Background
Multiple sclerosis (MS) is a neurological autoimmune disorder of the central nervous system (CNS), characterized by recurrent episodes of inflammatory demyelination, leading to a progressive deterioration of the neurological functions [
1]. Although recent studies have revealed that both innate and adaptive immune cells contribute to the pathogenesis and progression of MS and its in vivo model experimental autoimmune encephalomyelitis (EAE) [
2], there is little information about the mechanisms involved in the resolution of inflammation during remission, a process that leads to clinical improvement of patients. Besides the dampening role of IL-10- and TGF-β-producing regulatory cells [
3], just a handful of mechanisms have been associated with the resolution of the autoimmune neuroinflammation. Recent evidence has linked cytotoxic T-lymphocyte antigen-4 (CTLA-4), programmed death-1 (PD-1), T cell immunoreceptor with immunoglobulin and immunoreceptor tyrosine-based inhibition motif domains (TIGIT), and T cell immunoglobulin- and mucin domain-containing molecule 3 (TIM-3) as important immune regulatory pathways in the resolution of EAE, showing a strong genetic or functional correlation in MS (reviewed in [
4]). Interestingly, all the depicted pathways involve the inhibition of the activity of myelin-reactive T cells.
The macrophage galactose-type lectin (MGL; aka Clec10a or CD301) is a member of the C-type lectin receptor (CLR) family [
5,
6]. This family is composed of a variety of transmembrane and soluble receptors that recognize glycan structures in a Ca
2+-dependent manner through a common carbohydrate recognition domain (CRD) and play important roles in both innate and adaptive immune responses [
6]. MGL is expressed by myeloid antigen-presenting cells (APC) such as dendritic cells (DCs) and macrophages [
5]. Human MGL and its mouse ortholog MGL2 bind
O-glycans decorated with sTn or Tn antigens (sialylated and non-sialylated
N-acetylgalactosamine, αGalNAc), residues, and the LacdiNAc epitope (GalNAcβ1-4GlcNAc). In contrast, the mouse ortholog MGL1 binds Lewis-A and Lewis-X structures [
5,
7].
MGL binding to CD45 on activated CD4
+ T cells mediates effector T cell apoptosis, leading to the dampening of the inflammatory response [
8]; while in the APC compartment, MGL triggering induces IL-10 production and inhibits cell maturation and migration [
9‐
14]. It has been proposed that these effects are exploited by pathogens and tumors as immune escape mechanisms [
5,
7,
11,
13,
15,
16]. At the same time, the immunoregulatory properties of MGL are involved in dampening autoimmune diseases in different settings [
5,
10]. However, its role in autoimmune neuroinflammation has not been reported.
Here, we present data showing that MGL1, but not MGL2, participates in the resolution phase of the disease by inducing apoptosis of autoreactive T cells and diminishing the expression of pro-inflammatory cytokines and autoantibodies. Moreover, in human settings, we show that MGL expression is increased in active MS lesions and on alternatively activated microglia and macrophages which, in turn, induces the secretion of the immunoregulatory cytokine IL-10. Altogether, these data demonstrate that MGL acts as a negative regulator in autoimmune-induced neuroinflammation, uncovering a novel clue for neuroprotective therapeutic strategies with relevance in MS clinical applications.
Methods
Mice
MGL1-deficient mice (
Clec10a−/−; C57BL/6) [
17] were provided by the Consortium for Functional Glycomics (
www.functionalglycomics.org). C57BL/6 mice were purchased from Charles River Laboratories (Maastricht, The Netherlands). Mice were bred and housed at the Amsterdam Animal Research Center (VU/VUmc) under specific pathogen-free conditions. All experiments were approved by the Animal Experiments Committee of the VUmc and performed in accordance with the national and international guidelines and regulations.
