PILRα Negatively Regulates Mouse Inflammatory Arthritis This information is current as of July 6, 2014.

J Immunol 2014; 193:860-870; Prepublished online 16 June 2014; doi: 10.4049/jimmunol.1400045 http://www.jimmunol.org/content/193/2/860

Supplementary Material References Subscriptions Permissions Email Alerts

http://www.jimmunol.org/content/suppl/2014/06/16/jimmunol.140004 5.DCSupplemental.html This article cites 43 articles, 19 of which you can access for free at: http://www.jimmunol.org/content/193/2/860.full#ref-list-1 Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscriptions Submit copyright permission requests at: http://www.aai.org/ji/copyright.html Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/cgi/alerts/etoc

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 9650 Rockville Pike, Bethesda, MD 20814-3994. Copyright © 2014 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606.

Downloaded from http://www.jimmunol.org/ at Otterbein University on July 6, 2014

Yonglian Sun, Patrick Caplazi, Juan Zhang, Anita Mazloom, Sarah Kummerfeld, Gabriel Quinones, Kate Senger, Justin Lesch, Ivan Peng, Andrew Sebrell, Wilman Luk, Yanmei Lu, Zhonghua Lin, Kai Barck, Judy Young, Mariela Del Rio, Sophie Lehar, Vida Asghari, WeiYu Lin, Sanjeev Mariathasan, Jason DeVoss, Shahram Misaghi, Mercedesz Balazs, Tao Sai, Benjamin Haley, Philip E. Hass, Min Xu, Wenjun Ouyang, Flavius Martin, Wyne P. Lee and Ali A. Zarrin

The Journal of Immunology

PILRa Negatively Regulates Mouse Inflammatory Arthritis Yonglian Sun,* Patrick Caplazi,† Juan Zhang,* Anita Mazloom,‡ Sarah Kummerfeld,x Gabriel Quinones,‡ Kate Senger,* Justin Lesch,* Ivan Peng,* Andrew Sebrell,{ Wilman Luk,‖ Yanmei Lu,‖ Zhonghua Lin,* Kai Barck,# Judy Young,‖ Mariela Del Rio,** Sophie Lehar,†† Vida Asghari,** WeiYu Lin,* Sanjeev Mariathasan,†† Jason DeVoss,* Shahram Misaghi,‡‡ Mercedesz Balazs,* Tao Sai,{ Benjamin Haley,‡ Philip E. Hass,‡ Min Xu,* Wenjun Ouyang,* Flavius Martin,* Wyne P. Lee,* and Ali A. Zarrin*

I

mmune responses are modulated by a network of positiveand negative-regulatory mechanisms. Paired receptors consist of highly related activating and inhibitory receptors that are widely involved in the regulation of the immune system (1). Both inhibitory and activating receptors share high similarity in their extracellular domain, whereas their intracellular signaling domains are divergent (2, 3). Paired Ig-like type 2 receptor (PILR)a belongs to the Ig superfamily. Its intracellular domain contains two ITIMs that recruit SHP-1 and SHP-2 to trigger an inhibitory signaling cascade, resulting in reduced intracellular calcium mobilization (4, 5). PILRa and its gene-linked activating counterpart, PILRb, share highly similar extracellular domains, suggesting that they may recognize the same ligands (5–7). PILRb has a truncated cytoplasmic domain and a charged amino acid residue in its transmembrane region that associates with the

*Department of Immunology, Genentech, South San Francisco, CA 94080; † Department of Pathology, Genentech, South San Francisco, CA 94080; ‡ Department of Protein Chemistry, Genentech, South San Francisco, CA 94080; x Department of Bioinformatics and Computational Biology, Genentech, South San Francisco, CA 94080; {Department of Antibody Engineering, Genentech, South San Francisco, CA 94080; ‖Department of Biochemical and Cellular Pharmacology, Genentech, South San Francisco, CA 94080; #Department of Biomedical Imaging, Genentech, South San Francisco, CA 94080; **Department of Transgenic Technology, Genentech, South San Francisco, CA 94080; ††Department of Microbial Pathogenesis, Genentech, South San Francisco, CA 94080; and ‡‡Department of Early Stage Cell Culture, Genentech, South San Francisco, CA 94080 Received for publication January 9, 2014. Accepted for publication May 19, 2014. The microarray data presented in this article have been submitted to the National Center for Biotechnology Information’s Gene Expression Omnibus (http://www.ncbi. nlm.nih.gov/geo/query/acc.cgi?acc=GSE48780) under accession number GSE48780. Address correspondence and reprint requests to Dr. Ali A. Zarrin, Genentech, 1 DNA Way, South San Francisco, CA 94080. E-mail address: [email protected] The online version of this article contains supplemental material. Abbreviations used in this article: BM, bone marrow; BMDC, BM-derived DC; BMDM, BM-derived monocyte; CAIA, collagen Ab–induced arthritis; CLEC4G, C-type lectin domain family 4, member G; DC, dendritic cell; ES, embryonic stem; h, human; JCBV, joint cortical bone volume; m, mouse; micro-CT, microcomputed tomography; MPO, myeloperoxidase; OA, osteoarthritis; PILR, paired Ig-like type 2 receptor; PNBV, periosteal new bone volume; RA, rheumatoid arthritis; WT, wildtype. Copyright Ó 2014 by The American Association of Immunologists, Inc. 0022-1767/14/$16.00 www.jimmunol.org/cgi/doi/10.4049/jimmunol.1400045

ITAM-bearing DAP12 adaptor molecule to deliver activating signals (8). Human and mouse PILRa share only ∼40% homology (7), yet conserved residues mediate ligand interactions (9). PILRa and PILRb transcripts show similar tissue expression, with high levels in spleen, liver, and lung and lower levels in the small intestine (8). PILRa and PILRb are predominantly expressed in cells of the myelomonocytic lineage, including monocytes, macrophages, granulocytes, and monocyte-derived dendritic cells (DCs) (4, 8). Additionally, PILRb is expressed in NK cells (8). It was shown that PILRa binds to mouse CD99 (8), PILR-associating neural protein (10), and HSV-1 glycoprotein B (11). Specific sialylated O-linked glycans on ligands are required for their binding to PILRa (6, 10, 12). We recently identified two more binding partners of PILRa, neural proliferation differentiation and control-1 and collectin-12, and found that an evolutionarily conserved PILRa domain mediates its interaction with these diverse sialylated ligands (9). This suggests that a complex network of ligands might modulate cellular functions via PILRa. It was shown that PILRa binds to HSV-1 glycoprotein B and serves as a virus entry coreceptor (11, 13). Previous studies showed that PILRb-deficient and agonist anti-PILRa–treated mice show improved clearance of infectious reagents and improved survival (14, 15). A recent study showed that PILRa negatively regulates neutrophil recruitment during TLR-mediated inflammatory responses (16). However, the role of PILRa or PILRb in autoimmunity is largely unknown. Myeloid cells play a critical role in the pathophysiology of inflammation and autoimmunity. Rheumatoid arthritis (RA) is a systemic, inflammatory, autoimmune disorder manifested by chronic polyarthritis with synovial hyperplasia and joint destruction, resulting in pain, loss of joint function, and concomitant reduction of life quality (17, 18). Myeloid cells, including monocytes/macrophages and neutrophils, play an important role in various stages of arthritis development (19, 20). To investigate the role of PILRa in myeloid cell–mediated immune responses and its effects on autoimmune diseases, we studied its function in myeloid-driven models of RA, including collagen Ab–induced arthritis (CAIA) and K/BxN serum–transfer arthritis, using PILRadeficient mice and PILRa-specific mAbs. Murine CAIA is in-

Downloaded from http://www.jimmunol.org/ at Otterbein University on July 6, 2014

Paired Ig-like type 2 receptor (PILR)a inhibitory receptor and its counterpart PILRb activating receptor are coexpressed on myeloid cells. In this article, we report that PILRa, but not PILRb, is elevated in human rheumatoid arthritis synovial tissue and correlates with inflammatory cell infiltration. Pilra2/2 mice produce more pathogenic cytokines during inflammation and are prone to enhanced autoimmune arthritis. Correspondingly, engaging PILRa with anti-PILRa mAb ameliorates inflammation in mouse arthritis models and suppresses the production of proinflammatory cytokines. Our studies suggest that PILRa mediates an important inhibitory pathway that can dampen inflammatory responses. The Journal of Immunology, 2014, 193: 860–870.

