The immunodominant myeloperoxidase T-cellepitope induces local cell-mediated injury inantimyeloperoxidase glomerulonephritisJoshua D. Ooia, Janet Changa, Michael J. Hickeya, Dorin-Bogdan Borzab, Lars Fuggerc, Stephen R. Holdswortha,d,and A. Richard Kitchinga,d,e,1

aCenter for Inflammatory Diseases, Department of Medicine, Monash University, and Departments of dNephrology and ePediatric Nephrology, MonashMedical Centre, Clayton 3168, Victoria, Australia; bDepartments of Medicine and Pathology, Vanderbilt University School of Medicine, Nashville, TN 37232;and cMedical Research Council Human Immunology Unit, Weatherall Institute of Molecular Medicine, John Radcliffe Hospital, University of Oxford, OxfordOX3 9DS, United Kingdom

Edited by Emil R. Unanue, Washington University School of Medicine, St. Louis, MO, and approved August 9, 2012 (received for review June 19, 2012)

Microscopic polyangiitis is an autoimmune small-vessel vasculitisthat often manifests as focal and necrotizing glomerulonephritisand renal failure. Antineutrophil cytoplasmic Abs (ANCAs) specificfor myeloperoxidase (MPO) play a role in this disease, but the roleof autoreactive MPO-specific CD4+ T cells is uncertain. By screeningoverlapping peptides of 20 amino acids spanning the MPO mole-cule, we identified an immunodominant MPO CD4+ T-cell epitope(MPO409–428). Immunizing C57BL/6 mice with MPO409–428 inducedfocal necrotizing glomerulonephritis similar to that seen afterwhole MPO immunization, when MPOwas deposited in glomeruli.Transfer of an MPO409–428-specific CD4+ T-cell clone to Rag1−/−

mice induced focal necrotizing glomerulonephritis when glomeru-lar MPO deposition was induced either by passive transfer of MPO-ANCA and LPS or by planting MPO409–428 conjugated to a murineantiglomerular basement membrane mAb. MPO409–428 also in-duced biologically active anti-MPO Abs in mice. The MPO409–428

epitope has a minimum immunogenic core region of 11 aminoacids, MPO415–426, with several critical residues. ANCA-activatedneutrophils not only induce injury but lodged the autoantigenMPO in glomeruli, allowing autoreactive anti-MPO CD4+ cells toinduce delayed type hypersensitivity-like necrotizing glomerularlesions. These studies identify an immunodominant MPO T-cellepitope and redefine how effector responses can induce injuryin MPO-ANCA–associated microscopic polyangiitis.

autoimmunity | lymphocytes | T helper 1 cells | macrophages

Small-vessel vasculitis is often induced by autoimmunity toneutrophil granule proteins, predominantly myeloperoxidase

(MPO) and proteinase 3 (Pr3) (1), as well as lysosomal mem-brane protein-2 (2). Although there is some overlap, autoim-munity to MPO is strongly associated with microscopic poly-angiitis and reactivity to Pr3 results in granulomatosis withpolyangiitis (GPA). The kidneys are frequently affected by focaland segmental necrotizing glomerulonephritis (FNGN), leadingto rapidly progressive glomerulonephritis and end-stage renalfailure. Due to the presence of auto-Abs to MPO and Pr3, thisdisease is also known as antineutrophil cytoplasmic Ab (ANCA)-associated vasculitis (3). Morbidity and mortality rates remainhigh, with a 5-y survival rate of 46–85% in microscopic poly-angiitis (4), and most treatments have limited effectiveness andsignificant toxicities (5).Evidence for a pathogenic role for ANCA in microscopic

polyangiitis includes the use of plasma exchange as therapy,a case report of lung hemorrhage in a neonate following pla-cental transfer of MPO-ANCA, and other observations inhumans (6, 7). Moreover, ANCA can activate neutrophils andpromote their adhesion in vitro (8) and in vivo (9–12). The ad-hesion of neutrophils in target tissues, particularly the kidney,induces injury by means of the release of injurious oxidants andenzymes, including MPO itself (13). In addition, transferred anti-

MPO Abs can induce neutrophil- and complement-mediatedFNGN (14–18), enhanced by infection-related signals like LPS(14, 19, 20).Although there is a rationale for autoreactive CD4+ cells

contributing to the development of disease in microscopic pol-yangiitis, their role is less clear. There is evidence that MPO-ANCA production requires antigen-specific CD4+ T cells (21,22). Furthermore, autoreactive MPO-specific CD4+ T cells canbe induced experimentally in animals (23), MPO-specific T cellsthat produce IFN-γ are present in the peripheral blood ofhumans with acute ANCA-associated vasculitis (24–26), andurinary CD4+ effector/memory cells reflect disease activity (27).Effector/memory CD4+ T cells, together with macrophages,tissue factor, and fibrin, are present in glomeruli of patients withANCA-associated glomerulonephritis (28, 29). Finally, in a mu-rine model of anti-MPO FNGN, where autoimmunity to MPO isinduced and glomerulonephritis is triggered by injection of sheepanti-mouse glomerular basement membrane (GBM) Ab, CD4+

T-cell depletion during the effector phase attenuated disease(23). Based on this evidence, we hypothesize that MPO-specificeffector CD4+ cells are important in disease by localizing toglomeruli and inducing a delayed type hypersensitivity (DTH)-like lesion. Anti-MPO CD4+ cells may localize to glomeruli byrecognizing MPO within glomeruli, acting as a planted glomer-ular autoantigen deposited by ANCA-activated neutrophils thathave degranulated and/or formed neutrophil extracellular traps(NETs) (30).Although MPO’s B-cell epitopes have been the subject of

studies mapping them to areas within the heavy chain (31, 32),the T-cell epitopes of MPO are undefined. Identifying MPO’s T-cell epitopes is important in understanding the pathogenesis ofanti-MPO disease and would represent progress toward less toxictherapies focused on the autoimmune response. In the currentstudies, we have defined an immunodominant CD4+ T-cell MPOepitope and then used this epitope to test the hypothesis thatantigen-specific CD4+ T cells recognize both this epitope andMPO itself in glomeruli and induce FNGN. This immunodo-minant T-cell epitope exists across at least three different MHCII alleles and also can induce MPO-ANCA. Our studies redefineour understanding of anti-MPO disease to now include a distinct

Author contributions: J.D.O., S.R.H., and A.R.K. designed research; J.D.O. and J.C.performed research; M.J.H., D.-B.B., and L.F. contributed new reagents/analytic tools;J.D.O., J.C., S.R.H., and A.R.K. analyzed data; and J.D.O. and A.R.K. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.1To whom correspondence should be addressed. E-mail: richard.kitching@monash.edu.

See Author Summary on page 15547 (volume 109, number 39).

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1210147109/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1210147109 PNAS | Published online September 5, 2012 | E2615–E2624

IMMUNOLO

GY

PNASPL

US

Dow

nloa

ded

by g

uest

on

Oct

ober

8, 2

020

role for effector CD4+ T cells that recognize MPO, which areplanted in glomerular capillaries by MPO-ANCA–activatedneutrophils. The pathogenesis of effector responses in micro-scopic polyangiitis should now be considered to include a se-quential mix of types II (Ab-mediated) and IV (DTH-like)hypersensitivity.

