ARTHRITIS & RHEUMATOLOGY Vol. 66, No. 9, September 2014, pp 2368–2379 DOI 10.1002/art.38711 © 2014, American College of Rheumatology

Inhibition of Fucosylation Reshapes Inflammatory Macrophages and Suppresses Type II Collagen–Induced Arthritis Jun Li,1 Hui-Chen Hsu,2 Yana Ding,1 Hao Li,1 Qi Wu,2 PingAr Yang,2 Bao Luo,1 Amber L. Rowse,1 David M. Spalding,1 S. Louis Bridges Jr.,1 and John D. Mountz2 Objective. Fucosylation catalyzed by fucosyltransferases (FUTs) is an important posttranslational modification involved in a variety of biologic processes. This study was undertaken to determine the roles of fucosylation in rheumatoid arthritis (RA) and to assess the efficacy of reestablishing immune homeostasis with the use of 2-deoxy-D-galactose (2-D-gal), a fucosylation inhibitor. Methods. Quantitative polymerase chain reaction was performed to determine the expression of FUT genes in synovial tissue from RA and osteoarthritis (OA) patients and in fluorescence-activated cell–sorted cells from RA synovial fluid. The in vivo inhibitory effect of 2-D-gal was evaluated in a murine collagen-induced arthritis (CIA) model. The in vitro effects of 2-D-gal on differentiation of inflammatory macrophages, produc-

tion of cytokines, and antigen uptake, processing, and presentation functions were analyzed. Results. FUTs that are involved in terminal or subterminal fucosylation, but not those involved in core fucosylation or O-fucosylation, were up-regulated in RA compared to OA synovial tissue. The expression of terminal FUTs was highly positively correlated with the expression of TNF (encoding for tumor necrosis factor ␣). Terminal FUTs were predominantly expressed in M1 macrophages. In vivo, 2-D-gal treatment of mice precluded the development of CIA by reducing inflammatory macrophages and Th17 cells in the draining lymph nodes and decreasing the levels of TNF␣, interleukin-6 (IL-6), and antibodies to type II collagen in the serum. In vitro, treatment with 2-D-gal skewed the differentiation of M1 macrophages to IL-10–producing M2 macrophages. Furthermore, 2-D-gal significantly inhibited the antigen-presenting function of M1 macrophages. Conclusion. Terminal fucosylation is a novel hallmark of inflammatory macrophages. Inhibition of terminal FUTs reshapes the differentiation and functions of M1 macrophages, leading to resolution of inflammation in arthritis.

Supported by the Arthritis Foundation (grant to Dr. J. Li), the Lupus Research Institute (grant to Dr. Hsu), and Daiichi Sankyo Co., Ltd (grant to Dr. Mountz), the NIH (grants 1R01-AI-083705 to Dr. Hsu and 1R01-AI-071110 and P30-AR-048311 to Dr. Mountz), a Rheumatology Research Foundation grant to Dr. Mountz, and the Department of Veterans Affairs (Merit Review grant 1I01BX000600 to Dr. Mountz). Portions of the study were performed at the University of Alabama at Birmingham Rheumatic Diseases Core Center (Analytical Genomics and Transgenics Core, Comprehensive Flow Cytometry Core, and Analytic Imaging and Immunoreagents Core), supported by NIH grant P30-AR-048311, the Center for AIDS Research flow cytometry facility, supported by NIH grant P30-AI-027767, and the Center for Metabolic Bone Disease–Histomorphometry and Molecular Analysis Core Laboratory, supported by NIH grant P30AR-46031. 1 Jun Li, MD, PhD, Yana Ding, MD, PhD, Hao Li, PhD, Bao Luo, PhD, Amber L. Rowse, PhD, David M. Spalding, MD, S. Louis Bridges Jr., MD, PhD: University of Alabama at Birmingham; 2HuiChen Hsu, PhD, Qi Wu, BS, PingAr Yang, BS, John D. Mountz, MD, PhD: University of Alabama at Birmingham and Birmingham VA Medical Center, Birmingham, Alabama. Address correspondence to John D. Mountz, MD, PhD, University of Alabama at Birmingham, Division of Clinical Immunology and Rheumatology, Department of Medicine, SHEL Building, Suite 307, 1825 University Boulevard, Birmingham, AL 35294. E-mail: [email protected]. Submitted for publication February 3, 2014; accepted in revised form May 13, 2014.

Modification by L-fucose (6-deoxy-L-galactose) can be found typically in N- and O-linked glycans and glycolipids. Fucosylation is the final step in the biosynthesis of oligosaccharides and is considered to be a terminal glycan modification. Fucose-containing glycans play an important regulatory role in a wide range of physiologic and pathologic events, including cell and tissue development, angiogenesis, fertilization, cell adhesion, inflammation, tumor apoptosis, and tumor metastasis (1). It has been shown by lectin-binding assay that fucosylation of the IgG heavy chain and ␣1-acid glycoprotein is increased in rheumatoid arthritis (RA) patients compared to healthy individuals (2,3). However, the roles of fucosylation in the pathogenesis of RA have not been investigated. 2368

INHIBITION OF FUCOSYLATION IN ARTHRITIS

2369

Figure 1. Expression profile of fucosyltransferase (FUT) genes in human rheumatoid arthritis (RA) and osteoarthritis (OA) synovial tissue. A, Diagram of the fucosylated sites of human FUTs, including ␣(1,2)-, ␣(1,3/4)-, ␣(1,6)-, and O-FUTs. Fucosylated sites are shown on O- and N-glycans. B, Expression of FUTs, SLC35C1 (encodes for GDP-fucose transporter), and TNF (encodes for tumor necrosis factor ␣) in RA and OA synovial tissue. Results are the mean ⫾ SEM copy number for each gene of interest (GOI) (n ⫽ 14 patients per group), calculated as (GOI/Gapdh) ⫻ 105. Terminal and subterminal FUTs exhibiting significant differences in expression between RA and OA synovial tissue are shown in red, while O-FUTs and core FUTs not exhibiting significant differences between RA and OA synovial tissue are shown in blue. EGF ⫽ epidermal growth factor; TSRs ⫽ thrombospondin repeats.