The Journal of Immunology

mAbs was confirmed by ELISA and flow cytometry analysis. These mAbs do not cross-react with mouse PILRb. Hamster anti-PILRa mAb was used for staining, and mouse anti-PILRa mAb was used for in vivo and in vitro functional studies.

Gene expression analysis For gene expression analysis, total RNA was isolated using TRIzol reagent (Invitrogen). Total RNA (50 ng) was subjected to RT-PCR using Access RTPCR System (Promega). Primers used for amplification were as follows: b-actin, sense primer: 59-TACCTCATGAAGATCCTCA-39 and antisense primer: 59-TTCGTGGATGCCACAGGAC-39; PILRa, sense primer: 59GACTCCTGAGCACTCCAGTGTCCCTAGACAG-39 and antisense primer 59-AAGGTGACTCTAAGAGTGTCATCCAGACCA-39; and PILRb, sense primer: 59-GGAAATTCAGAAAGATACAACA-39 and antisense primer: 59CAGGTTCATCAGTGAAGGAT-39. For quantitative RT-PCR, total RNA (50 ng) was subjected to RT and real-time quantitative PCR assay using Access RT-PCR System (Promega), Perkin Elmer SYBR Green, and an ABI 7500 instrument. Gene-specific unlabeled primers used were as follows: b-actin, sense primer: 59-TACCTCATGAAGATCCTCA-39 and antisense primer: 59-TTCGTGGATGCCACAGGAC-39) PILRb, sense primer: 59CACCATGAGGAGCCCCTTCATCGTC-39 and antisense primer: 59TAGGACTATGACCAAAACTTTAGCCAGGAG-39.

Cell extraction and Western blot analysis

Materials and Methods Mice BALB/c mice were purchased from Charles River Laboratory. Pilra-deficient mice were generated by genOway and were maintained in-house. Mice were used at 6–10 wk of age. All animal experiments were performed in accordance with institutional animal care and use committee– approved protocols.

Generation of Pilra2/2 mice The genomic region containing the murine PILRa locus was isolated by PCR from 129Sv BAC genomic DNA (genOway, Lyon, France). The 6050-bp 59 homology arm corresponds to nt 137843309–137837259 cluster NC_000071.6 contig GRCm38.p2 in chromosome 5 of C57BL/6J. The 1872-bp 39 homology arm corresponds to nt 137830963–137829091 of the same region. This strategy allows for deletion of a 6.3-kb region comprising exons 1 and 2 of PILRa (containing ATG, Ig V-set domain, and extracellular domain). Neomycin-resistant clones were first screened by PCR using primers 59-ATGCTCCAGACTGCCTTGGGAAAAG-39 (neocassette) and antisense 59-CACTTCACAGTCTAGTCACTTCCGCACC-39 (outside 39 homology arm). Final clones were confirmed using primers 59-GCACCATGTCCTGAGGAAGAGTAGTTTGA-39 (within 59 homology arm) and 59-CACTTCACAGTCTAGTCACTTCCGCACC-39 (within 39 homology arm), which generates a 7.0-kb fragment in WT embryonic stem (ES) cells or a 3.8-kb fragment in neo-containing targeted ES cells. The expected neo-excised fragment of 2.1 kb was cloned and sequenced to confirm the desired recombination event for all ES clones. Gene targeting was further confirmed by Southern blot using a neomycin internal probe (SacI digest generating an 8.6-kb fragment) and a 39 external probe (SphI digest changing 11.0 kb to 6.0 kb). The 39 probe contains a 450-bp fragment between the 39 end of the targeting arm and the SphI site. A total of 6 of 646 clones carried correct integration. Two independent clones were microinjected into C57BL/6J-Tyrc-2J/J blastocysts to generate chimeras, which were bred to C57BL/6N mice (Charles River) to generate PILRa-knockout mice. Subsequently, the neomycin cassette was removed by breeding PILRa-knockout mice carrying the neomycin cassette with C57BL/6N Rosa-26-cre knockin mice (Taconic). The following primers were used to confirm the cre recombination and genotype: forward primer-1: 59-TGCAAACCACAGTTGGGTTGGCAACGGCTG-39 (located in exon 2, which is deleted in the knockout) and reverse primer: 59-GGAGGAGGCCAGAAGCCAACCTTTGCTATC-39 (within 39 targeting arm, which generates a 470-bp fragment in WT mice) or forward primer-2: 59-ATGACAGTTTGTGTCTGGATCAGAAC-39 (upstream of the loxp site) and the same reverse primer, which generates a 600-bp product from the knockout allele.

Bone marrow (BM) cells and splenocytes were harvested, and RBCs were lysed with ACK lysing buffer. Cells were washed twice with ice-cold PBS (pH 7.4) and then treated in Cell Extraction Buffer (Invitrogen) containing 1 mM PMSF and protease inhibitor mixture (Sigma). Lysates were incubated at 4˚C for 30 min, with vortexing at 10-min intervals, and the insoluble material was removed by centrifugation at 13,000 rpm for 10 min at 4˚C. Soluble extracts with Novex Tris-Glycine SDS sample buffer were separated by SDS-PAGE using 4–12% Bis-Tris gel and transferred to a PVDF/Filter Paper Sandwich (all from Invitrogen). Blots were blocked with 5% nonfat dried milk, 0.1% Tween-20 in TBS buffer and incubated with hamster anti-mouse PILRa mAb, followed by HRP-conjugated antihamster IgG (Jackson ImmunoResearch) for 1 h. Proteins were detected by chemiluminescence with an Amersham ECL Plus Western blotting system (GE Healthcare).

Immunohistochemistry Immunohistochemistry for PILRa was performed on formalin-fixed, paraffin-embedded tissue using in-house–generated Armenian hamster anti-mPILRa mAb and an automated system (VENTANA DISCOVERY XT, Ventana Medical Systems, Tucson, AZ). Briefly, 4-mm sections were deparaffinized, rehydrated, and subjected to Ag retrieval (Ventana Protease 2). Primary Ab (5 mg/ml) was incubated at 37˚C for 60 min. Binding was visualized using secondary anti-hamster IgG, Ventana Rabbit OmniMap, and Ventana DAB reagents (Ventana Medical Systems). Immunohistochemistry for myeloperoxidase (MPO) and F4/80 was performed on formalin-fixed, acid-decalcified, paraffin-embedded sections of murine arthritic paws to visualize infiltration by neutrophils or macrophages, respectively. For MPO, a rabbit polyclonal Ab A95-1 (LabVision/ Neomarker, Fremont, CA) was used on an automated platform (Ventana) using Ventana Rabbit OmniMap and DAB for visualization. For F4/ 80, a commercial rat Ab (Cl:A3-1; AbD Serotec, Raleigh, NC) was used on an automated platform (Leica Bond) and visualized with DAB using an ABC Peroxidase protocol.

Flow cytometry Single-cell suspensions were prepared from BM and peritoneal lavage fluid. After RBC lysis, cells were incubated with rat anti-mouse CD16/CD32 (Mouse BD Fc Block) and stained with in-house–generated hamster antimouse PILRa mAb, followed by allophycocyanin-conjugated anti-hamster IgG (R&D Systems). BM cells were further stained with PE-conjugated Ly6G and FITC-conjugated anti-CD11b. Peritoneal cells were stained with PE-conjugated Gr-1 and FITC-conjugated anti-CD11b or FITC-conjugated CD117 and PE-conjugated FcεRI (all from BD Pharmingen). Cell acquisition was performed on a FACSCalibur (Becton Dickinson, Mountain View, CA), and data were analyzed with FlowJo software.

Generation of mAbs Anti-PILRa mAbs were generated from Armenian hamsters and Pilra2/2 mice using immunogenic fusion proteins consisting of the extracellular domain of the mouse Pilra gene (GenBank accession number: NM_153510), including amino acid positions Met 1 to Val 197, as previously described (9). Mouse anti-PILRa mAbs are IgG1 isotype. The specificity of anti-PILRa

Induction of CAIA Arthritis was induced in 7–8-wk-old (age-matched for a given study) female C57BL/6N or BALB/c mice with an arthritogenic anti–type II collagen mAb mixture purchased from Chondrex. Pilra2/2 and WT littermate mice received 4 mg a mixture of anti–type II collagen mAbs i.v. in 400 ml

Downloaded from http://www.jimmunol.org/ at Otterbein University on July 6, 2014

duced by i.v. injection of mAbs against type II collagen, followed by i.p. injection of LPS. This model is widely used to study the pathogenesis of autoimmune arthritis and to determine efficacy of therapeutics (21–24). Myeloid cells, FcgRs (22), and proinflammatory cytokines, especially TNF-a and IL-1b (21), are indispensable for the development and maintenance of arthritis in this model (23). The K/BxN serum–transfer arthritis model is induced by transferring K/BxN serum into normal mice; LPS is not needed in this model. The K/BxN serum–transfer arthritis model shares multiple features with human RA, including symmetrical involvement of peripheral joints, pannus formation, synovial hyperplasia, and bone and cartilage degradation (25). We show that Pilra2/2 mice developed more severe arthritis than did wild-type (WT) mice in these two models, corroborating the inhibitory role of PILRa during inflammation. Accordingly, we found that administration of an anti-PILRa mAb attenuated mouse inflammatory arthritis and reduced the production of several proinflammatory cytokines, indicating that PILRa is part of a critical inflammatory axis regulating myeloid cell function.

