capture as inFig 1G. Data represent the mean 6 S.E.M. (one-way analysis of variance; *P , 0.05 for indicated comparisons, and not significant if undesignated). (C)The pharmacokinetic/pharmacodynamic relationship between PF-1355 concentrations and residual MPO activity in plasma in the mouse peritonitismodel are represented in the graph. Closed circles represent observed data. The line represents model-based characterization of the pharmacokinetic/pharmacodynamic relationship using an inhibitory Imaxmodel implemented in NONMEM 7.2 software (ICON, Dublin, Ireland) assuming a proportionalerror model and interindividual variability in baseline (nondrug-treated) MPO activity. The shaded region represents the 95% confidence intervalsaround the best fit. The IC50 was estimated to be 437 ng/ml (S.E.E. = 74 ng/ml; 1.4 mM). Active MPO at baseline (nondrug treated) was estimated to be599 ng/ml (S.E.E. = 87). Residual variability was estimated to be 0.36 (S.E.E. = 0.05). Zym, Zymosan.

MPO Inhibition Prevents Vasculitis and Anti-GBM Diseases 293

at ASPE

T Journals on O

ctober 12, 2020jpet.aspetjournals.org

Dow

nloaded from

edema in experimental immune complex vasculitis and wasassociated with the suppression of cytokines and infiltratingleukocytes.MPO Activity Is Essential for Disease Induction in

a Model of Anti-GBM Disease. Anti-GBM disease iscaused by antibodies produced against the noncollagenousa3 chain of type IV collagen, present in the GBMs and in lungalveoli (Hellmark and Segelmark, 2014). Renal symptomsinclude hematuria, proteinuria, and hemorrhage (Cui andZhao, 2011), whereas respiratory symptoms include hypoxiaand hemorrhage. Anti-GBM disease may rapidly progressinto kidney failure, respiratory distress, and death. Treat-ment typically includes apheresis to remove nephrotoxicantibodies, glucocorticoids, cyclophosphamide (Levy et al.,2001), and recently rituximab (Schless et al., 2009). To model

disease, heterologous sera enriched in anti-GBM antibodiesare typically administered to mice. Disease severity is usuallymild to moderate, with renal dysfunction evident shortly afterantibody administration. Disease is driven initially by theinnate immune system, which may be followed by an adaptiveimmune phase. Consistent with this, neutrophil depletion andMPO2/2 animals were protected during the acute phase(Feith et al., 1996; Odobasic et al., 2007); however, in onereport, MPO deficiency exacerbated symptoms of the autolo-gous disease phase, suggesting that MPO may regulate theadaptive response and/or promote resolution.Disease induction and severity are known to be dependent

on the specific lot of anti-GBM antibody and it is essential tocharacterize the disease course with each specific source. Assuch, we prepared kidney specimens 2 hours (Fig. 4, A and B;

Fig. 3. PF-1355 andMPO deficiency protects in immune complex–induced pulmonary vasculitis. Mice treated with vehicle (0) or PF-1355 (20 or 100mg/kg)before induction of pulmonary immune complex vasculitis, consisting of intravenous BSA and lung administration of anti-albumin antibodies. (A)Hematoxylin and eosin–stained lung tissue sections, arrowsmarking vessels, negative control (i), immune complex injurywith vehicle (ii), 20mg/kg PF-1355(iii), or 100 mg/kg PF-1355 (iv). The inset in (iii) illustrates peri-vessel leukocytes. (B) Edema area was quantified by histomorphometry of 10 vessels peranimal (two pooled experiments, n = 20 mice). (C) Neutrophil influx was characterized in bronchoalveolar lavage collected at the indicated times. (D and E)Plasma was collected to characterize the kinetics of MPO activity and mass. (F and G) Plasma cytokines analyzed 2 and 4 hours postinjury using multiplexenzyme-linked immunosorbent assay (Supplemental Fig. 2). For B–G, n = 6 mice per group; graphs depict mean 6 S.E.M. (*P , 0.05; **P , 0.01; ***P ,0.001; ****P , 0.0001, one-way analysis of variance was performed). ImmComp, immune complex; MCP-1, monocyte chemoattractant protein-1.