Fucosyltransferases (FUTs) are enzymes that catalyze the transfer of a fucose residue from the GDPfucose donor to the acceptor molecules. Thirteen FUT genes have been identified in humans. These FUTs catalyze ␣(1,2)-, ␣(1,3/4)-, ␣(1,6)-, and O-fucosylation (Figure 1A). The ␣(1,2)- and ␣(1,3/4)-FUTs catalyze the fucosylation that localizes at the terminal or subterminal of the acceptor molecules, whereas ␣(1,6)-FUTs catalyze the core fucosylation that occurs at the innermost moiety of N-glycan, and O-fucosylation, regulated by POFUT1/2, is the direct O-linkage of fucose to serine and threonine residues in the epidermal growth factor– like and thrombospondin type 1 repeat domains (1,4). Macrophages are the central players in the pathogenesis of RA and are considered to be important therapeutic targets in the management of RA (5–7). They produce a number of inflammatory mediators, and their professional antigen-presenting role has also been implicated in the pathogenesis of RA (5,8). Plasticity and flexibility are key characteristics of macrophages, which have been shown to be regulated at the transcriptional, posttranscriptional, and protein levels (9). In the present study, we discovered a novel posttranslational fucosylation mechanism that orchestrates the develop-

mental and functional plasticity of macrophages. We found that ␣(1,2)- and ␣(1,3/4)-FUTs were predominantly expressed in M1 inflammatory macrophages, and their expression was highly correlated with the levels of TNF (encoding for tumor necrosis factor ␣) in the synovial tissue of arthritis patients. The current study is the first to demonstrate that posttranslational terminal fucosylation is a hallmark of inflammatory macrophages and that the fucosylation inhibitor 2-deoxy-D-galactose (2-D-gal) can regulate the plasticity of inflammatory M1 macrophage differentiation and its functions, leading to an antiinflammatory M2 macrophage phenotype and the resolution of inflammation in arthritis. MATERIALS AND METHODS Animal studies. Mice. DBA/1J, C57BL/6, and myelin oligodendrocyte glycoprotein 35–55 peptide (MOG35–55)– specific 2D2 T cell receptor (TCR)–transgenic mice were obtained from The Jackson Laboratory. All animal procedures were approved by the University of Alabama at Birmingham (UAB) Institutional Animal Care and Use Committee. Induction of collagen-induced arthritis (CIA) and treatments. CIA was induced in DBA/1J mice that were 8–12 weeks old. Mice were immunized by intradermal administration of

2370

bovine type II collagen (CII) (Chondrex) emulsified in Freund’s complete adjuvant (CFA), followed by injection of bovine CII in Freund’s incomplete adjuvant on day 21, as described previously (10). The severity of arthritis was assessed daily (10). The mice received 2-D-gal (250 mg/kg body weight [BW]; Sigma), fucose (250 mg/kg BW; Sigma), or normal saline (as control) via intraperitoneal (IP) injections every 2 days, initiated on day 0 until about day 60, when the mice were killed for further analyses. Human studies. Patients. Synovial tissue samples from 14 RA patients and 14 osteoarthritis (OA) patients were obtained from the UAB Tissue Procurement Center, as described previously (11,12). For analysis of synovial fluid, 5 RA patients (mean age 53 years, range 38–79 years; mean duration of disease 14 years, range 6–22 years) were recruited from the UAB Rheumatology Clinic. All RA patients met the American College of Rheumatology 1987 revised criteria for RA (13). All samples of synovial tissue and synovial fluid were obtained for clinically indicated purposes. These studies were conducted in compliance with the Declaration of Helsinki and approved by the Institutional Review Board at UAB. All participants provided informed consent. Isolation of human RA synovial fibroblasts and synovial fluid mononuclear cells. Synovial fibroblasts and synovial fluid mononuclear cells were isolated from RA synovial tissue or RA synovial fluid using previously described methods (12,14). Quantitative reverse transcription–polymerase chain reaction (qRT-PCR). Isolation of RNA, synthesis of first-strand complementary DNA, and qRT-PCR analyses of human RA synovial fibroblasts were carried out as described previously (15). Details on all primers used in the present study are available at http://www.uab.edu/medicine/rheumatology/ faculty/33-area-2/primary-faculty/164-jun-li-md-phd. Flow cytometry and phospho-flow analyses. Single-cell suspensions were stained for identification of macrophages and macrophage subsets using mouse-specific antibodies, including fluorescein isothiocyanate (FITC)–conjugated antiCD11b (BD Biosciences), Alexa Fluor 647–conjugated anti–interleukin-23 p19 (anti–IL-23p19) (eBioscience), phycoerythrin (PE)–conjugated anti-TNF ␣ , and PE–Cy7– conjugated anti–Ly-6C. For identification of CD4 T cells and T cell subsets, cells were stained with FITC-conjugated antiCD4, PE–Cy7–conjugated anti-Thy1.2, allophycocyanin (APC)–conjugated anti–interferon-␥ (anti-IFN␥), and PEconjugated anti–IL-17. Intracellular and intranuclear staining was performed as described previously (15). For macrophages treated with 2-D-gal in vitro, cells were stained with Bio-ICAM followed by eFluor 450–conjugated streptavidin (eBioscience) and APC-conjugated anti-CD11b, FITC-conjugated antiCD80, and PE-conjugated anti-CD86. CD4 T cells from 2D2 mice were identified by staining with APC-conjugated antiCD4 and PE–Cy7–conjugated V␤11. DQ ovalbumin (OVA) (Invitrogen) was detected in the FITC channel. For macrophages loaded with green fluorescent protein (GFP)–labeled E␣ peptides, cells were stained with bio-YAe (eBioscience), followed by eFluor 450–conjugated streptavidin (eBioscience) and APC-conjugated anti-CD11b (BioLegend). To determine the degree of ERK-1/2 intracellular signaling, we performed phospho-flow studies according to the protocol from BD Biosciences. Cells were stained for intracell-