861

862 PBS on day 0, followed by 50 mg LPS i.p. in 100 ml PBS on day 3. To test the effect of anti-PILRa mAb on the development of CAIA, BALB/c mice received 2 mg Arthrogen-CIA Ab mixture i.v. in 200 ml PBS on day 0, followed by 50 mg LPS i.p. in 100 ml PBS on day 3. Mice were treated with 100 mg mouse anti-PILRa mAb or anti-ragweed mIgG1 isotype control i.p. daily in 100 ml sterile saline, starting on day 21. For K/BxN serum–transfer arthritis model, BALB/c mice were given 20 ml arthritogenic K/BxN serum on days 0 and 2 by i.v. injection and were treated with 100 mg mouse anti-PILRa mAb or anti-ragweed mIgG1 isotype control i.p. daily in 100 ml sterile saline starting on day 21. Swelling of the paws was monitored for 11–12 d. Each paw was assigned a score based on the following scale: 0 = no evidence of erythema and swelling; 1 = erythema and mild swelling confined to the midfoot (tarsal) or ankle; 2 = erythema and mild swelling extending from the ankle to the midfoot; 3 = erythema and moderate swelling extending from the ankle to the metatarsal joints; and 4 = erythema and severe swelling encompass the ankle, foot, and digits. The clinical score of each mouse is the sum of the four paw scores.

Histopathological assessment

FIGURE 1. Pilra2/2 mice generation and characterization. (A) Schematic diagram of PILRa-targeting construct. Exons 1 and 2 were replaced with loxp sites (triangles) flanked by the neomycin resistance gene (neo). The neomycin cassette was deleted by crossing the mice with the Rosa-26-cre knockin line. (B) Genomic DNA from WT ES cells or targeted ES cells were digested with SacI and probed with a neomycin internal probe (left panel). All independent clones detected a single band of an 8.6-kb fragment corresponding to a single integration event at the 59 homology arm. SphI genomic DNA digest was used to confirm expected homologous recombination at the 39 end using a genomic probe located between the 39 end of the targeting arm and the SphI site, which generated a shorter fragment of 6.0 kb compared with 11.0 kb in WT ES cells (right panel). (C) PILRa and PILRb expression in spleen and BM of WT and Pilra2/2 (KO) mice by RT-PCR (n = 3). (D) PILRb expression in spleen and BM of WT and Pilra2/2 (KO) mice by real-time PCR. (E) PILRa expression in spleen and BM of WT, Pilra+/2 (HET), and Pilra2/2 (KO) mice by Western blot. Data are representative of two or three experiments.

Microcomputed tomography imaging Microcomputed tomography (micro-CT) imaging was performed using a previously described micro-CT and automated analysis technique (26) to quantify bone destruction in the arthritis mouse model. Briefly, the paws were scanned ex vivo on a micro-CT 40 scanner (Scanco Medical, Br€uttisellen, Switzerland). The analysis algorithm automatically locates five metatarsophalangeal joints and three metacarpophalangeal joints and segments the cortical bone and periosteal new bone. Joint cortical bone volume (JCBV) is sensitive to cortical bone erosion at the joints, and periosteal new bone volume (PNBV) estimates the amount of periosteal new bone formation that results from a repair response to the bone erosion. The analysis technique was modified in such way that JCBV, as well as the PNBV estimate, was limited to the joints.

Measurement of cytokine concentration in arthritic hind paws Hind footpads were obtained from mice by cutting at the borderline of fur growth. The footpads were hammer smashed and shaken at 20 Hz for 30–60 min with 5-mm steel beads at 4˚C in ice-cold cell lysis buffer (Cell Signaling Technology) containing Protease Inhibitor Cocktail Set I (Calbiochem). Homogenized tissue was kept on ice for 30 min, vortexed occasionally, and then spun down at high speed for 15 min at 4˚C. Supernatants were collected, and all samples were frozen at 280˚C and thawed to room temperature before analysis. Cytokine levels were analyzed by ELISA (R&D System). The concentration of total protein in the supernatants was measured using a Pierce BCA kit and was normalized against the concentration of BSA. The concentration of cytokines was expressed as pg/mg of protein.

In vitro stimulation of BM-derived DCs BM-derived DCs (BMDCs) were generated as described elsewhere (27). In brief, femurs of mice were removed, and BM was flushed out with PBS containing 10% FCS. RBCs were lysed with ACK lysing buffer. Cells were cultivated in bacterial culture dishes for 9–10 d at 37˚C in a humidified

Downloaded from http://www.jimmunol.org/ at Otterbein University on July 6, 2014

Paws were removed, fixed in formalin, decalcified, and processed to paraffin sections stained with H&E. Sagittal sections were examined by light microscopy and scored on arbitrary scales ranging from 0 to 5 for features of arthritis, including number of affected joints, extent of inflammation, extent of soft tissue proliferation including pannus, cartilage injury, and bone remodeling. Immunoreactivity for MPO or F4/80 was evaluated on arthritic paws by microscopic examination of affected synovial structures (joints, tendon sheaths). Intensity of infiltration with immunoreactive cells was estimated using an arbitrary scoring scheme ranging from 0 to 3, according to the following criteria: 0 = none detected near or in synovial structures; 1 = few, scattered positive cells near or in synovial structures; 2 = frequent disseminated or loosely clustered positive cells near or in synovial structures; and 3 = locally extensive infiltration with numerous positive cells near or in synovial structures.

PILRa REGULATES INFLAMMATION

The Journal of Immunology atmosphere containing 5% CO2 using RPMI 1640 supplemented with 10% heat-inactivated FCS, 2 mM L-glutamine, 100 U/ml penicillin, 100 mg/ml streptomycin, 50 mM 2-ME (Sigma; Taufkirchen, Germany), and 20 ng/ml recombinant mouse (m)GM-CSF (R&D Systems). At days 9–10, the nonadherent cells were harvested and stimulated with mIgG (60 mg/ml; Jackson ImmunoResearch Laboratories) plus various concentrations of mouse anti-mouse PILRa mAb coated on Protein G Agarose beads (Roche) for 18 h. Cytokine production in supernatants was detected by ELISA (R&D) or Luminex (Bio-Rad).

PILRa and PILRb expression in synovial tissues of RA and osteoarthritis patients

FIGURE 2. Expression of PILRa in various tissues and cells. (A) Immunohistochemistry of normal murine tissues. Tissue sections from WT mouse spleen, liver, lung, and BM were stained with anti-PILRa Ab and compared with control tissues from Pilra2/2 mice. Spleen red pulp and tissue-resident macrophages, such as Kupffer cells in liver (inset) or exfoliated alveolar macrophages (inset) in lung, express PILRa. In BM, most PILRa+ cells are identified as granulocytes or their precursors. Scale bars, spleen = 500 mm; liver and lung = 50 mm; BM = 20 mm. (B) PILRa is expressed on murine granulocytes and monocytes from BM and peritoneal macrophages, neutrophils, and mast cells, but it is not expressed on spleen lymphocytes. Data are representative of three experiments.

Clara, CA). Samples were hybridized to GeneChip Human Genome U133 Plus 2.0 Arrays (Affymetrix). Arrays were washed and stained in the Affymetrix Fluidics station and scanned on a GeneChip scanner 3000. Expression signals were obtained using the Affymetrix GeneChip operating system and analysis software. Data have been deposited in the National Center for Biotechnology Information’s Gene Expression Omnibus under accession no. GSE48780 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi? acc=GSE48780). Computational analysis. Computational analysis was carried out using the R programming language (29) and libraries from the open-source Bioconductor project (30). Quality assessment of microarray data was conducted using the Bioconductor affy package (31), and the expression data were normalized using the robust multiarray average method (32, 33). PILRa and PILRb are each represented by two probes on the Affymetrix U133Plus chips. To compare the expression levels of PILRa and PILRb in RA and osteoarthritis (OA) samples, we used a linear model with terms for diagnosis (OA or RA) and inflammation status (inflamed or noninflamed) based on histology. Conceptually, this can be thought of as a t test comparing OA and RA expression with correction for inflammation status. The limma Bioconductor package was used for the computation (34).