294 Zheng et al.

at ASPE

T Journals on O

ctober 12, 2020jpet.aspetjournals.org

Dow

nloaded from

Fig. 4. MPO activity is essential for disease in-duction in a model of anti-GBM glomerulonephritis.Anti-GBM glomerulonephritis was induced by in-jection of sheep anti-rat GBM serum; normal sheepserum or PBS served as controls. (A–D) Kidneysections were prepared for fluorescent microscopy at2 hours (A and B; Supplemental Fig. 4, A and B) or21 days (C and D; Supplemental Fig. 4, C and D),specimens were costained for nuclei with Hoechst33342 (blue) and anti-sheep IgG (red) (A and C) andthe overlaid composite with anti-mouse IgG (green)(B and D). (E) Neutrophil accumulation after 2 hoursof sheep anti-GBM administration visualized byconfocal microscopy as follows: anti-sheep antibody(Alexa Fluor 647, purple), anti-MPO (Dylight 594,red), and anti-Ly6G (Dylight 488, green); costainingof anti-MPO and anti-Ly6G appears yellow. (F)Transmission electron micrographs prepared at21 days. The PBS control in (F) depicts normalglomerular architecture including uniform base-ment membrane (asterisk), segmentation of endo-thelial cells lining the capillary loops and podocytefoot processes lining the urinary space. (F–H) Sheepsera control (G) were generally similar to PBScontrols (F), whereas anti-GBM sera treatment(H), revealed evidence of subendothelial electrondense deposit, and swelling of podocytes accompa-nied by foot process effacement and fusion. Micewere pretreated with vehicle or PF-1355 beforeintravenous administration of PBS, control sheepserum, or anti-GBM serum and PF-1355 dosing(b.i.d.) was maintained until the end of the study.Twenty-four–hour urine was collected at the in-dicated times. (I) Urinarymicroalbumin and creatininewere determined by enzyme-linked immunosorbentassay and the UACR was calculated (n = 4–6 mice pergroup). (J–L) Similar results were obtained in twoindependent experiments. Plasma MPO activity (K)and protein levels (L) were measured 2 hours afteradministration of the indicated PBS, control sheepserum, or anti-GBM serum in animals treated withvehicle or PF-1355. Kidney neutrophil accumulationwas assessed in kidney specimens collected at 3 hours(J) (n = 6 per group), using anti-Ly6G andDAB stain asdescribed in Materials and Methods. Graphs depictmeans 6 S.E.M. One-way analysis of variance wasperformed adjusted for multiple comparisons. *P ,0.05; **P , 0.001; ****P , 0.0001 (anti-GBM–vehicleversus anti-GBM–PF-1355); ####P , 0.0001 (groupscompared were anti-GBM–vehicle versus control sheepsera–vehicle). CL, capillary loop; D, electron densedeposit, e, endothelial cell; f, foot process; NS, controlsheep serum; P, podocyte; UACR, urinary albumin/creatinine ratio; US, urinary space. Bar, 20 mm.