ResultsMPO409–428 Is the Immunodominant T-Cell Epitope of MPO. To defineCD4+ T-cell epitopes within MPO, we immunized groups ofC57BL/6 mice with pools of MPO 20-mers (from the mouseMPO sequence, with each peptide overlapping by 12 aa; TableS1). Separate groups of mice were injected with peptide poolsfrom the proregion of MPO (peptides 1–17), the light chain(peptides 18–31), or one of four quarters (H1–H4) of the heavychain (peptides 32–46, 47–61, 62–76, or 77–89). Splenocyteswere restimulated with each third of the relevant immunizingpool (e.g., splenocytes from the group immunized with peptides1–17 were restimulated with peptides 1–6, 7–12, or 13–17), andrecall responses were measured by IFN-γ enzyme-linked immu-nospot (ELISPOT) assay. The five peptide groups inducing thestrongest responses were peptide pools 7–12, 23–27, 52–56, 57–61,and 62–66 (Fig. 1A). Additional groups of mice were then immu-nized with one of these groups of peptides, and their draining lymphnode (LN) cells were restimulated with each of the individualpeptides from within the immunizing pool. The five individualpeptides that induced the strongest recall responses were peptides10, 24, 52, 57, and 61 (Fig. 1B). To determine the immunodominantMPO T-cell epitope, we then immunized separate groups of micewith each of these peptides and restimulated their draining LN cellsex vivo with either the immunizing peptide or recombinant mouseMPO using proliferation assays and ELISPOT assays for IFN-γ andIL-17A. Comparing recall responses, MPO peptide 52 (MPO409–428,PRWNGEKLYQEARKIVGAMV) induced the strongest re-sponses both to itself and to whole recombinant mouse MPO(Table 1). Proliferative responses were significantly higher thanthose induced by all other peptides. Although proinflammatorycytokine production after peptide 52 immunization was numer-ically greater than after all other peptides, peptide 61 also in-duced moderately strong IL-17A and IFN-γ production.Mice immunized with native mouse MPO (n = 5) developed

recall responses to peptide 52 but not to a control peptide, ov-albumin (OVA)323–339 (proliferation assay stimulation index:2.1 ± 0.2 vs. 1.0 ± 0.1, IFN-γ ELISPOT assay: 8.0 ± 1.8 spots, andIL-17A ELISPOT assay: 10.3 ± 1.7 spots; no samples restimu-lated with OVA323–339 showed any spots above media alone).Immunization with peptide 53 or 54, both of which overlap withpeptide 52 and induce recall responses by IFN-γ ELISPOT assayin mice immunized in peptide pool 52–56 (Fig. 1 A and B), didnot induce responses to peptide 52, 53, or 54 or to recombinantMPO (Fig. S1).

T-Cell Autoreactivity to MPO409–428 (MPO Peptide 52) Is Not MHC II-, I-Ab-, Restricted. Because there are no clear MHC class II associ-ations with anti-MPO glomerulonephritis, we sought to de-termine if MPO peptide 52 (MPO409–428) is immunoreactive inmice expressing different MHC II molecules. We immunizedBALB/c mice (expressing I-Ad) and humanized HLA-DRB1*15:01 transgenic (Tg) mice (without mouse MHC classII) (33) with the pool of MPO peptides 52–56. In both strains,peptide 52 was the immunogenic peptide (Fig. 2 A and B).BALB/c and HLA-DRB1*15:01 Tg mice immunized with mouseMPO peptide 52 developed autoreactive responses to both MPOpeptide 52 and to whole recombinant MPO (Fig. 2 C and D).

Defining the Core and Critical Residues of the MPO409–428 Epitope.Wethen further delineated the key areas of this T-cell epitope. Weidentified the critical peptide length for immunoreactivity by

immunizing C57BL/6 mice with the complete 20-aa MPO409–428(PRWNGEKLYQEARKIVGAMV) and measuring proliferativerecall responses to shortened 16-mers and 12-mers of MPO409–428(Fig. 3A). There were no detectable recall responses whenMPO409–428 was shortened from the C terminus to either a 16-mer, amino acids 409–424, or to a 12-mer, amino acids 409–420(Fig. 3A). However, recall responses with amino acids 413–428,containing the 16 C-terminal amino acids, were comparable torecall responses induced by full-length amino acids 409–428. Therecall response to the 12 C-terminal amino acids 417–428, al-though detectable, was reduced compared with MPO409–428. Todelineate the core immunoreactive residues of MPO409–428 fur-ther, mice were immunized with the 16-mer amino acids 413–428and recall responses to sequential 12-mers overlapping by 11 aawere measured. Restimulation of mice immunized with aminoacids 413–428 with the 12-mers containing amino acids 414–425

Fig. 1. MPO T-cell epitopes in C57BL/6 mice. (A) Overlapping peptides (8920-mers) spanning the entire mouse MPO sequence were used to immunizegroups of mice. Pro, proregion peptides (1–17); Light, light chain peptides(18–31). H1 (32–46), H2 (47–61), H3 (62–76), and H4 (77–89) are heavy chainpeptides. Splenocytes were restimulated ex vivo with each third of the im-munizing pool (x axis), and responses were measured by IFN-γ ELISPOT assay(mean number of spots minus baseline). Each dot represents the mean re-sponse from an individual mouse, and results are representative of two in-dependent experiments with a minimum of four mice per group. (B) Groupsof mice were immunized with one of the five strongest responding pools ofpeptides (peptides 7–12, 23–27, 52–56, 57–61, or 62–66) and draining LN cellswere restimulated ex vivo with individual peptides (x axis). Each dot representsthe mean response from an individual mouse, and results are representativeof two independent experiments with a minimum of four mice per group.

E2616 | www.pnas.org/cgi/doi/10.1073/pnas.1210147109 Ooi et al.

Dow

nloa

ded

by g

uest

on

Oct

ober

8, 2

020

and 415–426 induced recall responses comparable to thoseobtained by restimulation with the 16-mer immunogen aminoacids 413–428 (Fig. 3B). Therefore, the core immunoreactiveresidues of MPO409–428 are the 11 amino acids (415–425;KLYQEARKIVG). The critical residues within this 11-mer weredelineated by immunizing C57BL/6 mice with KLYQEARKIVGand restimulating draining LN cells with individual alanine-

substituted 11-mers [alanine itself (MPO420) was substituted bya serine]. Each 11-mer had one residue substituted by alanine insequential order. An individual amino acid was deemed a criticalresidue if its substitution by an alanine abrogated a positive exvivo proliferative response. The substitution of tyrosine, arginine,isoleucine, or valine defined these amino acids as the criticalresidues within this epitope (Fig. 3C).

MPO409–428 Immunization Induces Nephritogenic T-Cell Responses. Todetermine the nephritogenicity of the anti-MPO T-cell responsesgenerated by MPO409–428 immunization, a previously establishedmodel of anti-MPO–directed glomerulonephritis was used (23,34). In this model, C57BL/6 mice immunized with mouse MPOlose tolerance to MPO, but this autoimmunity, in itself, is notsufficient to induce glomerulonephritis, which is triggered byinjecting low-dose sheep anti-mouse GBM Ab. This Ab inducestransient glomerular neutrophil recruitment with MPO de-position in glomeruli. MPO409–428-immunized mice developedrenal injury with increased albuminuria and blood urea nitrogento a similar degree as mice immunized with whole native mouseMPO (Fig. 4 A and B). MPO409–428-immunized mice developedFNGN comparable to mice immunized with whole native mouseMPO, characterized by segmental glomerular necrosis and fibrindeposition within the glomerular tuft. Mice immunized withOVA323–339 and given anti-GBM Ab exhibited only mild lesions(Fig. 4 C–J). Analysis of glomerular cellular effectors showedincreased infiltration of CD4+ T cells and macrophages com-pared with control mice, similar to that seen after whole nativemouse MPO immunization (Fig. 4 K–M). Concurrently, a groupof mice (n = 6) immunized with the next highest respondingMPO peptide from Table 1 (peptide 61, MPO481–500) and givenanti-GBM Ab did not develop more proteinuria or histologicalglomerular injury than OVA323–339-immunized mice (6.1 ± 1.1vs. 5.3 ± 0.7 mg over 24 h and 16.1 ± 2.4 vs. 16.2 ± 2.0% ofglomeruli affected by necrosis, respectively).