PILRa ligand screen Recombinant human PILRa-Fc was screened for interacting partners using the Genentech human full-length cDNA expression library. COS-7 cells were transfected with 6 ng/ml individual cDNA expression vectors and FuGene 6 (Roche) in a 384-well format. Forty-eight hours after transfection, COS-7 cells were exposed to 10 ng/ml PILRa-Fc, followed by fixation in 4% formaldehyde. Protein–protein interactions were detected upon binding of an Alexa Fluor 488–conjugated anti-mIgG2a secondary Ab (Invitrogen) to the cell surface using an MD Isocyte plate scanner or, in follow-up experiments, using a GE IN Cell 2000 Analyzer. To confirm the binding of screening hit to PILRa, HEK 293T cells were transfected with C-type lectin domain family 4 (CLEC4G) cDNA, and transfectants were stained with human (h)PILRa-Fc/mPILRa-Fc, hPILRaR126A-Fc/mPILRaR133A-Fc, followed by FITC-conjugated antimIgG2a. Cell acquisition was performed on a FACSCalibur (Becton

Downloaded from http://www.jimmunol.org/ at Otterbein University on July 6, 2014

Patients and synovial tissues. Synovial tissues were obtained from RA subjects undergoing arthroplasty and/or synovectomy of affected joints (University of Michigan, two sequential cohorts). RA was diagnosed based upon the 1987 American College of Rheumatology criteria (28). Patients were treated using the standard of care for RA (nonsteroidal anti-inflammatory drugs and disease-modifying antirheumatic drugs), and some patients also were treated with biologics (adalimumab, etanercept, infliximab, anakinra, and rituximab). Patients were diagnosed with RA $3 y before surgery, and 70% of patients for whom data were available were positive for rheumatoid factor. All procedures to collect human specimens were performed under a protocol approved by the University of Michigan Institutional Review Board. Excised tissues were immediately snap-frozen in liquid nitrogen and stored at 280˚C. Each tissue was used for both histology and RNA extraction. For cryosectioning, samples were brought briefly to 220˚C, sectioned, and immediately returned to 280˚C to maintain RNA integrity. All tissues used for downstream studies were prospectively randomized during processing and sectioning, prior to expression analysis, to minimize technical batch effects in the data. RNA isolation. Frozen samples were weighed and homogenized in RLT (QIAGEN, Valencia, CA) + 2-ME (10 ml/ml) at a concentration of 100 mg/ ml. RNA was isolated using an RNeasy Mini Kit (QIAGEN) with oncolumn DNase digestion. Microarray hybridization. The protocols for preparation of cRNA and for array hybridization were followed as recommended by Affymetrix (Santa

863

864 Dickinson, Mountain View, CA), and data were analyzed with FlowJo software.

Statistical analysis Statistical significance was determined by the parametric Dunnett test for the comparison of joint cytokine concentrations, by the nonparametric Dunnett test for the comparison of arthritis scores and histopathological as well as micro-CT results, and by Student t test for the comparison of in vitro cytokine data. The p values , 0.05 were considered significant.

Results

Generation and characterization of Pilra2/2 mice To generate the PILRa-knockout mouse, exon 1 and exon 2 of the PILRa gene were deleted by homologous recombination in C57BL/6J ES cells (Fig. 1A), and the targeted locus was confirmed by Southern blot analysis using internal (Fig. 1B, left panel) or external (Fig. 1B, right panel) probes, as described in

Materials and Methods. Chimeric mice were bred to C57BL/6N mice to generate PILRa-deficient mice on a pure genetic background. Given the close proximity of PILRb, the neo-pgk cassette was removed by breeding PILRa-knockout mice to Rosa-26-cre knockin mice to avoid nonspecific effects on PILRb or other surrounding genes (35). RT-PCR analysis revealed that PILRa mRNA production was abolished in the spleen and BM of Pilra2/2 mice, whereas PILRb expression was not affected (Fig. 1C). Real time RT-PCR was done to better quantitate PILRb levels in WT and Pilra2/2 mice. These expression studies showed that PILRb expression in spleen and BM was not affected (Fig. 1D). PILRa protein was abundant in WT BM and was detected in spleen at lower levels. However, in Pilra2/2 mice, no PILRa protein was detected in either BM or spleen, and Pilra+/2 heterozygous mice displayed intermediate levels of protein (Fig. 1E). Pilra2/2 mice are viable, fertile, and have a normal lifespan. They did not show

Downloaded from http://www.jimmunol.org/ at Otterbein University on July 6, 2014

FIGURE 3. Pilra2/2 mice are more susceptible to arthritis. (A) Immunohistochemistry of inflamed joint tissue (CAIA model, day 11). Within the cavity of an inflamed joint, PILRa is expressed on neutrophils (arrows), the predominant cell type in CAIA. (B–E) C57BL/6N WT (n = 6) and Pilra2/2 mice (n = 8) were given arthritogenic anti–type II collagen mAb mixture to induce arthritis. (B) Pilra2/2 mice develop more severe CAIA. Pilra2/2 mice showed increased average clinical score (left panel), average daily clinical score (days 4–12, middle panel), and end point clinical score (day 12, right panel) compared with WT mice. (C) Hind paws, sagittal H&E. Representative images of the most severe lesions per group. Periarticular infiltration with inflammatory cells and formation of intraarticular exudate are obvious in Pilra2/2 mice (arrows) compared with WT mice, where these features are essentially absent. Scale bar, 200 mm. (D) Lesion scores for features of arthritis, including infiltration with inflammatory cells, fibroplasia, cartilage injury, and bone remodeling score per group on day 12, are lower in WT mice compared with Pilra2/2 mice. Data points represent average scores for two paws/mouse. Two limbs per animal, four to six sections/animal, and a total of 14 mice were examined. (E) Hind paws from mice harvested on day 12 were subjected to micro-CT analysis. Images show high resolution three-dimensional rendering of micro-CT scans. Quantification of bone integrity, including JCBV and PNBV are shown. Results are representative of two experiments. The microCT parameter estimates for Pilra2/2 mice were compared with the control group using the Dunnett test.

PILRa REGULATES INFLAMMATION

The Journal of Immunology

Pilra2/2 mice develop enhanced autoimmune arthritis To investigate the biological function of PILRa in myeloid cell– mediated immune responses and its role in autoimmune diseases, we made use of a CAIA model, which is a myeloid-dependent preclinical disease model. We observed an infiltration of PILRaproducing cells in mouse arthritic tissues, specifically periarticular soft tissue obtained from CAIA (Fig. 3A). We first tested how CAIA is affected by PILRa deficiency. As shown in Fig. 3B (left

panel), littermate WT mice started to develop clinical signs of arthritis 5 d after injection of the type II collagen–specific Ab mixture, 2 d after LPS administration, and reached a maximum clinical score of 3.8 6 2.6 (mean 6 SD) between days 7 and 9 before gradually resolving disease. Pilra2/2 mice showed similar kinetics of disease development as did WT mice, but they exhibited more severe disease, with maximum clinical scores of 8.0 6 2.6 (mean 6 SD). Pilra2/2 mice also had significantly increased average clinical scores and end point clinical scores compared with WT mice (Fig. 3B, middle and right panels). On termination of the experiment at day 12, histopathology confirmed the clinical observations. Compared with WT mice, Pilra2/2 mice had more severe disease, as shown by all evaluated parameters, including number of affected joints/paw, periarticular infiltration with mixed inflammatory cells, soft tissue proliferation (including pannus formation), cartilage injury, and bone remodeling (Fig. 3C, 3D). Consistent with the histological evidence for more extensive bone remodeling, micro-CT revealed more severe osteolysis (i.e., smaller JCBV), as well as increased new bone formation (i.e., greater PNBV), in Pilra2/2 mice compared with WT mice (Fig. 3E). These results support the regulatory role of PILRa in downregulating myeloid-dependent inflammation. PILRa deficiency results in an increase in intra-articular proinflammatory cytokine levels at the onset of arthritis To better understand how PILRa deficiency impacts CAIA, we tested inflammatory cytokine levels in joint tissue at the onset of disease. Onset of disease was similar in WT and Pilra2/2 mice, with the start of clinical signs of arthritis on day 5 following in-

FIGURE 4. Elevated levels of inflammatory cytokines were detected in joints of Pilra2/2 mice at the onset of disease. (A) C57BL/6N WT (n = 9) and Pilra2/2 mice (n = 9) were given arthritogenic anti–type II collagen mAb mixture to induce arthritis (CAIA). On day 5, hind footpads were obtained from mice for homogenization in cell lysis buffer containing protease inhibitors. Cytokine levels in lysate supernatants were analyzed by ELISA. The cytokine concentrations are expressed as pg/mg of protein. (B) Kinetics of neutrophils and macrophages in the early phase of CAIA show a similar degree of infiltration in arthritic joints of Pilra2/2 and WT mice. WT (dark gray bars, n = 4) and Pilra2/2 mice (light gray bars, n = 4) were given arthritogenic anti– type II collagen mAb mixture to induce arthritis. On days 3, 4, and 5, paws were collected for immunohistochemical staining of MPO and F4/80. Intensity of infiltration was estimated using an arbitrary scoring scheme ranging from 0 to 3. Bars represent mean of four arthritic paws. Error bars are SEM. Results are representative of two or three separate experiments. The p values are based on the parametric Dunnett test.