LI ET AL

ular pERK-1/2 using rabbit anti–ERK-1 (T202/Y204)/ERK-2 (T185/Y187) antibodies (R&D Systems), followed by Alexa Fluor 488–conjugated goat anti-rabbit IgG (Invitrogen). All flow cytometry data were acquired on a BD Biosciences LSRII flow cytometer. The data were analyzed using FlowJo software (Tree Star). Cell sorting. Unless specified otherwise, all reagents used for fluorescence-activated cell sorting were purchased from BioLegend. Human synovial fluid mononuclear cells were stained with PE-conjugated anti-CD16, PE–Cy7–conjugated anti-CD14, FITC-conjugated anti-CD68, PE-conjugated antiCD80, PE–Cy7–conjugated anti-CD4, PE-conjugated antiCD45RA, PerCP–Cy5.5–conjugated anti-CCR2, PE–Cy7– conjugated anti-CCR4, Alexa Fluor 700–conjugated antiCCR5, FITC-conjugated anti-CCR6, Pacific Blue–conjugated anti-CXCR3, and PE-conjugated anti-CD161 antibodies. The cells were then sorted into subsets of CD68⫹CD80⫹ (M1 macrophages), CD68⫹CD80⫺ (M2 macrophages), CD14⫹ CD16⫺ (classic monocytes), total CD4⫹ T cells, CD4⫹ CD45RA⫹ (naive CD4), CD4⫹CD45RO⫹ (memory CD4), CD4⫹CXCR3⫹CCR6⫺ (Th1) (16), and CD4⫹CXCR3⫺ CCR4⫹CCR6⫹CD161⫹ (Th17) (16), with purities of ⬎96% for each subset. Flow cytometry cell sorting was performed on a FACSAria II cell sorter (BD Biosciences). Histologic analysis of the mouse joints. All samples of joint tissue from the hind limbs of mice were processed and stained for histologic analysis using previously described methods (12,15,17). Generation of mouse bone marrow–derived macrophages, 2-D-gal treatment, peptide loading, and CD4 T cell proliferation assay. Mouse bone marrow–derived M1 and M2 macrophages were differentiated using granulocyte– macrophage colony-stimulating factor (GM-CSF) and M-CSF as described previously (18). For studies of GFP-labeled E␣ peptides, cells from C57BL/6 mice were used; for all other in vitro experiments, cells from C57BL/6 mice or DBA/1J mice were used. Cell counts were determined using a Cell Counting Kit 8 (Dojindo Molecular Technologies) and cell viability was estimated based on the ratio of the cell count in the experimental condition to that in the control condition. For studies of antigen-presenting function, mouse macrophages were loaded with 200 ␮g/ml OVA DQ (Invitrogen) for 30 minutes or 100 ␮g/ml GFP-labeled E␣ (19,20) for 16 hours prior to analysis. For CD4 T cell proliferation assay, M1 macrophages were pretreated with or without 2-D-gal for 2 days. The cells were then loaded with MOG35–55 and cultured for an additional 3 days with carboxyfluorescein succinimidyl ester–labeled CD4 T cells (Invitrogen) isolated from 2D2 TCR–transgenic mice. Enzyme-linked immunosorbent assay (ELISA). Cytokine levels in mouse and human bone marrow–derived macrophages were measured by ELISA according to the manufacturer’s instructions (BioLegend). The bone marrow–derived macrophages were stimulated with 200 ng/ml lipopolysaccharide for 16 hours before cytokine analysis. Statistical analysis. Statistical analyses were performed using Student’s 2-tailed t-test, one-way analysis of variance, and bivariate correlation analysis. P values less than 0.05 were considered statistically significant.

INHIBITION OF FUCOSYLATION IN ARTHRITIS

2371

Figure 2. Cellular patterns of expression of fucosyltransferases (FUTs) in human rheumatoid arthritis (RA) synovial fluid cells, and correlation of expression levels between FUTs and TNF (encoding for tumor necrosis factor ␣) in human RA synovial tissue. A, Expression of FUTs, including those involved in terminal, subterminal, core, and O-fucosylation, was evaluated in different sorted subpopulations of RA synovial fluid cells. Values are the mean ⫾ SEM expression in synovial tissue samples from 5 RA patients. ⴱ ⫽ P ⬍ 0.05 versus the other subpopulations, by one-way analysis of variance. B, Correlations between the expression of FUTs and the expression of TNF were assessed in RA and osteoarthritis (OA) synovial tissue (n ⫽ 14 samples per group). Values are the copy number for each gene of interest (GOI) in each cell subset or synovial tissue sample, calculated as (GOI/Gapdh) ⫻ 105.

RESULTS Up-regulation of terminal ␣(1,2)- and subterminal ␣(1,3/4)-FUTs in RA compared to OA synovial tissue. Thirteen FUTs involved in catalyzing ␣(1,2)-, ␣(1,3/4)-, ␣(1,6)-, and O-fucosylation have been identified in humans (Figure 1A). In synovial tissue samples obtained from RA patients, as compared to synovial tissue from OA patients, there was a statistically significant increase in the levels of most of the FUTs that catalyze terminal and subterminal fucosylation, including FUT1, FUT2, FUT4, FUT5, FUT6, FUT7, and FUT9. The levels of FUT10 and FUT11, whose functions are less well defined, were also significantly increased in RA patients (Figure 1B). Consistent with this, SLC35C1, which encodes the GDP-fucose transporter, was also up-regulated in RA synovial tissue (Figure 1B). In contrast, the expression of FUTs that catalyze core fucosylation and O-fucosylation, including FUT8 and POFUT2, was not increased in RA synovial tissue as compared to OA synovial tissue, although the expression of POFUT1 exhibited a marginally significant difference between RA and OA synovial tissue (Figure 1B). These results suggest that terminal and subterminal fucosylation, but not core fucosylation and O-fucosylation, might be associated with the pathogenesis of RA.

Role of terminal ␣ (1,2)- and subterminal ␣(1,3/4)-FUTs as hallmarks of M1 inflammatory macrophages, and strong positive correlation with the expression of TNF. The above findings suggest that ␣(1,2)- and ␣(1,3/4)-FUTs may share some common roles, unlike ␣(1,6)-FUT and O-fucosylation–related POFUTs, in the pathogenesis of RA. We therefore determined whether this is cell-type specific. RA synovial fluid cells (n ⫽ 5 patients) were sorted into classic monocytes, M1 and M2 macrophages, total CD4 T cells, naive CD4 T cells, memory CD4 T cells, and Th1 and Th17 cells (as defined in Materials and Methods). RA synovial fibroblasts were isolated from human RA synovial fragments or synovial tissue by culture (⬍3 passages). Real-time PCR revealed that terminal ␣(1,2)-FUTs and subterminal ␣(1,3/4)FUTs, including FUT1, FUT3, FUT7, and FUT9, were mainly expressed by M1 inflammatory macrophages, whereas ␣ (1,6)-FUT8 and O-fucosylation–related POFUT2 were mainly expressed by CD4 T cells and fibroblasts (Figure 2A). Among the CD4 T cell subsets, CD45RO⫹ memory CD4 T cells expressed significantly higher levels of POFUT2 and ␣(1,6)-FUT8, as compared to CD45RA⫹ naive CD4 T cells (Figure 2A). We consistently observed a 5–10-fold upregulation of the expression of terminal and subterminal

2372

LI ET AL

Figure 3. Effects of 2-deoxy-D-galactose (2-D-gal) and fucose on type II collagen (CII)–induced arthritis (CIA). A, CIA was induced in DBA/1J mice, followed by intraperitoneal injections of 2-D-gal or fucose (each 250 mg/kg body weight), or normal saline as control, every 2 days from day 0 to day 60. The clinical scores of arthritis (range 0–4 for each limb) were assessed daily. B, Serum levels of anti-CII antibodies, tumor necrosis factor ␣ (TNF␣), and interleukin-6 (IL-6) in each group were determined by enzyme-linked immunosorbent assay on day 14 of arthritis. C, The hind limbs of mice with CIA (representative gross images on the left) were assessed for histopathologic features of arthritis using hematoxylin and eosin staining (right). Bone erosion is indicated by arrows. Bars ⫽ 100 ␮m. D, Scores for the extent of bone erosion and synovitis were quantified in the knee joints of mice in each group. Values are the mean ⫾ SEM of 8 mice per group. ⴱ ⫽ P ⬍ 0.01 versus control.