Transfer of MPO409–428-Specific CD4+ T Cells into Rag1−/− Mice Induces

FNGN. To determine definitively whether T-cell autoreactivity tothis dominant MPO epitope is nephritogenic, we generatedmouse MPO409–428-specific T-cell clones from C57BL/6 mice.CD4+ lines from OT-II Tg mice (i.e., OVA323–339-specific) de-rived under the same conditions were control cells. We trans-ferred T-cell clones into naive Rag1−/− mice (MPO409–428-specificor OVA323–339-specific: 5 × 106 cells per mouse) and then im-munized mice with the clones’ cognate peptide. As before, MPOwas deposited in glomeruli using low-dose sheep anti-mouseGBM Ab and mice were culled 5 or 14 d after disease induction.At 5 d, mice that received the MPO409–428-specific T-cell clonedeveloped a DTH response to MPO injected intradermally 24 hbefore the end of the experiment, which was not detected in

Table 1. Immune responses induced by the five most immunogenic MPO peptides

Responses to immunizing peptide Responses to recombinant MPO

Peptide Proliferation, SI IFN-γ, spots IL-17A, spots Proliferation, SI IFN-γ, spots IL-17A, spots

10 2.3 ± 0.5 18 ± 4 13 ± 1 2.2 ± 0.2 21 ± 1 7 ± 124 3.8 ± 1.1 14 ± 2 12 ± 3 3.0 ± 0.3 19 ± 3 12 ± 152 9.2 ± 0.8* 64 ± 10** 67 ± 8** 4.3 ± 0.2* 39 ± 5*** 45 ± 5**57 2.4 ± 0.4 32 ± 5 33 ± 4 2.5 ± 0.2 22 ± 3 20 ± 461 5.6 ± 1.0 55 ± 10 64 ± 13 3.0 ± 0.3 32 ± 5 31 ± 3

Groups of mice were immunized with individual peptides, and immune responses to the immunizing peptideor to whole recombinant mouse MPO were measured ex vivo using an [3H]-thymidine proliferation assay andIFN-γ and IL-17A ELISPOT assays. SI, stimulation index; spots, mean number of spots minus baseline. Results arepresented as the mean ± SEM and are representative of three independent experiments, each with a minimumof five mice per group.*P < 0.05 vs. peptides 10, 24, 57, and 61; **P < 0.05 vs. peptides 10, 24, and 57; ***P < 0.05 vs. peptides 10 and 24.

Fig. 2. Immunogenicity of MPO peptide 52 (MPO409–428) is not restricted toMHC II I-Ab. The immunogenicity of MPO peptide 52 was examined in miceexpressing I-Ad (BALB/c) or human DR2 (DR2 Tg) (with the absence of murineMHC II and Tg expression of human HLA-DRB1*15:01). When immunizedwith the pool of peptides 52–56, peptide 52 is capable of inducing recallresponses in the context of I-Ad (BALB/c, n = 4) (A) or HLA-DR2 (DR2 Tg, n = 4)(B). When mice were immunized with peptide 52 (MPO409–428) (BALB/c, n = 6;DR2 Tg, n = 4), LN cells from both strains exhibited recall responses to boththe peptide itself (MPO peptide 52,●) and recombinant mouseMPO (rmMPO,○) compared with control OVA323–339 (□), as measured by proliferation (C andD). Each dot represents the mean response from an individual mouse. ***P <0.001. Results are representative of three independent experiments witha minimum of four mice per experiment. SI, stimulation index.

Ooi et al. PNAS | Published online September 5, 2012 | E2617

IMMUNOLO

GY

PNASPL

US

Dow

nloa

ded

by g

uest

on

Oct

ober

8, 2

020

control mice (footpad swelling: 0.4 ± 0.1 vs. 0.0 ± 0.0 mm).Splenocytes harvested at 14 d from mice that received MPO409–

428-specific clones showed recall responses to both MPO409–428and MPO (IFN-γ ELISPOT assay: 156 ± 33 and 170 ± 40spots, respectively).Mice receiving the MPO409–428-specific T-cell clone developed

progressive FNGN with increased albuminuria, blood urea ni-trogen, segmental glomerular necrosis, and fibrin depositioncompared with mice that received OVA323–339-specific T cellsand anti-GBM Ab (Fig. 5 A–F). There were increased CD4+ Tcells and macrophages, but not neutrophils, within glomeruli

compared with control clone recipients (Fig. 5 I–K). The degreeof perivascular and tubulointerstitial inflammation was also in-creased, with more CD4+ T cells and macrophages but similarnumbers of neutrophils in the tubulointerstitium (Fig. S2 andTable S2). Glomerular crescent formation was observed in micereceiving the MPO409–428-specific T-cell clone at day 14 (6 ± 2%of glomeruli affected) but not in mice receiving OVA323–339-specific T cells. In some mice, pulmonary histology showed signsof interstitial pneumonitis and lymphocytic bronchiolitis, butthese findings were inconsistent and not present in all clonetransfer studies.

Localizing MPO409–428 to Glomeruli Induces CD4+ T-Cell Clone-Mediated FNGN. To determine whether the presence of theT-cell epitope MPO409–428 itself in glomeruli can induce T-cell–mediated renal injury in mice with MPO409–428 CD4+ T cells, weinjected a biotinylated MPO409–428 peptide conjugated toa mouse anti-mouse monoclonal IgG1 (clone 8D1) specific forthe mouse noncollagenous domain of the α3 chain of type IVcollagen, which does not induce pathological albuminuria ormajor histological changes (35). Glomerular deposition of bio-tinylated MPO409–428 was demonstrated by direct immunofluo-rescence using streptavidin-FITC Ab (Fig. 6 A and B). Injectionof the MPO409–428-8D1 conjugate itself did not increase glo-merular neutrophil recruitment (Fig. 6C) or renal MPO activity(naive mice: 1.8 ± 0.2, 8D1 alone: 2.2 ± 0.5, 8D1-MPO409–428:2.0 ± 0.4; expressed as ΔA460/min × 10−3) at a 4-h time point.Compared with control Rag1−/− mice receiving OVA323–339-

specific cells and MPO409–428-8D1, mice receiving the MPO-specific clone (25 × 106 cells) followed by MPO409–428-8D1 de-veloped glomerular infiltrates of CD4+ T cells, macrophages,and neutrophils (Fig. 6 A–C) with FNGN (Fig. 6 H–K). Glo-merular crescent formation was observed in four of six micereceiving the MPO409–428-specific clone (3%, 5%, 5%, and 10%of glomeruli affected) but not in mice receiving OVA323–339-specific cells. These pathological abnormalities translated intofunctional injury with increased albuminuria and blood urea ni-trogen levels (Fig. 6 L and M). Similarly, perivascular in-flammation, tubulointerstitial injury, and infiltration of effectorcells were observed in mice that received the MPO409–428-specificclone (Table S2). These experiments show that MPO409–428 itselfin glomeruli, without the initial presence of neutrophils or en-dogenous MPO, is sufficient to induce effector MPO-specificT-cell localization and injury.

LPS and MPO-ANCA Recruit Neutrophils and Deposit MPO, Leading toCD4+ T-Cell–Mediated FNGN. Relapse of ANCA-associated vascu-litis is associated with infections (36, 37). At least part of theexplanation for these clinical observations may come from ex-perimental evidence showing that infection primes neutrophils[and potentially other cells (20)], allowing ANCA to bind to andactivate neutrophils (9, 19). This results in their recruitment totarget tissues, especially the kidney (9, 15). Modeling this processinvolves injecting LPS and transferring anti-MPO IgG (gener-ated in Mpo−/− mice) into mice. Although these Abs can induceinjury (14), recruiting neutrophils into glomeruli may also resultin the deposition of MPO within glomeruli, where effector CD4+

cells could potentially recognize MPO as a planted glomerularantigen. Having shown that MPO-specific CD4+ cells can rec-ognize the immunodominant MPO T-cell epitope, MPO409–428,in the glomerulus, we next tested the hypothesis that MPO fromanti-MPO Ab-activated neutrophils acts as an autoantigen intarget tissues, resulting in effector CD4+ cell-mediated injury.Mice were injected with LPS and anti-MPO IgG (generated byimmunizing Mpo−/− mice with murine MPO). Histologicalanalyses 3 h after injection showed [as previously demonstrated(20)] glomerular neutrophil recruitment after LPS and anti-MPO IgG (Fig. 7 A, C, and D). There was also increased renal

Fig. 3. Core immunogenic residues of MPO409–428. (A) C57BL/6 mice wereimmunized with MPO409–428 (MPO peptide 52), and reactivity to shortened16-mers and 12-mers was assessed by [3H]-thymidine proliferation in drain-ing LN cells. Ex vivo reactivity to the shortened 16-mers and 12-mers wascompared with reactivity induced by MPO peptide 52. (B) MPO413–428 wasused to immunize mice, reactivity to individual 12-mers was assessed by [3H]-thymidine proliferation in draining LN cells, and ex vivo reactivity to theindividual 12-mers was compared with MPO413–428. The core 11 amino acidresidues are the 11 amino acids shared by MPO414–425 and MPO415–426

