Igg Immunocomplexes Sensitize Human Monocytes for Inflammatory

IgG Immunocomplexes Sensitize Human Monocytes for Inflammatory Hyperactivity via Transcriptomic and Epigenetic Reprogramming in Rheumatoid Arthritis This information is current as of September 28, 2021. Qiao Zhong, Fang-Yuan Gong, Zheng Gong, Sheng-Hao Hua, Ke-Qin Zeng and Xiao-Ming Gao J Immunol published online 30 April 2018 http://www.jimmunol.org/content/early/2018/04/27/jimmun ol.1701756 Downloaded from

Supplementary http://www.jimmunol.org/content/suppl/2018/04/28/jimmunol.170175

Material 6.DCSupplemental http://www.jimmunol.org/

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication by guest on September 28, 2021

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2018 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Published April 30, 2018, doi:10.4049/jimmunol.1701756 The Journal of Immunology

IgG Immunocomplexes Sensitize Human Monocytes for Inﬂammatory Hyperactivity via Transcriptomic and Epigenetic Reprogramming in Rheumatoid Arthritis

Qiao Zhong,*,†,‡ Fang-Yuan Gong,* Zheng Gong,* Sheng-Hao Hua,* Ke-Qin Zeng,*,x and Xiao-Ming Gao*,{,‖

Prevalence of circulating immunocomplexes (ICs) strongly correlates with rheumatoid arthritis (RA) in humans. Deposits of IgG-ICs are abundant in affected joints of patients, yet molecular mechanisms for the pathogenic roles of such ICs are not fully understood. In this study, we present evidence that IgG-ICs precipitated from RA sera sensitized human monocytes for a long- lasting inﬂammatory functional state, characterized by a strong TNF-a response to cellular proteins representing damage-

associated molecular patterns and microbe-derived pathogen-associated molecular patterns. Importantly, plate-coated human Downloaded from IgG (a mimic of deposited IC without Ag restriction) exhibited a similarly robust ability of monocyte sensitization in vitro. The plate-coated human IgG–induced functional programming is accompanied by transcriptomic and epigenetic modification of various inflammatory cytokines and negative regulator genes. Moreover, macrophages freshly isolated from synovia of patients with RA, but not sera-negative arthropathy, displayed a signature gene expression profile highly similar to that of IC-sensitized human monocytes, indicative of historical priming events by IgG-ICs in vivo. Thus, the ability of IgG-ICs to drive sustainable

functional sensitization/reprogramming of monocytes and macrophages toward inﬂammation may render them key players in the http://www.jimmunol.org/ development of RA. The Journal of Immunology, 2018, 200: 000–000.

roduction of autoantibodies is the hallmark of systemic Inflammatory cytokines, predominantly TNF-a, IL-1b, IL-6, autoimmune disorders such as rheumatoid arthritis (RA), a and IL-8, are crucial players in the pathogenesis of RA (11). P chronic inflammatory disease of the joints occurring in Tissue macrophages and infiltrating monocytes are considered the ∼1% of the population worldwide. There is ample evidence sug- main source of inflammatory cytokines in affected joints of RA gesting that such autoantibodies are not merely side products of patients. Human CD14+ monocytes represent ∼5% of the WBCs. overactivation of the adaptive immune system but are important After transient circulation in the bloodstream, these highly plastic contributing factors in the pathogenesis process of the disease cells enter peripheral tissues (e.g., joint synovium) and can dif- by guest on September 28, 2021 (1, 2), particularly in the form of IgG immunocomplexes (ICs) ferentiate into dendritic cells, macrophages, or osteoclasts in a capable of interacting with myeloid cells through receptors for the contact-dependent manner. Both types of inflammatory cytokine– Fc fragment of Ig H chains (FcgRs) (3, 4). It has been shown that producing M1 and anti-inflammatory M2 macrophages are known the incidence and severity of erosive bone damage in RA posi- to be present in inflamed synovia of RA patients (12). Interest- tively correlate with the titers of anticitrullinated peptide Ab and ingly, recent advances indicate that costimulation through FcgRs rheumatoid factor (anti-IgG autoantibody) (5) and also that de- and TLRs could not only elicit inflammatory cytokine production posits of IgG-ICs are abundant in affected joints of RA patients by M2 macrophages but also drive M2–M1 macrophage conver- and in animal models of RA, in which autoantibodies are sion (13). More importantly, functional analysis and gene ex- demonstrably pathogenic (1, 2, 6). It is also evident that ICs pression profiling on freshly isolated synovial macrophages from containing RA-related IgG autoantibodies against vimentin, fi- RA patients revealed an altered biologic response to IL-10, with brinogen, or lactoferrin (LTF) are potent activators of myeloid attenuation of its anti-inflammatory function and a concomitant cells via simultaneous ligation of FcgRs and TLRs (2, 7–10). retention of IFN-g–like activating functions (14). Despite these

*Institute of Biology and Medical Sciences, School of Biology and Basic Medical The online version of this article contains supplemental material. Sciences, Soochow University, Suzhou 215123, China; †Department of Laboratory Abbreviations used in this article: ChIP, chromatin immunoprecipitation; cIgG, plate- Medicine, The Affiliated Suzhou Hospital of Nanjing Medical University, Suzhou x coated human IgG; CRT, calreticulin; DAMP, damage-associated molecular pattern; 215002, China; ‡Suzhou Municipal Hospital, Suzhou 215002, China; Department of DEG, differentially expressed gene; FDR, false discovery rate; GO, gene ontology; Rheumatology, The First Affiliated Hospital of Soochow University, Suzhou 215001, { GSEA, gene set enrichment analysis; HC-IC, IC precipitated from serum of healthy China; Jiangsu Key Laboratory of Infection and Immunity, Suzhou 215123, China; ‖ control; HMT, histone methyltransferase; IC, immunocomplex; LTF, lactoferrin; and Key Laboratory of Systemic Biomedical Study, Suzhou 215123, China LTF-IC, LTF-containing IC; Mo(cIgG), human monocyte primed with cIgG for ORCID: 0000-0002-6381-5632 (K.-Q.Z.). 24 h; Mo(cIgG/LPS), human monocyte primed with cIgG for 24 h followed by a 2 h LPS (10 ng/ml) stimulation; Mo(RPMI), human monocyte cultured in R10 Received for publication December 19, 2017. Accepted for publication April 4, 2018. medium alone for 24 h; Mo(RPMI/LPS), human monocyte cultured in R10 medium This work was supported by grants from the Ministry of Science and Technology alone for 24 h followed by a 2 h LPS (10 ng/ml) stimulation; NES, normalized (2017YFA0104502), the Ministry of Education (IRT1075), and the Natural Science enrichment score; PAMP, pathogen-associated molecular pattern; RA, rheumatoid ar- Foundation of China (31570868/31770942). thritis; RA-IC, IC precipitated from serum of RA patient; SF, synovial fluid; SpA, spondyloarthropathy; TSS, transcriptional start site. The sequences presented in this article have been submitted to the Gene Expression Omnibus (https://www.ncbi.nlm.nih.gov/geo/) under accession number GSE102728. Copyright Ó 2018 by The American Association of Immunologists, Inc. 0022-1767/18/$35.00 Address correspondence and reprint requests to Dr. Xiao-Ming Gao, School of Bi- ology and Basic Medical Sciences, Soochow University, 199 Ren-Ai Road, Suzhou 215123, China. E-mail address: [email protected]

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1701756 2 FUNCTIONAL REPROGRAMMING OF HUMAN MONOCYTES BY IgG-ICs important progresses of the recent decade, however, molecular Reagents and Abs mechanisms for the arthritogenic roles of RA-related ICs via LPS (Escherichia coli 0111:B4), zymosan (Z4250), human LTF, human monocyte/macrophage activation are still not fully understood. It IgM (no. I8260), IgA (no. I4036), and polymyxin B-agarose (no. P1411) has recently been shown that human monocytes and macrophages were from Sigma-Aldrich. Pam3CSK4 (no. tlrl-pms) and R406 were from can be trained, by exposure to Candida albicans or b-glucans, to InvivoGen. Human IVIG was from Sinopharm. Mouse IgG1 (ET901), exhibit a long-lasting enhanced functional state with a stronger IgG2b (MPC-11), mAbs against human FcgRI (CD64, 10.1) or FcgRIII (CD16, 3G8), and APC-labeled mAbs against human CD210 (3F9) were ability in proinflammatory cytokine production, phagocytosis, and from BioLegend. Mouse IgG2a (no. 16-4724-82) and mIgG3 (no. 14-4742-81) anti-infection defense, a phenomenon termed “trained immunity” were from eBioscience. Anti-human FcgRIIa mAb (CD32a, IV.3) was (15, 16). However, the question of whether deposited IgG-ICs from STEMCELL Technologies. Recombinant human IL-10 (no. 200-10) could prime monocytes/macrophages for sustainable trained was from PeproTech. Endotoxin-free recombinant human HMGB1 (no. HM-120) was from HMGBiotech. Fibrinogen prepared from human (heightened) inflammatory immunity has not been addressed. plasma (no. 341576-100MG) was from EMD Millipore. Mouse mAb In the current study, we demonstrate that exposure to circulating against human LTF (M860) and recombinant calreticulin (CRT) were IgG-ICs precipitated from sera of RA patients sensitized human prepared in house and depleted of endotoxin contamination by passing blood monocytes for an inflammatory functional state, character- through polymyxin B-agarose columns as previously described (8, 18, 19). ized by a strong TNF-a response to a variety of innate stimuli, LPS levels in these agents were confirmed to be below 0.058 ng/ml by the highly sensitive LPS ELISA kit (HEB526Ge; Cloud-Clone). including damage-associated molecular patterns (DAMPs) and pathogen-associated molecular patterns (PAMPs). This is sup- RT-PCR and ELISA ported by further findings that plate-coated human IgG (cIgG) RNA extraction was performed using Total RNA Isolation Kit II (V) and alone robustly primed human monocytes for long-lasting hyper- cDNA synthesis using First Strand cDNA Synthesis Kit (Fermentas). Downloaded from activity with significantly reduced phagocytosis ability. Tran- Quantitative RT-PCR was performed on StepOnePlus Real-Time PCR scriptomic analysis on the cIgG-sensitized monocytes [Mo(cIgG)], System (Applied Biosystems) using SYBR Green PCR Master Mix (Roche before and after LPS challenge, revealed characteristic changes in Life Science). mRNA levels were normalized to housekeeping gene expression, and folds were calculated and compared with human monocytes global gene expression profile as well as epigenetic modification cultured in R10 medium alone for 24 h [Mo(RPMI)]. ELISA kits for de- on various inflammatory cytokine and negative regulator genes. tection of TNF-a, IL-10, and SPP1 (osteopontin) were purchased from