FUTs in M1-differentiated inflammatory macrophages as compared to their precursors, classic monocytes (Figure 2A). This finding strongly indicates that the terminal ␣(1,2)-FUTs and subterminal ␣(1,3/4)-FUTs might play an indispensable role in the commitment, programming, and functional regulation of the M1 inflammatory macrophages, and that this may contribute to the pathogenesis of RA. We therefore examined whether the expression levels of ␣(1,2)-FUTs and ␣(1,3/4)-FUTs can be correlated with the expression levels of TNF in RA and OA synovial tissue. We identified a strong positive correlation between the expression levels of TNF and the expression levels of FUT1, FUT3, FUT7, and FUT9 (Figure 2B), as well as FUT2, FUT4, FUT5, FUT6, FUT10, FUT11, and POFUT1 (results not shown). In contrast, the expression of ␣ (1,6)-FUT8 and O-fucosylation–related POFUT2 was not correlated with the levels of TNF in either RA or OA synovial tissue (Figure 2B). Effects of inhibition of fucosylation on the development of CIA. Our results indicated that the expression of terminal FUTs is up-regulated in M1 inflammatory macrophages in RA. Therefore, it is conceivable that inhibition of these FUTs might have an impact on the development of arthritis. To investigate this, a fucose/ galactose analog, 2-D-gal, was used to inhibit fucosyla-

tion in a murine model of CIA. It has been shown that upon cellular uptake, 2-D-gal can be converted via the Leloir pathway to the activated UDP-2-deoxy-galactose (21,22). This compound can then compete with UDPgalactose for incorporation into glycan chains and prevent the formation of the Fuc␣(1,2)-gal linkage, making it an ␣(1,2)-fucosylation inhibitor (21,22). It has been demonstrated that 2-D-gal can inhibit the fucosylation of synapsin Ia/Ib (23). For these studies, CIA was induced in DBA/1J mice by immunization with bovine CII. The mice then received IP injections of 2-D-gal (250 mg/kg BW), fucose (250 mg/kg BW), or normal saline every 2 days, starting from day 0 until the mice were killed. Over the treatment period, we observed a highly significant inhibition of the development of CIA in mice receiving 2-D-gal compared to normal saline–treated control mice (Figure 3A). Interestingly, administration of equivalent amounts of fucose facilitated disease development and increased the severity of arthritis when compared to that in the control group (Figure 3A), which suggests that in mice under conditions of inflammation associated with upregulation of FUTs, a relative deficiency of fucose might occur. Serum levels of anti-CII antibodies, TNF␣, and IL-6 were determined in mice 2 weeks after priming with

INHIBITION OF FUCOSYLATION IN ARTHRITIS

2373

Figure 4. Effects of 2-deoxy-D-galactose (2-D-gal) and fucose on cellularity, inflammatory macrophage infiltration, and development of pathogenic CD4 T cells in the draining lymph nodes (LNs) of mice with type II collagen–induced arthritis (CIA). A, Total cell counts in the draining LNs of mice with CIA were determined in the 2-D-gal, fucose, and normal saline (control) treatment groups, as well as in untreated naive mice. B and C, Flow cytometry was performed to analyze inflammatory macrophages, including CD11b⫹Ly-6C⫹, CD11b⫹ tumor necrosis factor ␣ (TNF␣)– positive, and CD11b⫹ interleukin-23 (IL-23)–positive subsets in the draining LNs of mice in each group (B) or CD4 T cell subsets of IL-17A⫹ (Th17) T cells, interferon-␥ (IFN␥)–positive (Th1) T cells, and double-positive Th1/Th17 cells (C). D, The total numbers of cells in each inflammatory macrophage subset and CD4 T cell subset were determined in each group by multiplying the total cell count by the percentage of each cell subpopulation. Values are the mean ⫾ SEM of 8 mice per group. ⴱ ⫽ P ⬍ 0.05 versus control.

CII. The serum levels of all 3 were significantly decreased in the 2-D-gal–treated group compared to the control and fucose-treated groups of mice (Figure 3B). Moreover, histologic evaluation of the mouse joints revealed a significant decrease in synovial hyperplasia, inflammatory cell infiltration, and bone erosion, as indicated by hematoxylin and eosin staining, in 2-D-gal– treated mice compared to control mice. In fucosetreated mice, these histopathologic features of arthritis were more severe when compared to control mice (Figures 3C and D). Decreased M1 inflammatory macrophages and Th17 T cells by inhibition of fucosylation in mice with CIA. Selectin-dependent leukocyte adhesion regulated by Fut4/7 is one important function of fucosylation (24–27). Gene-knockout studies have shown that ␣(1,3)Fut4 and ␣(1,3)-Fut7 are essential for the recruitment of neutrophils and T cells to sites of inflammation and lymphocyte trafficking to secondary lymphoid organs. In addition, leukocytosis is the major phenotype in Fut4/7knockout mice (24–27). In mice treated with 2-D-gal, which inhibits the formation of Fuc␣(1,2)-gal, no significant leukocytosis was observed (results not shown). This suggests that selectin fucosylation might not be the mechanism contributing to the effects of 2-D-gal in mice with CIA. The total cell counts in the bone marrow and spleens of 2-D-gal–treated mice were not significantly

different from those in control mice, whereas there was a significant reduction of cellularity in the draining lymph nodes (LNs) of 2-D-gal–treated mice (Figure 4A). Because the expression of ␣(1,2)-FUTs is specific to M1 macrophages and exhibits much lower expression in monocytes and other cells, including different CD4 T cell subsets, it is possible that terminal fucosylation might play a critical role in M1 inflammatory macrophage development. Consistent with this, we found that there was a decrease in the percentages of CD11b⫹Ly-6C⫹, CD11b⫹TNF␣⫹, and CD11b⫹IL23⫹ M1 inflammatory macrophages in mice treated with 2-D-gal. Furthermore, the absolute cell counts of these inflammatory macrophages were significantly reduced in 2-D-gal–treated mice. In contrast, administration of fucose resulted in a significant increase in M1 inflammatory macrophages (Figures 4B and D, upper panel). Consistent with these inhibitory effects on M1 macrophages, 2-D-gal treatment also resulted in a significant decrease in the percentage and total number of CD4⫹IL-17A⫹ Th17 cells as well as CD4⫹IFN␥⫹ Th1 cells in mice with CIA (Figures 4C and D, lower panel). In contrast to the effects of 2-D-gal, treatment with fucose resulted in a significant increase in Th17 cells in mice with CIA (Figures 4C and D, lower panel). Impact of fucosylation inhibition on macrophage differentiation and plasticity. To explore whether fucosylation regulates the differentiation of inflammatory