(underlined) because both induced comparable reactivity to the immunizingMPO413–428. (C) To determine the critical residues, mice were immunized withthe core 11 amino acids, MPO415–425, and draining LN cells were restimulatedex vivo with alanine-substituted 11-mers. Alanine-substituted positions: +,MPO415–425; K, ALYQEARKIVG; L, KAYQEARKIVG; Y, KLAQEARKIVG; Q,KLYAEARKIVG; E, KLYQAARKIVG; A, KLYQESRKIVG; R, KLYQEAAKIVG; sec-ond K, KLYQEARAIVG; I, KLYQEARKAVG; V, KLYQEARKIAG; G, KLY-QEARKIVA; −, OVA323–339; MPO, whole recombinant mouse MPO. Reactivityto individual alanine-substituted 11-mers was compared with reactivity toMPO415–425 and showed that tyrosine (Y), arginine (R), isoleucine (I), and va-line (V) were critical to the response. *P < 0.05; ***P < 0.001. Results arerepresentative of two independent experiments, each with a minimum offour mice per group. Each dot represents the mean response from an in-dividual mouse. SI, stimulation index.

E2618 | www.pnas.org/cgi/doi/10.1073/pnas.1210147109 Ooi et al.

Dow

nloa

ded

by g

uest

on

Oct

ober

8, 2

020

MPO activity (Fig. 7B). Confocal microscopy using anti-CD45and MPO-specific Abs revealed the presence of leukocyte-asso-ciated MPO as well as extracellular MPO that had been de-posited in the glomerulus (Fig. 7E).Injecting LPS and anti-MPO Abs after transferring either 5 ×

106 or 25 × 106 MPO409–428-specific CD4+ T-cell clones inducedsignificant renal disease compared with transfer of LPS andOVA323–399-specific cells. MPO-specific effector T cells resultedin increased albuminuria, blood urea nitrogen, and FNGN withglomerular necrosis and fibrin deposition (Fig. 7 F–M). Renalinjury was more severe after 25 × 106 MPO-specific cells com-pared with 5 × 106 MPO-specific cells. Glomerular crescentformation was observed in three of six mice that received 25 ×106 MPO-specific cells (with 10%, 15%, and 45% of glomeruliaffected) but not in mice receiving 5 × 106 MPO-specific cells.There were also increases in glomerular infiltration of CD4+

T cells, macrophages, and neutrophils, similar to those observedwhen using the MPO409–428-8D1 conjugate to plant the T-cellepitope in glomeruli (Fig. 7 N–P). Perivascular and tubulointer-stitial injury followed a similar pattern (Table S2). Transfer of50 × 106 MPO409–428-specific cells resulted in markedly acceler-ated renal disease specific to mice receiving MPO-specific cells,and experiments were terminated at 6 d (Fig. S3).

MPO409–428 Induces Biologically Active MPO-ANCA. Indirect immu-nofluorescence using purified IgG on ethanol-fixed peritonealmouse neutrophils showed a p-ANCA pattern specific toMPO409–428-immunized mice and not seen in serum IgG fromOVA323–339-immunized mice or when IgG from MPO409–428-immunized mice was applied to neutrophils from Mpo−/− mice(Fig. 8 A–C). To determine if immunization with MPO409–428induced MPO-ANCA, we tested sera from MPO409–428-immu-nized C57BL/6 mice for MPO-specific IgG by ELISA. Sera fromOVA323–339- or whole native mouse MPO-immunized mice servedas negative and positive controls. MPO409–428-immunized micedeveloped MPO-ANCA IgG, albeit at lower titers compared withwhole native mouse MPO immunization (Fig. 8D). To determinewhether MPO-ANCA induced by MPO409–428 had functional ac-tivity, we transferred purified MPO-ANCA into LPS-primedC57BL/6 recipient mice. Compared with IgG from OVA323–339-immunized mice, serum IgG from MPO409–428-immunized miceresulted in increased glomerular neutrophil recruitment at 3 h, (Fig.8E), showing that these auto-Abs are biologically active.

DiscussionThe glomerulus is a common target in MPO-ANCA–associatedmicroscopic polyangiitis, with the severity of renal disease oftendefining the outcome. Although ANCA- and neutrophil-inducedinjury is important in microscopic polyangiitis, the presence ofCD4+ T cells (26) and the MPO itself (28, 30) within lesionsimplies that MPO-specific autoreactive T cells may localize toglomeruli and cause injury. The current studies demonstratea role for MPO-specific CD4+ cells in experimental ANCA-as-sociated FNGN and define an immunodominant CD4+-cellepitope of MPO (MPO409–428). Our findings support a modelwhereby both ANCA and autoreactive effector CD4+ cells areimportant in microscopic polyangiitis. After tolerance is lost toMPO (with MPO409–428 as a potential key epitope), antigen-

Fig. 4. MPO409–428 immunization induces nephritogenic autoimmunity.C57BL/6 mice were immunized with either OVA323–339 [control (n = 5), □],MPO409–428 (n = 6, ●), or native mouse MPO (n = 4, ○), and disease wastriggered by recruiting neutrophils to glomeruli with low-dose sheep anti-mouse GBM Ab. Mice immunized with an irrelevant antigen (OVA323–339)and injected with low-dose sheep anti-mouse GBM Ab developed minimalinjury, whereas mice immunized with MPO409–428 developed FNGN similarto that induced by immunization with whole-mouse MPO. Functional indicesof glomerular injury, albuminuria (A) and blood urea nitrogen (B), were

increased after MPO-specific immunization. Some glomeruli from MPO409–

428- and whole MPO–immunized mice exhibited segmental necrosis (H&E) (Cand E–G) with glomerular fibrin deposition (fibrin as the brown reactionproduct) (D and H–J). Photomicrographs of glomeruli and fibrin deposition aretaken at a magnification of 400×. (Scale bars: E–J, 10 μm.) Glomerular infiltratesof CD4+ T cells (K) andmacrophages (L) but not neutrophils (M) were increasedafter MPO409–428 immunization. gcs, glomerular cross section. *P < 0.05; **P <0.01; ***P < 0.001.

Ooi et al. PNAS | Published online September 5, 2012 | E2619

IMMUNOLO

GY

PNASPL

US

Dow

nloa

ded

by g

uest

on

Oct

ober

8, 2

020

specific CD4+ T cells provide T-cell help to autoreactive B cellsin producing MPO-ANCA. In many people who develop MPO-ANCA, these auto-Abs interact with and activate neutrophils,causing the neutrophils to be recruited to glomerular capillaries.Neutrophils mediate injury but, critically, also release the auto-antigen MPO, which then is present in glomeruli and can berecognized by autoreactive MPO-specific CD4+ cells. Glomer-ular localization of these damaging MPO-specific effector T cellsresults in the recruitment of effector leukocytes, the develop-ment of FNGN, and renal impairment.In contrast to other autoimmune diseases, there have been no

consistent MHC class II associations in microscopic polyangiitis.HLA typing studies in patients with ANCA-associated vasculitishave focused largely on GPA (most frequently marked by au-toimmunity to Pr3). In this disease, a variety of MHC class IIassociations have been found (reviewed in 38), with DPB1*04:01being the most consistent. There are, however no clear positiveassociations with microscopic polyangiitis, which usually fea-tures autoimmunity to MPO with MPO-ANCA. We thereforetested the immunogenicity of MPO409–428 in mice with I-Ad andmice with HLA-DRB1*15:01, finding that this peptide couldinduce immune responses in more than one MHC II haplotype.Further studies in C57BL/6 mice identified a minimum immu-nogenic peptide, MPO415–425, and several critical amino acidswithin this peptide.When glomerular disease was triggered with low-dose anti-