Importantly, synovial macrophages from RA synovia displayed a eBioscience. Standard curves were established using recombinant human http://www.jimmunol.org/ signature gene expression proﬁle similar to that of cIgG-sensitized cytokines, and the assay detection limit was 7.8 pg/ml for TNF-a, 2 pg/ml monocytes. We conclude that IC-mediated sustainable functional for IL-10, and 0.26 ng/ml for SPP1. sensitization/reprogramming of human monocytes/macrophages RNA sequencing toward inﬂammation is a major pathway for IgG-ICs to contribute to RA pathogenesis in humans. Total cellular RNA was extracted from cells, poly(A) enriched, fragmented, and converted into a library of Illumina-compatible sequencing templates with Illumina mRNA sequencing sample preparation kit according to the manufacturer’s instructions. Cluster formation, primer Materials and Methods hybridization, and single-end sequencing on HiSeq 2000 were carried

Patient sample collection and IC preparation out by Beijing Genomics Institute in Wuhan, China. Quantification of by guest on September 28, 2021 gene expression was performed using RSEM (20). Differential expres- Human blood specimens were obtained from age- and gender-matched sion was compared between two biological groups via EBSeq (21), with subjects with RA (n = 37) and from healthy subjects (n = 32) at the |log2(fold changes)| $1, false discovery rate (FDR) controlled at 0.05, First Affiliated Hospital of Soochow University. Patient information is and posterior possibility of being equivalent expression ,0.05. Func- listed in Supplemental Table I. For ICs preparation, serum samples were tional enrichment analysis of the differentially expressed genes (DEGs) thawed and immediately mixed with an equal volume of ice-cold 5% PEG was performed with BiNGO Cytoscape plugin (22). 6000 and incubated overnight at 4˚C as described by Elshafie et al. (17). The following day, serum samples were diluted three times with 2.5% PEG Chromatin immunoprecipitation sequencing 6000 in RPMI 1640 with phenol red. The diluted sera were layered on top of 2.5% PEG 6000 supplemented with 5% human serum albumin and For chromatin immunoprecipitation (ChIP) sequencing, cells were cross- centrifuged at 2100 3 g at 4˚C for 20 min. Supernatants were subsequently linked with 1% formaldehyde (Sigma-Aldrich) at room temperature for 3 removed and pellets dissolved in PBS into the initial serum volume. 10 min. Chromatin was fragmented to 200–500 bp with Bioruptor (20 30 s Synovial fluid (SF) was collected from six RA patients. Neutrophils (1.2– cycles, low power at 4˚C; Diagenode), cleared at 14,000 rpm at 4˚C for 20 9.7 3 109/l, mean 3.9 6 3.1 3 109/l), monocytes/macrophages (2.4–8.4 3 min and immunoprecipitated with 1 mgofaH3K4me3 Ab (no. 07-473; 109/l, mean 4.5 6 2.1 3 109/l), and lymphocytes (0.7–3.6 3 109/l, mean Merck Millipore) or aH3K27ac Ab (no. 07-449; Merck Millipore) and 40 ml 2.4 6 1.2 3 109/l) in the samples were counted using an automatic protein A/G plus Magnetic Beads (Merck Millipore). Then, the chromatin cell counter (XE5100; Sysmex). Protein concentration in the SF samples immunoprecipitated DNA was purified with MinElute PCR Purification Kits (46–58 mg/ml, mean 50 6 2.5 mg/ml) was quantitated using the bicin- (Qiagen). The libraries were prepared and sequenced with Illumina HiSeq choninic acid assay method (Thermo Scientific Pierce). TNF-a concen- 2000 platform. High-quality reads were uniquely mapped to human refer- tration in the RA-SF samples (1.5–4.7 ng/ml, mean 2.7 6 1.2 ng/ml) was ence sequence (GRCh37/hg19) using Bowtie2 within two mismatches (23). The 6 quantitated using ELISA kits. Cells and debrides in the pooled RA-SF H3K4me3 enrichment around transcriptional start sites (TSSs) (TSS 2kb) were depleted by centrifugation prior to use. was compared and visualized via ngs.plot on the selective genes (24). ChIP sequencing peaks were visualized by uploading BED files into the University of California, Santa Cruz genome browser as custom tracks (25). Human monocyte isolation and flow cytometry Statistical analysis Ficoll-Paque (Sigma-Aldrich) was used to isolate PBMCs from healthy volunteers by differential centrifugation. Monocytes were sorted from PBMCs All statistical analyses were performed with Prism software (GraphPad Soft- using human CD14 magnetic labeled beads (MACS; Miltenyi Biotec) ware). Student t test was used for comparison between two groups. Wilcoxon according to the manufacturer’s instructions. The resultant cells were over paired rank test was used for comparison of paired samples. A p value ,0.05 95% CD14 positive and viable, as evidenced by flow cytometric analysis and was considered statistically significant. Gene set enrichment analysis (GSEA) microscopic observation. Purified monocytes were cultured in R10 medium: was performed to determine whether the up- and downregulated gene sets in RPMI 1640 medium supplemented with 10% (v/v) autologous serum, human monocytes primed with cIgG for 24 h [Mo(cIgG)] are also enriched in 25 penicillin/streptomycin (100 U/ml), L-glutamine (2 mM), and 2-ME (5 3 10 M) samples of RA and seronegative spondyloarthropathy (SpA) synovial mac- at 37˚C and 5% CO2. Monocytes were stained with fluorescence-labeled rophages deposited by Antoniv and Ivashkiv (14) (GSE10500). IFN-g up- and Abs in PBS containing 1% BSA and 0.1% sodium azide. After washout, downregulated gene sets were retrieved from Molecular Signatures Database the fluorescence was assessed by flow cytometry (FACSCanto II; BD (GSE11864). Normalized enrichment score (NES) and FDR q values were Biosciences), and data were analyzed with FlowJo software (FlowJo). Molecular Signatures Database–calculated according to the GSEA statistics (26). The Journal of Immunology 3

Study approval This study was approved by the Ethics Committee of Soochow Uni- versity Medical School, Suzhou, China. The methods were carried out in accordance with the guidelines of Soochow University. Written informed consent was obtained from all participants prior to inclusion in the study.