Downloaded from http://www.jimmunol.org/ at Otterbein University on July 6, 2014

any appreciable developmental defects in myeloid or lymphoid cell populations (Supplemental Table I). Previous work showed that PILRa transcripts are relatively high in spleen, liver, and lung (8). We tested PILRa protein expression in these tissues by immunohistochemistry (Fig. 2A) using Pilra2/2 mice as a negative control and in-house–generated Armenian hamster anti-mPILRa mAb, which did not show cross-reactivity to mPILRb. In the spleen, PILRa+ cells are mainly localized in the red pulp, consistent with expression predominantly by myeloid cells. In liver and lung, PILRa is expressed in tissue-resident macrophages, such as Kupffer cells or exfoliated alveolar macrophages, respectively. In BM, most PILRa+ cells are identified as granulocytes or their precursors. These data corroborate the notion that PILRa is expressed primarily in cells of the myeloid lineage. Correspondingly, PILRa protein is detected on granulocytes (CD11b+Ly6G+) and monocytes (CD11b+Ly6G2) from BM, as well as peritoneal cavity macrophages (CD11b+Gr-12), neutrophils (CD11b+Gr-1+), and mast cells (CD117+FcεRI+), but not on spleen lymphocytes, including T, B, NKT, and NK cells (Fig. 2B).

865

866

Engagement of PILRa suppresses the production of inflammatory cytokines and chemokines To confirm the regulatory role of PILRa in modulating cellular function, we tested how the production of cytokines and chemokines is affected in Pilra2/2 cells or upon its engagement in vitro. We used various PILRa-expressing cell types, such as BMDCs, BM-derived monocytes (BMDMs), and neutrophils, which are implicated in the pathology of anti-CAIA. First, we tested how these cells respond to LPS, which is used to activate myeloid cells in the same CAIA model. The results showed that Pilra2/2 BMDCs produced increased TNF-a and MCP-1 in response to LPS but not IL-1b, IL-6, or KC (Supplemental Fig. 2). A subtle trend of cytokine elevation was seen in neutrophils and BMDMs (Supplemental Fig. 2). The state of cellular hyperresponsiveness, as measured by cytokines/chemokines, appears to vary in different PILRa-expressing cells. BMDCs were more sensitive to this pathway, which might be explained by the high expression of PILRa (Fig. 5A) in these cells. We then asked how engagement of PILRa receptor in BMDCs modulated cytokine production. To couple stimulatory and putative inhibitory pathways, we used mIgG-coated beads as the stimulus. WT and Pilra2/2 BMDCs produced similar levels of the proinflammatory cytokine TNF-a and chemokine MCP-1 in response to mIgG-coated beads (Supplemental Fig. 2). However, stimulation of BMDCs with beads coated with anti-mPILRa mAb, but not isotype control, showed a dose-dependent inhibition of TNF and MCP-1 production in WT cells. The inhibitory effect of antiPILRa Ab was not seen in the BMDCs derived from Pilra2/2 cells, again confirming that the inhibitory effect is PILRa dependent (Fig. 5B). This result indicates that engagement of PILRa downregulates proinflammatory cytokine and chemokine production when it is proximal to an activating receptor, such as FcgR.

FIGURE 5. Engaging PILRa on BMDCs reduces cytokine production. (A) BMDCs express PILRa. WT and Pilra2/2 BMDCs were stained with anti-mPILRa mAb or isotype control. (B) WT and Pilra2/2 BMDCs were stimulated with mIgG plus various concentrations of mouse anti-mPILRa mAb–coated Protein G Agarose beads for 18 h, and cytokine production in the supernatants was measured. Results are representative of three experiments. *p , 0.05, Student t test.

A recent study showed that PILRa associates with its ligands in cis on neutrophils (16). Because BMDCs express both PILRa and its ligands (Supplemental Fig. 3A), we predicted that PILRa would interact with its ligands in cis, and this may play a role in negatively regulating BMDC function. If this were the case, dissociation of the interaction between PILRa and its ligands would increase BMDC cytokine production. We showed previously that sialic acid is an essential component of PILRa ligands in primary cells, and treatment of primary cells with sialidase A, which cleaves sialic acid from surface proteins, abolished their binding to PILRa (9). Therefore, we used sialidase A treatment to dissociate PILRa–ligand interactions on BMDCs. We found that sialidase A treatment increased cytokine production by BMDCs in response to mIgG coated on protein G beads, suggesting that the interaction of sialylated ligands with their receptors inhibits cellular function (Supplemental Fig. 3B). Given the fact that PILRa is able to interact with a network of ligands expressed by different tissues, including hematopoietic cells other than myeloid cells (9), cis and trans interactions may both be involved in the regulation of PILRa-expressing myeloid cell function. Anti-PILRa mAb treatment attenuates arthritis We then tested whether treatment with anti-PILRa mAb affects inflammation in vivo using the CAIA model. Because C57BL/6N WT mice develop only mild disease, we used BALB/c mice instead for anti-PILRa mAb treatment. We found that anti-PILRa mAb treatment significantly reduced inflammation compared with isotype control treatment (Fig. 6A). Isotype control–treated mice started to develop clinical signs of arthritis 4 d after injection of type II collagen–specific Ab mixture or 2 d after LPS administration and reached maximum clinical scores of 14.6 6 2.2 (mean 6 SD) between days 5 and 7. By contrast, anti-PILRa mAb–

Downloaded from http://www.jimmunol.org/ at Otterbein University on July 6, 2014

jection of type II collagen–specific Abs. At this time point, hind paw lysates of Pilra2/2 mice exhibited significantly increased production of proinflammatory cytokines, including IL-1b and IL-6, as well as neutrophil and monocyte chemokines KC and MCP-1 (Fig. 4A). The increased inflammatory cytokine and chemokine production could be due to either increased inflammatory cell number or activity. It was shown that PILRa plays an important role during acute inflammatory responses in regulating neutrophilic migration (16). We then asked how cellular infiltrates are impacted in the ankle joints at early time points each day from day 3 to day 5 using immunohistochemistry MPO to stain neutrophils or F4/80 to localize macrophages (Fig. 4B). Neutrophils were slightly reduced in PILRa-knockout joints on day 3; however, their numbers were transiently increased on day 4 compared with WT mice. Macrophages in PILRa-knockout joints also showed a similar trend on day 3; however, their numbers were slightly increased on day 5. Overall, the changes in macrophage and neutrophil cellularities were not statistically significant between WT and Pilra2/2 mice (Fig. 4B). The slight reduction in neutrophil/macrophages in PILRa-knockout mice on day 3 might be secondary to the systemic effect of LPS on the increased migration of these cells to other tissues, such as liver, in these mice (16). In addition, we did not observe any difference between WT and Pilra2/2 neutrophil and macrophage cellular infiltrates in an acute localized inflammatory model of Staphylococcus aureus (Supplemental Fig. 1). These observations support the notion that the lack of PILRa has a global effect on the severity of arthritis, mainly through enhanced cytokine production. In summary, our data suggest that the joint-infiltrating Pilra2/2 myeloid cells are functionally hyperresponsive, further highlighting the regulatory role of PILRa in myeloid cells.

