of July 6, 2016.This information is current as Pancreatitis

Human IgG4-Related Autoimmuneof Murine Autoimmune Pancreatitis and

Production Are Prominent FeaturesαIFN-Plasmacytoid Dendritic Cell Activation and

WatanabeOkazaki, Warren Strober, Tsutomu Chiba and TomohiroAkifumi Takaori-Kondo, Masatoshi Kudo, Kazuichi Kiyomi Mizugishi, Kazushige Uchida, Norimitsu Kadowaki,Masahiro Shiokawa, Yuzo Kodama, Toshiharu Sakurai, Yasuyuki Arai, Kouhei Yamashita, Katsutoshi Kuriyama,

http://www.jimmunol.org/content/195/7/3033doi: 10.4049/jimmunol.1500971August 2015;

2015; 195:3033-3044; Prepublished online 21J Immunol 

MaterialSupplementary

1.DCSupplemental.htmlhttp://www.jimmunol.org/content/suppl/2015/08/20/jimmunol.150097

Referenceshttp://www.jimmunol.org/content/195/7/3033.full#ref-list-1

, 20 of which you can access for free at: cites 57 articlesThis article

Subscriptionshttp://jimmunol.org/subscriptions

is online at: The Journal of ImmunologyInformation about subscribing to

Permissionshttp://www.aai.org/ji/copyright.htmlSubmit copyright permission requests at:

Email Alertshttp://jimmunol.org/cgi/alerts/etocReceive free email-alerts when new articles cite this article. Sign up at:

Print ISSN: 0022-1767 Online ISSN: 1550-6606. Immunologists, Inc. All rights reserved.Copyright © 2015 by The American Association of9650 Rockville Pike, Bethesda, MD 20814-3994.The American Association of Immunologists, Inc.,

is published twice each month byThe Journal of Immunology

at kinki university medical library on July 6, 2016

http://ww

w.jim

munol.org/

Dow

nloaded from

at kinki university medical library on July 6, 2016

http://ww

w.jim

munol.org/

Dow

nloaded from

The Journal of Immunology

Plasmacytoid Dendritic Cell Activation and IFN-a ProductionAre Prominent Features of Murine Autoimmune Pancreatitisand Human IgG4-Related Autoimmune Pancreatitis

Yasuyuki Arai,* Kouhei Yamashita,* Katsutoshi Kuriyama,† Masahiro Shiokawa,†

Yuzo Kodama,† Toshiharu Sakurai,‡ Kiyomi Mizugishi,* Kazushige Uchida,x

Norimitsu Kadowaki,* Akifumi Takaori-Kondo,* Masatoshi Kudo,‡ Kazuichi Okazaki,x

Warren Strober,{ Tsutomu Chiba,† and Tomohiro Watanabe†,{,‖

The abnormal immune response accompanying IgG4-related autoimmune pancreatitis (AIP) is presently unclear. In this study, we

examined the role of plasmacytoid dendritic cell (pDC) activation and IFN-a production in this disease as well as in a murine model

of AIP (MRL/Mp mice treated with polyinosinic-polycytidylic acid). We found that the development of AIP in treated MRL/Mp

mice occurred in parallel with pancreatic accumulation of pDCs producing IFN-a, and with pDC depletion and IFN-a-blocking

studies, we showed that such accumulation was necessary for AIP induction. In addition, we found that the pancreas of treated

MRL/Mp mice contained neutrophil extracellular traps (NETs) shown previously to stimulate pDCs to produce IFN-a. Consistent

with these findings, we found that patients with IgG4-related AIP also exhibited pancreatic tissue localization of IFN-a–expressing

pDCs and had significantly higher serum IFN-a levels than healthy controls. In addition, the inflamed pancreas of these patients

but not controls also contained NETs that were shown to be capable of pDC activation. More importantly, patient pDCs cultured

in the presence of NETs produced greatly increased levels of IFN-a and induced control B cells to produce IgG4 (but not IgG1) as

compared with control pDCs. These data suggest that pDC activation and production of IFN-a is a major cause of murine AIP; in

addition, the increased pDC production of IFN-a and its relation to IgG4 production observed in IgG4-related AIP suggest that

this mechanism also plays a role in the human disease. The Journal of Immunology, 2015, 195: 3033–3044.

Immunoglobulin G4–related disease (IgG4-RD) is a newlyestablished chronic inflammatory disorder characterized bymassive infiltration of IgG4-expressing plasma cells into mul-

tiple organs, elevated serum IgG4 levels, and storiform fibrosis (1, 2).Various inflammatory disorders, such as those formerly diagnosed asautoimmune pancreatitis (AIP), sialadenitis, and autoimmune chol-angitis, are now regarded as organ-specificmanifestations of systemicIgG4-RD (1–4). Although elevated serum IgG4 levels define thepresence of IgG4-RD–associated inflammation in a variety of organs,and thus it is essential for disease diagnosis, the pathophysiologicalrole of this elevated Ig subtype is poorly understood. On one hand,elevated IgG4 levelsmaybeanepiphenomenonassociatedwith IgG4-RD–associated chronic inflammation, because this IgG subtypeexhibits limited ability to bind to complement and FcgRs (5, 6) andhas been shown to have anti-inflammatory activity in experimentalmyasthenia gravis (7). On the other hand, enhanced IgG4 levels mayplay a vital immunopathologic role in this disease, because IgG4Absagainst autoantigens have also been shown to participate in thepathogenesis of autoimmunity (8, 9).Adaptive immune responses involving T cell–derived cytokines

such as IL-4, IL-10, and IL-13 that have been found to promote IgG4production by B cells (5, 10, 11) may be involved in IgG4-RD–associated IgG4 production, because the number of Th2 cells andregulatory T cells (Tregs) producing these cytokines are increased inIgG4-RD patient lesions and peripheral blood (12–14). However,innate immune responses that activate B cells in a T cell–independentmanner to induce Ig production upon exposure to microbial-associated molecular patterns (MAMPs) and damage-associatedmolecular patterns (15) may also be involved in the enhanced IgG4production. This possibility is suggested by our previous studies inwhichwe reported that TLR and nucleotide-binding oligomerizationdomain-like receptor activation in monocytes (16) and basophils(17) from patients with IgG4-RD enhance IgG4 production byB cells from healthy control individuals via production of BAFF.

*Department of Hematology and Oncology, Graduate School of Medicine, Kyoto Univer-sity, Kyoto 606-8507, Japan; †Department of Gastroenterology and Hepatology, GraduateSchool of Medicine, Kyoto University, Kyoto 606-8507, Japan; ‡Department of Gastro-enterology and Hepatology, Kinki University School of Medicine, Osaka 589-8511, Japan;xDepartment of Gastroenterology and Hepatology, Kansai Medical University, Osaka 573-1191, Japan; {Mucosal Immunity Section, Laboratory of Host Defenses, National Instituteof Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD 20892;and ‖Center for Innovation in Immunoregulative Technology and Therapeutics, GraduateSchool of Medicine, Kyoto University, Kyoto 606-8507, Japan

Received for publication April 27, 2015. Accepted for publication July 23, 2015.

This work was supported by Grants-in-Aid for Scientific Research 25293172 and15K15370 from the Japan Society for the Promotion of Science, the Japanese Societyof Gastroenterology, the Kato Memorial Trust for Nambyo Research, Special Coordina-tion Funds from the Ministry of Education, Culture, Sports, Science and Technology ofJapan, and Astellas Pharma Inc. in Creation of Innovation Centers for Advanced Inter-disciplinary Research Areas Program. This work was also supported by Health and LaborSciences Research Grants for Research on Intractable Diseases from the Ministry ofHealth, Labor and Welfare, Japan (to T.W.) and by Grant-in-Aid for Scientific Research15K09499 from the Japan Society for the Promotion of Science (to K.Y.).

Address correspondence and reprint requests to Dr. Tomohiro Watanabe or Dr. KouheiYamashita, Department of Gastroenterology and Hepatology, Graduate School of Medi-cine, Kyoto University, 54 Shogoin Kawahara-cho, Sakyo-ku, Kyoto 606-8507, Japan(T.W.) or Department of Hematology and Oncology, Graduate School of Medicine, KyotoUniversity, 54 Shogoin Kawahara-cho, Sakyo-ku, Kyoto 606-8507, Japan (K.Y.). E-mailaddresses: tmhrwtnb@kuhp.kyoto-u.ac.jp (T.W.) or kouhei@kuhp.kyoto-u.ac.jp (K.Y.)

The online version of this article contains supplemental material.

Abbreviations used in this article: AIP, autoimmune pancreatitis; BDCA, blood den-dritic cell Ag; IFNAR, type I IFN receptor; IgG4-RD, IgG4-related disease; LF, lacto-ferrin; MAMP, microbial-associated molecular pattern; MPO, myeloperoxidase; MSU,monosodium urate; NET, neutrophil extracellular trap; pDC, plasmacytoid dendriticcell; PDCA, pDC Ag; poly (I:C), polyinosinic-polycytidylic acid; Siglec, sialic acid–binding Ig-like lectin; SLE, systemic lupus erythematosus; Treg, regulatory T cell.