Reagents
MGL1 and MGL2 were generated by fusing the extracellular domain of each lectin to a human IgG1-Fc tail as described [
7], fluorescein isothiocyanate (FITC)-annexin V staining (BD Biosciences), unlabeled anti-MGL (clones 18E4 and 1G6.6 as described [
8]), phycoerythrin (PE)-anti-MGL (Miltenyi, REA586), anti-mMHC II (clone LN3, kind gift from Amsterdam UMC, location VUmc Amsterdam, department of pathology, the Netherlands), and anti-P2Y12R (Anaspec).
Generation 2 cystamine core Gal and GalNac dendrimers
Gal-Gal or GalNac-Gal (Elicityl) were conjugated to generation 2 cystamine core PAMAM dendrimers (Sigma-Aldrich) via reductive amination. Briefly, 1 equivalent of dendrimer was mixed with 40 equivalents of glycan in 200 μl acetic acid-dimethyl sulfoxide mixture (20% and 80%, respectively, both from Sigma-Aldrich). After thoroughly mixing, 80 equivalents of 2-picoline-borane complex (Sigma-Aldrich) was added and incubated at 65 °C for 3 h. For dendrimer glycation, a mixture of dichloromethane-diethyl ether (1:1, both from Sigma-Aldrich) was added and thoroughly vortexed. Glycated dendrimers were pelleted by centrifugation at 14000 rpm for 1 min. The supernatant was discarded, and the pellet was washed 3 times with diethyl ether (Sigma-Aldrich) and pelleted again by centrifugation. Pellet was dissolved in MilliQ water and lyophilized. Then, lyophilized glycated dendrimers were dissolved and purified over Superdex 200 10/300 GL column (GE Healthcare Life Sciences) on an Ultimate 3000 HPLC system (ThermoFisher).
Mouse T cells
Naïve CD4+ T cells (CD62L+ CD44low) were isolated from mouse spleens using the naive CD4+ T Cell Isolation Kit (Miltenyi Biotec) according to the manufacturer’s protocol and stimulated (2 × 106 cells/ml) with plate-bound anti-CD3 (3 μg/ml; clone 145-2C11; BD Biosciences) and soluble anti-CD28 (1 μg/ml; clone 37.51; BD Biosciences) for 3 days.
Cell death assays
Activated CD4+ T cells (1 × 106 cells/ml) were incubated with 10 μg/ml MGL1− and/or MGL2-Fc in RPMI for different time periods. Apoptotic cells were identified by AV/7-AAD staining. Cell death was determined as the difference in the percentage of AV and 7-AAD double-negative cells between polarized cells treated with or without each stimulus. For ex vivo experiments with cells from mice with EAE, draining lymph node (DLN) cells were stimulated ON in the presence or absence of 10 μg/ml MGL1− and/or MGL2-Fc in RPMI 10% FCS. Then, cells were stained and analyzed as stated above.
Induction and assessment of EAE in mice
EAE was induced in 8–12-week-old female Clec10a−/− and wild-type mice (C57BL/6) by s.c. immunization with 150 μg mouse MOG35–55 (MEVGWYRSPFSRVVHLYRNGK; Cambridge Research Biochemicals) in complete Freund’s adjuvant (CFA) supplemented with 4 mg/ml Mycobacterium tuberculosis (H37RA; Difco). Control animals were injected with a 1:1 PBS/CFA mixture. All animals received 200 ng pertussis toxin (Sigma) i.p. on days 0 and 2. Mice were examined daily for signs of EAE and scored as follows: 0, no clinical signs; 0.5, half limp tail; 1, complete limp tail; 1.5, lack of toe-spreading reflex; 2, half hind limb weakness; 2.5, hind limb weakness; 3, half hind limb paralysis; 3.5, incomplete hind limb paralysis; 4, complete hind limb paralysis; 4.5, diaphragmatic paralysis/paralysis of (one of the) front legs; and 5, death by EAE. At day 27, proliferation was determined in antigen-specific splenocytes and draining lymph node (DLN) cells by [3H]-thymidine incorporation following ex vivo restimulation with 25 μg/ml MOG35–55. Cytokine production was determined in supernatants following 72-h antigen restimulation by ELISA. The IL-17 ELISA kit was from R&D.