PILRa REGULATES INFLAMMATION

The Journal of Immunology

FIGURE 6. Anti-PILRa mAb treatment attenuates arthritis development. (A–C) BALB/c mice were given arthritogenic anti–type II collagen mAb mixture to induce arthritis (CAIA) and were treated with anti-PILRa mAb (n = 5). (A) Anti-PILRa mAb–treated mice exhibited reduced disease development compared with isotype control. AntiPILRa–treated mice showed decreased average clinical score (left panel), average daily clinical score (days 4–11, middle panel), and end point clinical score (day 11, right panel) compared with isotype control–treated mice. (B) Pathological scores of isotype control and anti-PILRa mAb– treated mice on day 11. Data points represent average scores for two paws/mouse. Two limbs per animal, four to six sections/animal, and a total of 10 mice were examined. (C) Hind paws from mice harvested on day 11 were subjected to micro-CT analysis. Images show high-resolution three-dimensional rendering of micro-CT scans. Quantification of bone integrity, including JCBV and PNBV. (D) Anti-PILRa mAb treatment inhibited disease development in a K/BxN serum–transfer arthritis model. BALB/c mice were given 20 ml of arthrogenic K/BxN serum on days 0 and 2 by i.v. injection and were treated with anti-PILRa mAb (n = 6). Mean clinical score (left panel), average daily clinical score (middle panel), and day-7 clinical score (right panel) are shown. In vivo arthritis scores and micro-CT parameter estimates for the treatment groups were compared with the control group using the Dunnett test. Results are representative of two experiments.

anti-PILRa mAb–treated mice exhibited a reduced maximum clinical score of 8.5 6 4.0 (mean 6 SD) between days 5 and 7 (Fig. 6D, left panel). Anti-PILRa mAb–treated mice also showed significantly reduced average clinical scores and terminal clinical scores compared with isotype control–treated mice (Fig. 6D, middle and right panels). Additional studies are required to understand how PILRa is relevant in the context of additional inflammatory diseases. PILRa is expressed in human inflammatory arthritic tissues and correlates with monocyte and DC levels To test whether PILRa is involved in human inflammatory arthritis, we assessed the expression patterns of PILRa and its counterpart PILRb using mRNA microarrays on human synovial tissue. Samples were obtained during joint resection surgery from OA and RA patients. Sections were taken for both histological observation and mRNA extraction. The Affymetrix expression microarrays included two probes each for PILRa and PILRb. PILRa consistently showed significantly higher expression in the RA patients compared with the OA patients, including a correction for inflammation status (p , 1e27) (Fig. 7A, upper panels). In contrast, PILRb showed a modestly higher level of expression in OA patients (p , 0.02) (Fig. 7A, lower panels). The differential expression of PILRa and PILRb in inflammatory RA suggests that PILRa might function as a negative-feedback regulatory loop to dampen inflammation. Cell-specific gene signatures can be used as an estimate of cellularity (40). For each cell type, we calculated a summary gene set score

Downloaded from http://www.jimmunol.org/ at Otterbein University on July 6, 2014

treated mice exhibited delayed disease onset, which gradually increased to a much reduced maximum clinical score of 4.4 6 4.3 (mean 6 SD) by the end of the study. Anti-PILRa mAb–treated mice also showed significantly reduced average clinical scores and end point clinical scores compared with isotype control– treated mice (Fig. 6A, middle and right panels). In accordance with in vivo findings, histopathological assessment of hind paws at the termination of the study at day 11 showed significantly decreased scores for all scored features in anti-PILRa mAb–treated mice compared with isotype control Ab–treated mice (Fig. 6B). Finally, micro-CT imaging of paws at day 11 showed that antiPILRa mAb–treated mice exhibited less severe bone remodeling compared with isotype control Ab–treated mice (Fig. 6C), as measured by JCBV and PNBV (p , 0.01 for both parameters). The observation that anti-PILRa treatment attenuates arthritis development, whereas Pilra2/2 mice show enhanced disease, suggests that the anti-PILRa mAb does not block the receptor but rather engages it. To measure the impact of PILRa modulation on other arthritis models, we tested the effect of anti-PILRa treatment on mouse K/BxN serum–transfer arthritis. The K/BxN serum–transfer model is a widely accepted model of inflammatory arthritis that shares features of human RA (25). In K/BxN mice, the induction of disease does not require LPS, and this disease depends highly on the innate cells in a TNF-independent fashion (36–39). Mice started to develop clinical signs of arthritis 1 d after K/BxN serum transfer. Isotype control–treated mice reached maximum clinical scores of 14.7 6 1.0 (mean 6 SD), whereas

867

868 using a quartile-trimmed mean of the normalized probe set values present in the gene set. This summary score provides an estimate for the relative abundance of the given cell type across samples. Correspondingly, we found that PILRa expression was highly correlated with the tissue-specific gene sets obtained from different inflammatory cell types, such as DCs (Spearman r = +0.7, p , 2e216), neutrophils (r = +0.5, p , 2e25), and monocytes (r = 0.8, p , 2e2 16) in RA samples (Fig. 7B, upper panels). In contrast, PILRb expression was minimally inversely correlated with DC-specific and monocyte-specific genes (r = 20.2 for both, p , 0.01), and it was

PILRa REGULATES INFLAMMATION not significantly correlated with neutrophil-specific genes (r = 20.1, p , 0.3) (Fig. 7B, lower panels). These trends held true when we considered all samples together or RA and OA samples in isolation. Our human expression study shows that PILRa expression may dissociate from PILRb, and this inhibitory pathway is overrepresented in myeloid cells in human RA. Diverse ligands are involved in PILRa pathway We (9) and other investigators (8, 10, 41) reported that PILRa engaged multiple ligands with common molecular determinants

Downloaded from http://www.jimmunol.org/ at Otterbein University on July 6, 2014

FIGURE 7. PILRa-inhibitory versus PILRb-activating pathways in human RA joints and their correlation with pathogenic cell types. Synovial samples were taken from RA (n = 81) and OA (n = 121) patients during joint resection surgery. (A) The Affymetrix expression microarrays included two probes each for PILRa and PILRb. PILRa expression is significantly higher (p , 1e207) in RA patients compared with OA patients; this also holds when corrected for inflammation status. PILRb expression is modestly higher in OA patients compared with RA patients (p , 0.02), correcting for inflammation status. (B) The Affymetrix expression microarrays of PILRa and PILRb. In RA patients, PILRa expression is significantly correlated with DC-specific genes (r = +0.7, p , 2e216), neutrophil-specific genes (r = +0.5, p , 2e25), and monocytes (r = +0.8, p , 2e216). This trend also holds true for all of the inflammatory RA subset of patients. In contrast, PILRb expression was not positively correlated with DC, neutrophil, or monocyte gene expression (r = 20.2, p = 0.01; r = 20.1, p = 0.2; and r = 20.2, p , 0.01). Scores were calculated using a quartile-trimmed mean of the normalized probe set values present in the gene set.

The Journal of Immunology (9). Using recombinant hPILRa-Fc to screen a high-throughput human full-length cDNA expression library, we identified an additional PILRa ligand CLEC4G [previously named LSECtin (42)] (Fig. 8A). We confirmed its binding to both human and mouse PILRa by transfecting HEK 293T cells with this putative ligand and staining with human and mouse PILRa-Fc for analysis by flow cytometry (Fig. 8B). Consistent with our previous findings that an evolutionarily conserved PILRa arginine site (hPILRaR126A and mPILRaR133A) is essential for its interaction with diverse sialylated ligands, we found that hPILRaR126AFc and mPILRaR133A-Fc lost their binding activity to these ligands (Fig. 8B), suggesting that they all share the same sialylated PILRa binding domain. mPILRb-Fc also bound to CLEC4G (Fig. 8B, far right panel). Thus, a complex network of ligands using a common molecular determinant may modulate the PILRa pathway in different tissues.