GBM Ab, MPO409–428 peptide-immunized mice (but notOVA323–339-immunized mice) developed significant FNGN,showing that immunity to this peptide induced FNGN. MPO isa major constituent of neutrophils and is released when neu-trophils induce inflammation, contributing to tissue injury bygenerating damaging oxidants (13, 39). However, in microscopicpolyangiitis following ANCA-induced glomerular neutrophil lo-calization, neutrophils release MPO that is deposited in glo-meruli (28, 30). Immunologically, the glomerulus becomes a sitewhere MPO, the autoantigen, is present and available for antigenrecognition by effector T cells.Therefore, after generating T helper (Th) 1 IFN-γ–secreting

MPO409–428-specific clones that induced dermal DTH in vivo, weperformed a series of transfer experiments demonstrating thatMPO-specific CD4+ cells induce FNGN after MPO has beendeposited in the glomerular microvasculature. This antigen-specific disease (not found in recipients of OVA-specific CD4+ Tcells) was characterized by segmental glomerular necrosis; glo-merular fibrin deposition; CD4+ T-cell, macrophage, and neu-trophil recruitment; and functional renal injury (albuminuria andraised blood urea nitrogen). Tubulointerstitial infiltrates andinjury, in the form of tubular dilation, necrosis, and protein castformation, as well as perivascular inflammation were present inmice receiving MPO-specific CD4+ cells.Because injection of anti-GBM Ab induces neutrophil re-

cruitment, increased renal MPO activity, and free MPO withinthe glomerulus, we initially injected these Abs to trigger disease(34, 39–41). However, this strategy of inducing glomerular MPOdeposition is confounded by anti-GBM Ab-induced neutrophilrecruitment and the enzymatic effects of MPO, which, together,complicate the study of MPO as an autoantigen. By defining animmunodominant MPO peptide, we have defined a role forMPO as a planted autoantigen distinct from its effects as aninjurious effector molecule. To determine whether MPO409–428itself could be recognized as a nephritogenic peptide withinglomeruli, we used a modification of our previously publisheddelivery system (35). By conjugating MPO409–428 to a non-immunogenic mouse anti-mouse GBM monoclonal IgG1 Ab, wecould plant MPO409–428 in glomeruli. Compared with some othermurine IgG subclasses, the murine IgG1 subclass has limitedcapacity to cause injury itself, because it fixes complement poorlyand has a relatively low affinity for leukocyte-activating Fcγ

Fig. 5. Transfer of T-cell clones specific for MPO409–428 induces progressiveFNGN. T-cell clones specific for MPO409–428 (5 × 106) were transferred intoRag1−/− mice (●), disease was triggered 7 d later with low-dose sheep anti-mouse GBM Ab, and experiments were ended after a further 5 d (n = 6) or 14d (n = 4). Control Rag1−/− mice that received sheep anti-mouse GBM Ab witha CD4+ T-cell line specific for OVA323–339 (○) (n = 6, 5 d; n = 4, 14 d) developedminimal injury, but mice given MPO409–428-specific cells developed pro-gressive functional injury with pathological albuminuria (A), increased bloodurea nitrogen levels (B), and FNGN (C) with glomerular fibrin deposition (D).Histologically, glomeruli from control mice at 14 d were near normal (E)compared with those of mice receiving MPO409–428-specific cells (F). (Scalebars: E and F, 20 μm.) Progressive increases in glomerular CD4+ T cells (G) andmacrophages (H) but not neutrophils (I) over time were seen in recipients ofMPO-specific T cells. gcs, glomerular cross section. *P < 0.05; **P < 0.01;***P < 0.001.

E2620 | www.pnas.org/cgi/doi/10.1073/pnas.1210147109 Ooi et al.

Dow

nloa

ded

by g

uest

on

Oct

ober

8, 2

020

receptors (42). At the dose that we used, the Ab itself inducesminimal leukocyte recruitment and no albuminuria (35).Therefore, compared with polyclonal sheep IgG or anti-MPOAbs, it served as a relatively inactive carrier that could localizeMPO409–428 to glomeruli. When effector CD4+ cells were trans-ferred into mice, only MPO-specific cells localized to glomeruliand induced significant FNGN with renal impairment.Studies in patients with ANCA-associated vasculitis show

associations between infection and disease relapse and activity,suggesting a role for infection in the pathogenesis of these dis-eases (36, 37). Having shown that MPO-specific T cells couldrecognize MPO409–428 within glomeruli to induce injury, wetriggered disease with transferred anti-MPO Abs and LPS,a prototypic infection-related signal. This is likely to be moreanalogous to the “clinical” situation, where enzymatically activeand autoantigenic MPO lodges in glomeruli when it is releasedby MPO-ANCA–activated neutrophils. LPS has activatingeffects on both neutrophils and intrinsic glomerular cells, in-cluding induction of the production of neutrophil chemo-attractants by glomerular endothelial cells (20). Although, asexpected, there was some relatively mild injury induced by theinitial LPS and anti-MPO Ab-induced glomerular neutrophil re-cruitment, MPO-specific T cells caused significant and dose-de-pendent FNGN, with the highest dose of cells resulting in rapidlyprogressive renal failure in an antigen-specific manner and pre-mature termination of experiments. These experiments show thatthe humoral and cellular arms of the autoimmune response toMPO collaborate to induce ANCA-associated vasculitis.There are several ways by which MPO, an abundant protein in

neutrophils, can become lodged in glomeruli in ANCA-associ-ated glomerulonephritis. Although ANCA-induced neutrophilrecruitment results in neutrophil degranulation, ANCA alsopromotes NET formation (30). These NETs are extracellular,include both MPO and Pr3, and are present in glomeruli inANCA-associated glomerulonephritis. In addition, neutrophilmicroparticles, which are present in ANCA-associated vasculitis,contain MPO and can localize to endothelial cells (43). Itremains to be determined which cell type presents MPO409–428 toeffector CD4+ cells in glomeruli, allowing antigen-specific localMPO recognition. Leukocytes traditionally present antigens toeffector CD4+ cells. Dendritic cells (rare in glomeruli) ormonocyte/macrophages may be important in ingesting MPOpresent within glomeruli to allow glomerular T-cell recognitionof antigen, whereas neutrophils themselves can also expressMHC II and present antigens under some circumstances (44).Alternatively, intrinsic glomerular cells could be important, be-cause they can both internalize MPO (45) and be in contact withintracapillary MPO-specific CD4+ T cells. Studies in murineglomerulonephritis induced by a planted foreign antigen supporta role for glomerular cells in both antigen recognition and acti-vation of effector CD4+ cells (46–48). Chimeric mice lackingMHC II expression (46) or the costimulatory molecule CD40(47) in renal tissue cells did not develop severe T cell-mediatedglomerulonephritis, and in the same model, glomerular CD80and CD86 contributed to injury (48).

Fig. 6. Planting MPO409–428 in glomeruli induces CD4+ cell-mediated FNGN.Injecting a mouse anti-mouse α3 chain of type IV collagen [α3(IV)NC1] IgG1mAb (clone 8D1) conjugated to biotinylated MPO409–428 into naive Rag1−/−

mice results in deposition of biotinylated MPO409–428 peptide in glomeruli.After 4 h, a streptavidin-FITC Ab showed no biotin in glomeruli of miceinjected with unconjugated 8D1 (n = 6) (A), but clear biotin signal wasobserved within glomeruli of mice injected with the 8D1-biotinylatedMPO409–428 conjugate (n = 6) (B). The tubular fluorescence in both panels isendogenous biotin within renal tubules. (C) Injection of 8D1-biotinylatedMPO409–428 conjugate (●) did not result in an increase in glomerular neu-trophils compared with naive (□) or unconjugated 8D1 mAb alone in re-cipient (○) Rag1−/− mice at 4 h. MPO409–428-specific CD4+ T-cell clones (25 ×106) were transferred into Rag1−/− mice (●), and disease was triggered 7 d

later by targeting MPO409–428 to glomeruli via MPO409–428 conjugated toclone 8D1. Mice were culled 14 d (n = 6) after triggering disease. ControlRag1−/− mice were injected with the OVA323–339-specific CD4+ T-cell linefollowed by the 8D1-MPO409–428 conjugate (○) (n = 6). The presence ofMPO409–428 in glomeruli resulted in glomerular localization of MPO-specificbut not OVA-specific T cells (D), with macrophage and neutrophils in glo-meruli (E and F). Recruitment of these leukocytes resulted in FNGN withglomerular fibrin deposition (G–J). The 8D1-MPO409–428 conjugate with OVA-specific CD4+ cells did not induce functional renal injury, but MPO409–428-specific CD4+ cells resulted in both albuminuria (K) and renal impairment (L).**P < 0.01; ***P < 0.001. (Scale bars: A, B, I, and J, 20 μm.)