Copyright� 2015 by The American Association of Immunologists, Inc. 0022-1767/15/$25.00

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1500971

at kinki university medical library on July 6, 2016

http://ww

w.jim

munol.org/

Dow

nloaded from

Plasmacytoid dendritic cells (pDCs) are another potentialsource of BAFF in IgG4-RD, because these cells produce IFN-a,a cytokine known to enhance BAFF production (18, 19). The in-volvement of pDCs in IgG4-RD immunopathogenesis aligns withthe fact that pDCactivation leading to increased IFN-aproductionhas been postulated to be a major mechanism of autoantibodyproduction in systemic lupus erythematosus (SLE). In this dis-ease, FcR-mediated endocytosis of DNA/RNA-containing im-mune complexes are thought to activate pDCs expressing TLR7and TLR9 and the pDCs so activated then induce autoantibodies(20, 21). However, little is known regarding the potential in-volvement of pDC-mediated IFN-a production in IgG4-RD de-velopment.In the current study, we show that increased number of pDCs

producing IFN-a populate the inflamed pancreas in both the murinemodel of AIP and patients with IgG4-relatedAIP. In themurinemodel,we demonstrate by pDC depletion and IFN-a receptor blocking thatthese cells are amajor cause of the pancreatitis. Furthermore, in boththe murine model and humans with IgG4-related AIP, the inflamedpancreas contained neutrophil extracellular traps (NETs) that couldbe shown to activate pDCs from patients with IgG4-related AIP toproduce IFN-a and BAFF, which leads to production of IgG4by B cells. Overall, these data provide strong evidence that pDCactivation and IFN-a production are prominent features of IgG4-related AIP.

Materials and MethodsInduction of AIP in MRL/Mp mice

Female MRL/Mp mice were purchased from Japan SLC (Shizuoka, Japan).All mice were bred at Kyoto University animal facility under specificpathogen-free conditions. This study was approved by the Kyoto Universityethics review boards for animal experiments. Six-week-old femaleMRL/Mpmice were treated by i.p. injections with 100 mg polyinosinic-polycytidylicacid [poly (I:C); InvivoGen, San Diego, CA] twice a week for a total of 14 or16 injections to induce experimentalAIP (22, 23). In some experiments,micewere treatedwithAb against bonemarrow stromal cell Ag 2 (100mg, 120G8;Bioceros, Utrecht, the Netherlands) and Ab against IFN-ab receptor(100 mg, anti–type I IFN receptor [IFNAR] Ab, MAR1-5A3; BD Bio-sciences, San Jose,CA) (24) prior to eachpoly (I:C) injection todeplete pDCsand inhibit type I IFN–mediated signaling pathways, respectively. Controlmicewere treatedwith either 100mg rat IgG (Sigma-Aldrich, St. Louis,MO)or mouse IgG (Sigma-Aldrich).

Cytokine and chemokine array analysis

Pancreatic lysates were prepared as described previously (25). Pancreaticlysates (100mg) and serum (100ml) were applied to a commercialmembranewith boundmouse cytokine array panelA (R&DSystems,Minneapolis,MN)to screen for cytokine and chemokine profiles.

Flow cytometry analysis of the intrapancreatic immune cells

The pancreas was minced with scissors and shaken in RPMI 1640 buffercontaining 10% FBS, 25 mmol/l HEPES, and 3 mg/ml collagenase (Wako,Osaka, Japan) at 37˚C and 180 rpm for 30 min. The cell suspension wasspun at 50 3 g for 30 s to remove debris, followed by centrifugation with30%Percoll solution (GEHealthcare, Tokyo, Japan) at 2000 rpm for 30minat room temperature. Isolated pancreatic immune cells were stained withFITC, PE, or allophycocyanin-conjugated B220 Ab (eBioscience, SanDiego, CA), pDC Ag (PDCA)-1 (eBioscience), Gr-1 Ab (eBioscience),CD3 Ab (eBioscience), CD11b Ab (BD Biosciences), CD19 Ab (Bio-Legend, San Diego, CA), sialic acid–binding Ig-like lectin (Siglec)-H Ab(BioLegend), CD11c Ab (BioLegend), rat IgG (BioLegend), or hamsterIgG (BioLegend). Flowcytometric analysiswas performed using anAccuriC6 flow cytometer (BD Biosciences) and CFlow Plus software (BD Bio-sciences).

Serum cytokine assay

Serum was subjected to separate ELISAs for a murine cytokine Ab array,IFN-a, IFN-b, BAFF (R&D Systems), and dsDNA Ab titer (Shibayagi Co.,Gunma, Japan). Serum levels of anti-lactoferrin (LF) Ab titer weremeasuredas previously described (26).

Histology and immunohistochemistry

H&E staining was performed on paraffin-embedded pancreas samples.Histopathological evaluation of pancreatic lesions was performed to assessthe degree of inflammatory cell infiltration (22, 23, 27). Inflammatory infil-trates were scored as follows: 0, pancreas without mononuclear cell infil-tration; 1, mononuclear cell aggregation and/or infiltration within theinterstitium, with no parenchymal destruction; 2, focal parenchymal de-struction with mononuclear cell infiltration; 3, diffuse parenchymal de-struction but some intact parenchymal residue retained; and 4, almost allpancreatic tissue, except the pancreatic islets, destroyed or replaced withfibrosis or adipose tissue (22). Immunofluorescence staining was performedusing anti-amylase (Sigma-Aldrich), anti–PDCA-1 (BioLegend), and anti-myeloperoxidase (MPO; Abcam, Eugene, OR) Abs followed by the incu-bation with matched secondary Abs (anti-rabbit or rat IgG–Alexa Fluor 488/546; Life Technologies, Carlsbad, CA) (28). Depletion of pDCs was con-firmed by the immunohistochemical analysis by using anti–PDCA-1 Ab(BioLegend) or anti-Siglec H Ab (Abcam). In some analyses, dsDNA inNETs and nuclei were stained with Sytox Orange (Life Technologies)(28, 29).

Patient characteristics

We enrolled patients who were recently diagnosed with IgG4-RD andexhibited clinical features of AIP at Kyoto University Hospital and KansaiMedicalUniversityHospital between January 2007 and September 2013.Allhuman protocols were in compliance with the Declaration of Helsinki andapproved by Kyoto University ethics committee. Written informed consentwasobtainedfromeachpatient.Twentypatientswereincluded(16malesand4females; age range, 42–83 y;median, 68 y). Resected pancreas samples wereobtained from 4 patients, peripheral blood cell samples were obtained from13 patients, and serum samples were obtained from 15 patients. All patientswere diagnosed with IgG4-related AIP according to the Japanese diagnosticcriteria (30). Patient serum IgG4 levels were much higher (range 226–1490mg/dl; median 528.5 mg/dl) than the upper normal limit (135 mg/dl). Inaddition to healthy control samples, serum samples from patients with al-coholic chronic pancreatitis (n = 10) were also evaluated (all males; agerange, 49–73 y; median 67 y).

Evaluation of serum IFN-a, BAFF, NETs, and anti-LF Ab titer

Serum samples from patients with IgG4-related AIP before therapeutic in-tervention (n = 11), healthy controls (n = 10), and patients with chronicpancreatitis (n = 10) were subjected to ELISA for measuring IFN-a (R&DSystems) and BAFF (Antigenix America, Huntington Station, NY). SerumNETs was quantified from the same samples using the Quant-iT PicoGreendsDNAReagent (LifeTechnologies) (28). Serumanti-LFAb levelswere alsomeasured by ELISA (n = 15) (26).

Immunofluorescence staining of human pancreas samples

Paraffin-embeddedpancreas tissuesamplesresectedfrompatientswithIgG4-relatedAIP (n=4), patientswith pancreatic cancer (as controls; sectionswerefree of tumor invasion as confirmed by H&E; n = 8), and patients with al-coholic chronic pancreatitis (n = 9) were prepared. Immunofluorescencestaining was performed using anti-CD123 (Miltenyi Biotec, Auburn, CA),anti-CD303 (blood DC Ag 2 [BDCA2]; Miltenyi Biotec), anti-IgG4(Southern Biotechnology Associates, Birmingham, AL), anti–IFN-a (Mil-tenyi Biotec), anti-BAFF (R&D Systems), and anti-MPO (Abcam) Abs,followed by incubation with matched secondary Abs (anti-mouse/rabbitIgG–Alexa Fluor 488/546; Life Technologies) (28). In some analyses,dsDNA in NETs and nuclei were stained with Sytox Orange (Life Tech-nologies) (28, 29).

Observation and measurement of NET formation

Humanneutrophils (23 107/ml)were isolated fromperipheral blood, andNET formation was visualized after stimulation with monosodium urate(MSU) crystals (InvivoGen; 50 mg/ml), rabbit anti-human LF IgG Ab(10 mg/ml; Sigma-Aldrich), or control rabbit IgG (10 mg/ml; Sigma-Aldrich). Quantitative analysis was performed by counting the numberof NET-forming neutrophils per field under confocal laser scanningfluorescence microscopy (data are the mean from five randomly selectedfields) (29).

In vitro coculture and evaluation of cytokines and Igs in culturesupernatants

Neutrophils were isolated from the peripheral blood of healthy controlsand patients with IgG4-related AIP. CD19-positive B cells and pDCs wereisolated from PBMCs using CD19 and CD303 (BDCA2) MicroBeads

3034 PLASMACYTOID DENDRITIC CELLS AND AUTOIMMUNE PANCREATITIS

at kinki university medical library on July 6, 2016

http://ww

w.jim

munol.org/

Dow

nloaded from

(Miltenyi Biotec), respectively. The purity of pDCs was .90% as assessedby flow cytometric analysis (data not shown). Neutrophils (1 3 106/ml)were incubated with MSU crystals (50 mg/ml) or rabbit anti-LF IgG Ab(10 mg/ml; Sigma-Aldrich) for 3 h. Purified B cells (13 106/ml) and pDCs(13106/ml)wereadded to thewells and incubatedfor7d. Insomeexperiments,DNase (3 U/ml; Sigma-Aldrich), anti-IFNAR Ab (1 mg/ml; PBL Biomedi-cal Laboratories, Piscataway, NJ), or control IgG (1 mg/ml; Sigma-Aldrich)was added to the coculture. Culture supernatants were collected and sub-jected toELISAs formeasuring IFN-a, BAFF, and IgG1, IgG4, or IgA (16, 17).