In this study, we reported that PILRa represents a potent inhibitory receptor that negatively regulates the function of myeloid cells to control inflammation in mouse arthritis models. PILRa is predominantly expressed by macrophages, DCs, and neutrophils, all of which are important in the pathogenesis of arthritis. Using PILRa-knockout mice in a CAIA model, we demonstrate that the absence of PILRa leads to dysregulation of inflammatory processes in inflamed tissues, resulting in increased inflammatory cytokine production and severity of disease. Conversely, engaging PILRa with an anti-PILRa mAb reduces disease severity in both CAIA and K/BxN serum–transfer models. In preclinical arthritis models, as well as in human inflammatory arthritis, PILRa gene expression profiling supports the notion that this receptor acts in a counter-regulatory fashion to dampen inflammation. It is interesting that PILRa (inhibitory receptor), but not PILRb (activating receptor), expression is positively correlated with monocytes or neutrophil gene signatures in human RA, suggesting that this pathway is highly regulated via its ligands to possibly keep these inflammatory cells in check. The usefulness of the PILRa pathway in human RA or additional inflammatory models requires addition studies. A complex network of ligands expressed in a variety of tissues has been implicated in PILRa interactions (8–10). Although we do not fully understand the consequence of the interaction of PILRa

FIGURE 8. CLEC4G is a novel ligand of PILRa. (A) Recombinant hPILRa-Fc was screened for interacting partners using a Genentech human full-length cDNA expression library. COS-7 cells were transfected with individual cDNA expression vector and then exposed to PILRa-Fc, followed by fixation. Protein–protein interactions were detected with an Alexa Fluor 488–conjugated anti-mouse IgG2a secondary Ab and scanned with an MD Isocyte plate scanner. (B) HEK 293T cells were transfected with expression vectors for CLEC4G and then stained with h/mPILRa-Fc, hPILRaR126A/mPILRa-R133A-Fc, or mPILRb-Fc (black line) or isotype control IgG (gray line).

with individual ligands, it is interesting that a conserved domain in PILRa is critical for its interaction with all known ligands, a property shared with the Siglec family of inhibitory receptors (9, 43). A recent study showed that PILRa interacting domain in cis engaged its ligands in the same cell (16). The new ligand CLEC4G is expressed in macrophages, it might be able to interact in cis with PILRa. CLEC4G is implicated in Ebola virus entry receptor, as well as clearance and presentation of various Ags (44). Given the complexity of the PILRa receptor/ligand network, PILRa appears to have a diverse function in that it fine tunes the immune response depending on inflammatory cues, similar to other paired receptors (45). Additional studies are required to determine how cis/trans interactions of PILRa ligands are modulated during inflammation where this pathway is upregulated. In the context of inflammatory arthritis, the main effect of PILRa knockout is identified as an enhanced production of proinflammatory cytokines, resulting in more severe disease. At the onset of arthritis, PILRa-knockout mice produced higher levels of proinflammatory cytokines, such as IL-6 and IL-1b, as well as chemokines, such as KC and MCP-1, in joint tissues, even though both WT and Pilra2/2 cellular infiltrates in arthritic joints were similar in WT and Pilra2/2 mice by histopathology. This suggests that, in the absence of PILRa, myeloid cells infiltrating the joints are hyperresponsive, supporting a regulatory role for PILRa in these cells. Although the myeloid cells in joint tissues at the onset of disease are hyperresponsive in terms of proinflammatory cytokine production, we did not detect robust hyperresponsiveness of PILRa-expressing cells in the absence of PILRa in vitro. Pilra2/2 BMDCs produced mildly, but consistently, increased levels of TNF and MCP-1 in response to LPS, whereas WT and Pilra2/2 BMDCs produced similar levels of these cytokines in response to mIgG binding to Protein G beads (Supplemental Fig. 2). For BMDMs, Pilra2/2 mice showed a trend toward increased proinflammatory cytokine and chemokine production in response to LPS (Supplemental Fig. 2). Therefore, the cumulative effect of this augmented response may explain the robust phenotype in the context of experimental arthritis. The inhibitory effect of PILRa receptor mediated by Ab cross-linking in vitro suggests that, when proximal, stimulatory pathways, such as FcgR, can be diminished. PILRb was shown to promote cytokine production in myeloid cells, and PILRb2/2 mice and BMDMs produced fewer proinflammatory cytokines (14). In addition, PILRa is able to regulate neutrophil infiltration during acute inflammation via modulation of integrin activation in an acute inflammatory peritonitis model (16), in which PILRa-knockout mice show increased infiltration of neutrophils upon thioglycollate challenge (16). In contrast, in CAIA we did not see a difference between WT and PILRa-knockout mice with regard to the number of neutrophils and macrophages in arthritic joints. In support of this finding, we did not observe any difference between WT and Pilra2/2 neutrophil and macrophage cellular infiltrates in an acute localized inflammatory model of S. aureus (Supplemental Fig. 1). Despite these differences, both studies suggest that PILRa plays a negative regulatory role in the inflammatory process, likely by complex mechanisms under different inflammatory cues. Our findings suggest that, in the context of sterile inflammation, PILRa is a regulatory receptor that modulates inflammatory cytokine responses; however, we cannot rule out that PILRa might have additional functions outside of cytokine production to affect inflammatory responses. Anti-PILRa treatment reduced inflammatory arthritis. Consistent with this observation, engaging PILRa on BMDCs, which highly express PILRa on their surface, inhibits their function when a stimulating receptor, such as FcgR, is coengaged. Thus,

Downloaded from http://www.jimmunol.org/ at Otterbein University on July 6, 2014

Discussion

869

870 PILRa-knockout and anti-PILRa–agonizing studies complement each other in demonstrating the regulatory effect of this factor in two rodent models of RA. Similarly, engagement of PILRa reduces serum or bronchoalveolar lavage fluid levels of proinflammatory cytokines, including IL-1b, TNF, and IL-6, in a model of S. aureus–induced pneumonia (14). Further studies are required to understand what counter-regulatory receptors might specifically interact with PILRa to modulate its function in myeloid cells and how much this pathway contributes to other inflammatory diseases. The contribution of the PILRa pathway may vary, depending on the type of myeloid cell, inflammation type and stage, stimulation strength, and the ligand involved. Thus, PILRa is emerging as an important regulatory pathway in myeloid cells and, therefore, is an attractive target in arthritis or potentially other inflammatory diseases.

Disclosures References 1. Kuroki, K., A. Furukawa, and K. Maenaka. 2012. Molecular recognition of paired receptors in the immune system. Front. Microbiol. 3: 429. 2. Lanier, L. L. 2001. Face off—the interplay between activating and inhibitory immune receptors. Curr. Opin. Immunol. 13: 326–331. 3. Taylor, L. S., S. P. Paul, and D. W. McVicar. 2000. Paired inhibitory and activating receptor signals. Rev. Immunogenet. 2: 204–219. 4. Fournier, N., L. Chalus, I. Durand, E. Garcia, J. J. Pin, T. Churakova, S. Patel, C. Zlot, D. Gorman, S. Zurawski, et al. 2000. FDF03, a novel inhibitory receptor of the immunoglobulin superfamily, is expressed by human dendritic and myeloid cells. J. Immunol. 165: 1197–1209. 5. Mousseau, D. D., D. Banville, D. L’Abbe´, P. Bouchard, and S. H. Shen. 2000. PILRalpha, a novel immunoreceptor tyrosine-based inhibitory motif-bearing protein, recruits SHP-1 upon tyrosine phosphorylation and is paired with the truncated counterpart PILRbeta. J. Biol. Chem. 275: 4467–4474. 6. Tabata, S., K. Kuroki, J. Wang, M. Kajikawa, I. Shiratori, D. Kohda, H. Arase, and K. Maenaka. 2008. Biophysical characterization of O-glycosylated CD99 recognition by paired Ig-like type 2 receptors. J. Biol. Chem. 283: 8893–8901. 7. Wilson, M. D., J. Cheung, D. W. Martindale, S. W. Scherer, and B. F. Koop. 2006. Comparative analysis of the paired immunoglobulin-like receptor (PILR) locus in six mammalian genomes: duplication, conversion, and the birth of new genes. Physiol. Genomics 27: 201–218. 8. Shiratori, I., K. Ogasawara, T. Saito, L. L. Lanier, and H. Arase. 2004. Activation of natural killer cells and dendritic cells upon recognition of a novel CD99-like ligand by paired immunoglobulin-like type 2 receptor. J. Exp. Med. 199: 525–533. 9. Sun, Y., K. Senger, T. K. Baginski, A. Mazloom, Y. Chinn, H. Pantua, K. Hamidzadeh, S. R. Ramani, E. Luis, I. Tom, et al. 2012. Evolutionarily conserved paired immunoglobulin-like receptor a (PILRa) domain mediates its interaction with diverse sialylated ligands. J. Biol. Chem. 287: 15837–15850. 10. Kogure, A., I. Shiratori, J. Wang, L. L. Lanier, and H. Arase. 2011. PANP is a novel O-glycosylated PILRa ligand expressed in neural tissues. Biochem. Biophys. Res. Commun. 405: 428–433. 11. Satoh, T., J. Arii, T. Suenaga, J. Wang, A. Kogure, J. Uehori, N. Arase, I. Shiratori, S. Tanaka, Y. Kawaguchi, et al. 2008. PILRalpha is a herpes simplex virus-1 entry coreceptor that associates with glycoprotein B. Cell 132: 935–944. 12. Wang, J., Q. Fan, T. Satoh, J. Arii, L. L. Lanier, P. G. Spear, Y. Kawaguchi, and H. Arase. 2009. Binding of herpes simplex virus glycoprotein B (gB) to paired immunoglobulin-like type 2 receptor alpha depends on specific sialylated O-linked glycans on gB. J. Virol. 83: 13042–13045. 13. Fan, Q., and R. Longnecker. 2010. The Ig-like v-type domain of paired Ig-like type 2 receptor alpha is critical for herpes simplex virus type 1-mediated membrane fusion. J. Virol. 84: 8664–8672. 14. Banerjee, A., F. Stevenaert, K. Pande, E. Haghjoo, S. Antonenko, D. M. Gorman, M. Sathe, T. K. McClanahan, R. Pierce, S. P. Turner, et al. 2010. Modulation of paired immunoglobulin-like type 2 receptor signaling alters the host response to Staphylococcus aureus-induced pneumonia. Infect. Immun. 78: 1353–1363. 15. Tato, C. M., B. Joyce-Shaikh, A. Banerjee, Y. Chen, M. Sathe, S. E. Ewald, M. R. Liu, D. Gorman, T. K. McClanahan, J. H. Phillips, et al. 2012. The myeloid receptor PILRb mediates the balance of inflammatory responses through regulation of IL-27 production. PLoS ONE 7: e31680. 16. Wang, J., I. Shiratori, J. Uehori, M. Ikawa, and H. Arase. 2013. Neutrophil infiltration during inflammation is regulated by PILRa via modulation of integrin activation. Nat. Immunol. 14: 34–40. 17. Lee, D. M., and M. E. Weinblatt. 2001. Rheumatoid arthritis. Lancet 358: 903–911. 18. Klareskog, L., A. I. Catrina, and S. Paget. 2009. Rheumatoid arthritis. Lancet 373: 659–672.