Ooi et al. PNAS | Published online September 5, 2012 | E2621

IMMUNOLO

GY

PNASPL

US

Dow

nloa

ded

by g

uest

on

Oct

ober

8, 2

020

Both effector Th1 cells and Th17 cells are important in ex-perimental glomerulonephritis induced by a planted foreignantigen (35, 49–51), whereas in autoimmune diseases, cell-me-diated injury can be mediated by either Th1 cells, Th17 cells, orboth Th1 and Th17 cells (52–54). We have shown that either Th1cells or Th17 cells can induce injury in glomerulonephritis (35),

and evidence exists for the involvement of both Th1 and Th17cells in human immune renal injury (55, 56). Although somehuman observational studies and our recent experimental datasupport a role for Th17 cells and IL-17A in ANCA-associatedglomerulonephritis (34, 57, 58), other studies implicate Th1 cells(59). The current studies, in showing that injury can be mediated

Fig. 7. Coinjection of LPS and anti-MPO Abs deposits MPO in glomeruli and triggers cell-mediated glomerular injury and FNGN. Injecting LPS and anti-MPOAbs into naive Rag1−/− mice (n = 4 in each group) results in neutrophil recruitment (A, C, and D) and increased renal MPO activity after 4 h (B). (Scale bars: Cand D, 20 μm.) (E) When assessed by confocal microscopy, free [nonleukocyte (CD45)-associated] MPO could be detected in glomeruli (CD45, red; MPO, green;nonleukocyte-associated MPO in green). (Scale bar: 5 μm.) The dotted line shows the outline of the glomerulus. MPO409–428-specific CD4

+ T-cell clones weretransferred into Rag1−/− mice (5 × 106, n = 6; 25 × 106, n = 6; ●), and disease was triggered 7 d later by injecting LPS and anti-MPO Abs. Mice were culled aftera further 14 d. Control groups were Rag1−/− mice injected with LPS and anti-MPO Abs, followed by an OVA323–339-specific CD4

+ T-cell line (5 × 106, n = 4; 25 ×106, n = 5; ○). Mice with OVA-specific T cells injected with LPS and anti-MPO Abs showed only mild injury with modest albuminuria (F), no renal impairment(G), and low proportions of glomeruli affected by segmental necrosis (H) and fibrin deposition (I). Renal injury was markedly increased by MPO409–428-specificCD4+ T cells in a dose-dependent manner. Representative photomicrographs of glomeruli show only mild glomeruli hypercellularity in control mice (J) butFNGN in mice receiving MPO-specific CD4+ cells (K). A similar pattern was seen with glomerular fibrin deposition [control cells (L) and MPO-specific CD4+ cells(M)]. (Scale bars: J–M, 25 μm.) Glomerular leukocyte recruitment showed a dose-dependent increase in CD4+ cells (N) and antigen-specific increases inmacrophages (O) and neutrophils (P). gcs, glomerular cross section. *P < 0.05; **P < 0.01; ***P < 0.001.

E2622 | www.pnas.org/cgi/doi/10.1073/pnas.1210147109 Ooi et al.

Dow

nloa

ded

by g

uest

on

Oct

ober

8, 2

020

by IFN-γ–secreting Th1 cells, support a model wherein both Th1and Th17 cells mediate injury in anti-MPO disease.The MPO409–428 epitope can also induce MPO-ANCA with

some biological activity, because transfer of pooled IgG fromMPO409–428-immunized mice with LPS could induce modestglomerular neutrophil recruitment. The target areas for humanMPO-ANCA have been identified in regions within the MPOheavy chain (MPO387–745, corresponding to mouse MPO360–718)(31), which include our T-cell epitope, although it is not neces-sary for autoreactive T-cell and B-cell epitopes to be from similarparts of the autoantigen. Targeted antigen-specific therapy isa long-term aim in treating autoimmune disease. The identifi-cation of an important epitope within MPO provides a platformfor further work aimed at developing antigen-specific therapies,given the apparently relatively restricted range of autoantigens inmicroscopic polyangiitis with MPO-ANCA.When considered with published data on the role of ANCA in

disease, our studies demonstrate that tissue injury in microscopicpolyangiitis is mediated by the following series of events. Fol-lowing neutrophil priming and activation by the coordinate ac-tion of infection-related signals and MPO-ANCA, neutrophilsare recruited to glomeruli. Here, they initiate injury not only byreleasing injurious mediators but by depositing MPO, the auto-antigen, in glomerular capillaries. The presence of MPO withinthese small vessels allows effector MPO-specific CD4+ T cells tolocalize to glomeruli and induce a DTH-like necrotizing glo-merulonephritis. Therefore, the pathogenesis of microscopicpolyangiitis includes distinct and important roles for both MPO-

ANCA–activated neutrophils and autoreactive effector CD4+ Tcells that recognize MPO within glomeruli.

Materials and MethodsMPO Peptides, Proteins, and Mice. Peptide libraries for MPO T-cell epitopescreening assays were synthesized as PepSets (Mimotopes). Peptides were 20aa long and overlapped by 12 aa. Peptide sequences are based on NationalCenter for Biotechnology Information reference NP_034954 (mouse MPO;Table S1). Individual peptide immunization and restimulation assays wereperformed with >90% pure peptides by HPLC (Mimotopes or AusPep).OVA323–339 was purchased from Auspep. Native mouse MPO was purifiedfrom a mouse cell line, 32Dcl3, and recombinant mouse MPO was generatedusing a baculovirus system, as previously described (60). C57BL/6 and BALB/cmice were obtained from Monash Animal Services, Monash University.Rag1−/−, humanized MHC class II−/− HLA-DRB1*15:01 Tg (33), Mpo−/−, andOT-II mice were bred at Monash Medical Centre Animal Facilities. Male mice,aged 6–8 wk, were used for experiments and kept in specific pathogen-freeconditions at Monash Medical Centre Animal Facilities.

Identification of the MPO T-Cell Epitopes. To identify the T-cell immunogenicregions of MPO, groups of C57BL/6 mice were initially immunized s.c. withpeptide pools that correspond to the proregion of the molecule (peptides 1–17), the light chain (peptides 18–31), or one of four quarters (H1–H4) of theheavy chain (peptides 32–46, 47–61, 62–76, and 77–89; 10 μg of peptide permouse) in Freund’s incomplete adjuvant (Sigma). Spleens were harvested 6d postimmunization. Splenocytes were stimulated ex vivo with each third ofthe immunizing pool, and responses were measured by IFN-γ ELISPOT assay.Groups of mice were then immunized s.c. in the base of the tail with one ofthe peptide pools 7–12, 23–27, 52–56, 57–61, or 62–66 (10 μg of peptide permouse) in Freund’s complete adjuvant. Draining LN cells were harvested 10d postimmunization. Finally, C57BL/6 mice were immunized s.c. in the baseof tail with one of the five peptides that induced the strongest responses(peptides 10, 24, 52, 57, and 61; 10 μg per mouse) in Freund’s completeadjuvant. Draining LN cells were harvested 10 d postimmunization. Re-activity to individual peptides and to whole recombinant MPO was com-pared by [3H]-thymidine proliferation and IFN-γ and IL-17A ELISPOT assays.

Murine Model of anti-MPO–Directed Glomerulonephritis. Mice were immu-nized s.c. in the base of tail with 100 μg ofMPO409–428, OVA323–339, or 40 μg ofnative mouse MPO in Freund’s complete adjuvant, followed by a similar s.c.injection 7 d later in the back of the neck in Freund’s incomplete adjuvant.Seven days later, mice received an i.v. injection of 1 mg of sheep anti-mouseGBM Ab on 2 d consecutively. Injury was assessed 4 d after the last i.v.injection.