2374

LI ET AL

Figure 5. Treatment with 2-deoxy-D-galactose (2-D-gal) regulates the plasticity of inflammatory macrophages. A, The viability of mouse bone marrow (BM)–derived M1 inflammatory macrophages (M␾) and human rheumatoid arthritis (RA) synovial M1 inflammatory macrophages (along with human RA synovial fibroblasts for comparison) was assessed in cultures of cells incubated with granulocyte–macrophage colony-stimulating factor (GM-CSF) and treated without (control) or with a high dose (15 mM) of 2-D-gal or fucose for 7 days. Representative findings from light microscopy (left) and cell count assays assessing cell viability (right) are shown. B, The viability of RAW264.7 cells was assessed in cultures of cells treated with various concentrations of 2-D-gal or osmolarity controls. C, M1 and M2 macrophages were prepared using GM-CSF and M-CSF, respectively, without or with various concentrations of 2-D-gal for 7 days. The viability of the cells was determined on day 7. D, M1 and M2 macrophages, prepared and treated as described in C, were stimulated with 200 ng/ml of lipopolysaccharide (LPS) 16 hours before analysis. The effects of 2-D-gal on the levels of interleukin-10 (IL-10) (left) and interferon-␥ (IFN␥) (middle) were assessed in the culture supernatants on day 7. Flow cytometry histograms show the effects of 2-D-gal on the levels of pERK-1/2 in M1 macrophages (right). Values are the mean ⫾ SEM of 3 samples per group. ⴱ ⫽ P ⬍ 0.05; ⴱⴱ ⫽ P ⬍ 0.01 versus untreated control.

M1 macrophages, mouse bone marrow–derived M1 macrophages were polarized by incubation with GMCSF and then treated with or without 15 mM 2-D-gal or fucose from day 0. On day 6, 2-D-gal–treated mouse bone marrow cells failed to survive and differentiate, as indicated by the findings on light microscopy and cell count assay (Figure 5A, upper panels). Similar results were observed in human RA synovial fluid macrophages treated with 2-D-gal (Figure 5A, middle panels). In contrast, treatment with 2-D-gal did not have an impact on the survival of human RA synovial fibroblasts (Figure 5A, lower panels). Interestingly, treatment with fucose did not have a significant effect on either mouse or human cells (Figure 5A), possibly because of the redundancy of endogenous fucose generated by the cells. Treatment with 2-D-gal at a concentration of 15 mM could potentially alter the extracellular osmolarity to a significant extent, and thus alter the phenotype of macrophages. However, this possibility was ruled out,

since equal molar concentrations of fucose, glucose, NaCl, and glycerol did not have the capability of reducing the viability of RAW264.7 macrophages when compared to the effects of 2-D-gal (Figure 5B). Our findings therefore suggest that the observed effect was specific to 2-D-gal and M1 inflammatory macrophages. Further analysis of the impact of various concentrations of 2-D-gal on the differentiation of M1 inflammatory macrophages and M2 antiinflammatory macrophages showed that lower concentrations of 2-D-gal (0.01, 0.1, and 1 mM) did not exhibit an effect on the viability of either M1 or M2 macrophages (Figure 5C). However, at higher concentrations (5 and 15 mM), 2-D-gal selectively reduced the cell number of M1 macrophages without inhibiting the growth of M2 macrophages (Figure 5C). These data suggest that higher concentrations of 2-D-gal can specifically preclude the differentiation of macrophages under conditions of inflammatory M1 polarization.

INHIBITION OF FUCOSYLATION IN ARTHRITIS

2375

Figure 6. Effects of 2-deoxy-D-galactose (2-D-gal) on the antigen uptake, processing, and presentation functions of inflammatory macrophages. A, Mouse bone marrow (BM)–derived M1 macrophages were differentiated by using granulocyte–macrophage colony-stimulating factor (GM-CSF) for 6 days, followed by treatment with various doses of 2-D-gal for 2 days. On day 8, the viability of the cells was assessed. B, Flow cytometry analysis was used to assess the ovalbumin (OVA) DQ–labeled antigen uptake (upper) and green fluorescent protein (GFP)–labeled E␣ antigen uptake and/or processing (lower) in 2-D-gal–treated M1 macrophages. C, The expression of CD86 (upper) and intracellular adhesion molecule (ICAM) (lower) was assessed by flow cytometry analysis in GM-CSF–polarized BM-derived macrophages 2 days after treatment with or without 2-D-gal. D, The proliferative response of CD4 T cells from 2D2 T cell receptor (TCR)–transgenic mice was assessed by flow cytometry analysis. M1 macrophages were pretreated with or without 2-D-gal for 2 days. The cells were then loaded with myelin oligodendrocyte glycoprotein 35–55 peptide and cultured for an additional 3 days with carboxyfluorescein succinimidyl ester (CFSE)–labeled CD4 T cells isolated from 2D2 TCR–transgenic mice. Values are the mean ⫾ SEM of 3 samples per group. APC ⫽ antigen-presenting cell. ⴱ ⫽ P ⬍ 0.05; ⴱⴱ ⫽ P ⬍ 0.01, versus the other conditions.

To determine whether 2-D-gal is capable of regulating cytokine expression in macrophages without altering the cell viability, we focused on the lower concentrations of 2-D-gal (0, 0.01, 0.1, and 1 mM), which did not reduce cell density. Strikingly, 2-D-gal induced, in a dose-dependent manner, an augmentation of antiinflammatory IL-10 production in macrophages under conditions of both M1 and M2 polarization (Figure 5D, left). IL-10 has been shown to be a critical cytokine that suppresses arthritis development (28,29). There was a gradual decline in the generation of proinflammatory IFN␥ in 2-D-gal–treated cells within the same dose range (Figure 5D, middle). Because of the polarizing effects of these 2 cytokines on macrophages (30), a reduction of IFN␥ and augmentation of IL-10 production from macrophages by 2-D-gal can further drive them toward the M2 antiinflammatory phenotype, in an autocrine manner. Our results strongly indicated that inhibition of fucosylation can regulate macrophage plasticity and skew the differentiation of macrophages toward an antiinflammatory phenotype. It has been shown that activation of ERKs is

one critical factor modulating the expression of IL-10 in macrophages (31). Consistent with this, phospho-flow cytometry analysis indicated that 2-D-gal treatment upregulated the expression of pERK-1/2 in M1 inflammatory macrophages (Figure 5D, right). Suppression of antigen processing and presentation in inflammatory macrophages by fucosylation inhibition. The antigen-presentation function is one hallmark of M1 macrophages, and there are dynamic interactions of macrophages with T cells during antigen presentation (32). Glycosylation has previously been implicated in major histocompatibility complex (MHC) processing and antigen presentation. Specifically, the loss of N-glycan on class II MHC significantly impairs the ability of antigen-presenting cells (APCs) to present bacterial polysaccharides to T cells (33). To determine whether and how the inhibition of fucosylation affects the antigen-presenting functions of macrophages, antigen uptake, processing, and presentation in the context of class II MHC was investigated. Bone marrow–derived M1 macrophages were generated in vitro by incubating mouse bone marrow–derived cells with GM-CSF. Fully