MPO Inhibition Prevents Vasculitis and Anti-GBM Diseases 295

at ASPE

T Journals on O

ctober 12, 2020jpet.aspetjournals.org

Dow

nloaded from

Supplemental Fig. 4, A and B) and 21 days (Fig. 4, C and D;Supplemental Fig. 4, C and D) after administration of sheepanti-rat GBM (Fig. 4, A–D) or PBS controls (SupplementalFig. 4, A–D). Anti-sheep antibody staining was found to bespecific to the GBM (Fig. 4, A and B, red), and absent incontrols (Supplemental Fig. 4, A and B). Mouse IgGaccumulated and colocalized with the anti-sheep antibodyafter 21 days (Fig. 4D, green), consistent with the adaptiveimmune response described by others (Feith et al., 1996;Odobasic et al., 2007). Kidneys were collected 2 hours afteranti-GBM or control sheep sera treatment and were preparedfor confocal analysis. Ly6G/MPO-positive neutrophils werecommonly seen in glomeruli of anti-GBM–treated animals(Fig. 4E) but rarely in control animals (Supplemental Fig. 4E).Glomerular ultrastructure analyzed by TEM revealed electron-rich deposits and thickened basement membrane in anti-GBM–treated animals (Fig. 4H), but not in mice treated withPBS or control sheep sera (Fig. 4, F and G, respectively). This isconsistent with the primary deposition of the sheep anti-GBMantibody as well as the accumulation of mouse anti-sheep IgG(Fig. 4B). Glomeruli from the anti-GBM–treated animalsdisplayed pathologic structural changes that included podocyteswelling, foot effacement and fusion, and dilation of the urinaryspace (Fig. 4H).Having established MPO-positive neutrophils and glomer-

ular pathology after anti-GBM treatment, we characterizedthe role of MPO activity in the kidney after anti-GBMadministration using PF-1355 (50mg/kg) administered 1 hourbefore anti-GBM treatment with continued b.i.d. administra-tion until the end of study. Kidney injury with an elevatedurinary albumin/creatinine ratio was evident by 24 hours andremained so over 3 weeks (Fig. 4I). PF-1355 treatment protectedmice from both the acute and adaptive disease phases and MPOdeficiency was similarly protective in this study (Fig. 4I). Thus,MPO activity is a critical component of the experimental anti-GBM disease process.Anti-GBM disease is an immune complex vasculitis of the

kidney. As in lung vasculitis (Fig. 3, D and E), we detectedincreased circulating plasma MPO levels after anti-GBMimmune complex injury (Fig. 4, K and L), whereas no changein plasma MPO was observed in mice treated with controlsheep sera. MPO activity was reduced in animals that receivedPF-1355 treatment (Fig. 4K), while total MPO was unaffected(Fig. 4L). Since MPO deficiency was previously reported toreduce neutrophil accumulation in the kidney after anti-GBMtreatment inmice (Odobasic et al., 2007), we examinedwhetherPF-1355 treatment similarly blocked neutrophil accumulationin anti-GBM disease. Kidney sections were stained with anti-Ly6G antibody and developed with anti-rat–conjugated HRPand DAB stain. DAB-positive area showed an increase (9.3times) in neutrophil accumulation 3 hours after anti-GBMadministration, and this was attenuated .70% by PF-1355treatment (Fig. 4J). MPO inhibition reduced neutrophil accu-mulation in experimental anti-GBM glomerulonephritis, as itdid in pulmonary immune complex vasculitis.

DiscussionLeukocyte-produced oxidants were linked to the pathogen-

esis of vasculitis in patients and in experimental modelsdecades ago (Sacks et al., 1978), and Minota et al. (1999)noted that increased MPO during disease exacerbations were

coincident with MPO antibody depletion, suggesting that MPOactivity may be involved in the pathogenesis of autoimmunevasculitis. Consistent with this, there is evidence for hypochlorite-modified proteins in patients with glomerulonephritis (Gröneet al., 2002). Despite this evidence, the lack of specific toolsimpeded efforts to define pathogenic mechanisms (Warren et al.,1990), as any treatment that reduces H2O2 will correspondinglyreduce MPO activity. Thus, the role of H2O2 as a primaryoxidant or as simply the catalyst for MPO activity had beenindefinable. The converse is not true; MPO is a significantconsumer of H2O2 and MPO deficiency elevates H2O2 levelsfrom zebrafish to humans (Gerber et al., 1996; Pase et al.,2012). Except for PF-1355 reported here, the only MPO-specificinhibitors were just recently described (Tidén et al., 2011) andnow offer the potential to elucidate the role of MPO activity inphysiology and disease.Here we describe the in vitro and in vivo pharmacology of