Statistical analyses

For statistical analyses, t tests were performed using GraphPad Prism(GraphPad Software, La Jolla, CA). Differences with p , 0.05 were con-sidered statistically significant.

ResultsType I IFN production and pDC infiltration characterize thepancreatitis induced by poly (I:C) in MRL/Mp mice, a model ofAIP

In our initial studies, we used a well-established animal model ofhuman AIP to identify the type of immune cells and cytokines that

might be involved in the development of human AIP (27, 31).

Consistent with previous reports (27, 31), MRL/Mp mice treated

with poly (I:C) exhibited several important histological features of

pancreatitis that were similar to human AIP. Thus, after treatment

with poly (I:C), the pancreatic tissue of MRL/Mp mice exhibited

a massive infiltration of immune cells, destruction of the pancreatic

acinar architecture, and pancreatic fibrosis (Fig. 1A); consistentwith

these changes, treated mice exhibited significantly higher pancrea-

titis inflammatory scores than untreated mice (Fig. 1A). In addition,

treated mice exhibited a marked elevation in serum autoantibody

levels, including anti-dsDNA Ab and anti-LF Ab (Fig. 1B). Ac-

companying flow cytometric analysis revealed that treated mice as

compared with untreated mice had a marked increase in the number

of intrapancreatic CD3+ T cells, B220+ B cells, Gr-1+ neutrophils,

and CD11b+ macrophages (Fig. 1C).In further studies, we evaluated the cytokines and chemokines

produced in mice with murine AIP. For this purpose, we analyzed

serum or pancreatic lysates obtained from MRL/Mp mice with or

without poly (I:C) treatment with cytokine and chemokine arrays.

This analysis showed, as expected, that serum levels of prototypical

proinflammatory cytokines such as TNF-a, IFN-g, and IL-6 were

elevated in mice treated with poly (I:C). More interestingly, this

analysis also showed that poly (I:C) injection was accompanied by

amarked increase in a group of five chemokines, CXCL9, CXCL10,

CXCL11, CCL2, and CCL5 in both serum and pancreatic lysates

that are tightly regulated by type I IFNs (Fig. 1D) (25, 32). These

findings prompted us to measure IFN-a and IFN-b serum levels in

mice treated with poly (I:C), and indeed, we found that poly (I:C)

injection led to amarked elevation of IFN-a and IFN-b serum levels

(Fig. 1E). These elevations also relate to the fact that serum levels of

BAFF, a cytokine induced by type I IFN and considered to be in-

volved in the immunopathogenesis of human IgG4-related AIP (16,

17), were also elevated upon injection of mice with poly (I:C).We reasoned that pDCs in the inflamed pancreas of poly (I:C)–

treated mice might be the source of the increased type I IFN pro-duction in these mice because pDCs are known to produce highlevels of this cytokine (20). To examine this possibility, we con-ducted flow cytometric analysis of pancreatic infiltrates derivedfrom treated and untreated mice and indeed found that pDCs, definedas B220dimPDCA-1+ cells (33, 34), were significantly increased inthe pancreatic infiltrate of mice treated with poly (I:C) (Fig. 1F, leftand middle panels). Consistent with previous reports regarding thecell-surface markers of pDCs (35, 36), triple-color stainingsrevealed that pancreatic B220dimPDCA-1+ pDCs were positive for

both Siglec-H and CD11c and negative for CD19 (SupplementalFig. 1). In addition, accompanying immunofluorescence analysisof pancreatic tissue from treated and untreated mice revealed in-creased accumulation of PDCA-1+ pDCs in the pancreas of thetreatedmice (Fig. 1F, right panel). These findings led us to concludethat pDCs are greatly increased in the inflamed pancreas of micetreated with poly (I:C). In addition, as shown by pDC depletionstudies described below, these pDCs are the major source of thegreatly increased type I IFN produced by these treated mice.Finally,we turnedour attention to the possiblemechanismof pDC

stimulation in the inflamed pancreas of poly (I:C)–treated mice.Whereas poly (I:C) stimulation of pathogen recognition receptors inpDCs is likely to be one source of stimulation at least initially, tworecent studies addressing the pathogenic roles of pDCs in SLE haveprovided evidence that neutrophil-associated NETs containing self-DNA as well as neutrophil-derived proteins are potent inducers ofIFN-a production of pDC that is then involved in autoantibodyproduction (37, 38). Based on these studies, plus the facts that thepancreatitis inMRL/Mpmice treated with poly (I:C) contains Gr-1+

neutrophils and that serum Ab titer against LF, prototypical NETcomponent (39), was elevated, we determined if NETs are present inthe inflamed pancreatic tissue. In fact, we found that NETs asdefined by colocalization of extracellular dsDNA and MPO werepresent in the pancreatic tissue of mice treated with poly (I:C), butnot in control mice (Fig. 1G). These data thus suggest the possi-bility that pDCs producing IFN-a in the pancreas of MRL/Mpmice are being stimulated by NETs. Of interest, in the studiesmentioned above (37, 38), type I IFN is an inducer of NETs so thatthe latter is part of a positive-feedback loop that is both involved intype I IFN induction and is itself induced by type I IFN.

Depletion of pDCs prevents AIP in MRL/Mp mice

ThepresenceofpDCsandtypeIIFNinexperimentalAIPinMRL/Mpmice documented above led us to evaluate the role of these com-ponents in development of the AIP. To address this question, wedetermined the effect ofmultiple injections of 120G8Ab, anAb thathas previously been shown to deplete pDCs (40), on AIP develop-ment. We found that repeated injections of 120G8 Ab were in facteffective in greatly reducing the number and percentages of pDCs inthe pancreas of MRL/Mp mice as compared with mice treated withcontrol Ab (Fig. 2A). The depletion of pDCs was further confirmedby the tissue-staining studies in which treatment with 120G8 Abmarkedly reduced pancreatic infiltration of PDCA-1+ or Siglec-H+

cells (Supplemental Fig. 2). In addition, such depletion of pDCsprevented AIP development as assessed by pathological analysis(Fig. 2B) and reductions in the numbers of infiltrating T cells,B cells, andCD11b+macrophages in the pancreas (Fig. 2C). Perhapsmore importantly, depletion of pDCs led to a marked decrease insystemic type I IFN–related responses, as assessed by greatly di-minishedproduction of IFN-a andBAFF (Fig. 2D). In addition, suchdepletion caused greatly reduced pancreatic NETs formation (Fig.2E) as well as a significant reduction in serum Ab titers against LF,a prototypical NET component (Fig. 2F). These data thus suggestedthat pDCs play a critical role in the pathogenesis of AIP inMRL/Mpmice treated with poly (I:C).

IFN-a–mediated signaling pathway blockade prevents AIP inMRL/Mp mice

The pDC depletion studies above showed that depletion resulted ina significant reduction of type I IFN production in mice with ex-perimentalAIPbutdidnotdeterminehowsuchreductionoftypeIIFNled to reduced pancreatitis. To answer this question, we determinedthe effect of treating mice with neutralizing Ab against IFNAR(24) on the development of experimental AIP.

The Journal of Immunology 3035

at kinki university medical library on July 6, 2016

http://ww

w.jim

munol.org/

Dow

nloaded from

FIGURE 1. Massive increase in pDCs in the pancreas ofMRL/Mpmice treated with poly (I:C). MRL/Mpmice (n = 10) received i.p. injections with 100mg

poly (I:C) twice a week for 8 wk (a total of 16 injections). Mice were sacrificed 3 h after the last injection. Nontreated mice (n = 5) were used as controls. (A)

Pancreas pathology by H&E staining of nontreated MRL/Mp mice (left panel) and poly (I:C)–treated MRL/Mpmice (middle panel). Pancreatitis scores were

assessedusing the scoring systemdescribed in theMaterials andMethods (right panel). (B) Serumautoantibody levels (anti-dsDNAAb in the left panel andanti-

LFAb in the right panel) in mice treatedwith poly (I:C) and control mice (n = 5 each). The anti-LFAb titer was defined as the ratio to the average controlmouse

titer (arbitrary units [a.u.]). (C) Flow cytometric analysis of pancreatic immune cells. The numbers and percentages of CD3+ T cells, B220+ B cells, Gr-1+

neutrophils, and CD11b+ macrophages are shown together with representative dot plots and histograms. The data shown were obtained from MRL/Mp mice

treated with poly (I:C) (n = 10) and nontreated MRL/Mp mice (n = 5). (D) Serum and pancreatic lysate cytokine and chemokine arrays. Array analyses were

performed using serum or pancreatic lysates. The relative expression of cytokines and chemokines in serum and pancreatic lysates are shown together with the

representative images. The data presentedwere obtained fromMRL/Mpmice treatedwith poly (I:C) (n= 4) and nontreatedMRL/Mpmice (n= 4). Note that the

duplicate dots and red squares indicate chemokines related to type I IFNproduction. (E) Serum levels of IFN-a, IFN-b, andBAFF in (Figure legend continues)

3036 PLASMACYTOID DENDRITIC CELLS AND AUTOIMMUNE PANCREATITIS

at kinki university medical library on July 6, 2016

http://ww

w.jim

munol.org/

Dow

nloaded from

We found that such treatment of MRL/Mp mice inhibited AIPdevelopment to about the same extent as pDC depletion (Fig. 3A)

and coincided with the fact that the number of pDCs in the pan-

creas was also significantly reduced, indicating that type I IFN

production is involved in the pancreatic accumulation of pDCs

(Fig. 3B). Consistent with this reduction in the number of pancreatic

pDCs, serum levels of IFN-a and BAFF were significantly reduced

in mice treated with the anti-IFNAR Ab, as was the case with mice

treated with pDC-depleting Ab (Fig. 3C). Moreover, pancreatic

NET formation and elevation of the serum anti-LF Ab titer was

inhibited in mice treated with anti-IFNARAb (Fig. 3D). These data

have several implications. First, they show that production of

type I IFN (particularly IFN-a) is a critical factor in the development

of experimental murine AIP. Second, they show that type I IFN

production is required for the maintenance of the increased pan-

creatic infiltration of pDCs necessary for AIP development. The

most likely explanation for this latter effect is that type I IFN is also

necessary for NET generation, a possible activator of pDCs.