Results Monocyte sensitization by RA-related circulating ICs To assess if RA-related ICs could program human monocytes for a sustainable proinflammatory phenotype, ICs precipitated from sera of RA patients (RA-ICs) and ICs precipitated from sera of healthy controls (HC-ICs) were compared for the ability to sensitize CD14+ peripheral blood monocytes for hyperresponsiveness in vitro. Although treatment with RA-ICs or HC-ICs alone induced little or no detectable TNF-a production by human monocytes, the cells pretreated with RA-ICs, but not HC-ICs, became significantly more responsive in terms of TNF-a production to cellular proteins released from necrotic cells that are Downloaded from known to be a major source of DAMPs responsible for triggering inflammatory responses in vivo (Fig. 1A). SF from RA patients FIGURE 1. Human monocyte sensitization by RA-related ICs. Freshly has been shown to contain necrotic proteins capable of activating prepared CD14+ blood monocytes were primed with IgG-ICs PEG precipi- monocytes/macrophages (27). Interestingly, RA-IC–primed hu- tated from sera of healthy controls (HC) (n = 32) or RA patients (n =37) man monocytes were highly responsive to stimulation of RA-SF for 24 h. After washes, the cells were restimulated for 24 h with or without soluble fractions of necrotic protein (1 mg/ml) or RA-SF (1 mg/ml) (A), (equal proportion mixture of SF samples from six RA patients) http://www.jimmunol.org/ followed by ELISA quantitation of TNF-a in the culture supernatant. (B) in vitro (Fig. 1A). Moreover, these sensitized cells were similarly Freshly prepared monocytes were primed with IgG-ICs from sera of HCs hyperresponsive to individual proteins representing DAMPs (fi- (n = 12) or RA patients (n = 12) for 24 h. After washes, the cells were brinogen from human plasma, recombinant human CRT, and restimulated for 24 h with CRT (1 mg/ml), HMGB1 (1 mg/ml), fibrinogen HMGB1) as well as a variety of pathogen-derived PAMPs, in- (10 mg/ml), LPS (10 ng/ml), Pam3CSK4 (10 mg/ml), or zymosan (100 ng/ml). cluding LPS, zymosan, and Pam3CSK4 (Fig. 1B). Each dot represents one donor. (C) Monocytes were primed with LTF-IC, It is also of interest to note that freshly isolated circulating OVA-containing IC (OVA-IC), BSA-containing ICs (BSA-IC), or a com- monocytes from RA patients were more responsive than those from bination of the Ag- and isotype-matched control Ab pairs (Ag/Ab, 10:10 HCs in terms of TNF-a production following stimulation by innate mg/ml) for 24 h. After washes, the cells were restimulated for 24 h with or stimuli (data not shown), which is in agreement with previously without HMGB1 (1 mg/ml), or LPS (10 ng/ml), or RA-SF (1 mg/ml). Data by guest on September 28, 2021 6 published works comparing the responsiveness of blood monocytes (mean SEM) are representative of three independent experiments with three biological replicates.*p , 0.05, **p , 0.01, ***p , 0.001. (or mononuclear cells) from patients with RA or systemic lupus erythematosus and HCs (28–30). The enhanced functional state of circulating monocytes from such patients could, at least partially, be the result of priming/sensitization by circulating ICs. We have re- Freshly fractionated human monocytes from healthy donors were cently reported that prevalence of anti-LTF IgG autoantibodies and stimulated with or without (RPMI), cIgG, cIgM, cIgA, or LPS for LTF-containing ICs (LTF-IC) positively correlate with RA in hu- 24 h, followed by quantitation of cytokines in the culture super- mans and also that ICs between human LTF and specific IgG au- natant and assessment of responsiveness of the cells to subsequent toantibodies from RA patients or mouse anti-human LTF mAb stimuli. Unlike LPS, which strongly triggered the release of TNF-a (M860, of IgG1 subclass) are able to elicit strong TNF-a and IL-1b and IL-10 by human monocytes, cIgGs induced no detectable responses in human monocytes/macrophages through simultaneous immediate cytokine responses (Fig. 2A). Expectedly, LPS toler- ligation of TLR4 and FcgRIIa (8). In this study, we further dem- ance occurred in the LPS-stimulated monocytes, as evidenced by onstrate that an LTF-IC was also able to drive human monocytes their attenuated TNF-a synthesis in subsequent functional assays into a sustainable hyperactive functional state, whereas ICs between (Fig. 2B). By contrast, however, Mo(cIgG), but not human chicken OVA or BSA with respectively specific mouse mAbs of monocytes primed with cIgM or cIgA, became markedly more IgG1 subtype were unable to do the same (Fig. 1C). These results responsive to LPS in terms of TNF-a production compared imply that RA-related ICs may be able to facilitate proinflammatory with Mo(RPMI) (Fig. 2B). Mo(cIgG) was also hyperresponsive to cytokine productions by licensing inflammatory hyperactivity to RA-SF, cellular proteins from necrotic cells, as well as individual monocytes/macrophages in vivo. protein preparations of human CRT, HMGB1, and fibrinogen (Fig. 2C). Strikingly, however, Mo(cIgG), but not any of the Sustained hyperactivity of human monocytes after control groups, lost completely the ability to produce IL-10 upon cIgG priming further stimulation (Fig. 2B, 2D). More importantly, the proin- Heterogeneity of ICs precipitated from RA sera was of consid- flammatory functional state of Mo(cIgG) was long lasting; after erable concern when conducting detailed investigation on mo- a 5–7 d resting period following cIgG priming, the cells remained lecular mechanisms of IC sensitization of myeloid cells. LTF-IC, or hyperresponsive in terms of TNF-a production and virtually any other TLR agonist autoantigen-containing ICs, was not an ideal unresponsive in IL-10 production compared with Mo(RPMI) substitute either because coligation of TLRs and FcgRs would (Fig. 2E). complicate IC-induced signaling in the cell. Therefore, we cIgG was very robust in training human monocytes for employed cIgG as a mimic of deposited IgG-ICs without Ag re- hyperresponsiveness. Even at pictogram range, human IgG in strictions in the following functional and transcriptomic studies. plate-bound form was able to sensitize human monocytes for 4 FUNCTIONAL REPROGRAMMING OF HUMAN MONOCYTES BY IgG-ICs Downloaded from http://www.jimmunol.org/ FIGURE 2. Enhanced responsiveness of cIgG-primed monocytes to DAMPs and PAMPs. (A) Freshly isolated CD14+ human blood monocytes were cultured for 24 h in R10 medium in microtiter tissue culture plates precoated with 10 mg/ml IVIG (cIgG), human IgM (cIgM), or IgA (cIgA). Control cells were cultured in R10 medium alone (RPMI) or LPS (1 mg/ml). TNF-a and IL-10 in the culture supernatant were quantitated by ELISAs. (B) After cIgG priming, monocytes were detached, washed, and then stimulated for 24 h with, or without (Medium), LPS (10 ng/ml). (C) Mo(cIgG) and Mo(RPMI) were stimulated for 24 h with CRT (1 mg/ml), HMGB1 (1 mg/ml), fibrinogen (10 mg/ml), RA-SF (1 mg/ml), or necrotic proteins (1 mg/ml). (D) Kinetics of TNF-a and IL-10 production by Mo(cIgG) and Mo(RPMI) after LPS (10 ng/ml) stimulation. (E) Washed Mo(RPMI) and Mo(cIgG) were allowed to rest in R10 medium for 1, 3, 5, or 7 d and then stimulated with LPS (10 ng/ml) for 24 h. Each pair of dots represents one donor (n = 5). Data (mean 6 SEM) are representative of at least three independent experiments of three biological replicates. *p , 0.05, **p , 0.01, ***p , 0.001. by guest on September 28, 2021 sustained hyperactivity, whereas soluble IgG (as much as 100 mg/ (Fig. 4A). mAb against human CD32 (FcgRII), but not that ml) was unable to do so (Fig. 3A). Besides TNF-a, IgG-IC–sen- against CD16 or CD64, almost completely reversed the cIgG- or sitized human monocytes were also hyperactive in IL-6 and IL-8 LTF-IC–mediated monocyte sensitization (Fig. 4B, 4C). Of the production upon LPS stimulation, as evidenced by ELISA quan- FcgRs expressed by human myeloid cells, FcgRIIa is the only one titation of these cytokines in the culture supernatant (data not possessing an ITAM in its cytoplasmic domain, which could ac- shown), which was further confirmed by quantitative PCR results tivate signaling cascades via SRC family kinase and Syk (3, 4). showing augmented expression of mRNA for TNF-a, IL-1b, IL-6, Consistently, the Syk-specific inhibitor R406 (31) dose-dependently IL-8, and IL-1RN in Mo(cIgG) following LPS stimulation blocked cIgG-mediated human monocyte priming (Fig. 4D). (Fig. 3B). Note that IL-10 mRNA expression in Mo(cIgG), but not Together, these results suggest that IgG-ICs sensitize human Mo(RPMI), was significantly inhibited in parallel experiments. monocytes for functional hyperactivity through crosslinking cell Although IL-1b was undetectable in the culture supernatant of surface FcgRIIa (CD32a) and Syk-dependent activation signaling. LPS-stimulated Mo(cIgG), intracellular synthesis of the protein Transcriptomic profiling of cIgG-sensitized monocytes was observed by flow cytometric analysis (Fig. 3C). FACS analysis also revealed strong suppression of membrane-bound CD14 To gain further molecular insight of IgG-IC sensitization of human (LPS coreceptor) expression in IC-sensitized monocytes monocytes, we performed transcriptomic analysis on Mo(cIgG) (Fig. 3D), which was confirmed in additional flow cytometry before and after LPS challenge. Freshly fractionated monocytes experiments by saturating FcgRs and fixing the cells (data not from three unrelated healthy donors were cultured for 24 h in R10 shown). Moreover, the ability of Mo(cIgG) to engulf pathogen- medium for preparation of Mo(RPMI) and Mo(cIgG). Half of the derived particles was significantly inhibited (Fig. 3E). To- cultures were stimulated with LPS for 2 h, resulting in additional gether, these results mark functional deviation away from groups of human monocytes cultured in R10 medium alone for 24 h defense and scavenge. Note that LPS stimulation of human followed by a 2 h LPS (10 ng/ml) stimulation [Mo(RPMI/LPS)] and monocytesresultedinenhancedCD14expressionanden- human monocytes primed with cIgG for 24 h followed by a 2 h LPS hanced phagocytosis ability in the same experiments. (10 ng/ml) stimulation [Mo(cIgG/LPS)], respectively. Cells of the four groups from each donor were individually processed for RNA IC-mediated monocyte sensitization is dependent on extraction, and the resultant RNA was subjected to genome-wide FcgRIIa-Syk transcriptomic analysis by high-throughput RNA sequencing Consistent with earlier observations that Fc fragments of mouse (Fig. 5A). The transcriptomic profiles thus obtained were sub- IgG1 and IgG2a bound human FcgRIIa with relatively high affinity jected to unbiased principal component analysis, which showed a (3), mouse IgG of subclasses IgG1 and IgG2a, but not IgG2b and clear segregation of paired groups, indicative of their distinct gene IgG3, were as effective as human IVIG in monocyte priming expression profiles (Fig. 5B). The Journal of Immunology 5 Downloaded from