PF-1355, definingMPO activity as a central mediator of diseasepathology in distinct models of immune complex vasculitis. Theloss of residual plasma MPO activity in peritonitis and vas-culitis after oral administration of PF-1355 confirmed irrevers-ible target inhibition. Mechanism-based MPO inhibition byPF-1355 requires MPO catalysis (Fig. 1D) and these data areconsistent withMPO originating at sites of local inflammationclearing through the circulation. It is also possible that thestrong inflammatory stimulus used in these models leads tospillover activation of MPO in the circulation and that this isinhibited by the systemic exposure of PF-1355. Consistentwith the latter, circulating neutrophil numbers actuallyincrease approximately 3-fold in peritonitis (SupplementalFig. 5). Therefore, to varying degrees, both mechanisms maycontribute to elevated MPO plasma levels and depend on thespecific inflammatory stimulus and system affected.The lung is an important target organ in vasculitis. We

found that PF-1355 treatment reduced circulating MPOactivity nearly 90%, with an approximately 50% reductionof pulmonary edema; this effect is similar to that of potentanti-inflammatory treatments such as dexamethasone andanti–TNF-a (Warren et al., 1990; Warren, 1991). AlthoughPF-1355 treatment did not lower MPO mass levels at thetimes sampled in other inflammatory models, we did observea transient reduction in total MPO levels at 2 hours inpulmonary immune complex disease (Fig. 3E). The mecha-nism for this reduction was not defined by our study, but wespeculate that this may relate to a described function of MPOactivity in promoting NET formation. Similar to the results ofParker et al. (2012), we observed that MPO inhibition reducedphorbol 12-myristate 13-acetate–induced NET formation inhuman neutrophils in vitro (data not shown); and one mightexpect this to translate into a delay in the appearance ofplasmaMPO during acute inflammation in vivo with PF-1355treatment, a possibility we are currently examining.Neutrophil influx was blocked in immune complex disease

models by PF-1355 treatment and in pulmonary vasculitis; thiswas accompanied by sharp reductions in circulating cytokineshaving neutrophil and monocyte chemoattractant activityincluding TNF-a, monocyte chemoattractant protein-1 (CCL2),macrophage inflammatory protein-2 (CXCL2), and KC (CXCL1).The suppression of TNF-a is particularly noteworthy, becauseanti–TNF-a therapy was shown to be efficacious in immunecomplex vasculitis (Warren, 1991). These data indicate thatcytokine suppression is likely an important component of

296 Zheng et al.

at ASPE

T Journals on O

ctober 12, 2020jpet.aspetjournals.org

Dow

nloaded from

therapeutic activity associated with MPO inhibition. Inagreement with this, it was found that chronic treatmentusing another MPO inhibitor (TX-1; 3-isobutyl-2-thioxo-7H-purin-6-one) reduced modest increase of cytokine RNAs inthe lung tissue of the smoking guinea pig chronic obstructivepulmonary disease model (Churg et al., 2012). However, toour knowledge, the studies presented here are the first toshow that MPO inhibition actually suppressed circulatingcytokines induced by an acute and robust inflammatorystimulus. The specific mechanism(s) by which MPO exacer-bates inflammation and elevates cytokines remains animportant question, but is hypothesized to be linked toMPO-generated oxidation products, which include a varietyof isoprostenoids, lipid peroxides, and lipoproteins proposedto be proinflammatory downstream mediators (Zhang et al.,2002; Poliakov et al., 2003; Vasilyev et al., 2005; Huanget al., 2014).We found that MPO activity is critical in driving pathology in

anti-GBM disease and this is consistent with our results in theMPO2/2mice (Supplemental Fig. 6). This is also consistent withthe results of Rehan et al. (1984), who recognized the con-tribution of neutrophils and H2O2 in this disease, and ourresults now provide a mechanistic basis for the observed H2O2

requirement. Because MPO deficiency actually elevates H2O2

levels and since PF-1355 does not effectively scavengeH2O2 (Fig.1E), our results indicate that H2O2 drives disease primarily bysupporting MPO activity. These results vary from those ofOdobasic et al. (2007), who observed acute protection inMPO2/2