Pancreatic infiltration of IFN-a producing pDCs in patientswith IgG4-related AIP

The studies of experimental AIP in MRL/Mpmice discussed aboveallowed us to probe the mechanism of human IgG4-related AIP; inparticular, they allowed us to determine the relevance of pDC-mediated IFN-a signaling pathways in this disease.In initial studies, we measured IFN-a levels in the serum of pa-

tients with IgG4-related AIP and control individuals. Serum IFN-alevels were markedly elevated in patients with IgG4-related AIP ascomparedwith healthy controls or patients with chronic pancreatitis(Fig. 4A). Consistent with previous reports (41, 42), this elevationcorrelated with elevated serum BAFF levels (Fig. 4A).We next conducted immunohistochemical studies of pancreatic

tissue from patients with IgG4-related AIP who had undergonepancreatectomy (n = 4) as well as control pancreatic tissue frompatients with pancreatic cancer (noncancerous tissue) (n = 8) andwith alcoholic chronic pancreatitis (n = 9) to determine levels ofpancreatic pDC infiltration in these diseases. The pancreatic tissues

mice treatedwith poly (I:C) and controlmice (n=5 each). (F) Pancreatic accumulation of pDCs.Representative dot plots offlowcytometric analyses; pancreatic

immune cells were stained with B220 and PDCA-1. pDCs were defined as B220dimPDCA-1+ cells (left panel). The total number of pancreatic pDCs in mice

treatedwith poly (I:C) and controlmice (n=5 each) (middle panel). Immunofluorescence stainings showing pancreatic infiltration of pDCsvisualized as PDCA-

1+ cells. Pancreatic tissueswere stainedwith amylase (green) and PDCA-1 (red) (right panel). Nucleiwere counterstainedwithDAPI. (G) Immunofluorescence

detection of NETs visualized by extracellular colocalization of MPO and dsDNA. NET formation was indicated by white arrows. Results are expressed as the

mean6 SE. *p, 0.05, **p, 0.01.

FIGURE 2. Depletion of pDCs prevents AIP in MRL/Mp mice treated with poly (I:C). MRL/Mp mice were treated with the pDC-depleting Ab (100 mg,

120G8; n = 5) or control Ab (100mg; n= 5) prior to each poly (I:C) injection.Mice in each group received i.p. injectionswith 100mg poly (I:C) twice aweek for

a total of 14 times. (A) Confirmation of pDC depletion by flow cytometry. Representative dot plots of flow cytometric analyses for the detection of pDCs are

shown (left panel). The number of pDCs in the pancreaswas determined as described in Fig. 1 (right panel). (B) Prevention ofAIP development by the depletion

of pDCswith 120G8Ab. Representative images ofH&E staining are shown in the left panel. The pancreatic pathology ofMRL/Mpmice treatedwith 120G8Ab

or a control Ab was evaluated as described in Fig. 1 (right panel). (C) Flow cytometric analysis of pancreatic immune cells. The numbers and percentages of

CD3+T cells, B220+B cells, Gr-1+ neutrophils, andCD11b+macrophageswere determined as described in Fig. 1. (D) Serum levels of IFN-a, IFN-b, andBAFF

inmice treatedwith 120G8Ab and controlAb. (E) Immunofluorescence detection ofNETs visualized by extracellular colocalization ofMPOand dsDNA.NET

formationwas indicated bywhite arrows. (F) Serum levels of anti-LFAb titerwere determined as described in Fig. 1 (arbitrary units [a.u.]).Results are expressed

as the mean6 SE. *p, 0.05, **p, 0.01.

The Journal of Immunology 3037

at kinki university medical library on July 6, 2016

http://ww

w.jim

munol.org/

Dow

nloaded from

of patients with IgG4-related AIP were distinguished from thecontrol tissues by the fact that they displayed storiform fibrosis(Fig. 4B, left panel) as well as a massive infiltration of inflammatorycells enriched in IgG4+ plasma cells (Fig. 4B, right panel). Thesecharacteristics were in line with the published Japanese criteria forthe diagnosis of IgG4-related AIP (30). More importantly, thepancreatic tissues from patients with IgG4-related AIP, but notcontrol pancreas tissues or chronic pancreatitis tissues, containedeasily identified cells bearingCD123 andBDCA2, both ofwhich areestablished markers of human pDCs (43) (Fig. 4C, SupplementalFig. 3). In addition, the pancreatic BDCA2+ pDCs from patients

with IgG4-related AIP expressed both IFN-a and BAFF (Fig. 4D).These data suggest that pDCs are a major component of the pan-creatic lesion in patientswith IgG4-relatedAIP; in addition, becausethese cells express IFN-a, it is likely that pDC-derived IFN-a re-sponses are one of the prominent features of IgG4-related AIP, as isthe case in the murine AIP model described above.

NET formation in patients with IgG4-related AIP

Having confirmed that pDCs producing IFN-a are present in theinflamed pancreas of both experimental AIP and human IgG4-related AIP, we hypothesized that AIP in humans was similar to

FIGURE 3. Blockade of IFN-a–mediated signaling pathways prevents

AIP in MRL/Mp mice treated with poly (I:C). MRL/Mp mice were treated

with a type I IFN receptor–neutralizingAb (100mg, anti-IFNARAb;n=5) or

a controlAb (100mg; n=4) prior to each poly (I:C) injection.Mice fromeach

group received i.p. injections with 100 mg poly (I:C) twice a week for a total

of 14 times. (A) Prevention ofAIP development by anti-IFNARAb injection.

Representative images by H&E staining are shown in the left panel. The

pancreatic pathology of MRL/Mp mice treated with anti-IFNAR Ab or

control Ab was evaluated as described in Fig. 1 (right panel). (B) Repre-

sentative dot plots of flow cytometric analyses for the detection of pDCs (left

panel). The number of pDCs in the pancreas was determined as described in

Fig. 1 (right panel). (C) Serum levels of IFN-a, IFN-b, and BAFF in mice

treated with anti-IFNAR Ab and control Ab. (D) Immunofluorescent detec-

tion of NETs visualized by extracellular colocalization ofMPO and dsDNA.

NET formation is indicated bywhite arrows. Serum levels of anti-LFAb titer

were determined as described in Fig. 1 (arbitrary units [a.u.]). Results are

expressed as the mean6 SE. *p, 0.05, **p, 0.01. FIGURE 4. Serum levels of IFN-a and pancreatic localization of IFN-a–

producingpDCs inpatientswith IgG4-relatedAIP. (A) Serumlevels of IFN-a

and BAFF in patients with IgG4-related AIP (n = 11), healthy controls (n =

10), and patients with chronic pancreatitis (n = 10). Each dot corresponds to

the value of an assayed sample, and black bars indicate the mean values. (B)

Massive infiltration of immune cells and storiform fibrosis in the pancreatic

tissue specimens from patients with IgG4-related AIP (H&E staining, top

panel). Accumulation of IgG4+ plasma cells in the pancreas of patients with

IgG4-related AIP was also observed (inset in top panel). (C) Localization of

CD123+BDCA2+ pDCs in IgG4-related AIP pancreas tissue samples. (D)

Localization of pDCs expressing IFN-a and BAFF as determined by double

staining for BDAC2 with either IFN-a or BAFF in pancreatic tissues of

patientswith IgG4-relatedAIP.One representative image from four pancreas

specimens from patients with IgG4-related AIP is presented (C and D).

*p, 0.05, **p, 0.01.