FIGURE 3. Inflammatory cytokine expression and phagocytosis of cIgG-sensitized monocytes. (A) Freshly isolated CD14+ human blood monocytes were cultured in triplicated wells either precoated with IVIG of increasing concentrations (0–100 mg/ml) or supplemented with soluble IVIG of the same concentration range in R10 medium for 24 h. After detaches and washes, the cultured monocytes were restimulated for 24 h with, or without (Medium), LPS (10 ng/ml), and TNF-a concentration in the culture supernatant was determined by ELISAs. (B) Mo(cIgG) and Mo(RPMI) were stimulated with LPS for 0, 2, 4, and 6 h, and the mRNA expression of TNF-a, IL-6, IL-8, IL-10, IL-1b, and IL1RN was analyzed by RT-PCR. The results are expressed as http://www.jimmunol.org/ relative expression to HPRT. (C) Mo(cIgG) and Mo(RPMI) were stimulated with LPS for 24 h followed by intracellular staining of IL-1b. Cells stained with isotype control Abs (filled histogram) were included as negative controls. (D) Freshly isolated CD14+ human blood monocytes were primed for 24 h with, or without (RPMI), either cIgG or LPS in R10 medium followed by staining with fluorescence-labeled Abs against human CD14 and FACS analysis. Flow cytometric plots are representative of six independent experiments. (E) Human blood monocytes were cultured for 24 h in R10 medium in plates precoated with 10 mg/ml IVIG (cIgG). Cells treated with, or without (RPMI), LPS (1 mg/ml) for 24 h were included as controls. After washes, opsonized FITC- zymosan was added to the cultures (final concentration 10 mg/ml) and incubated for 1 h, followed by flow cytometric quantitation of FITC-positive cells for assessment of phagocytosis percentage. Data (mean 6 SEM) are from three independent experiments. *p , 0.05, **p , 0.01, ***p , 0.001, unpaired, two-tailed Student t test. by guest on September 28, 2021 The DEGs between Mo(cIgG) and Mo(RPMI) were identified growth factor supporting cell proliferation and differentiation and subjected to gene ontology (GO) enrichment analysis, which but also an important inflammatory cytokine capable of aug- was visualized using BiNGO Cytoscape plugin (Fig. 5C). Down- menting TLR agonist responsiveness of human and murine regulated DEGs were mainly clustered in categories of regulation macrophages (37, 38). The upregulated cytokine genes also of immune response, apoptosis, phagocytosis, and cell activation. included TNFSF14 (LIGHT) and SPP1 (osteopontin) that are Downregulation of the apoptosis genes underlies the long survival known to promote osteoclastogenesis and bone resorption in time of Mo(cIgG). Suppression of the immune regulation response inflammatory arthritis (39, 40). SPP1 has also been shown to be genes signifies less stringent control on the inflammatory products. critical for the induction of pathological Th17 response in RA Inhibition of the phagocytosis genes is consistent with the reduced (41). Production of SPP1 by Mo(cIgG) was confirmed by phagocytosis ability of the cIgG-primed cells (Fig. 3E). The distinct ELISAs using blood monocyte samples from six unrelated clusters of genes significantly upregulated in Mo(cIgG) include donors (Fig. 5E). chemotaxis, angiogenesis, cell adhesion molecules, regulation of actin cytoskeleton organization, signaling and kinase regulation, Transcriptomic characterization of the Mo(cIgG) response and histone modification enzymes, which are related to immune cell to LPS infiltration, edema, and swelling under inflammatory lesions (32), For transcriptomic characterization of Mo(cIgG) response to innate cell differentiation, as well as epigenetic modification (33, 34). stimuli, a set of DEGs comparing Mo(cIgG/LPS) and Mo(RPMI/ Strongly upregulated or downregulated genes in Mo(cIgG) are LPS) was generated and visualized using BiNGO Cytoscape plugin also shown in Fig. 5D. Strong inhibition of the CD14 gene (Fig. 6A). Remarkably, a strong cluster of regulation on inflam- (Fig. 5D) further underlies the loss of CD14 (LPS coreceptor) matory response (GO cluster 14 in Fig. 6A) consisting of both up- expression in Mo(cIgG), as shown in Fig. 3D. Inhibition of the and downregulated genes was observed, reflecting a distinctive TLR4 genes for chemokine receptors CCR2 and CXCR4 was also evi- (LPS) response pattern of Mo(cIgG) compared with Mo(RPMI). As dent. CCR2 is a receptor for monocyte chemotactic proteins, illustrated in Fig. 6B, these genes can be further hierarchically whereas CXCR4 is a 7-transmembrane–domain G protein–coupled grouped into the following three subclusters: 1) inherited genes from receptor capable of binding CXCL12 (stromal cell–derived factor 1), cIgG priming, 2) enhanced LPS-inducible genes, and 3) suppressed which is expressed by various tissue cells, including fibroblast-like LPS-inducible genes. The inherited subcluster includes cIgG- synoviocytes (35). Interestingly, downmodulation of the CXCR1, upregulated inflammatory genes (e.g., CSF1, IL-1RN, SPP1, CXCR2, and CXCR4 genes in SF monocytes of RA patients has CCL22) that were virtually unaffected by subsequent LPS been documented (36). stimulation, whereas the enhanced subcluster contains inflam- Of the cytokine genes, only a few were significantly en- matory genes (e.g., TNF-a, IL-6, IL-23, CCL1) not fully acti- hanced, with CSF1 as the most notable. CSF1 is not only a vated in Mo(cIgG) but strongly induced by LPS. The suppressed 6 FUNCTIONAL REPROGRAMMING OF HUMAN MONOCYTES BY IgG-ICs

analysis for selected cytokine genes. As illustrated in Fig. 7C and Supplemental Fig. 1, Mo(cIgG) displayed significant enrichment of H3K4me3 modifications on the promoter regions of inflammatory cytokine genes (TNF-a, IL-6, IL-8, and IL-1RN), whereas genes for negative regulators such as NFKBIA, PTGER2, PTGER4, and PTGIR exhibited significantly diminished H3K4me3 modification. There was no significant H3K27ac enrichment on the genes for TNF-a and IL-6, except IL-8 and IL-1RN (Fig. 7D). H3K4me3 modification around TSSs strongly correlates with gene transcription activity. To further verify the correlation between inflammatory hyperreactivity and histone methylation in Mo(cIgG), a genome landscape ChIP sequencing was performed. Within the cluster of regulation of inflammatory response genes in Mo(cIgG/LPS) shown in Fig. 6B, the subclusters of enhanced and inherited, but not the suppressed, showed clear H3K4me3 enrichment around the TSSs (Fig. 7E). Collectively, the gain and loss of H3K4me3 modifications on inflammatory and inflammation regulatory genes, respectively,

may partially account for IgG-IC–mediated programming of Downloaded from monocytes for more active transcription for inﬂammatory mediators and less supply of negative regulators and also underline the fact that the functional phenotype induced by IC sensitization is very stable.