animals but worsening chronic disease in their anti-GBM study.The basis for this difference is not clear; however, it is recognizedthat different sources of anti-GBM antibody can have variabledisease severity and this may be related to the different outcomesbetween the two studies.Anti-GBM differs from the pulmonary vasculitis models in

important ways, including the presence of an endogenousantigen in kidney and lung tissues. Although leukocyterecruitment typically occurs in postcapillary venules in aprocess involving endothelial cell expression of P-selectin,glomerular endothelial cells express little P-selectin andneutrophil rolling and adhesion are not typically observed atthis site (Ley and Gaehtgens, 1991; Jung and Ley, 1997), witha notable exception being postnephrotoxic anti-GBM serumexposure. However, the role of P-selectin in this process hasbeen controversial. Mayadas et al. (1996) suggested thatP-selectin deficiency actually exacerbated anti-GBM diseaseand increased neutrophil accumulation, whereas others haveused intravital microscopy and showed that the anti-GBMkidney accumulated neutrophils involving P-selectin providedby platelets in trans, rather than by endothelial cell expressionof P-selectin (Kuligowski et al., 2006). MPO activity has beenimplicated in bidirectional platelet neutrophil activation inprocesses involving NET formation (Caudrillier et al., 2012).We speculate that MPO inhibition may limit NET formationand attenuate platelet neutrophil aggregation, thereby reducingaccumulation and activation of neutrophils in the affected tissue,a possibility that is currently under investigation.In summary, we demonstrated that MPO activity plays

a fundamental role in preclinical models of immune complexvasculitis diseases including anti-GBM disease. We alsoreported data illustrating that MPO activity is responsiblefor promoting inflammatory exacerbation accompanied byleukocyte recruitment and cytokine induction. Thus, an MPO

inhibitor could have a broad effect on the treatment ofvascular inflammatory conditions, particularly those result-ing from autoimmune responses, such as anti-GBM disease,systemic lupus erythematosus, and ANCA vasculitis.

Acknowledgments

The authors thank Sergey Fillipov for MPO insight, DonaldVanleeuwen and Rebecca Conrad for developing the MPO bloodassay, Carol Menard and Ingrid Stock for MPO assays, SamanthaSpath for MPO and TPO assays, Angela Wolford and StevenGernhardt for bioanalytical support, Dan Wydlicka and Dan Bowlesfor PF-1355 synthesis, Andrew Robertson and Alan Opsahl forhistology and image analysis, Tim Rolph for support and criticalfeedback, and John Kalliel for MPO activity assays.

Authorship Contributions

Participated in research design: Zheng, Warner, Ruggeri, Skoura,Ahn, Kalgutkar, Bonin, Okerberg, Bell, Johnson, Buckbinder.

Conducted experiments: Zheng, Warner, Su, Cortes, Skoura, Ward,Sun, Bobrowski, Kawabe, Bonin, Coskran, Bell, Kapoor.

Contributed new reagents or analytic tools: Ruggeri.Performed data analysis: Zheng, Warner, Ward, Ahn, Kalgutkar,

Maurer, Bonin, Zhang, Coskran, Johnson, Buckbinder.Wrote or contributed to the writing of the manuscript: Zheng,

Warner, Ruggeri, Skoura, Ward, Ahn, Kalgutkar, Maurer, Bonin,Coskran, Buckbinder.

References

Bart�unková J, Tesar V, and Sedivá A (2003) Diagnostic and pathogenetic role ofantineutrophil cytoplasmic autoantibodies. Clin Immunol 106:73–82.

Branzk N and Papayannopoulos V (2013) Molecular mechanisms regulating NETosisin infection and disease. Semin Immunopathol 35:513–530.

Caudrillier A, Kessenbrock K, Gilliss BM, Nguyen JX, Marques MB, Monestier M,Toy P, Werb Z, and Looney MR (2012) Platelets induce neutrophil extracellulartraps in transfusion-related acute lung injury. J Clin Invest 122:2661–2671.