2376

differentiated macrophages were then treated with various concentrations of 2-D-gal on day 6 through day 8, which did not alter the viability of the cells (Figure 6A). For analysis of antigen uptake and processing, M1 macrophages were pulsed with OVA DQ, which exhibits fluorescence upon proteolytic degradation. Uptake and/or proteolysis of OVA DQ by M1 macrophages were diminished following treatment with 2-D-gal at 0.01 and 0.1 mM, and the reductions were significant in the presence of 1, 5, and 15 mM of 2-D-gal (Figure 6B, upper panels). To further investigate the effect of 2-D-gal on antigen uptake, processing, class II MHC loading, and surface presentation, a recombinant protein, GFPlabeled E␣, consisting of amino acids 46–74 of the I-Ed␣ class II MHC subunit and a GFP tag, was utilized (19,34). The inherent GFP was used to detect the presence of the peptide after uptake. The processed I-E␣ peptide can be detected using the Y-Ae monoclonal antibody, which is specific for the complex composed of the I-E␣ peptide 52–68 (pE␣) bound to the I-Ab class II MHC. In the presence of low doses of 2-D-gal (0.01, 0.1, and 1 mM), there was a significant decrease in the processing and presentation of the I-E␣ peptide, as indicated by a decreased percentage of surface Y-Ae⫹ cells (Figure 6B, lower panels). Treatment with higher doses of 2-D-gal (5 and 15 mM) resulted in decreased antigen uptake, as indicated by the percentage of GFP⫹ cells (Figure 6B, lower panels). Initial contact between APCs and T cells is established between adhesion molecules, including lymphocyte function–associated antigen 1 on T cells and intercellular adhesion molecule 1 (ICAM-1) on APCs (35). Interaction between the B7 costimulatory molecules CD80 and CD86 on APCs and CD28 on T cells has been shown to provide a critical signal for T cell activation, while the lack of this signal results in T cell anergy (36). Interestingly, the expression of CD86 and ICAM by mouse bone marrow–derived macrophages was reduced with 2-D-gal treatment. This occurred in a dose-dependent manner, with significant reductions in the levels of CD86 and ICAM in the presence of 2-D-gal at concentrations of 1, 5, and 15 mM (Figure 6C). In contrast, the expression of CD80 was not changed (results not shown). Consistently, the ability of APCs loaded with MOG35–55 peptide to stimulate the proliferation of CD4 T cells from 2D2 mice was reduced or abrogated when the APCs were pretreated with 2-D-gal (Figure 6D). These findings suggest that 2-D-gal not only suppresses antigen uptake and presentation, but also disrupts the interac-

LI ET AL

tion between APCs and T cells by down-regulating adhesion and costimulatory molecules on APCs. DISCUSSION The fucosylated carbohydrate structure plays an important role in host–microbe interactions, selectindependent leukocyte adhesion, tissue development, and tumor metastasis (1,4). However, how fucosylation is involved in the pathogenesis of RA is not known. Our results indicate that in RA synovial tissue, as compared to OA synovial tissue, there was a significant increase in the expression of FUT1, FUT2, FUT4, FUT5, FUT6, FUT7, FUT9, FUT10, and FUT11. In contrast, FUT8 and FUTs 12 and 13 (POFUT1 and POFUT2) were not up-regulated in the synovial tissue from patients with RA. The expression of terminal and subterminal FUTs was highly correlated with that of TNF in the synovial tissue. Fucosylation inhibition skewed the M1 inflammatory macrophages to a M2 antiinflammatory phenotype. These results strongly suggest that subsets of FUTs are required for commitment to and maintenance of the proinflammatory phenotypes of inflammatory M1 macrophages. Specific FUTs can have overlapping enzymatic activity. FUT1 and FUT2 both catalyze the addition of fucose to the galactose moiety of Gal␤(1,4)-GlcNAc or Gal␤(1,3)-GlcNAc structures, which are usually localized at the termini of the glycan. To investigate the roles of FUT1 and FUT2 in RA, the Fut1/2 doubleknockout mouse model is one potential tool for carrying out the studies, because Fut1 or Fut2 single-knockout does not completely eliminate the activity of ␣(1,2)fucosyltransferases. However, this approach is not feasible due to the lethal effects on embryos in doubleknockout mice. Therefore, in the current study, we took advantage of utilizing the fucose analog 2-D-gal, which can inhibit the formation of the Fuc␣(1,2)-gal linkage (23), and thereby fulfilled the goal of ruling out the activity of both FUT1 and FUT2. Treatment with 2-D-gal led to a dramatic inhibition of arthritis in mice with CIA, whereas fucose facilitated the development and severity of CIA. Serum levels of TNF␣, IL-6, and anti-CII antibodies were significantly reduced by 2-D-gal. Immunization with CII plus CFA promoted an immune response in the draining LNs of mice, including induction of inflammatory macrophages, antigen processing and presentation to T cells, and development of adaptive Th1, Th17, and anti-CII responses (37–39). Our data suggested that the inhibition of terminal fucosylation with 2-D-gal significantly

INHIBITION OF FUCOSYLATION IN ARTHRITIS

suppressed the immune responses in mice following priming with CII plus CFA. One key scenario that was altered by 2-D-gal might be the antigen presentation of M1 macrophages. In the mice with CIA treated with 2-D-gal, we consistently observed a remarkable reduction in Th17 cells in the draining LNs, in addition to the reduction in inflammatory macrophages. It was unlikely that 2-D-gal had a direct impact on Th17 cells, since the levels of FUT1 and FUT2 were low in Th17 cells. We have previously reported that there is a bidirectional interaction between M1 inflammatory macrophages and Th17 cells in RA and other autoimmune diseases (12,40). Therefore, it is possible that the decrease in Th17 cells was indirectly caused by the elimination of M1 inflammatory macrophages. It is also possible that other unidentified mechanisms, such as microbiota, might be involved, because fucose has been shown to be an important signal that regulates bacterial intestinal colonization (41). Macrophages exhibit remarkable diversity and plasticity in response to environmental cues, undergoing a “spectrum” of activation, including M1 (classic or proinflammatory) and M2 (alternative or antiinflammatory) as two extremes (42). It has been shown that the phenotypes of polarized M1 and M2 macrophages can be reversed in vitro and in vivo, which is associated with the persistence and resolution of inflammation dynamically (43,44). A network of transcription factors, signaling molecules, epigenetic mechanisms, and posttranscriptional regulators contributes to the activation and plasticity of macrophages, including interferon regulatory factors/STATs, NF-␬B, the suppressor of cytokine signaling family of proteins, peroxisome proliferator– activated receptors, Kruppel-like factors 2 and 4, c-Myc, hypoxia-inducible factors 1␣ and 2␣, histone demethylase, Jumonji histone demethylase 3, and microRNA-155 (9). The current study identified fucosylation as a novel and key mechanism regulating macrophage plasticity. Our preliminary data, generated using Ulex Europaeus Agglutinin I pull-down followed by liquid chromatography mass spectrometry, further suggested that fucosylation of histones might be involved in this process. Thirteen FUT genes have been identified in mammals. They are responsible for ␣(1,2)-, ␣(1,3)-, ␣(1,4)-, ␣(1,6)-, and O-fucosylation on various surfacelocalized and secreted molecules (4,45). Yao and colleagues have demonstrated that O-fucosylation of Notch catalyzed by POFUT1 regulates lymphoid and myeloid homeostasis (46). The present study was focused on the effect of 2-D-gal, which inhibits the formation of