A Resemblance of transcriptomic signature of RA synovial FIGURE 4. FcgRIIa-Syk signaling in cIgG-sensitized monocytes. ( ) http://www.jimmunol.org/ macrophages with cIgG-sensitized monocytes Human monocytes were cultured for 24 h in R10 medium in culture plates precoated with mouse IgG1, IgG2a, IgG2b, or IgG3 at 10 mg/ml. RA synovial macrophages, mostly derived from infiltrating blood Control cells were cultured in uncoated plates with medium alone monocytes, play a pivotal role in RA pathogenesis and exhibit (RPMI). Human monocytes were treated for 24 h with either plate-coated hyperresponsiveness to TLR stimulation by producing a vast B C IVIG (cIgG) ( )orLTF-IC(10mg/ml) ( ) in the presence or absence of amount of inflammatory cytokines (51). It was of interest to see m mouse mAb (10 g/ml) against human CD16, CD32, or CD64. Irrelevant whether such hyperactive macrophages from RA synovia have mouse Abs of IgG1 (mIgG1) and IgG2b (mIgG2b) subclasses were in- a transcriptomic signature similar to that of cIgG-sensitized cluded as isotype controls. (D) Human monocytes were treated for 24 h with, or without (RPMI), cIgG in the presence of graded concentrations monocytes. Using Affymetrix microarrays and RT-PCR tech- of R406. TNF-a concentration in the culture supernatant was determined niques, Antoniv and Ivashkiv (14) performed gene expression by guest on September 28, 2021 by ELISAs. Data (mean 6 SEM) are representative of three independent profile analysis on freshly isolated synovial macrophages from experiments with three biological replicates. **p , 0.01, ***p , 0.001, patients with RA or SpA and observed a strong IFN-g signature unpaired, two-tailed Student t test. (the expression of 277 IFN-g–inducible genes was elevated in these cells relative to control macrophages). These data, subcluster mainly consists of negative regulator genes that were available in the Gene Expression Omnibus database significantly suppressed during Mo(cIgG) response to LPS as (GSE10500), allowed us to perform GSEA for gene expression compared with Mo(RPMI), such as NF-kB activation feedback profile comparison between Mo(cIgG) and synovial macro- regulators [FOXO3 (42), CITED2 (43), NR4A3 (44), and phages from patients with RA or SpA. Interestingly, the gene ADORA2B (45)], IFN stimulation inhibition (IFNB1) (46), re- expression profile of RA synovial macrophages showed a robust ceptors for PGE2 (PTGER4), and IL-10 (IL-10RA). This sup- correlation with the upregulated (NES: 1.633, FDR q =0.014) 2 pressedactivationofsuchregulatory genes is at least partially and downregulated (NES: 1.78, FDR q = 0.013) genes in Mo responsible for the hyperresponsiveness of cIgG-sensitized (cIgG) (Fig. 8A), which was much stronger than the correlation monocytes. For instance, a significant inhibition on the between the signature genes in synovial macrophages and IFN- PTGER4 and IL-10R genes implies attenuated PGE2- and IL-10– g–inducible genes in human macrophages (upregulated, NES: 2 mediated inflammation control (14, 47, 48). Reduction of IL-10R 0.222, FDR q = 0.520; downregulated, NES: 0.274, FDR q = (CD210) expression in Mo(cIgG), confirmed by flow cytometric 0.129). Fig. 8B lists the major enriched signature genes of RA analysis (Fig. 6C), led to resistance to IL-10–mediated suppression synovial macrophages in common with those of Mo(cIgG), on TNF-a production by the cells (Fig. 6D). including 1) inflammatory mediators CSF1, IL-8, IL-1RN, CCL2, CCL3, CCL7, TNFSF14, and SPP1; 2) MAPK signal- H3K4me3 modification correlates with hyperactivity of ing and AP-1 (FOSB and FOSL1); 3) cell differentiation–re- cIgG-sensitized monocytes lated transcription factors (C2H2, bZIP, and EGR) and ETS2; 4) Histone methylation and acetylation are major epigenetic modi- histone modification enzymes (SETD3, DOT1L, and KDM6B); fications pivotal to regulation of inflammatory gene transcription and 5) as in Mo(cIgG), the IL-10R, CXCR2, CD14, and HLA- (49). In general, H3K4me3 marks promoters (15), whereas DMB genes were strongly suppressed in RA synovial macro- H3K27ac marks active promoters and enhancers (50). Interest- phages. ingly, cIgG sensitization of monocytes led to increased expression of histone methyltransferase (HMT) genes and was susceptible to Discussion inhibition of the HMT-specific inhibitor 59-methylthioadenosine In the current study, after exposure to RA-ICs, LTF-IC, or cIgG, (Fig. 7A, 7B). We next performed ChIP assays targeting H3K4me3 human monocytes adopted a highly enhanced inflammatory and H3K27ac modifications in Mo(cIgG), followed by ChIP-PCR functional state capable of making exacerbated inflammatory The Journal of Immunology 7 Downloaded from http://www.jimmunol.org/ by guest on September 28, 2021

FIGURE 5. Transcriptomic profile of cIgG-sensitized monocytes. (A) Schematic representation of sample preparation for RNA sequencing using blood samples from three unrelated healthy donors. (B) Principal component analysis on the transcriptome data sets from Mo(RPMI), Mo(cIgG), Mo(cIgG/LPS), and Mo(RPMI/LPS). (C) GO enrichMap of DEGs between Mo(cIgG) and Mo(RPMI) were generated by BiNGO Cytoscape plugin. Enriched GO terms as well as gene numbers from the upregulated and downregulated DEGs are indicated. (D) Heatmap showing signature genes expressed in Mo(cIgG) under the categories as indicated. Reads per kilobase per million values were normalized among the groups. (E) Human blood monocytes were treated with cIgG for 24 h, and culture supernatant was harvested for ELISA quantitation of SPP1. Blood samples from six donors were individually assayed; each pair of dots represents one donor. **p , 0.01, Wilcoxon paired rank test. cytokine responses to subsequent stimuli of DAMPs or PAMPs. from certain TLR agonist–containing ICs of soluble form that The sustainable inflammatory phenotype of IC-primed monocytes exhibit monocyte/macrophage–activating properties (7–10), soluble is licensed by transcriptomic and epigenetic modifications. Apart IgG-ICs in general are not comparable to deposited ICs in 8 FUNCTIONAL REPROGRAMMING OF HUMAN MONOCYTES BY IgG-ICs

FIGURE 6. Inﬂammatory gene expression of Mo(cIgG) in response to LPS. (A) GO enrichMap of DEGs between Mo(cIgG/LPS) and Mo(RPMI/ LPS) were generated by BiNGO Cyto- scape plugin. Enriched GO terms as well as gene numbers from the upregulated and downregulated DEGs are indicated. Downloaded from (B) Heatmap showing signature DEGs from the regulation of inﬂammatory response genes in Mo(cIgG/LPS) compared with Mo(RPMI/LPS). Reads per kilobase per million values were normalized among the groups. The genes are further grouped into three subclus- http://www.jimmunol.org/ ters as indicated. (C) Flow cytometry plot of IL-10R (CD210) expression in Mo(cIgG) and Mo(RPMI). (D) Mo(cIgG) and Mo(RPMI) were stimulated with LPS (10 ng/ml) in the presence of increasing concentrations of recombinant human IL-10 for 24 h followed by ELISA quantitation of TNF-a in the culture supernatant. Results are expressed by guest on September 28, 2021 as percentage inhibition of TNF-a production compared with no IL-10 control using the following formula: (1-IL-10 group/no IL-10 control)%. Data (mean 6 SEM of three biological replicates) are representative of three independent experiments. *p , 0.05, **p , 0.01.

monocyte training (Fig. 3A). In an in vivo situation, monocytes macrophages used in Antoniv and Ivashkiv’s work (14) were from are most likely sensitized by immobilized ICs just as they infiltrate RA-SF rather than synovial tissues. It is possible that the mac- peripheral tissues (e.g., affected RA synovia) and then very rophages of synovial tissue, and not those from SF, play a major quickly differentiate into macrophages. Interestingly, monocyte- role in RA pathogenesis. Further in situ studies evaluating the derived macrophages (differentiated in the presence of CSF1) functional status of tissue macrophages in affected RA synovia in could also be sensitized by cIgG for a hyperactive inflammatory relation to IgG-IC deposition is warranted. state in vitro (Supplemental Fig. 2), implying that synovial mac- As shown in our pathophysiological hypothesis presented rophages in RA patients are also susceptible to IgG-IC–mediated in Fig. 9, the initial steps of IgG-IC–mediated monocyte/ training. The remarkable resemblance of the gene expression macrophage training include FcgRIIa (CD32a) cross-linking profile of RA synovial macrophages to that of Mo(cIgG) provides and Syk-dependent signaling. Of the FcgRs expressed by hu- additional evidence for this notion. It should be noted that man myeloid cells, FcgRIIa is evidently the most important to RA The Journal of Immunology 9 Downloaded from http://www.jimmunol.org/ by guest on September 28, 2021