T-Cell Clone Transfer Models. Each Rag1−/− mouse received 5–50 × 106

(MPO409–428- or OVA323–339-specific cells) T cells, followed by s.c. injection of100 μg of the cognate peptide in Freund’s complete adjuvant. MPO wasdeposited in glomeruli 7 d posttransfer by (i) i.v. injection of 1 mg of sheepanti-mouse GBM Ab, (ii) i.v. injection of 150 μg of 8D1 mAb conjugated toMPO409–428, or (iii) i.p. injection of LPS (0.5 μg/g) and i.v. injection of proteinG-purified anti-MPO IgG Ab (50 μg/g) derived from immunizing Mpo−/− mice(20). Injury was assessed at day 5 or at day 14 posttransfer. To conjugateMPO409–428 to 8D1 mAb, 5 mg/mL (grown in-house) was reacted with 0.1 mg/mL N-succinimidyl-6-maleimido-caproate (Sigma) for 2 h. MPO409–428 (10 mg/mL) was combined with 8D1 mAb at 10-fold molar excess for 3 h, the re-action was stopped by adding 2 mM cysteine, and unconjugated peptidewas removed by dialysis in PBS. Conjugation was confirmed by Western blotusing streptavidin-HRP Abs (BD Biosciences).

Additional methods are detailed in SI Materials and Methods.

Statistics. Where there were three or more groups (the majority of experi-ments), including experiments across different time points, one-way ANOVAfollowed by Tukey’s posttest was used to assess differences. Where therewere only two groups, a Student t test was used. Means and SEMs are shown(*P < 0.05, **P < 0.01, and ***P < 0.001).

Study Approval. These studies were conducted in strict accordance with theAustralian code of practice for the care and use of animals for scientificpurposes by the National Health and Medical Research Council of Australia.Animal studies were approved by the Monash University Animal EthicsCommittee.

Fig. 8. MPO409–428-immunized mice develop biologically active MPO-ANCA.(A–C) Representative micrographs (400×) of ethanol-fixed peritoneal neu-trophils stained by indirect immunofluorescence with purified serum IgG.(Scale bars: 1 μm.) The result is representative of two individual experimentswith IgG purified from pooled sera. (D) Serum from C57BL/6 mice immunizedwith MPO409–428 develops detectable levels of MPO-specific IgG, as measuredby serum IgG ELISA (n = 6 per group). nmMPO, native mouse myeloperox-idase. (E ) Purified serum IgG was transferred into LPS-primed syngeneicrecipients, and glomerular recruitment of neutrophils was assessed 3 h later(n = 4 per group). □, serum from OVA323-339 immunized C57BL/6 mice; ●,serum from MPO409-428 immunized C57BL/6 mice; ○, serum from nativemouse myeloperoxidase immunized mice. ***P < 0.001.

Ooi et al. PNAS | Published online September 5, 2012 | E2623

IMMUNOLO

GY

PNASPL

US

Dow

nloa

ded

by g

uest

on

Oct

ober

8, 2

020

ACKNOWLEDGMENTS. We thank Prof. J. Dowling for advice on pulmonaryhistology; Prof. F. Carbone for the anti-Vα Ab; Dr. C. Lo, A. Li, and C. Lo fortechnical assistance; and Dr. H. Braley (Commonwealth Serum Laboratories,

Australia) for technical advice on 8D1 conjugation. These studies were fundedby National Health and Medical Research Council of Australia Program Grant334067 and Project Grant 1008849.

1. Falk RJ, Jennette JC (1988) Anti-neutrophil cytoplasmic autoantibodies with specificityfor myeloperoxidase in patients with systemic vasculitis and idiopathic necrotizingand crescentic glomerulonephritis. N Engl J Med 318:1651–1657.

2. Kain R, et al. (2008) Molecular mimicry in pauci-immune focal necrotizingglomerulonephritis. Nat Med 14:1088–1096.

3. Jennette JC, Falk RJ (1997) Small-vessel vasculitis. N Engl J Med 337:1512–1523.4. Corral-Gudino L, Borao-Cengotita-Bengoa M, Del Pino-Montes J, Lerma-Márquez JL

(2011) Overall survival, renal survival and relapse in patients with microscopic polyangiitis:A systematic review of current evidence. Rheumatology (Oxford) 50:1414–1423.

5. Bosch X, Guilabert A, Espinosa G, Mirapeix E (2007) Treatment of antineutrophilcytoplasmic antibody associated vasculitis: A systematic review. JAMA 298:655–669.

6. Bansal PJ, Tobin MC (2004) Neonatal microscopic polyangiitis secondary to transfer ofmaternal myeloperoxidase-antineutrophil cytoplasmic antibody resulting in neonatalpulmonary hemorrhage and renal involvement. AnnAllergy Asthma Immunol 93:398–401.

7. Jayne DR, et al.; European Vasculitis Study Group (2007) Randomized trial of plasmaexchange or high-dosage methylprednisolone as adjunctive therapy for severe renalvasculitis. J Am Soc Nephrol 18:2180–2188.

8. Williams JM, et al. (2007) Antineutrophil cytoplasm antibody-stimulated neutrophiladhesion depends on diacylglycerol kinase-catalyzed phosphatidic acid formation. JAm Soc Nephrol 18:1112–1120.

9. Kuligowski MP, et al. (2009) Antimyeloperoxidase antibodies rapidly induce alpha-4-integrin-dependent glomerular neutrophil adhesion. Blood 113:6485–6494.

10. Little MA, et al. (2005) Antineutrophil cytoplasm antibodies directed againstmyeloperoxidase augment leukocyte-microvascular interactions in vivo. Blood 106:2050–2058.

11. Nolan SL, et al. (2008) Mechanisms of ANCA-mediated leukocyte-endothelial cellinteractions in vivo. J Am Soc Nephrol 19:973–984.

12. Schreiber A, Xiao H, Falk RJ, Jennette JC (2006) Bone marrow-derived cells aresufficient and necessary targets to mediate glomerulonephritis and vasculitis inducedby anti-myeloperoxidase antibodies. J Am Soc Nephrol 17:3355–3364.

13. van der Veen BS, de Winther MP, Heeringa P (2009) Myeloperoxidase: Molecularmechanisms of action and their relevance to human health and disease. AntioxidRedox Signal 11:2899–2937.

14. Xiao H, et al. (2002) Antineutrophil cytoplasmic autoantibodies specific formyeloperoxidase cause glomerulonephritis and vasculitis in mice. J Clin Invest 110:955–963.

15. Xiao H, et al. (2005) The role of neutrophils in the induction of glomerulonephritis byanti-myeloperoxidase antibodies. Am J Pathol 167(1):39–45.

16. Xiao H, Schreiber A, Heeringa P, Falk RJ, Jennette JC (2007) Alternative complementpathway in the pathogenesis of disease mediated by anti-neutrophil cytoplasmicautoantibodies. Am J Pathol 170(1):52–64.

17. Huugen D, et al. (2007) Inhibition of complement factor C5 protects against anti-myeloperoxidase antibody-mediated glomerulonephritis in mice. Kidney Int 71:646–654.

18. Schreiber A, et al. (2009) C5a receptor mediates neutrophil activation and ANCA-induced glomerulonephritis. J Am Soc Nephrol 20:289–298.

19. Huugen D, et al. (2005) Aggravation of anti-myeloperoxidase antibody-inducedglomerulonephritis by bacterial lipopolysaccharide: Role of tumor necrosis factor-alpha. Am J Pathol 167(1):47–58.

20. Summers SA, et al. (2010) Intrinsic renal cell and leukocyte-derived TLR4 aggravateexperimental anti-MPO glomerulonephritis. Kidney Int 78:1263–1274.

21. Brouwer E, et al. (1991) Predominance of IgG1 and IgG4 subclasses of anti-neutrophilcytoplasmic autoantibodies (ANCA) in patients with Wegener’s granulomatosis andclinically related disorders. Clin Exp Immunol 83:379–386.