2377

Fuc␣(1,2)-gal linkage (23). In mice that received longterm treatment with 2-D-gal for 4 months, we did not observe a significant change in the levels of bone marrow–derived myeloid progenitors, including common myeloid progenitors and granulocyte–macrophage progenitors, when compared to those in control mice (results not shown). The size of the thymus, counts of different subsets of thymocytes, including CD4 SP, CD8 SP, DN1, DN2, DN3, and DN4, were not altered by long-term 2-D-gal treatment (results not shown). Therefore, it is unlikely that 2-D-gal interfered with the O-fucosylation of Notch molecules. One of the most important functions of fucosylation is cell adhesion and migration. Interaction between selectins and the fucosylated structures on their ligands is essential in this process. Two FUTs, FUT4 and FUT7, are responsible for the ␣(1,3)-fucosylated reaction on the selectin ligand (24,27). The expression of both FUT4 and FUT7 is up-regulated in RA synovial tissue compared to OA synovial tissue, which suggests that FUT4 and FUT7 might play a role in the pathogenesis of RA. However, it is unlikely that 2-D-gal is also having an effect on these 2 FUTs, based on the following reasons: 1) 2-D-gal prevents the ␣(1,2)-fucosylated reaction to the galactose moiety of Gal␤(1,4)-GlcNAc or Gal␤(1,3)GlcNAc structures, and 2) blood leukocytosis, a major phenotype of Fut4⫺/⫺ and Fut7⫺/⫺ mice, was not observed in mice treated long-term with 2-D-gal. This study was the first to demonstrate that terminal fucosylation is one hallmark of M1 inflammatory macrophages and plays a determining role in their lineage commitment and differentiation. Fucosylation inhibition can functionally skew the M1 inflammatory macrophages toward an antiinflammatory phenotype, leading to the resolution of inflammation by reshaping their cytokine profiles, suppressing the expression of costimulatory molecules, and inhibiting antigen uptake and presenting functions. The results of the present study not only extend our understanding of macrophage plasticity to a posttranslational level, but also suggest that fucosylation is a novel therapeutic target for the management of patients with RA. ACKNOWLEDGMENT We thank Dr. Marc K. Jenkins (University of Minnesota) for kindly providing the GFP-labeled E␣ peptide. AUTHOR CONTRIBUTIONS All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved

2378

LI ET AL

the final version to be published. Dr. Mountz had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study conception and design. J. Li, Hsu, Mountz. Acquisition of data. J. Li, Hsu, Ding, H. Li, Wu, Yang, Luo, Rowse, Spalding, Bridges, Mountz. Analysis and interpretation of data. J. Li, Hsu, Ding, H. Li, Wu, Yang, Luo, Rowse, Spalding, Bridges, Mountz.

ROLE OF THE STUDY SPONSOR Daiichi Sankyo Co. had no role in the study design or in the collection, analysis, or interpretation of the data, the writing of the manuscript, or the decision to submit the manuscript for publication. Publication of this article was not contingent upon approval by Daiichi Sankyo Co.

REFERENCES 1. Ma B, Simala-Grant JL, Taylor DE. Fucosylation in prokaryotes and eukaryotes. Glycobiology 2006;16:158R–84R. 2. Gornik I, Maravic G, Dumic J, Flogel M, Lauc G. Fucosylation of IgG heavy chains is increased in rheumatoid arthritis. Clin Biochem 1999;32:605–8. 3. Ryden I, Pahlsson P, Lundblad A, Skogh T. Fucosylation of alpha1-acid glycoprotein (orosomucoid) compared with traditional biochemical markers of inflammation in recent onset rheumatoid arthritis. Clin Chimica Acta 2002;317:221–9. 4. Becker DJ, Lowe JB. Fucose: biosynthesis and biological function in mammals. Glycobiology 2003;13:41R–53R. 5. Szekanecz Z, Koch AE. Macrophages and their products in rheumatoid arthritis. Curr Opin Rheumatol 2007;19:289–95. 6. Hamilton JA, Tak PP. The dynamics of macrophage lineage populations in inflammatory and autoimmune diseases [review]. Arthritis Rheum 2009;60:1210–21. 7. Gierut A, Perlman H, Pope RM. Innate immunity and rheumatoid arthritis. Rheum Dis Clin North Am 2010;36:271–96. 8. Kinne RW, Stuhlmuller B, Burmester GR. Cells of the synovium in rheumatoid arthritis: macrophages. Arthritis Res Ther 2007;9: 224. 9. Sica A, Mantovani A. Macrophage plasticity and polarization: in vivo veritas. J Clin Invest 2012;122:787–95. 10. Rosloniec EF, Cremer M, Kang A, Myers LK. Collagen-induced arthritis. In: Coligan JE, Kruisbeek AM, Margulies DH, Shevach EM, Strober W, editors. Current protocols in immunology. Somerset (NJ): John Wiley & Sons; 2001. Chapter 15: Unit 15.5. 11. Xu X, Hsu HC, Chen J, Grizzle WE, Chatham WW, Stockard CR, et al. Increased expression of activation-induced cytidine deaminase is associated with anti-CCP and rheumatoid factor in rheumatoid arthritis. Scand J Immunol 2009;70:309–16. 12. Li J, Yang P, Wu Q, Li H, Ding Y, Hsu HC, et al. Death receptor 5–targeted depletion of interleukin-23–producing macrophages, Th17, and Th1/17 associated with defective tyrosine phosphatase in mice and patients with rheumatoid arthritis. Arthritis Rheum 2013;65:2594–605. 13. Arnett FC, Edworthy SM, Bloch DA, McShane DJ, Fries JF, Cooper NS, et al. The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis. Arthritis Rheum 1988;31:315–24. 14. Ichikawa K, Liu W, Fleck M, Zhang H, Zhao L, Ohtsuka T, et al. TRAIL-R2 (DR5) mediates apoptosis of synovial fibroblasts in rheumatoid arthritis. J Immunol 2003;171:1061–9. 15. Li J, Hsu HC, Yang P, Wu Q, Li H, Edgington LE, et al. Treatment of arthritis by macrophage depletion and immunomodulation: testing an apoptosis-mediated therapy in a humanized death receptor mouse model. Arthritis Rheum 2012;64:1098–109.