3038 PLASMACYTOID DENDRITIC CELLS AND AUTOIMMUNE PANCREATITIS

at kinki university medical library on July 6, 2016

http://ww

w.jim

munol.org/

Dow

nloaded from

that in mice in that NET formation in the pancreas also contrib-utes to pDC activation and subsequent IFN-a production.To examine this possibility, we initially performed immunohis-

tochemicalstudies todeterminewhetherNETformationoccursintheinflamed pancreas of IgG4-related AIP patients. To this end, westained pancreatic tissue to identify both extracellular MPO anddsDNA using a method similar to that employed in the detection ofNETs in the inflamedpancreasofMRL/Mpmice treatedwithpoly (I:C). We found that pancreatic tissue from patients with IgG4-relatedAIP does contain extracellular NET structures characterized bydsDNA colocalized with MPO, whereas pancreatic tissues fromcontrols and patients with chronic pancreatitis did not contain thesestructures (Fig. 5A). Serum NET concentrations, as determined bymeasuring serum dsDNA, were also higher in patients with IgG4-related AIP than in either healthy controls or patients with chronicpancreatitis, albeit with borderline significance (Fig. 5B). Thesedata suggest that NET formation is in fact induced in pancreaticlesions of patients with IgG4-related AIP.Recent studies have shown that neutrophils can be stimulated to

produceNETsby immune complexes composedof proteins releasedby the neutrophils and autoantibodies recognizing these proteins or

simply by the autoantibodies alone if they recognize neutrophilsurface Ags (44). In addition, Okazaki et al. (26) have reported thatAb titers against LF, a NET component protein, were elevated inpatients with AIP as well as in the mice with experimental AIP de-scribed above.With this information in hand, we next considered thepossibility that NET formation in IgG4-related AIP pancreatic tissueis induced, at least in part, byLF–anti-LFAb immune complexes.Wefound first that the serum anti-LFAb titer was increased in the serumof patients with IgG4-related AIP and that this increase was espe-cially prominent in Abs of the IgG4 subclass (i.e., the IgG subclassassociated with this disease) (Fig. 5C). We then found that exposureof neutrophils from normal individuals to polyclonal anti-human LFAb induced the formation of NETs to the same extent as exposure toMSUcrystals, awell-establishedNETinducer (Fig. 5D, 5E). Thus, itseemed likely that NET formation in IgG4-related AIP pancreatictissue is due, in part, to release ofLFand the formationofLF–anti-LFAb immune complexes (38, 45) (see further discussion below).

IgG4 production by NET-stimulated pDC

Having established that neutrophils in pancreatic lesions of patientswith IgG4-related AIP produce NETs as well as the potential

FIGURE 5. NET formation in patients with IgG4-

related AIP. (A) Detection of pancreatic NET formation

in samples obtained from patients with IgG4-related

AIP. Resected pancreatic specimens from patients with

IgG4-related AIP (n = 4), chronic pancreatitis (n = 9),

and healthy controls (n = 8) were prepared. NETs were

identified by immunofluorescence staining of MPO and

dsDNA (Sytox Orange). White arrows show extracel-

lular dsDNA (deposition of NETs). Note that NET for-

mation was observed only in the IgG4-related AIP

specimen (bottom panels). (B) Serum NET levels in

samples from patients with IgG4-related AIP (n = 11),

patients with chronic pancreatitis (n = 10), and healthy

controls (n = 10). Each dot corresponds to the value of an

assayed sample, and black bars indicate themeanvalues.

(C) Serum anti-LFAb titers in IgG4-relatedAIP (n= 15),

patients with chronic pancreatitis (n = 10), and healthy

controls (n = 14). LF-specific IgG, IgG1, and IgG4 Ab

titers were determined by ELISA (arbitrary units [a.u.]).

(D and E) NET induction by anti-LF Ab and MSU

crystals. Neutrophils (23 107/ml) isolated from healthy

controls (n = 5) were stimulated with MSU (50 mg/ml),

anti-LFAb (10mg/ml), or control IgG (10mg/ml) for 3 h

to induceNET formation.NETreleasewas visualized by

confocal microscopy (white arrows) (D) and quantified

by counting the number of NET-forming cells per field

(E). Results are expressed as the mean6 SE. *p, 0.05,

**p, 0.01.

The Journal of Immunology 3039

at kinki university medical library on July 6, 2016

http://ww

w.jim

munol.org/

Dow

nloaded from

relation between such NET formation and patient anti-LF Ab pro-duction, we next turned our attention to the role of NETs in pDCfunction. In initial studies along these lines, we determined the ca-pacity of peripheral blood CD19+ B cells from healthy controlindividuals to produce IgG4 when cocultured with neutrophils and/or pDCs in the presence or absence of NETs induced by the additionof MSU crystals to the culture. The results showed that in thepresence of MSU-induced NETs and pDCs, healthy control B cellsproduced significantly greater amounts of IgG4 (but not IgG1 orIgA) than when cultured with pDCs alone (Fig. 6A). This enhancedIgG4B cell responsewas accompanied by amarked increase in bothIFN-a and BAFF production (Fig. 6B). The latter were likely dueto NET-induced IFN-a production, because the increases wereblocked by DNase inhibition of NET formation or by neutraliza-tion of IFN-a–mediated signaling pathways by anti-IFNAR Ab(Fig. 6C). In addition, the NETs themselves were not directly in-volved in IgG4 production because addition of IFN-a or BAFF tocultures ofBcells fromcontrol individuals in the absence ofNETsorpDCs also led to enhanced IgG4 production (data not shown).Very similar results were obtainedwith cocultures inwhichNETs

were induced by a polyclonal anti-LF IgGAb instead ofMSU in theabsence of the addition of exogenous LF (Fig. 6D, 6E). Presumablythe latter can occur because, as shownpreviously, the neutrophils arecapable of producing LF that can then react with anti-LFAb to formNET-stimulating immune complexes (46, 47). It should be noted,however, that IgG4 subclass isolated from patient serum did not

have the capacity to induce NETs in vitro, probably because theconcentration of anti-LF Ab in this preparation was too low (datanot shown).Taken together, these coculture studies provide evidence that

NETs inducedbyMSUoranti-LFAb lead to IgG4Abproductionbycontrol B cells through a pDC IFN-a–mediated signaling pathway.It should be noted that such IFN-a/BAFF mediated IgG4 pro-duction could be due to either expansion of a pre-existing IgG4B cell subset or to selective IgG4 class-switch differentiation plusexpansion.

pDC activation by NETs plays an essential role in IgG4generation

Although the above coculture studies using healthy control samplesidentified the important role of the neutrophil–pDC interaction inIgG4 production, they did not specifically address the contributionsof the patient and control neutrophils or pDCs to such production.To answer this question, we evaluated IgG4 production in

cocultures containing neutrophils, pDCs, and B cells from bothcontrols andpatientswith IgG4-relatedAIP invariouscombinations.The results indicated that IgG4 production by control B cells wasmarkedly enhanced by cocultures containing pDCs from patientswith IgG4-related AIP in the presence of MSU-induced NETs ascompared with cultures containing pDCs from control individuals,regardless of the source of the NET-producing neutrophils (Fig.7A). In addition, IFN-a and BAFF production were also markedly

FIGURE 6. Enhanced IgG4 production by pDCs in response to NETs. (A and B) Ig production by B cells cocultured with NET-activated pDCs. Neutrophils

(Neut), pDCs, andB cellswere isolated from the peripheral blood of healthy controls (n= 6).Neutrophils (13 106/ml)were incubatedwithMSUcrystals for 3 h

to induce NET formation, after which pDCs (1 3 106/ml) and B cells (1 3 106/ml) were added to the coculture for 7 d and Ig production in the culture

supernatants was measured (A). IFN-a and BAFF production in the culture supernatants was measured by ELISA (B). (C) Inhibition of IgG4 production by an

anti–IFN-ab receptor (IFNAR) Ab (1mg/ml) and DNase (3 U/ml) in the coculture composed of MSU-induced NETs, pDCs, and B cells from healthy controls

(n=6). Productionof IgG4, IFN-a, andBAFF in culture supernatantswasmeasuredbyELISA.Productionof IgG4or cytokines in the presenceof controlAband

the absence of anti-IFNARAborDNasewas defined as 100% in each sample. (D andE) Ig production byB cells coculturedwith pDCs activated byNETs,which

were induced by an anti-LFAb. (D) Neutrophils were incubated with anti-LFAb (10 mg/ml) or MSU (50 mg/ml) for 3 h to induce NET formation, after which

pDCs andBcellswere added as described in (A). (E)Neutrophilswere incubatedwith anti-LFAborMSUfor 3 h to induceNET formation, afterwhichpDCs and

B cells were added in the presence of anti-IFNARAb andDNase for 7 d (n=4). IgG4 and IFN-a productionwasmeasured in culture supernatants. Production of

IgG4 or cytokines in the presence of control Ab and the absence of anti-IFNARAb or DNasewas defined as 100% in each sample. Results are expressed as the

mean6 SE. *p, 0.05, **p, 0.01.

3040 PLASMACYTOID DENDRITIC CELLS AND AUTOIMMUNE PANCREATITIS

at kinki university medical library on July 6, 2016

http://ww

w.jim

munol.org/

Dow

nloaded from

elevated in supernatants from cocultures containing pDCs frompatients with IgG4-related AIP, but not those containing pDCs fromcontrol individuals (Fig. 7A).Consistentwith the fact that control and patient neutrophils had an

equal capacity to support IgG4 production in the coculture system,neutrophils isolated from healthy controls and patients with IgG4-related AIP exhibited equivalent amounts of NET formation uponexposure to MSU crystals (Fig. 7B). In addition, inhibition of NETformation by DNase or blockade of IFN-a–mediated signalingpathways by anti-IFNARAb significantly reduced IgG4 productionin a coculture system comprised of control B cells, patient-derivedneutrophils, and patient-derived pDCs (Fig. 7C). Finally, IgG4production by control B cells was also enhanced in the presence ofNETs induced by anti-LF Ab when the cells were cocultured withpDCs from patients with IgG4-related AIP (Fig. 7D), and this en-hancement was again inhibited by the addition of DNase or by anti-IFNAR Ab (Fig. 7E). Taken together, these data suggest that pDCs,rather than neutrophils, have an indispensable role in the generationof IgG4 responses by B cells in patients with IgG4-related AIP and

that NET-triggered pDC activation followed by IFN-a production isrequired for the development of IgG4-related AIP.