FIGURE 7. Histone modification in cIgG-sensitized monocytes. (A) Expression profiles of the histone modification enzymes genes in Mo(cIgG) and Mo(RPMI) are compared, including histone acetyltransferase (HAT), histone deacetylase (HDAC), HMT, and histone demethylase (KDM). SETD3, fold change (fc) = 1.689, p = 6.76 3 1029. DOT1L, fc = 2.353, p = 0.001. PRMT1, fc = 1.583, p = 0.010. PRMT5, fc = 1.656, p = 1.72 3 1027.(B) Freshly prepared human monocytes were primed with cIgG in the presence of 1 mM 59-methylthioadenosine (cIgG/MTA) for 24 h, followed by LPS (10 ng/ml) restimulation and ELISA quantitation of TNF-a in the culture supernatant. Unprimed monocytes (RPMI) also received LPS stimulation as a control. (C) Comparison of H3K4me3 modifications of the promoter regions of inflammatory cytokine (TNF-a, IL-6, IL-8, IL-1RN) and regulator (NFKBIA, PTGER2, PTGER4, and PTGIR) genes in Mo(cIgG) and Mo(RPMI) by quantitative PCR using chromatin immunoprecipitated DNA as templates. (D) Comparison of H3K27ac modifications of the promoter regions of inflammatory cytokine genes (TNF-a, IL-6, IL-8, IL-1RN) is also shown. (E) H3K4me3 signals (read counts per million mapped reads) in the inherited, enhanced, and suppressed subclusters of inflammatory genes of Mo(cIgG) and Mo(RPMI). Data (mean 6 SEM) are representative of three independent experiments with three biological replicates. *p , 0.05, **p , 0.01, ***p , 0.001. pathogenesis. In mouse models of autoimmune arthritis, human highest proinflammatory response to the RA-specific ICs FcgRIIa–transgenic mice displayed a higher susceptibility to containing anticitrullinated peptide Ab (37). As a pivotal regu- collagen-induced arthritis (52). In a passive Ab transfer model, all lator in monocyte/macrophage responsiveness, CSF1 is capable FcgRIIa–transgenic mice developed arthritis, whereas none of the of recruiting ETS2 that could bind a TNF-a gene promoter and control animals were affected (53). Our transcriptomic analysis facilitate its transcription (55). Importantly, cIgG sensitization uncovers detailed insights into the molecular mechanisms caused a significant increase of H3K4me3 modifications at the for downstream events after FcgRIIa engagement by IgG-ICs histone tails of promoters of inflammatory cytokine genes (e.g., (Fig. 9B). CSF1 production by IC-primed cells is a key factor in TNF-a,IL-1b, IL-6), whereas H3K4me3 modifications on genes this process, which forms an autocrine loop with constitutively of certain key negative regulators, such as NFKBIA, PTGER2, expressed CSF1 receptor on myeloid cells and then drives and PTGER4, were decreased. This would result in stably re- monocyte–macrophage differentiation and sustains their sur- moving the rate-limiting chromatin barriers at both distal and vival. This is in line with an earlier work, showing that mono- proximal regulatory regions, leading to a long-lasting hyperac- cytes cultured on immobilized IgG were protected from tive inflammatory phenotype. apoptosis because of FcgR-induced CSF1 release (54) and Another important event underlying the sustained hyper- also that M-CSF–oriented human macrophages present the responsiveness of Mo(cIgG) is the loss of expression of 10 FUNCTIONAL REPROGRAMMING OF HUMAN MONOCYTES BY IgG-ICs Downloaded from

FIGURE 8. Similarity of signature gene expression between RA synovial macrophages and Mo(cIgG). (A) GSEA comparison for gene expression profiles of Mo(cIgG) (generated in this study) and synovial macrophages from patients with RA or SpA (data set from GSE10500, Ref. 14). NES and FDR q values were calculated according to the GSEA statistics. (B) Heatmap showing partial signature genes commonly expressed in synovial macrophages from patients with RA or SpA and Mo(cIgG). Signaling intensity of individual RA (n = 5) and SpA (n = 3) samples was used as expression values and normalized as indicated; each column represents gene expression from an individual. Fold change (average of three individual samples) of the corresponding genes in Mo(cIgG) relative to control macrophages Mo(RPMI) are shown and normalized as indicated. http://www.jimmunol.org/ anti-inflammatory cytokine IL-10 as well as reduced expression of provide a tonic signaling through FcgRIIa to drive monocytes/ IL-10R. It has previously been documented that RA synovial macrophages into a hyperactive functional state but not neces- macrophages were resistant to the anti-inflammatory effect of sarily immediate TNF-a release; and Type III, TLR agonist– IL-10 (14, 56) and also that ligation of FcgRs by ICs inhibited containing ICs (represented by LTF-IC and DNA-containing IL-10 activation of the Jak-Stat signaling pathway in human ICs) that are capable of triggering synergistic signaling via macrophages (56). It is reasonable to suggest that suppression of FcgRIIaandTLRs,leadingtoastrongTNF-a response as well IL-10 biosynthesis and generation of macrophages refractory to as sustainable hyperactivity of monocytes/macrophages. There is by guest on September 28, 2021 IL-10 can contribute to pathogenesis of inflammatory diseases evidence showing that synergistic activation between FcgRIIa characterized by production of ICs. In the case of murine bone and TLRs could not only efficiently activate monocytes/ marrow–derived macrophages, however, E-IgG (rabbit IgG– macrophages but also transform inhibitory M2 macrophage in- opsonized sheep erythrocytes) ligation of FcgRs along with LPS to an activated phenotype (M1-like), thereby contributing to the stimulation led to strongly enhanced IL-10 and abrogation of pathogenesis of inflammatory autoimmune diseases (12, 13). ICs inflammatory cytokine IL-12 production, whereas TNF-a re- from sera of RA patients likely contain higher levels of Type III sponsiveness was unaffected (57). In our hands, murine ICs (7–10) and are therefore more powerful in monocyte/ monocytes/macrophages were markedly less sensitive to hyper- macrophage activation and sensitization. activity training by IgG-ICs than human cells (data not shown). Clearly, IgG-ICs stand out as a class of potent and unique This could be explained by the lack of activating FcgRIIa and stimulators for monocyte/macrophage functional reprogram- high expression of the inhibitory receptor FcgRIIb in the mouse ming. The trained immunity of monocytes induced by exposure cells (3, 58). to C. albicans or b-glucans is mediated through the Dectin-1/ There have been some controversies in the literature about the Raf-1 signaling pathway (16), which is different from that effect of IgG-ICs on monocyte/macrophage immunobiological (FcgRIIa/Syk) triggered by cIgG. b-glucan–trained mono- activities. This could at least be partially explained by the cytes displayed stronger ability in phagocytosis, indicative of considerable differences in size, solubility, and biologic char- more active roles in immunological defense and scavenging acteristics of IC preparations employed by different laborato- (15, 16), whereas IgG-IC–sensitized monocytes appear to be ries. Considerably different IC preparations have been used steered away from defense and debrite scavenge. Glycolysis is under the term of “ICs.” These include the following: cIgG, upregulated in b-glucan–trained monocytes (59) but not seen boiled IgG, IgG-opsonized erythrocytes (or bacteria and in Mo(cIgG) (data not shown). In contrast, LPS-stimulated microbeads), ICs containing either nonbiologically active Ags monocytes/macrophages functionally differ from the IgG-IC– (e.g., OVA and BSA) or TLR agonist Ags (e.g., dsDNA, cit- trained cells in that reduced responsiveness (LPS tolerance) im- rullinated fibrinogen, and LTF) (7–10, 13, 48, 56, 57). We mediately follows strong inflammatory cytokine production and propose to classify IgG-ICs into the following three types enhanced phagocytosis ability. according to their biological and biophysical characteristics in In summary, the results presented in this study demonstrate future studies: Type I, small soluble ICs containing nonbiolog- that IgG-ICs, likely produced as a result of an overexuberant ically active Ags (represented by OVA-containing ICs and BSA- response to autoantigens, infectious pathogens, or noninfec- containing ICs) that are generally weak in their ability to trigger tious allergens, could prime monocytes/macrophages to a sustain- signaling via FcgRs; Type II, large, insoluble, and amorphous able hyperactive state capable of producing high levels of ICs (represented by cIgG or IgG-coated particles) that can proinflammatory cytokines. Using transcriptomic and epigenomic The Journal of Immunology 11 Downloaded from http://www.jimmunol.org/ by guest on September 28, 2021

FIGURE 9. Hypothesis on the molecular mechanisms of IgG-IC–mediated hyperactivity training of monocytes/macrophages. (A) Monocyte/ macrophage hyperactivity training by IgG-ICs in RA. Incoming monocytes and resident macrophages in RA synovia are primed by deposited IgG-ICs (step 1) and adapt a sustained hyperactive functional state (step 2); when encountering subsequent stimuli such as DAMPs or PAMPs, these cells produce an exacerbated amount of inflammatory cytokines (particularly TNF-a) (step 3). This chain of events may lead to overt osteoclastogenesis and progressive inflammatory damage to the joint tissues in RA. (B) Putative molecular mechanisms for IgG-IC–mediated hyperactivity training. Monocyte/ macrophage functional programming is triggered by IgG-IC binding to FcgRIIa and subsequent signaling through Syk, MAPK, and AP-1, which leads to CSF1 production and formation of a autocrine loop facilitating the proliferation/differentiation and survival of the cells. Elevated ETS2 could bind the TNF-a gene promoter that facilitates its subsequent transcription (55). Activation of HMT gene (e.g., SETD3) expression (also see Figs. 5D, 8B) results in H3K4me3 modification (see Fig. 7C, 7E) on the histone tail of inflammatory cytokine and regulator genes, thereby resulting in a sustained state of proinflammatory reactivity to a range of stimuli. analysis, we provide in this study a detailed mechanism of the References molecular events leading to the sensitization of these cells by the 1. Bax, M., T. W. Huizinga, and R. E. Toes. 2014. The pathogenic potential of IgG-ICs. Importantly, this information is closely reflected in clinical autoreactive antibodies in rheumatoid arthritis. Semin. Immunopathol. 36: 313– 325. RA. Thus, our study not only advances the understanding of the 2. Wegner, N., K. Lundberg, A. Kinloch, B. Fisher, V. Malmstro¨m, M. Feldmann, molecular mechanism of this pervasive immunological event but also and P. J. Venables. 2010. Autoimmunity to specific citrullinated proteins gives indicates potential targets for therapeutic intervention of a range of the first clues to the etiology of rheumatoid arthritis. Immunol. Rev. 233: 34–54. 3. Bruhns, P. 2012. Properties of mouse and human IgG receptors and their con- key inflammatory autoimmune diseases. tribution to disease models. Blood 119: 5640–5649. 4. Guilliams, M., P. Bruhns, Y. Saeys, H. Hammad, and B. N. Lambrecht. 2014. The function of Fcg receptors in dendritic cells and macrophages. [Published erratum appears in 2014 Nat. Rev. Immunol. 14: 349.] Nat. Rev. Immunol. 14: Acknowledgments 94–108. We thank Prof. F. Y. Liew of the Institute of Immunology, University of 5. Szodoray, P., Z. Szabo´, A. Kapitańy, A. Gyetvai, G. Lakos, S. Szańto´,G.Szu¨cs, Glasgow, U.K., for helpful discussion and critical comments on the and Z. Szekanecz. 2010. Anti-citrullinated protein/peptide autoantibodies in association with genetic and environmental factors as indicators of disease article. outcome in rheumatoid arthritis. Autoimmun. Rev. 9: 140–143. 6. Monach, P. A., W. Hueber, B. Kessler, B. H. Tomooka, M. BenBarak, B. P. Simmons, J. Wright, T. S. Thornhill, M. Monestier, H. Ploegh, et al. 2009. A broad screen for targets of immune complexes decorating arthritic joints Disclosures highlights deposition of nucleosomes in rheumatoid arthritis. Proc. Natl. Acad. The authors have no financial conflicts of interest. Sci. USA 106: 15867–15872. 12 FUNCTIONAL REPROGRAMMING OF HUMAN MONOCYTES BY IgG-ICs