22. Yoshida M, et al. (2005) In vitro production of myeloperoxidase anti-neutrophilcytoplasmic antibody and establishment of Th1-type T cell lines from peripheral bloodlymphocytes of patients. Clin Exp Rheumatol 23:227–230.

23. Ruth AJ, et al. (2006) Anti-neutrophil cytoplasmic antibodies and effector CD4+ cellsplay nonredundant roles in anti-myeloperoxidase crescentic glomerulonephritis. J AmSoc Nephrol 17:1940–1949.

24. Griffith ME, Coulthart A, Pusey CD (1996) T cell responses to myeloperoxidase (MPO)and proteinase 3 (PR3) in patients with systemic vasculitis. Clin Exp Immunol 103:253–258.

25. Chavele KM, et al. (2010) Regulation of myeloperoxidase-specific T cell responsesduring disease remission in antineutrophil cytoplasmic antibody-associated vasculitis:The role of Treg cells and tryptophan degradation. Arthritis Rheum 62:1539–1548.

26. Brouwer E, Stegeman CA, Huitema MG, Limburg PC, Kallenberg CG (1994) T cellreactivity to proteinase 3 and myeloperoxidase in patients with Wegener’sgranulomatosis (WG). Clin Exp Immunol 98:448–453.

27. Abdulahad WH, Kallenberg CG, Limburg PC, Stegeman CA (2009) Urinary CD4+effector memory T cells reflect renal disease activity in antineutrophil cytoplasmicantibody-associated vasculitis. Arthritis Rheum 60:2830–2838.

28. Brouwer E, et al. (1994) Neutrophil activation in vitro and in vivo in Wegener’sgranulomatosis. Kidney Int 45:1120–1131.

29. CunninghamMA, Huang XR, Dowling JP, Tipping PG, Holdsworth SR (1999) Prominenceof cell-mediated immunity effectors in “pauci-immune” glomerulonephritis. J Am SocNephrol 10:499–506.

30. Kessenbrock K, et al. (2009) Netting neutrophils in autoimmune small-vessel vasculitis.Nat Med 15:623–625.

31. Erdbrügger U, et al. (2006) Mapping of myeloperoxidase epitopes recognized byMPO-ANCA using human-mouse MPO chimers. Kidney Int 69:1799–1805.

32. Suzuki K, et al. (2007) Analysis of risk epitopes of anti-neutrophil antibody MPO-ANCA in vasculitis in Japanese population. Microbiol Immunol 51:1215–1220.

33. Rich C, et al. (2004) Myelin oligodendrocyte glycoprotein-35-55 peptide inducessevere chronic experimental autoimmune encephalomyelitis in HLA-DR2-transgenicmice. Eur J Immunol 34:1251–1261.

34. Gan PY, et al. (2010) Th17 cells promote autoimmune anti-myeloperoxidaseglomerulonephritis. J Am Soc Nephrol 21:925–931.

35. Summers SA, et al. (2009) Th1 and Th17 cells induce proliferative glomerulonephritis.J Am Soc Nephrol 20:2518–2524.

36. Stegeman CA, et al. (1994) Association of chronic nasal carriage of Staphylococcusaureus and higher relapse rates in Wegener granulomatosis. Ann Intern Med 120(1):12–17.

37. de Lind van Wijngaarden RA, et al. (2008) Hypotheses on the etiology ofantineutrophil cytoplasmic autoantibody associated vasculitis: The cause is hidden,but the result is known. Clin J Am Soc Nephrol 3:237–252.

38. Willcocks LC, Lyons PA, Rees AJ, Smith KG (2010) The contribution of genetic variationand infection to the pathogenesis of ANCA-associated systemic vasculitis. Arthritis ResTher 12:202.

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

40. Kitching AR, et al. (2002) The requirement for granulocyte-macrophage colony-stimulating factor and granulocyte colony-stimulating factor in leukocyte-mediatedimmune glomerular injury. J Am Soc Nephrol 13:350–358.

41. Kuligowski MP, Kitching AR, Hickey MJ (2006) Leukocyte recruitment to the inflamedglomerulus: A critical role for platelet-derived P-selectin in the absence of rolling. JImmunol 176:6991–6999.

42. Nimmerjahn F, Ravetch JV (2005) Divergent immunoglobulin g subclass activitythrough selective Fc receptor binding. Science 310:1510–1512.

43. Hong Y, et al. (2012) Anti-neutrophil cytoplasmic antibodies stimulate release ofneutrophil microparticles. J Am Soc Nephrol 23:49–62.

44. Culshaw S, Millington OR, Brewer JM, McInnes IB (2008) Murine neutrophils presentClass II restricted antigen. Immunol Lett 118(1):49–54.

45. Yang JJ, et al. (2001) Internalization of proteinase 3 is concomitant with endothelialcell apoptosis and internalization of myeloperoxidase with generation of intracellularoxidants. Am J Pathol 158:581–592.

46. Li S, Kurts C, Köntgen F, Holdsworth SR, Tipping PG (1998) Major histocompatibilitycomplex class II expression by intrinsic renal cells is required for crescenticglomerulonephritis. J Exp Med 188:597–602.

47. Ruth AJ, Kitching AR, Semple TJ, Tipping PG, Holdsworth SR (2003) Intrinsic renal cellexpression of CD40 directs Th1 effectors inducing experimental crescenticglomerulonephritis. J Am Soc Nephrol 14:2813–2822.

48. Odobasic D, et al. (2005) Glomerular expression of CD80 and CD86 is required forleukocyte accumulation and injury in crescentic glomerulonephritis. J Am Soc Nephrol16:2012–2022.

49. Kitching AR, Holdsworth SR, Tipping PG (1999) IFN-gamma mediates crescentformation and cell-mediated immune injury in murine glomerulonephritis. J Am SocNephrol 10:752–759.

50. Paust HJ, et al. (2009) The IL-23/Th17 axis contributes to renal injury in experimentalglomerulonephritis. J Am Soc Nephrol 20:969–979.

51. Steinmetz OM, et al. (2011) The Th17-defining transcription factor RORγt promotesglomerulonephritis. J Am Soc Nephrol 22:472–483.

52. Korn T, Bettelli E, Oukka M, Kuchroo VK (2009) IL-17 and Th17 Cells. Annu RevImmunol 27:485–517.

53. Luger D, et al. (2008) Either a Th17 or a Th1 effector response can driveautoimmunity: Conditions of disease induction affect dominant effector category. JExp Med 205:799–810.

54. Jäger A, Dardalhon V, Sobel RA, Bettelli E, Kuchroo VK (2009) Th1, Th17, and Th9effector cells induce experimental autoimmune encephalomyelitis with differentpathological phenotypes. J Immunol 183:7169–7177.

55. Holdsworth SR, Kitching AR, Tipping PG (1999) Th1 and Th2 T helper cell subsetsaffect patterns of injury and outcomes in glomerulonephritis. Kidney Int 55:1198–1216.

56. Kitching AR, Holdsworth SR (2011) The emergence of TH17 cells as effectors of renalinjury. J Am Soc Nephrol 22:235–238.

57. AbdulahadWH, Stegeman CA, Limburg PC, Kallenberg CG (2008) Skewed distributionof Th17 lymphocytes in patients with Wegener’s granulomatosis in remission.Arthritis Rheum 58:2196–2205.

58. Nogueira E, et al. (2010) Serum IL-17 and IL-23 levels and autoantigen-specific Th17cells are elevated in patients with ANCA-associated vasculitis. Nephrol Dial Transplant25:2209–2217.

59. Lúdvíksson BR, et al. (1998) Active Wegener’s granulomatosis is associated with HLA-DR+ CD4+ T cells exhibiting an unbalanced Th1-type T cell cytokine pattern: Reversalwith IL-10. J Immunol 160:3602–3609.

60. Apostolopoulos J, Ooi JD, Odobasic D, Holdsworth SR, Kitching AR (2006) Theisolation and purification of biologically active recombinant and native autoantigensfor the study of autoimmune disease. J Immunol Methods 308(1–2):167–178.

E2624 | www.pnas.org/cgi/doi/10.1073/pnas.1210147109 Ooi et al.

Dow

nloa

ded

by g

uest

on

Oct

ober

8, 2

020