DiscussionThe immunologicmechanismsunderlying thepathogenesisof IgG4-RD as well as the increased production of the IgG4 subclass thatcharacterizes this inflammatory disease are still poorly defined. In thisstudy,weaddressthisquestionwithstudiesofmicewithexperimentalAIP, MRL/Mp mice treated with poly (I:C), and studies of patientswith IgG4-related AIP. The studies of the treated MRL/Mp miceshowed that experimental AIP is characterized by a chronicallyinflamed pancreas exhibiting destruction of acinar architecture,massive infiltration of immune cells, and fibrosis; all findings wereshared by human AIP. Although the pathogenic roles of proin-flammatory cytokines such as IFN-g, IL-6, and TNF-a are impli-cated in the development of such experimental AIP (31, 48), theroles of type I IFNs in this model had been poorly defined. In thisstudy, we showed that experimental AIP is associated with thepancreatic infiltration of pDCs that have either migrated to or have

FIGURE 7. Enhanced NET-induced pDC activation in patients with IgG4-related AIP. (A) Coculture experiments composed of pDCs, B cells, and MSU-

inducedNETswere performed as described in Fig. 6A.Neutrophils isolated fromhealthy controls (Ctrl; n= 6) or patients with IgG4-relatedAIP (Pt; n= 8)were

subjected toNETinductionbyMSUasdescribed inFig. 6. pDCs isolated fromhealthy controls (Ctrl) or patientswith IgG4-relatedAIP (Pt)were subjected to the

coculture assay togetherwithB cells isolated fromhealthy controls (Ctrl B). IgG4, IFN-a, andBAFF production in the coculture supernatant was determined by

ELISA. (B) Neutrophils (Neut) isolated from healthy controls (Ctrl; n = 5) and patients with IgG4-related AIP (Pt; n = 5) were subjected to NET induction by

MSU, as described in Fig. 5. NET formation was determined by counting the number of NET-forming cells under high-power field. (C) Inhibition of IgG4

production by the addition of an anti–IFN-ab receptor Ab (IFNARAb) and DNase. B cells from healthy controls (Ctrl B; n = 6) were cocultured with pDCs (Pt

pDCs) andMSU-inducedNETs (PtNeut) derived frompatientswith IgG4-relatedAIP (n=6) in the presence of anti-IFNARAborDNase for 7 d, as described in

Fig. 6C. Productionof IgG4or cytokines in the presence of controlAb and absence of anti-IFNARAborDNasewas defined as 100% in each sample. (D andE) Ig

productionbyhealthycontrolBcells (CtrlB) coculturedwith pDCs frompatientswith IgG4-relatedAIP (Pt pDCs) activatedbyNETs,whichwere inducedbyan

anti-LF Ab. IgG4-related AIP patient-derived neutrophils (Pt Neut) were incubated with anti-LF Ab (10 mg/ml) or MSU (50 mg/ml) for 3 h to induce NET

formation, after which Pt pDCs and Ctrl B cells were added (D). Pt Neut were incubated with anti-LFAb for 3 h to induce NET formation, after which Pt pDCs

andCtrlBcellswere added in thepresenceof the anti-IFNARAbandDNase for 7d (n=4) (E). IgG4and IFN-aproduction in culture supernatantsweremeasured

as described in Fig. 6. Production of IgG4 or cytokines in the presence of control Ab and absence of anti-IFNAR Ab or DNase was defined as 100% in each

sample. Results are expressed as the mean6 SE. *p, 0.05, **p, 0.01.

The Journal of Immunology 3041

at kinki university medical library on July 6, 2016

http://ww

w.jim

munol.org/

Dow

nloaded from

been expanded in this organ; these cells are the likely source ofa markedly increased production of IFN-a. We confirmed the im-portance of this pDC-mediated IFN-a production to the develop-ment of the pancreatic inflammation by showing that systemicadministration of pDC-depleting Ab (120G8 Ab) and neutralizingAb against IFNAR led to almost complete amelioration of disease.In parallel with the murine AIP findings, studies of IgG4-related

AIP disclosed that human AIP was also associated with pDCs lo-calized in the pancreas producing IFN-a. Whether or not suchproduction is a major driver of the human pancreatic inflammationas in themousemodel is still not established and awaits the outcomeof clinical trials with anti–IFN-aAb or anti-IFNARAb. In any case,we could show that patient-derivedperipheral blood pDCs enhancedIgG4 Ab production by B cells through an IFN-a–mediated sig-naling mechanism. Thus, a central feature of both experimentalmurine AIP and human IgG4-related AIP pathogenesis is the acti-vation of pDCs producing IFN-a, and in the case of the humandisease, this pDC abnormality is a factor in the excessive IgG4production that characterizes this disease.Our studies of both the murine model of AIP and patients with

IgG4-related AIP disclosed that the inflamed pancreas harboredNETs (i.e., weblike structures composed of chromatin, moleculesderived from granules, and some cytoplasmic proteins that aregenerated by stimulated and dying neutrophils) (21). In addition, thetiter of autoantibody against LF, a NET component protein, ismarkedly elevated in the serum of mice with murine experimentalAIP and patients with IgG4-related AIP. These structures in otherstudies have been shown to be involved in pDC activation (34, 38)and were therefore assumed to have a similar role in AIP. In the caseof human IgG4-related AIP, the role of NETs in activation of pDCswas actively studied using pDCs isolated from the circulation aswellasNETsgenerated invitrowith eitherMSUor anti-LFAb. In supportof the hypothesis that NETs are stimulators of pDC production ofIFN-a, we found that cocultures of pDCs with neutrophils inducedto express NETs by either MSU or anti-LFAb lead to stimulation ofpDCproduction of IFN-a andBAFFaswell as the induction of IgG4production by normal B cells also present in the coculture. Thus,assuming that the induced NETs in the culture mimic those in thepancreas, these in vitro data offer convincing evidence that NETstructures are in fact an important stimulatory component in thepancreas in bothmurine and humanAIP. However, it is also possiblethat other signaling mechanisms are also present, such as directstimulation of pDCs by TLR7 and TLR9 ligands associated withcommensal flora that have gained access to the inflamed pancreas.Finally, it should also be noted that pDCs in patients with IgG4-related AIP have other as-yet-undisclosed abnormalities that ren-der them more susceptible to NET activation than normal pDCs.Although in the current study, we showed that MSU and anti-LF

Ab were capable of initiating IFN-a production by pDCs, it is notclear that these stimulants are actually those that are operative in theinduction of this cytokine in patients with IgG4-related AIP. Thedeposition of MSU into appendicular joints has an essential role inthe development of acute gout (49), and recent studies have pro-vided evidence that MSU-induced NETs initiate inflammatory re-sponses in this disease (50, 51). Whether a similar mechanism isapplied to IgG4-related AIP requires further studies to demonstratethat MSU deposition occurs in IgG4-related pancreatic lesions aswell as evaluation of whether uric acid-clearing agents have effi-cacy in patients with IgG4-relatedAIP. The relevance of anti-LFAbto IFN-a production in IgG4-related AIP is somewhat more likelythan MSU, because serum anti-LF Ab titers were elevated in theserum of patients with IgG4-RD and experimental AIP. However,these datamaymerely indicate that neutrophil function as a generalsource of autoantigens in IgG4-related AIP, and any of several of

these autoantigens can trigger the formation of immune complexesthat stimulate NET formation. In this regard, we speculate thatneutrophil release of autoantigens and subsequent formation ofimmune complexes comprise a positive-feedback loop that favorsNET induction, ultimately leading to the release of NETAgs withthe capacity to stimulate pDCs to produce IFN-a. However, con-firmation of this idea requires future studies that evaluate serumAbprofiles against various types of NET proteins in IgG4-related AIP;in addition, studies that evaluate whether immune complexes, in-cluding LF–anti-LF Ab immune complexes, can stimulate pDCsindependently of NETs are also needed.Our coculture experiments using neutrophils and pDCs from

healthy controls and patients with IgG4-related AIP incubated to-gether invarious combinations clearly showed that pDCs, rather thanneutrophils,were indispensable for thegenerationof IgG4 responsesin patients with IgG4-related AIP. In most instances, pDCs serveimportant host-defense functions, particularly in relation to viralinfection (20). This protection is accomplished by their ability tosense microbial nucleic acids by TLR7 and TLR9 and then produceIFN-a, a cytokine that leads to adaptive generation of Th1 and Th17responses (20, 52). However, evidence has recently been obtainedthat such IFN-a production could also be involved in initiation ofvarious diseases. In particular, excessive IFN-a production by pDCsin response to self-DNA has been implicated in the pathogenesis ofautoimmune diseases such as psoriasis and SLE (20, 53). In the caseof SLE, such IFN-a production is considered to arise from NETrelease of autoimmune complexes that activate pDCs in the sameway suggested in this study to occur in IgG4-related AIP. However,despite these mechanistic similarities, the clinicopathological fea-tures of these two diseases are completely different. In particular,enhancement of IgG4Ab responses and storiformfibrosis formationare observed in IgG4-related AIP and not in SLE, whereas wide-spread organ and tissue damage from autoantibody deposition isseen in SLE but not in IgG4-related AIP. This discrepancy can bepartially explained by the presence of a significant population ofpatientswith SLEhaving autoantibodies against BAFF that partiallyblunt certain autoimmune responses including IgG4 production(54). In addition, there is likely to be a difference between SLE andIgG4-related AIP in Treg responses. Thus, whereas Treg responsesare increased in IgG4-relatedAIP (13, 14), they are defective in SLE(55, 56). Given that pDCs have been shown to induce and supportTreg function (53, 57), we assume that pDC activation in IgG4-related AIP is accompanied by the induction of Treg-derived cyto-kines that limit the production of destructive autoantibodies.We previously showed that TLR and nucleotide-binding oligo-

merization domain-like receptor activation of monocytes andbasophils isolatedfrompatientswithIgG4-relatedAIPenhanceIgG4production byB cells from healthy controls (16, 17). This correlatedwith evidence that monocytes and basophils activated by MAMPscould be involved in the development of IgG4-related AIP. Inthe current study, we identified another type of innate immuneresponse mediated by neutrophils and pDCs in the developmentof IgG4-related AIP. Distinct from the MAMP-activated pathwaysmentioned above, this pathway is triggered by the activation ofneutrophils byendogenousdanger signals, suchasdamage-associatedmolecular patterns (MSU crystals) and autoantibodies (anti-LFAb).At the moment, it is not clear if these pathways are largely separateand thus that responses to either microbial Ags or endogenous fac-tors can trigger the abnormal innate immune response leading toIgG4-related AIP or, alternately, these pathways are interrelated.The latter possibility is suggested by the fact that both pathwaysresult in enhancement of BAFF production by innate cells and thesubsequent production of IgG4 by B cells (16, 17). Clinical obser-vations showing that serum BAFF levels are significantly higher in