Churg A, Marshall CV, Sin DD, Bolton S, Zhou S, Thain K, Cadogan EB, Maltby J,Soars MG, Mallinder PR, et al. (2012) Late intervention with a myeloperoxidaseinhibitor stops progression of experimental chronic obstructive pulmonary disease.Am J Respir Crit Care Med 185:34–43.

Cui Z and Zhao MH (2011) Advances in human antiglomerular basement membranedisease. Nat Rev Nephrol 7:697–705.

Dammacco F, Battaglia S, Gesualdo L, and Racanelli V (2013) Goodpasture’s disease:a report of ten cases and a review of the literature. Autoimmun Rev 12:1101–1108.

Feith GW, Assmann KJ, Bogman MJ, van Gompel AP, Schalkwijk J, and Koene RA(1996) Different mediator systems in biphasic heterologous phase of anti-GBMnephritis in mice. Nephrol Dial Transplant 11:599–607.

Franck T, Kohnen S, Boudjeltia KZ, Van Antwerpen P, Bosseloir A, Niesten A, GachO, Nys M, Deby-Dupont G, and Serteyn D (2009) A new easy method for specificmeasurement of active myeloperoxidase in human biological fluids and tissueextracts. Talanta 80:723–729.

Fuchs TA, Kremer Hovinga JA, Schatzberg D, Wagner DD, and Lämmle B (2012)Circulating DNA and myeloperoxidase indicate disease activity in patients withthrombotic microangiopathies. Blood 120:1157–1164.

Gaffo AL (2013) Thrombosis in vasculitis. Best Pract Res Clin Rheumatol 27:57–67.Gerber CE, Kuçi S, Zipfel M, Niethammer D, and Bruchelt G (1996) Phagocyticactivity and oxidative burst of granulocytes in persons with myeloperoxidase de-ficiency. Eur J Clin Chem Clin Biochem 34:901–908.

Goldman R, Claycamp GH, Sweetland MA, Sedlov AV, Tyurin VA, Kisin ER, TyurinaYY, Ritov VB, Wenger SL, Grant SG, et al. (1999) Myeloperoxidase-catalyzedredox-cycling of phenol promotes lipid peroxidation and thiol oxidation in HL-60cells. Free Radic Biol Med 27:1050–1063.

Gröne HJ, Gröne EF, and Malle E (2002) Immunohistochemical detection ofhypochlorite-modified proteins in glomeruli of human membranous glomerulone-phritis. Lab Invest 82:5–14.

Halbwachs L and Lesavre P (2012) Endothelium-neutrophil interactions in ANCA-associated diseases. J Am Soc Nephrol 23:1449–1461.

Hansson M, Olsson I, and Nauseef WM (2006) Biosynthesis, processing, and sortingof human myeloperoxidase. Arch Biochem Biophys 445:214–224.

Hellmark T and Segelmark M (2014) Diagnosis and classification of Goodpasture’sdisease (anti-GBM). J Autoimmun 48-49:108–112.

Huang Y, DiDonato JA, Levison BS, Schmitt D, Li L, Wu Y, Buffa J, Kim T,Gerstenecker GS, Gu X, et al. (2014) An abundant dysfunctional apolipoproteinA1 in human atheroma. Nat Med 20:193–203.

Johnson RJ, Klebanoff SJ, Ochi RF, Adler S, Baker P, Sparks L, and Couser WG(1987) Participation of the myeloperoxidase-H2O2-halide system in immune com-plex nephritis. Kidney Int 32:342–349.

Jung U and Ley K (1997) Regulation of E-selectin, P-selectin, and intercellular ad-hesion molecule 1 expression in mouse cremaster muscle vasculature. Microcir-culation 4:311–319.

Kettritz R (2012) How anti-neutrophil cytoplasmic autoantibodies activate neu-trophils. Clin Exp Immunol 169:220–228.