7. Harre, U., D. Georgess, H. Bang, A. Bozec, R. Axmann, E. Ossipova, 31. Weinblatt, M. E., A. Kavanaugh, M. C. Genovese, T. K. Musser, E. B. Grossbard, P. J. Jakobsson, W. Baum, F. Nimmerjahn, E. Szarka, et al. 2012. Induction of and D. B. Magilavy. 2010. An oral spleen tyrosine kinase (Syk) inhibitor for osteoclastogenesis and bone loss by human autoantibodies against citrullinated rheumatoid arthritis. N. Engl. J. Med. 363: 1303–1312. vimentin. J. Clin. Invest. 122: 1791–1802. 32. Elshabrawy, H. A., Z. Chen, M. V. Volin, S. Ravella, S. Virupannavar, and 8. Hu, L., X. Hu, K. Long, C. Gao, H. L. Dong, Q. Zhong, X. M. Gao, and S. Shahrara. 2015. The pathogenic role of angiogenesis in rheumatoid arthritis. F. Y. Gong. 2017. Extraordinarily potent proinflammatory properties of Angiogenesis 18: 433–448. lactoferrin-containing immunocomplexes against human monocytes and mac- 33. Barozzi, I., M. Simonatto, S. Bonifacio, L. Yang, R. Rohs, S. Ghisletti, and rophages. Sci. Rep. 7: 4230. G. Natoli. 2014. Coregulation of transcription factor binding and nucleosome 9. Sokolove, J., X. Zhao, P. E. Chandra, and W. H. Robinson. 2011. Immune occupancy through DNA features of mammalian enhancers. Mol. Cell 54: 844– complexes containing citrullinated fibrinogen costimulate macrophages via Toll- 857. like receptor 4 and Fcg receptor. Arthritis Rheum. 63: 53–62. 34. Glass, C. K., and G. Natoli. 2016. Molecular control of activation and priming in 10. Marshak-Rothstein, A., and I. R. Rifkin. 2007. Immunologically active auto- macrophages. Nat. Immunol. 17: 26–33. antigens: the role of toll-like receptors in the development of chronic inflam- 35. Burger, J. A., N. J. Zvaifler, N. Tsukada, G. S. Firestein, and T. J. Kipps. 2001. matory disease. Annu. Rev. Immunol. 25: 419–441. Fibroblast-like synoviocytes support B-cell pseudoemperipolesis via a stromal 11. McInnes, I. B., and G. Schett. 2007. Cytokines in the pathogenesis of rheumatoid cell-derived factor-1- and CD106 (VCAM-1)-dependent mechanism. J. Clin. arthritis. Nat. Rev. Immunol. 7: 429–442. Invest. 107: 305–315. 12. Udalova, I. A., A. Mantovani, and M. Feldmann. 2016. Macrophage heteroge- 36. Bru¨hl, H., K. Wagner, H. Kellner, M. Schattenkirchner, D. Schlo¨ndorff, and neity in the context of rheumatoid arthritis. Nat. Rev. Rheumatol. 12: 472–485. M. Mack. 2001. Surface expression of CC- and CXC-chemokine receptors on 13. Vogelpoel, L. T., I. S. Hansen, T. Rispens, F. J. Muller, T. M. van Capel, leucocyte subsets in inflammatory joint diseases. Clin. Exp. Immunol. 126: 551– M. C. Turina, J. B. Vos, D. L. Baeten, M. L. Kapsenberg, E. C. de Jong, and 559. J. den Dunnen. 2014. Fc gamma receptor-TLR cross-talk elicits pro- 37. Clavel, C., L. Ceccato, F. Anquetil, G. Serre, and M. Sebbag. 2016. Among inflammatory cytokine production by human M2 macrophages. Nat. Commun. human macrophages polarised to different phenotypes, the M-CSF-oriented cells 5: 5444. present the highest pro-inflammatory response to the rheumatoid arthritis- 14. Antoniv, T. T., and L. B. Ivashkiv. 2006. Dysregulation of interleukin-10- specific immune complexes containing ACPA. Ann. Rheum. Dis. 75: 2184–2191. dependent gene expression in rheumatoid arthritis synovial macrophages. Ar- 38. Sweet, M. J., C. C. Campbell, D. P. Sester, D. Xu, R. C. McDonald, thritis Rheum. 54: 2711–2721. K. J. Stacey, D. A. Hume, and F. Y. Liew. 2002. Colony-stimulating factor-1 Downloaded from 15. Saeed, S., J. Quintin, H. H. Kerstens, N. A. Rao, A. Aghajanirefah, F. Matarese, suppresses responses to CpG DNA and expression of toll-like receptor 9 but S. C. Cheng, J. Ratter, K. Berentsen, M. A. van der Ent, et al. 2014. Epigenetic enhances responses to lipopolysaccharide in murine macrophages. J. Immunol. programming of monocyte-to-macrophage differentiation and trained innate 168: 392–399. immunity. Science 345: 1251086. 39. Edwards, J. R., S. G. Sun, R. Locklin, C. M. Shipman, I. E. Adamopoulos, 16. Quintin, J., S. Saeed, J. H. A. Martens, E. J. Giamarellos-Bourboulis, D. C. Ifrim, N. A. Athanasou, and A. Sabokbar. 2006. LIGHT (TNFSF14), a novel mediator C. Logie, L. Jacobs, T. Jansen, B. J. Kullberg, C. Wijmenga, et al. 2012. Candida of bone resorption, is elevated in rheumatoid arthritis. Arthritis Rheum. 54: albicans infection affords protection against reinfection via functional reprog- 1451–1462.

ramming of monocytes. Cell Host Microbe 12: 223–232. 40. Asou, Y., S. R. Rittling, H. Yoshitake, K. Tsuji, K. Shinomiya, A. Nifuji, http://www.jimmunol.org/ 17. Elshaﬁe, A. I., E. Ahlin, L. Mathsson, G. ElGhazali, and J. Ro¨nnelid. 2007. D. T. Denhardt, and M. Noda. 2001. Osteopontin facilitates angiogenesis, ac- Circulating immune complexes (IC) and IC-induced levels of GM-CSF are in- cumulation of osteoclasts, and resorption in ectopic bone. Endocrinology 142: creased in sudanese patients with acute visceral Leishmania donovani infection 1325–1332. undergoing sodium stibogluconate treatment: implications for disease patho- 41. Chen, G., X. Zhang, R. Li, L. Fang, X. Niu, Y. Zheng, D. He, R. Xu, and genesis. J. Immunol. 178: 5383–5389. J. Z. Zhang. 2010. Role of osteopontin in synovial Th17 differentiation in 18. Hu, X. M., Y. R. Xu, R. Yan, S. L. Sun, H. L. Dong, J. Wang, and X. M. Gao. rheumatoid arthritis. Arthritis Rheum. 62: 2900–2908. 2015. A novel murine anti-lactoferrin monoclonal antibody activates human 42. Snoeks, L., C. R. Weber, J. R. Turner, M. Bhattacharyya, K. Wasland, and polymorphonuclear leukocytes through membrane-bound lactoferrin and TLR4. S. D. Savkovic. 2008. Tumor suppressor Foxo3a is involved in the regulation of BioMed Res. Int. 2015: 285237. lipopolysaccharide-induced interleukin-8 in intestinal HT-29 cells. Infect. 19. Hong, C., X. Qiu, Y. Li, Q. Huang, Z. Zhong, Y. Zhang, X. Liu, L. Sun, P. Lv, Immun. 76: 4677–4685. and X. M. Gao. 2010. Functional analysis of recombinant calreticulin fragment 43. Lou, X., S. Sun, W. Chen, Y. Zhou, Y. Huang, X. Liu, Y. Shan, and C. Wang.