3042 PLASMACYTOID DENDRITIC CELLS AND AUTOIMMUNE PANCREATITIS

at kinki university medical library on July 6, 2016

http://ww

w.jim

munol.org/

Dow

nloaded from

patients with IgG4-related AIP than in normal individuals (41, 42)support the importance of this cytokine in the immunopathogenesisof this disease.In conclusion, our findings show that bothmurine AIP and human

IgG4-related AIP are characterized by greatly increased pDC acti-vation and IFN-a production and that, in the case of the murinedisease, this could be shown to be a critical mechanism underlyingthe pancreatic inflammation. In addition, both the murine and hu-man AIP were associated with the presence of pancreatic NETs,neutrophil-derived structures that in the case of the human diseasewere shown with in vitro studies of induced NETs to be a cause ofpDC activation as well as increased IFN-a and BAFF production.Finally, this NET-induced IFN-a and BAFF production by pDCswas shown to be responsible, at least in part, for B cell production ofthe signature abnormality of this form of autoimmunity in humans,IgG4-RD.

AcknowledgmentsWe thank Dr. Kyoichi Takaori (Kyoto University) for providing the resected

pancreas samples.

DisclosuresThe authors have no financial conflicts of interest.

References1. Stone, J. H., Y. Zen, and V. Deshpande. 2012. IgG4-related disease. N. Engl.

J. Med. 366: 539–551.2. Yamamoto, M., H. Takahashi, and Y. Shinomura. 2014. Mechanisms and as-

sessment of IgG4-related disease: lessons for the rheumatologist. Nat. Rev.Rheumatol. 10: 148–159.

3. Kamisawa, T., and A. Okamoto. 2006. Autoimmune pancreatitis: proposal ofIgG4-related sclerosing disease. J. Gastroenterol. 41: 613–625.

4. Kamisawa, T., S. T. Chari, M. M. Lerch, M. H. Kim, T. M. Gress, andT. Shimosegawa. 2013. Recent advances in autoimmune pancreatitis: type 1 andtype 2. Gut 62: 1373–1380.

5. Aalberse, R. C., S. O. Stapel, J. Schuurman, and T. Rispens. 2009. Immuno-globulin G4: an odd antibody. Clin. Exp. Allergy 39: 469–477.

6. Kawa, S., K. Kitahara, H. Hamano, Y. Ozaki, N. Arakura, K. Yoshizawa,T. Umemura, M. Ota, S. Mizoguchi, Y. Shimozuru, and S. Bahram. 2008. Anovel immunoglobulin-immunoglobulin interaction in autoimmunity. PLoS One3: e1637.

7. van der Neut Kolfschoten, M., J. Schuurman, M. Losen, W. K. Bleeker,P. Martınez-Martınez, E. Vermeulen, T. H. den Bleker, L. Wiegman, T. Vink,L. A. Aarden, et al. 2007. Anti-inflammatory activity of human IgG4 antibodiesby dynamic Fab arm exchange. Science 317: 1554–1557.

8. Hussain, A., T. Pankhurst, M. Goodall, R. Colman, R. Jefferis, C. O. Savage, andJ. M. Williams. 2009. Chimeric IgG4 PR3-ANCA induces selective inflamma-tory responses from neutrophils through engagement of Fcgamma receptors.Immunology 128: 236–244.

9. Tomas, N. M., L. H. Beck, Jr., C. Meyer-Schwesinger, B. Seitz-Polski, H. Ma,G. Zahner, G. Dolla, E. Hoxha, U. Helmchen, A. S. Dabert-Gay, et al. 2014.Thrombospondin type-1 domain-containing 7A in idiopathic membranous ne-phropathy. N. Engl. J. Med. 371: 2277–2287.

10. Jeannin, P., S. Lecoanet, Y. Delneste, J. F. Gauchat, and J. Y. Bonnefoy. 1998.IgE versus IgG4 production can be differentially regulated by IL-10. J. Immunol.160: 3555–3561.

11. Punnonen, J., G. Aversa, B. G. Cocks, A. N. McKenzie, S. Menon, G. Zurawski,R. de Waal Malefyt, and J. E. de Vries. 1993. Interleukin 13 induces interleukin4-independent IgG4 and IgE synthesis and CD23 expression by human B cells.Proc. Natl. Acad. Sci. USA 90: 3730–3734.

12. Akitake, R., T. Watanabe, C. Zaima, N. Uza, H. Ida, S. Tada, N. Nishida, andT. Chiba. 2010. Possible involvement of T helper type 2 responses to Toll-likereceptor ligands in IgG4-related sclerosing disease. Gut 59: 542–545.

13. Miyoshi, H., K. Uchida, T. Taniguchi, S. Yazumi, M. Matsushita, M. Takaoka,and K. Okazaki. 2008. Circulating naıve and CD4+CD25high regulatory T cellsin patients with autoimmune pancreatitis. Pancreas 36: 133–140.

14. Zen, Y., T. Fujii, K. Harada, M. Kawano, K. Yamada, M. Takahira, andY. Nakanuma. 2007. Th2 and regulatory immune reactions are increased inimmunoglobin G4-related sclerosing pancreatitis and cholangitis. Hepatology45: 1538–1546.

15. Cerutti, A., I. Puga, and M. Cols. 2011. Innate control of B cell responses. TrendsImmunol. 32: 202–211.

16. Watanabe, T., K. Yamashita, S. Fujikawa, T. Sakurai, M. Kudo, M. Shiokawa,Y. Kodama, K. Uchida, K. Okazaki, and T. Chiba. 2012. Involvement of acti-vation of toll-like receptors and nucleotide-binding oligomerization domain-likereceptors in enhanced IgG4 responses in autoimmune pancreatitis. ArthritisRheum. 64: 914–924.

17. Watanabe, T., K. Yamashita, T. Sakurai, M. Kudo, M. Shiokawa, N. Uza,Y. Kodama, K. Uchida, K. Okazaki, and T. Chiba. 2013. Toll-like receptor ac-tivation in basophils contributes to the development of IgG4-related disease.J. Gastroenterol. 48: 247–253.

18. Mackay, F., and P. Schneider. 2009. Cracking the BAFF code. Nat. Rev.Immunol. 9: 491–502.

19. Tezuka, H., Y. Abe, J. Asano, T. Sato, J. Liu, M. Iwata, and T. Ohteki. 2011.Prominent role for plasmacytoid dendritic cells in mucosal T cell-independentIgA induction. Immunity 34: 247–257.

20. Gilliet, M., W. Cao, and Y. J. Liu. 2008. Plasmacytoid dendritic cells: sensingnucleic acids in viral infection and autoimmune diseases. Nat. Rev. Immunol. 8:594–606.

21. Liu, Z., and A. Davidson. 2012. Taming lupus-a new understanding of patho-genesis is leading to clinical advances. Nat. Med. 18: 871–882.

22. Asada, M., A. Nishio, T. Akamatsu, J. Tanaka, K. Saga, M. Kido, N. Watanabe,K. Uchida, T. Fukui, K. Okazaki, and T. Chiba. 2010. Analysis of humoralimmune response in experimental autoimmune pancreatitis in mice. Pancreas39: 224–231.

23. Kanno, H., M. Nose, J. Itoh, Y. Taniguchi, and M. Kyogoku. 1992. Spontaneousdevelopment of pancreatitis in the MRL/Mp strain of mice in autoimmunemechanism. Clin. Exp. Immunol. 89: 68–73.

24. Shahangian, A., E. K. Chow, X. Tian, J. R. Kang, A. Ghaffari, S. Y. Liu,J. A. Belperio, G. Cheng, and J. C. Deng. 2009. Type I IFNs mediate developmentof postinfluenza bacterial pneumonia in mice. J. Clin. Invest. 119: 1910–1920.

25. Tsuji, Y., T. Watanabe, M. Kudo, H. Arai, W. Strober, and T. Chiba. 2012.Sensing of commensal organisms by the intracellular sensor NOD1 mediatesexperimental pancreatitis. Immunity 37: 326–338.

26. Okazaki, K., K. Uchida, M. Ohana, H. Nakase, S. Uose, M. Inai, Y. Matsushima,K. Katamura, K. Ohmori, and T. Chiba. 2000. Autoimmune-related pancreatitisis associated with autoantibodies and a Th1/Th2-type cellular immune response.Gastroenterology 118: 573–581.

27. Schwaiger, T., C. van den Brandt, B. Fitzner, S. Zaatreh, F. Kraatz, A. Dummer,H. Nizze, M. Evert, B. M. Broker, M. C. Brunner-Weinzierl, et al. 2014. Au-toimmune pancreatitis in MRL/Mp mice is a T cell-mediated disease responsiveto cyclosporine A and rapamycin treatment. Gut 63: 494–505.