MPO Inhibition Prevents Vasculitis and Anti-GBM Diseases 297

at ASPE

T Journals on O

ctober 12, 2020jpet.aspetjournals.org

Dow

nloaded from

Klinke A, Nussbaum C, Kubala L, Friedrichs K, Rudolph TK, Rudolph V, Paust HJ,Schröder C, Benten D, Lau D, et al. (2011) Myeloperoxidase attracts neutrophils byphysical forces. Blood 117:1350–1358.

Kuligowski MP, Kitching AR, and Hickey MJ (2006) Leukocyte recruitment to theinflamed glomerulus: a critical role for platelet-derived P-selectin in the absence ofrolling. J Immunol 176:6991–6999.

Kutter D (1998) Prevalence of myeloperoxidase deficiency: population studies usingBayer-Technicon automated hematology. J Mol Med (Berl) 76:669–675.

Lefkowitz DL, Mills K, Morgan D, and Lefkowitz SS (1992) Macrophage activationand immunomodulation by myeloperoxidase. Proc Soc Exp Biol Med 199:204–210.

Lefkowitz DL, Roberts E, Grattendick K, Schwab C, Stuart R, Lincoln J, Allen RC,Moguilevsky N, Bollen A, and Lefkowitz SS (2000) The endothelium and cytokinesecretion: the role of peroxidases as immunoregulators. Cell Immunol 202:23–30.

Lekstrom-Himes JA and Gallin JI (2000) Immunodeficiency diseases caused bydefects in phagocytes. N Engl J Med 343:1703–1714.

Levy JB, Turner AN, Rees AJ, and Pusey CD (2001) Long-term outcome of anti-glomerular basement membrane antibody disease treated with plasma exchangeand immunosuppression. Ann Intern Med 134:1033–1042.

Ley K and Gaehtgens P (1991) Endothelial, not hemodynamic, differences are responsiblefor preferential leukocyte rolling in rat mesenteric venules. Circ Res 69:1034–1041.

Mayadas TN, Mendrick DL, Brady HR, Tang T, Papayianni A, Assmann KJ, WagnerDD, Hynes RO, and Cotran RS (1996) Acute passive anti-glomerular basementmembrane nephritis in P-selectin-deficient mice. Kidney Int 49:1342–1349.

Metzler KD, Fuchs TA, Nauseef WM, Reumaux D, Roesler J, Schulze I, Wahn V,Papayannopoulos V, and Zychlinsky A (2011) Myeloperoxidase is required forneutrophil extracellular trap formation: implications for innate immunity. Blood117:953–959.

Minota S, Horie S, Yamada A, Iwamoto M, Yoshio T, Mimori A, Masuyama J,and Kano S (1999) Circulating myeloperoxidase and anti-myeloperoxidase anti-body in patients with vasculitis. Scand J Rheumatol 28:94–99.

Odobasic D, Kitching AR, Semple TJ, and Holdsworth SR (2007) Endogenous mye-loperoxidase promotes neutrophil-mediated renal injury, but attenuates T cellimmunity inducing crescentic glomerulonephritis. J Am Soc Nephrol 18:760–770.

Parker H, Dragunow M, Hampton MB, Kettle AJ, and Winterbourn CC (2012)Requirements for NADPH oxidase and myeloperoxidase in neutrophil extracellulartrap formation differ depending on the stimulus. J Leukoc Biol 92:841–849.

Pase L, Layton JE, Wittmann C, Ellett F, Nowell CJ, Reyes-Aldasoro CC, Varma S,Rogers KL, Hall CJ, Keightley MC, et al. (2012) Neutrophil-delivered myeloper-oxidase dampens the hydrogen peroxide burst after tissue wounding in zebrafish.Curr Biol 22:1818–1824.

Poliakov E, Brennan ML, Macpherson J, Zhang R, Sha W, Narine L, Salomon RG,and Hazen SL (2003) Isolevuglandins, a novel class of isoprostenoid derivatives,function as integrated sensors of oxidant stress and are generated by myeloper-oxidase in vivo. FASEB J 17:2209–2220.