19. Gordon, S., and P. R. Taylor. 2005. Monocyte and macrophage heterogeneity. Nat. Rev. Immunol. 5: 953–964. 20. Ne´meth, T., and A. Mo´csai. 2012. The role of neutrophils in autoimmune diseases. Immunol. Lett. 143: 9–19. 21. Kagari, T., H. Doi, and T. Shimozato. 2002. The importance of IL-1 beta and TNF-alpha, and the noninvolvement of IL-6, in the development of monoclonal antibody-induced arthritis. J. Immunol. 169: 1459–1466. 22. Kagari, T., D. Tanaka, H. Doi, and T. Shimozato. 2003. Essential role of Fc gamma receptors in anti-type II collagen antibody-induced arthritis. J. Immunol. 170: 4318–4324. 23. Tanaka, D., T. Kagari, H. Doi, and T. Shimozato. 2006. Essential role of neutrophils in anti-type II collagen antibody and lipopolysaccharide-induced arthritis. Immunology 119: 195–202. 24. Banda, N. K., J. M. Thurman, D. Kraus, A. Wood, M. C. Carroll, W. P. Arend, and V. M. Holers. 2006. Alternative complement pathway activation is essential for inflammation and joint destruction in the passive transfer model of collageninduced arthritis. J. Immunol. 177: 1904–1912. 25. Ditzel, H. J. 2004. The K/BxN mouse: a model of human inflammatory arthritis. Trends Mol. Med. 10: 40–45. 26. Barck, K. H., W. P. Lee, L. J. Diehl, J. Ross, P. Gribling, Y. Zhang, K. Nguyen, N. van Bruggen, S. Hurst, and R. A. Carano. 2004. Quantification of cortical bone loss and repair for therapeutic evaluation in collagen-induced arthritis, by micro-computed tomography and automated image analysis. Arthritis Rheum. 50: 3377–3386. 27. Lutz, M. B., N. Kukutsch, A. L. Ogilvie, S. Ro¨ssner, F. Koch, N. Romani, and G. Schuler. 1999. An advanced culture method for generating large quantities of highly pure dendritic cells from mouse bone marrow. J. Immunol. Methods 223: 77–92. 28. Arnett, F. C., S. M. Edworthy, D. A. Bloch, D. J. McShane, J. F. Fries, N. S. Cooper, L. A. Healey, S. R. Kaplan, M. H. Liang, H. S. Luthra, et al. 1988. The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis. Arthritis Rheum. 31: 315–324. 29. R Core Team. 2012. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing. Vienna, Austria. 30. Gentleman, R. C., V. J. Carey, D. M. Bates, B. Bolstad, M. Dettling, S. Dudoit, B. Ellis, L. Gautier, Y. Ge, J. Gentry, et al. 2004. Bioconductor: open software development for computational biology and bioinformatics. Genome Biol. 5: R80. 31. Gautier, L., L. Cope, B. M. Bolstad, and R. A. Irizarry. 2004. affy—analysis of Affymetrix GeneChip data at the probe level. Bioinformatics 20: 307–315. 32. Bolstad, B. M., R. A. Irizarry, M. Astrand, and T. P. Speed. 2003. A comparison of normalization methods for high density oligonucleotide array data based on variance and bias. Bioinformatics 19: 185–193. 33. Irizarry, R. A., B. Hobbs, F. Collin, Y. D. Beazer-Barclay, K. J. Antonellis, U. Scherf, and T. P. Speed. 2003. Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics 4: 249–264. 34. Smyth, G. K. 2005. Limma: linear models for microarray data. Springer, New York. 35. Cogne´, M., R. Lansford, A. Bottaro, J. Zhang, J. Gorman, F. Young, H. L. Cheng, and F. W. Alt. 1994. A class switch control region at the 39 end of the immunoglobulin heavy chain locus. Cell 77: 737–747. 36. Wipke, B. T., Z. Wang, W. Nagengast, D. E. Reichert, and P. M. Allen. 2004. Staging the initiation of autoantibody-induced arthritis: a critical role for immune complexes. J. Immunol. 172: 7694–7702. 37. Bruhns, P., A. Samuelsson, J. W. Pollard, and J. V. Ravetch. 2003. Colonystimulating factor-1-dependent macrophages are responsible for IVIG protection in antibody-induced autoimmune disease. Immunity 18: 573–581. 38. Ji, H., K. Ohmura, U. Mahmood, D. M. Lee, F. M. Hofhuis, S. A. Boackle, K. Takahashi, V. M. Holers, M. Walport, C. Gerard, et al. 2002. Arthritis critically dependent on innate immune system players. Immunity 16: 157–168. 39. Ji, H., A. Pettit, K. Ohmura, A. Ortiz-Lopez, V. Duchatelle, C. Degott, E. Gravallese, D. Mathis, and C. Benoist. 2002. Critical roles for interleukin 1 and tumor necrosis factor alpha in antibody-induced arthritis. J. Exp. Med. 196: 77–85. 40. Abbas, A. R., D. Baldwin, Y. Ma, W. Ouyang, A. Gurney, F. Martin, S. Fong, M. van Lookeren Campagne, P. Godowski, P. M. Williams, et al. 2005. Immune response in silico (IRIS): immune-specific genes identified from a compendium of microarray expression data. Genes Immun. 6: 319–331. 41. Satoh, T., and H. Arase. 2008. HSV-1 infection through inhibitory receptor, PILRalpha. Uirusu 58: 27–36. 42. Liu, W., L. Tang, G. Zhang, H. Wei, Y. Cui, L. Guo, Z. Gou, X. Chen, D. Jiang, Y. Zhu, et al. 2004. Characterization of a novel C-type lectin-like gene, LSECtin: demonstration of carbohydrate binding and expression in sinusoidal endothelial cells of liver and lymph node. J. Biol. Chem. 279: 18748–18758. 43. Crocker, P. R., J. C. Paulson, and A. Varki. 2007. Siglecs and their roles in the immune system. Nat. Rev. Immunol. 7: 255–266. 44. Dominguez-Soto, A., L. Aragoneses-Fenoll, E. Martin-Gayo, L. Martinez-Prats, M. Colmenares, M. Naranjo-Gomez, F. E. Borras, P. Munoz, M. Zubiaur, M. L. Toribio, et al. 2007. The DC-SIGN-related lectin LSECtin mediates antigen capture and pathogen binding by human myeloid cells. Blood 109: 5337– 5345. 45. Akkaya, M., and A. N. Barclay. 2013. How do pathogens drive the evolution of paired receptors? Eur. J. Immunol. 43: 303–313.

Downloaded from http://www.jimmunol.org/ at Otterbein University on July 6, 2014

The authors have no financial conflicts of interest.

PILRa REGULATES INFLAMMATION