16. Cohen CJ, Crome SQ, MacDonald KG, Dai EL, Mager DL, Levings MK. Human Th1 and Th17 cells exhibit epigenetic stability at signature cytokine and transcription factor loci. J Immunol 2011;187:5615–26. 17. Wu Y, Liu J, Feng X, Yang P, Xu X, Hsu HC, et al. Synovial fibroblasts promote osteoclast formation by RANKL in a novel model of spontaneous erosive arthritis. Arthritis Rheum 2005;52: 3257–68. 18. Lari R, Fleetwood AJ, Kitchener PD, Cook AD, Pavasovic D, Hertzog PJ, et al. Macrophage lineage phenotypes and osteoclastogenesis: complexity in the control by GM-CSF and TGF-␤. Bone 2007;40:323–36. 19. Itano AA, McSorley SJ, Reinhardt RL, Ehst BD, Ingulli E, Rudensky AY, et al. Distinct dendritic cell populations sequentially present antigen to CD4 T cells and stimulate different aspects of cell-mediated immunity. Immunity 2003;19:47–57. 20. Pape KA, Catron DM, Itano AA, Jenkins MK. The humoral immune response is initiated in lymph nodes by B cells that acquire soluble antigen directly in the follicles. Immunity 2007;26: 491–502. 21. Bullock S, Potter J, Rose SP. Effects of the amnesic agent 2-deoxygalactose on incorporation of fucose into chick brain glycoproteins. J Neurochem 1990;54:135–42. 22. Holden HM, Rayment I, Thoden JB. Structure and function of enzymes of the Leloir pathway for galactose metabolism. J Biol Chem 2003;278:43885–8. 23. Murrey HE, Gama CI, Kalovidouris SA, Luo WI, Driggers EM, Porton B, et al. Protein fucosylation regulates synapsin Ia/Ib expression and neuronal morphology in primary hippocampal neurons. Proc Natl Acad Sci U S A 2006;103:21–6. 24. Maly P, Thall A, Petryniak B, Rogers CE, Smith PL, Marks RM, et al. The ␣(1,3)fucosyltransferase Fuc-TVII controls leukocyte trafficking through an essential role in L-, E-, and P-selectin ligand biosynthesis. Cell 1996;86:643–53. 25. Wagers AJ, Lowe JB, Kansas GS. An important role for the ␣1,3 fucosyltransferase, FucT-VII, in leukocyte adhesion to E-selectin. Blood 1996;88:2125–32. 26. Smithson G, Rogers CE, Smith PL, Scheidegger EP, Petryniak B, Myers JT, et al. Fuc-TVII is required for T helper 1 and T cytotoxic 1 lymphocyte selectin ligand expression and recruitment in inflammation, and together with Fuc-TIV regulates naive T cell trafficking to lymph nodes. J Exp Med 2001;194:601–14. 27. Homeister JW, Thall AD, Petryniak B, Maly P, Rogers CE, Smith PL, et al. The ␣(1,3)fucosyltransferases FucT-IV and FucT-VII exert collaborative control over selectin-dependent leukocyte recruitment and lymphocyte homing. Immunity 2001;15:115–26. 28. Huang QQ, Koessler RE, Birkett R, Perlman H, Xing L, Pope RM. TLR2 deletion promotes arthritis through reduction of IL-10. J Leukoc Biol 2013;93:751–9. 29. Lubberts E, Joosten LA, Helsen MM, van den Berg WB. Regulatory role of interleukin 10 in joint inflammation and cartilage destruction in murine streptococcal cell wall (SCW) arthritis: more therapeutic benefit with IL-4/IL-10 combination therapy than with IL-10 treatment alone. Cytokine 1998;10:361–9. 30. Zizzo G, Cohen PL. IL-17 stimulates differentiation of human anti-inflammatory macrophages and phagocytosis of apoptotic neutrophils in response to IL-10 and glucocorticoids. J Immunol 2013;190:5237–46. 31. Saraiva M, O’Garra A. The regulation of IL-10 production by immune cells. Nat Rev Immunol 2010;10:170–81. 32. Underhill DM, Bassetti M, Rudensky A, Aderem A. Dynamic interactions of macrophages with T cells during antigen presentation. J Exp Med 1999;190:1909–14. 33. Ryan SO, Bonomo JA, Zhao F, Cobb BA. MHCII glycosylation modulates Bacteroides fragilis carbohydrate antigen presentation. J Exp Med 2011;208:1041–53. 34. Matz MV, Fradkov AF, Labas YA, Savitsky AP, Zaraisky AG,

INHIBITION OF FUCOSYLATION IN ARTHRITIS

35. 36. 37. 38. 39.

40.

Markelov ML, et al. Fluorescent proteins from nonbioluminescent Anthozoa species. Nat Biotechnol 1999;17:969–73. Shaw AS, Dustin ML. Making the T cell receptor go the distance: a topological view of T cell activation. Immunity 1997;6:361–9. Harding FA, McArthur JG, Gross JA, Raulet DH, Allison JP. CD28-mediated signalling co-stimulates murine T cells and prevents induction of anergy in T-cell clones. Nature 1992;356:607–9. Mauri C, Williams RO, Walmsley M, Feldmann M. Relationship between Th1/Th2 cytokine patterns and the arthritogenic response in collagen-induced arthritis. Eur J Immunol 1996;26:1511–8. Leung BP, McInnes IB, Esfandiari E, Wei XQ, Liew FY. Combined effects of IL-12 and IL-18 on the induction of collageninduced arthritis. J Immunol 2000;164:6495–502. Kai H, Shibuya K, Wang Y, Kameta H, Kameyama T, TaharaHanaoka S, et al. Critical role of M. tuberculosis for dendritic cell maturation to induce collagen-induced arthritis in H-2b background of C57BL/6 mice. Immunology 2006;118:233–9. Li J, Hsu HC, Mountz JD. The dynamic duo: inflammatory M1 macrophages and Th17 cells in rheumatic diseases. J Orthop Rheumatol 2013;1:1–4.

2379

41. Pacheco AR, Curtis MM, Ritchie JM, Munera D, Waldor MK, Moreira CG, et al. Fucose sensing regulates bacterial intestinal colonization. Nature 2012;492:113–7. 42. Mosser DM, Edwards JP. Exploring the full spectrum of macrophage activation. Nat Rev Immunol 2008;8:958–69. 43. Martinez FO, Helming L, Gordon S. Alternative activation of macrophages: an immunologic functional perspective. Annu Rev Immunol 2009;27:451–83. 44. Guiducci C, Vicari AP, Sangaletti S, Trinchieri G, Colombo MP. Redirecting in vivo elicited tumor infiltrating macrophages and dendritic cells towards tumor rejection. Cancer Res 2005;65: 3437–46. 45. Moloney DJ, Haltiwanger RS. The O-linked fucose glycosylation pathway: identification and characterization of a uridine diphosphoglucose: fucose-␤1,3-glucosyltransferase activity from Chinese hamster ovary cells. Glycobiology 1999;9:679–87. 46. Yao D, Huang Y, Huang X, Wang W, Yan Q, Wei L, et al. Protein O-fucosyltransferase 1 (Pofut1) regulates lymphoid and myeloid homeostasis through modulation of Notch receptor ligand interactions. Blood 2011;117:5652–62.