Rat acute EAE
We used EAE data acquired from an independent study performed in our laboratory, and the acute EAE was induced as described previously [
18]. For microarray analysis, 1 μg of total RNA was linearly amplified (at ServiceXS) by T7 RNA amplification, and Cy3 or Cy5 was incorporated during the cDNA synthesis according to the manufacturer’s instructions (Agilent Technologies). Equal amounts of Cy3- and Cy5-labeled samples were hybridized 17 h on a rat Agilent oligo microarray. For the cerebellum, samples of two EAE animals and two CFA control animals per time point were hybridized separately in a loop-style experimental setup, using four microarrays per time point. Because of the small sample size, the samples of the brainstem of two animals per time point were pooled after RNA isolation and hybridized in a direct dye swap, using two microarrays per time point. The arrays were scanned with an Agilent G2565AA dual-laser microarray scanner. The resulting images were analyzed with the Agilent Feature Extraction Software (
www.agilent.com). In brief, in the first step, outliers were detected, then the values were corrected for background and normalized using the linear/Lowess method as described in the Agilent feature extraction manual. The resulting intensities of the spots were used for the calculation of absolute difference and ratios for EAE vs. CFA control animals. The data analysis was performed using the Spotfire software for functional genomics, selecting genes by filtering on ratio and difference. We considered a gene up- or downregulated if the change in gene expression was visible in all four different hybridizations with a ratio EAE vs. CFA control of minimal 1.5. For further analysis, clinical scores were normalized. Minimum clinical scores were set at 0%, whereas maximum clinical scores were set at 100%. Similarly, fold changes in mRNA expression were normalized for all individual genes, with a minimum fold change in the course of EAE set at 0% and maximum fold change at 100%. Subsequently, relative least square differences (variance score Σ(clin. score − gen. score)2/clin. score) between normalized clinical scores and normalized fold changes were calculated for each gene. This parameter allowed filtering of data on the basis of variations in gene expression with respect to the clinical scores.
Determination of anti-MOG IgG levels
The blood was drawn at 27 dpi, and antigen-specific serum antibody titers were measured by ELISA. Briefly, ELISA plates were coated with 10 μg/ml MOG35–55 in PBS, and the MOG-specific IgG serum antibody titer was measured using rabbit anti-mouse IgG HRP-linked antibody (DAKO) and biotinilated goat anti-mouse IgG1 and anti-mouse IgG2c (Jackson ImmunoResearch). The end point dilution was 2 times of the blank value.
Brain tissue
In collaboration with The Netherlands Brain Bank (Amsterdam, The Netherlands, coordinator Dr. I. Huitinga), we used human post-mortem brain tissue from three non-neurological controls and eight MS patients (see [
19] for patient details). The study was approved by the institutional ethics review board (VU University Medical Center, Amsterdam, The Netherlands), and all donors or their next of kin provided written informed consent for brain autopsy and use of material and clinical information for research purposes. Lesion types were determined by proteolipid protein (PLP) and MHC II staining (Additional file
1: Figure S1).