Rehan A, Johnson KJ, Wiggins RC, Kunkel RG, and Ward PA (1984) Evidence for therole of oxygen radicals in acute nephrotoxic nephritis. Lab Invest 51:396–403.

Sacks T, Moldow CF, Craddock PR, Bowers TK, and Jacob HS (1978) Oxygen radicalsmediate endothelial cell damage by complement-stimulated granulocytes. An invitro model of immune vascular damage. J Clin Invest 61:1161–1167.

Schless B, Yildirim S, Beha D, Keller F, and Czock D (2009) Rituximab in two cases ofGoodpasture’s syndrome. NDT Plus 2:225–227.

Short AK, Esnault VLM, and Lockwood CM (1995) Anti-neutrophil cytoplasm anti-bodies and anti-glomerular basement membrane antibodies: two coexisting distinctautoreactivities detectable in patients with rapidly progressive glomerulonephritis.Am J Kidney Dis 26:439–445.

Thomas GM, Carbo C, Curtis BR, Martinod K, Mazo IB, Schatzberg D, Cifuni SM, FuchsTA, von Andrian UH, Hartwig JH, et al. (2012) Extracellular DNA traps are associatedwith the pathogenesis of TRALI in humans and mice. Blood 119:6335–6343.

Tidén AK, Sjögren T, Svensson M, Bernlind A, Senthilmohan R, Auchère F, NormanH, Markgren PO, Gustavsson S, Schmidt S, et al. (2011) 2-thioxanthines aremechanism-based inactivators of myeloperoxidase that block oxidative stressduring inflammation. J Biol Chem 286:37578–37589.

Vasilyev N, Williams T, Brennan ML, Unzek S, Zhou X, Heinecke JW, Spitz DR,Topol EJ, Hazen SL, and Penn MS (2005) Myeloperoxidase-generated oxidantsmodulate left ventricular remodeling but not infarct size after myocardial in-farction. Circulation 112:2812–2820.

Ward J, Spath SN, Pabst B, Carpino PA, Ruggeri RB, Xing G, Speers AE, Cravatt BF,and Ahn K (2013) Mechanistic characterization of a 2-thioxanthine myeloperox-idase inhibitor and selectivity assessment utilizing click chemistry—activity-basedprotein profiling. Biochemistry 52:9187–9201.

Warner RL, Beltran L, Younkin EM, Lewis CS, Weiss SJ, Varani J, and Johnson KJ(2001) Role of stromelysin 1 and gelatinase B in experimental acute lung injury.Am J Respir Cell Mol Biol 24:537–544.

Warren JS (1991) Disparate roles for TNF in the pathogenesis of acute immune com-plex alveolitis and dermal vasculitis. Clin Immunol Immunopathol 61:249–259.

Warren JS, Yabroff KR, Mandel DM, Johnson KJ, and Ward PA (1990) Role of O2- inneutrophil recruitment into sites of dermal and pulmonary vasculitis. Free RadicBiol Med 8:163–172.

Yang R, Hellmark T, Zhao J, Cui Z, Segelmark M, Zhao MH, and Wang HY (2007)Antigen and epitope specificity of anti-glomerular basement membrane antibodiesin patients with goodpasture disease with or without anti-neutrophil cytoplasmicantibodies. J Am Soc Nephrol 18:1338–1343.

Zhang R, Brennan ML, Shen Z, MacPherson JC, Schmitt D, Molenda CE, and HazenSL (2002) Myeloperoxidase functions as a major enzymatic catalyst for initiation oflipid peroxidation at sites of inflammation. J Biol Chem 277:46116–46122.

Address correspondence to: Leonard Buckbinder, Cardiovascular andMetabolic Diseases Research Unit, Pfizer World Wide Research and De-velopment, 610 Main Street, 003/302, Cambridge, MA 02139. E-mail: leonard.buckbinder@pfizer.com

298 Zheng et al.

at ASPE

T Journals on O

ctober 12, 2020jpet.aspetjournals.org

Dow

nloaded from