28. Arai, Y., K. Yamashita, K. Mizugishi, T. Watanabe, S. Sakamoto, T. Kitano,T. Kondo, H. Kawabata, N. Kadowaki, and A. Takaori-Kondo. 2013. Serumneutrophil extracellular trap levels predict thrombotic microangiopathy afterallogeneic stem cell transplantation. Biol. Blood Marrow Transplant. 19: 1683–1689.

29. Arai, Y., Y. Nishinaka, T. Arai, M. Morita, K. Mizugishi, S. Adachi, A. Takaori-Kondo, T. Watanabe, and K. Yamashita. 2014. Uric acid induces NADPHoxidase-independent neutrophil extracellular trap formation. Biochem. Biophys.Res. Commun. 443: 556–561.

30. Okazaki, K., K. Uchida, M. Koyabu, H. Miyoshi, and M. Takaoka. 2011. Recentadvances in the concept and diagnosis of autoimmune pancreatitis and IgG4-related disease. J. Gastroenterol. 46: 277–288.

31. Nishio, A., M. Asada, K. Uchida, T. Fukui, T. Chiba, and K. Okazaki. 2011. Therole of innate immunity in the pathogenesis of experimental autoimmune pan-creatitis in mice. Pancreas 40: 95–102.

32. Watanabe, T., N. Asano, S. Fichtner-Feigl, P. L. Gorelick, Y. Tsuji, Y. Matsumoto,T. Chiba, I. J. Fuss, A. Kitani, and W. Strober. 2010. NOD1 contributes to mousehost defense against Helicobacter pylori via induction of type I IFN and activationof the ISGF3 signaling pathway. J. Clin. Invest. 120: 1645–1662.

33. Gregorio, J., S. Meller, C. Conrad, A. Di Nardo, B. Homey, A. Lauerma, N. Arai,R. L. Gallo, J. Digiovanni, and M. Gilliet. 2010. Plasmacytoid dendritic cellssense skin injury and promote wound healing through type I interferons. J. Exp.Med. 207: 2921–2930.

34. Doring, Y., H. D. Manthey, M. Drechsler, D. Lievens, R. T. Megens,O. Soehnlein, M. Busch, M. Manca, R. R. Koenen, J. Pelisek, et al. 2012. Auto-antigenic protein-DNA complexes stimulate plasmacytoid dendritic cells topromote atherosclerosis. Circulation 125: 1673–1683.

35. Loschko, J., S. Heink, D. Hackl, D. Dudziak, W. Reindl, T. Korn, andA. B. Krug. 2011. Antigen targeting to plasmacytoid dendritic cells via Siglec-Hinhibits Th cell-dependent autoimmunity. J. Immunol. 187: 6346–6356.

36. Pelayo, R., J. Hirose, J. Huang, K. P. Garrett, A. Delogu, M. Busslinger, andP. W. Kincade. 2005. Derivation of 2 categories of plasmacytoid dendritic cellsin murine bone marrow. Blood 105: 4407–4415.

37. Garcia-Romo, G. S., S. Caielli, B. Vega, J. Connolly, F. Allantaz, Z. Xu,M. Punaro, J. Baisch, C. Guiducci, R. L. Coffman, et al. 2011. Netting neu-trophils are major inducers of type I IFN production in pediatric systemic lupuserythematosus. Sci. Transl. Med. 3: 73ra20.

38. Lande, R., D. Ganguly, V. Facchinetti, L. Frasca, C. Conrad, J. Gregorio,S. Meller, G. Chamilos, R. Sebasigari, V. Riccieri, et al. 2011. Neutrophils ac-tivate plasmacytoid dendritic cells by releasing self-DNA-peptide complexes insystemic lupus erythematosus. Sci. Transl. Med. 3: 73ra19.

39. Mayadas, T. N., X. Cullere, and C. A. Lowell. 2014. The multifaceted functionsof neutrophils. Annu. Rev. Pathol. 9: 181–218.

40. Asselin-Paturel, C., G. Brizard, J. J. Pin, F. Briere, and G. Trinchieri. 2003.Mouse strain differences in plasmacytoid dendritic cell frequency and functionrevealed by a novel monoclonal antibody. J. Immunol. 171: 6466–6477.

41. Kiyama, K., D. Kawabata, Y. Hosono, K. Kitagori, N. Yukawa, H. Yoshifuji,K. Omura, T. Fujii, and T. Mimori. 2012. Serum BAFF and APRIL levels inpatients with IgG4-related disease and their clinical significance. Arthritis Res.Ther. 14: R86.

The Journal of Immunology 3043

at kinki university medical library on July 6, 2016

http://ww

w.jim

munol.org/

Dow

nloaded from

42. Yamanishi, H., T. Kumagi, T. Yokota, N. Azemoto, M. Koizumi, Y. Kobayashi,M. Abe, H. Murakami, Y. Hiasa, B. Matsuura, et al. 2011. Clinical significanceof B cell-activating factor in autoimmune pancreatitis. Pancreas 40: 840–845.

43. Boiocchi, L., S. Lonardi, W. Vermi, S. Fisogni, and F. Facchetti. 2013. BDCA-2(CD303): a highly specific marker for normal and neoplastic plasmacytoiddendritic cells. Blood 122: 296–297.

44. Kaplan, M. J., and M. Radic. 2012. Neutrophil extracellular traps: double-edgedswords of innate immunity. J. Immunol. 189: 2689–2695.

45. Kessenbrock, K., M. Krumbholz, U. Schonermarck, W. Back, W. L. Gross,Z. Werb, H. J. Grone, V. Brinkmann, and D. E. Jenne. 2009. Netting neutrophilsin autoimmune small-vessel vasculitis. Nat. Med. 15: 623–625.

46. Barrientos, L., V. Marin-Esteban, L. de Chaisemartin, V. L. Le-Moal, C. Sandre,E. Bianchini, V. Nicolas, M. Pallardy, and S. Chollet-Martin. 2013. An improvedstrategy to recover large fragments of functional human neutrophil extracellulartraps. Front. Immunol. 4: 166.

47. Behnen, M., C. Leschczyk, S. Moller, T. Batel, M. Klinger, W. Solbach, andT. Laskay. 2014. Immobilized immune complexes induce neutrophil extracel-lular trap release by human neutrophil granulocytes via FcgRIIIB and Mac-1.J. Immunol. 193: 1954–1965.

48. Fitzner, B., S. A. Holzhueter, S. Ibrahim, H. Nizze, and R. Jaster. 2009.Interferon-gamma treatment accelerates and aggravates autoimmune pancreatitisin the MRL/Mp-mouse. Pancreatology 9: 233–239.

49. Knight, J. S., C. Carmona-Rivera, and M. J. Kaplan. 2012. Proteins derived fromneutrophil extracellular traps may serve as self-antigens and mediate organdamage in autoimmune diseases. Front. Immunol. 3: 380.

50. Mitroulis, I., K. Kambas, A. Chrysanthopoulou, P. Skendros, E. Apostolidou,I. Kourtzelis, G. I. Drosos, D. T. Boumpas, and K. Ritis. 2011. Neutrophil

extracellular trap formation is associated with IL-1b and autophagy-relatedsignaling in gout. PLoS One 6: e29318.

51. Schauer, C., C. Janko, L. E. Munoz, Y. Zhao, D. Kienhofer, B. Frey, M. Lell,B. Manger, J. Rech, E. Naschberger, et al. 2014. Aggregated neutrophil extra-cellular traps limit inflammation by degrading cytokines and chemokines. Nat.Med. 20: 511–517.

52. Shi, G., B. P. Vistica, L. F. Nugent, C. Tan, E. F. Wawrousek, D. M. Klinman,and I. Gery. 2013. Differential involvement of Th1 and Th17 in pathogenicautoimmune processes triggered by different TLR ligands. J. Immunol. 191:415–423.

53. Reizis, B., M. Colonna, G. Trinchieri, F. Barrat, and M. Gilliet. 2011. Plasma-cytoid dendritic cells: one-trick ponies or workhorses of the immune system?Nat. Rev. Immunol. 11: 558–565.

54. Price, J. V., D. J. Haddon, D. Kemmer, G. Delepine, G. Mandelbaum,J. A. Jarrell, R. Gupta, I. Balboni, E. F. Chakravarty, J. Sokolove, et al. 2013.Protein microarray analysis reveals BAFF-binding autoantibodies in systemiclupus erythematosus. J. Clin. Invest. 123: 5135–5145.

55. Valencia, X., C. Yarboro, G. Illei, and P. E. Lipsky. 2007. Deficient CD4+CD25high T regulatory cell function in patients with active systemic lupuserythematosus. J. Immunol. 178: 2579–2588.

56. Yang, J., X. Yang, H. Zou, Y. Chu, and M. Li. 2011. Recovery of the immunebalance between Th17 and regulatory T cells as a treatment for systemic lupuserythematosus. Rheumatology (Oxford) 50: 1366–1372.

57. Ito, T., M. Yang, Y. H. Wang, R. Lande, J. Gregorio, O. A. Perng, X. F. Qin,Y. J. Liu, and M. Gilliet. 2007. Plasmacytoid dendritic cells prime IL-10-producing T regulatory cells by inducible costimulator ligand. J. Exp. Med.204: 105–115.

3044 PLASMACYTOID DENDRITIC CELLS AND AUTOIMMUNE PANCREATITIS

at kinki university medical library on July 6, 2016

http://ww

w.jim

munol.org/

Dow

nloaded from