39-272: implications for immunobiological activities of calreticulin in health and 2011. Negative feedback regulation of NF-kB action by CITED2 in the nucleus. by guest on September 28, 2021 disease. J. Immunol. 185: 4561–4569. J. Immunol. 186: 539–548. 20. Li, B., and C. N. Dewey. 2011. RSEM: accurate transcript quantification from RNA- 44. Calvayrac, O., R. Rodrı´guez-Calvo, I. Martı´-Pamies, J. Alonso, B. Ferrań, Seq data with or without a reference genome. BMC Bioinformatics 12: 323. S. Aguilo´, J. Crespo, A. Rodrı´guez-Sinovas, C. Rodrı´guez, and J. Martıńez-Gon- 21. Leng, N., J. A. Dawson, J. A. Thomson, V. Ruotti, A. I. Rissman, B. M. Smits, za´lez. 2015. NOR-1 modulates the inflammatory response of vascular smooth J. D. Haag, M. N. Gould, R. M. Stewart, and C. Kendziorski. 2013. EBSeq: an muscle cells by preventing NFkB activation. J. Mol. Cell. Cardiol. 80: 34–44. empirical Bayes hierarchical model for inference in RNA-seq experiments. 45. Cso´ka, B., Z. H. Ne´meth, P. Rosenberger, H. K. Eltzschig, Z. Spolarics, Bioinformatics 29: 1035–1043. P. Pacher, Z. Selmeczy, B. Koscso´, L. Himer, E. S. Vizi, et al. 2010. A2B 22. Maere, S., K. Heymans, and M. Kuiper. 2005. BiNGO: a Cytoscape plugin to adenosine receptors protect against sepsis-induced mortality by dampening ex- assess overrepresentation of gene ontology categories in biological networks. cessive inflammation. J. Immunol. 185: 542–550. Bioinformatics 21: 3448–3449. 46. Crouse, J., U. Kalinke, and A. Oxenius. 2015. Regulation of antiviral T cell 23. Langmead, B., and S. L. Salzberg. 2012. Fast gapped-read alignment with responses by type I interferons. Nat. Rev. Immunol. 15: 231–242. Bowtie 2. Nat. Methods 9: 357–359. 47. Gomez, P. F., M. H. Pillinger, M. Attur, N. Marjanovic, M. Dave, J. Park, 24. Shen, L., N. Shao, X. Liu, and E. Nestler. 2014. ngs.plot: quick mining and C. O. Bingham, III, H. Al-Mussawir, and S. B. Abramson. 2005. Resolution of visualization of next-generation sequencing data by integrating genomic data- inflammation: prostaglandin E2 dissociates nuclear trafficking of individual NF- bases. BMC Genomics 15: 284. kappaB subunits (p65, p50) in stimulated rheumatoid synovial fibroblasts. J. 25. Kent, W. J., C. W. Sugnet, T. S. Furey, K. M. Roskin, T. H. Pringle, A. M. Zahler, Immunol. 175: 6924–6930. and D. Haussler. 2002. The human genome browser at UCSC. Genome Res. 12: 48. Zhang, Y., S. Liu, J. Liu, T. Zhang, Q. Shen, Y. Yu, and X. Cao. 2009. Immune 996–1006. complex/Ig negatively regulate TLR4-triggered inflammatory response in mac- 26. Subramanian, A., P. Tamayo, V. K. Mootha, S. Mukherjee, B. L. Ebert, rophages through Fc gamma RIIb-dependent PGE2 production. J. Immunol. 182: M. A. Gillette, A. Paulovich, S. L. Pomeroy, T. R. Golub, E. S. Lander, and 554–562. J. P. Mesirov. 2005. Gene set enrichment analysis: a knowledge-based approach 49. Bernstein, B. E., E. L. Humphrey, R. L. Erlich, R. Schneider, P. Bouman, for interpreting genome-wide expression profiles. Proc. Natl. Acad. Sci. USA J. S. Liu, T. Kouzarides, and S. L. Schreiber. 2002. Methylation of histone H3 102: 15545–15550. Lys 4 in coding regions of active genes. Proc. Natl. Acad. Sci. USA 99: 8695– 27. Lopa, S., M. J. Leijs, M. Moretti, E. Lubberts, G. J. van Osch, and 8700. Y. M. Bastiaansen-Jenniskens. 2015. Arthritic and non-arthritic synovial fluids 50. Creyghton, M. P., A. W. Cheng, G. G. Welstead, T. Kooistra, B. W. Carey, modulate IL10 and IL1RA gene expression in differentially activated primary E. J. Steine, J. Hanna, M. A. Lodato, G. M. Frampton, P. A. Sharp, et al. 2010. human monocytes. Osteoarthritis Cartilage 23: 1853–1857. Histone H3K27ac separates active from poised enhancers and predicts devel- 28. Salazar-Aldrete, C., M. Galań-Dıéz, E. Fernańdez-Ruiz, P. Ninõ-Moreno, opmental state. Proc. Natl. Acad. Sci. USA 107: 21931–21936. L. Estrada-Capetillo, C. Abud-Mendoza, E. Layseca-Espinosa, L. Baranda, and 51. Huang, Q., Y. Ma, A. Adebayo, and R. M. Pope. 2007. Increased macrophage R. Gonza´lez-Amaro. 2013. Expression and function of dectin-1 is defective in activation mediated through toll-like receptors in rheumatoid arthritis. Arthritis monocytes from patients with systemic lupus erythematosus and rheumatoid Rheum. 56: 2192–2201. arthritis. J. Clin. Immunol. 33: 368–377. 52. Van de Velde, N. C., P. L. Mottram, M. S. Powell, B. Lim, R. Holmdahl, and 29. Lacerte, P., A. Brunet, B. Egarnes, B. Ducheˆne, J. P. Brown, and J. Gosselin. P. M. Hogarth. 2010. Transgenic mice expressing human FcgammaRIIa have 2016. Overexpression of TLR2 and TLR9 on monocyte subsets of active rheu- enhanced sensitivity to induced autoimmune arthritis as well as elevated Th17 matoid arthritis patients contributes to enhance responsiveness to TLR agonists. cells. Immunol. Lett. 130: 82–88. Arthritis Res. Ther. 18: 10. 53. Tan Sardjono, C., P. L. Mottram, N. C. van de Velde, M. S. Powell, D. Power, 30. Kowalski, M. L., A. Wolska, J. Grzegorczyk, J. Hilt, M. Jarzebska, R. F. Slocombe, I. P. Wicks, I. K. Campbell, S. E. McKenzie, M. Brooks, et al. M. Drobniewski, M. Synder, and M. Kurowski. 2008. Increased responsiveness to 2005. Development of spontaneous multisystem autoimmune disease and hy- toll-like receptor 4 stimulation in peripheral blood mononuclear cells from patients persensitivity to antibody-induced inflammation in Fcgamma receptor IIa- with recent onset rheumatoid arthritis. Mediators Inflamm. 2008: 132732. transgenic mice. Arthritis Rheum. 52: 3220–3229. The Journal of Immunology 13

54. Marsh, C. B., R. P. Pomerantz, J. M. Parker, A. V. Winnard, E. L. Mazzaferri, Jr., 57. Gerber, J. S., and D. M. Mosser. 2001. Reversing lipopolysaccharide toxicity by N. Moldovan, T. W. Kelley, E. Beck, and M. D. Wewers. 1999. Regulation of ligating the macrophage Fc gamma receptors. J. Immunol. 166: 6861–6868. monocyte survival in vitro by deposited IgG: role of macrophage colony- 58. Bolland, S., and J. V. Ravetch. 2000. Spontaneous autoimmune disease in Fc stimulating factor. J. Immunol. 162: 6217–6225. (gamma)RIIB-deﬁcient mice results from strain-speciﬁc epistasis. Immunity 55. Kra¨mer, B., K. Wiegmann, and M. Kro¨nke. 1995. Regulation of the human TNF 13: 277–285. promoter by the transcription factor Ets. J. Biol. Chem. 270: 6577–6583. 59. Cheng, S. C., J. Quintin, R. A. Cramer, K. M. Shepardson, S. Saeed, V. Kumar, 56. Ji, J. D., I. Tassiulas, K. H. Park-Min, A. Aydin, I. Mecklenbrauker, E. J. Giamarellos-Bourboulis, J. H. Martens, N. A. Rao, A. Aghajanirefah, et al. A. Tarakhovsky, L. Pricop, J. E. Salmon, and L. B. Ivashkiv. 2003. Inhibition of 2014. mTOR- and HIF-1a-mediated aerobic glycolysis as metabolic basis for interleukin 10 signaling after Fc receptor ligation and during rheumatoid ar- trained immunity. [Published erratum appears in 2014 Science 346: aaa1503.] thritis. J. Exp. Med. 197: 1573–1583. Science 345: 1250684. Downloaded from http://www.jimmunol.org/ by guest on September 28, 2021