Microglia isolation and culture
As described by Beaino et al. [
19], briefly, 5 to 10 g of brain white matter was obtained at autopsy, and microglia isolation procedure was performed within 4 to 24 h. Single-cell suspensions were prepared using 0.05% trypsin (Sigma-Aldrich). Cells were then filtered through a 100-μm nylon mesh (BD Bioscience), centrifuged, and the pellet was resuspended in a gradient buffer (3.56 g/l Na
2HPO
4·2H
2O, 0.78 g/l NaH
2PO
4·H
2O, 8 g/l NaCl, 4 g/l KCl, 2 g/l d-(+)-glucose, and 2 g/l BSA, pH 7.4) and centrifuged for 35 min at 1200×
g. The cell debri-myelin layer was removed. The cell pellet was treated with red blood cell lysis buffer (8.3 g/l NH
4Cl and 1 g/l KHCO
3, pH 7.4) for 10 min at 4 °C. Cells were then suspended and cultured in DMEM/Ham nutrient mixture F10 (1:1), 1% penicillin-streptomycin-glutamine, and 25 ng/ml granulocyte-macrophage colony-stimulating factor (GM-CSF is only added for the first 2 days). After 7 days, the microglia were polarized into M1 phenotype by priming with 1 × 10
3 U/ml of rhIFN-γ (U-Cytech) for 24 h, followed by the addition of 10 ng/ml of
Escherichia coli LPS (InvivoGen) to the medium for 24 h. M2 microglia were polarized by adding 10 ng/ml of IL-4 (Immunotools) for 48 h. Untreated cells are referred to as M0.
Macrophage differentiation and culture
Human peripheral blood mononuclear cells were isolated from heparinized human peripheral blood from healthy donors by density gradient centrifugation on Lymphoprep™ (Stemcell Technologies), and monocytes were isolated using CD14-MACS microbeads (Miltenyi Biotec) according to the manufacturer’s protocol. All blood donors gave informed consent. Macrophages were generated by culturing monocytes for 5 days in RPMI 1640 medium (Invitrogen) containing 10% fetal calf serum, 2 mM glutamine 50 U/ml penicillin, and 50 μg/ml streptomycin (all from Lonza), in the presence of 50 ng/ml M-CSF (Miltenyi). On day 5, macrophages were polarized into M1 phenotype by priming with 50 ng/ml of rhIFN-γ (Peprotech) for 24 h, followed by the addition of 10 ng/ml of E. coli LPS (Sigma) to the medium for 24 h. M2 macrophages were polarized by adding 20 ng/ml of IL-4 (Immunotools) for 48 h. Untreated cells are referred to as M0. For MGL triggering experiments, macrophages (50.000) were plated on 96-well F-bottom plates coated with 1 μM Tn3−, gal-dendrimers, or uncoated wells in the presence or absence of 10 ng/ml LPS. Supernatants were harvested after 18 h, and IL-10 levels were determined by ELISA (eBioscience).
Flow cytometry
For surface expression analysis, cells were incubated with primary antibodies according to the manufacturer’s instructions. For all flow data analyses, debris and dead cells were excluded by 7-AAD staining. Cells were acquired on a FACSCalibur™ (BD Biosciences) or on a CyAn ADP (Beckman Coulter) and analyzed with FlowJo software (Tree Star).
MGL-binding assays
For analysis of MGL ligand expression, splenocytes or DLN cells were incubated for 30 min at 37 °C with MGL1− or MGL2-Fc (10 μg/ml) in a solution of 20 mM Tris-HCl, pH 7.4, 150 mM NaCl, 2 mM MgCl2, 1 mM CaCl2, and 0.5% (w/v) bovine serum albumin (BSA), followed by staining with secondary FITC-labeled anti-human Fc (Jackson ImmunoResearch) and analysis on a FACSCalibur.
Immunohistochemistry and immunofluorescence staining
For human tissue staining, air-dried frozen sections (5 μm) were fixed with acetone (10 min at RT). For single immunohistochemistry staining of PLP, MHC-II, and MGL, we used the protocol described previously by Beaino et al. [
19] using Envision-HRP (Dako) and 3,3′-diaminobenzidine (Dako) as a detection method. For immunofluorescence staining on human tissue, the sections were fixed with acetone then blocked for non-specific binding with goat serum (10%) for 20 min at RT. The sections were then incubated with the primary antibody in PBS/1% serum overnight at 4 °C. The sections were then washed with PBS three times for 5 min and incubated with fluorescently labeled secondary antibody in PBS/1% serum or for 1 h at RT. The nuclei were stained with Hoechst (1/1000) for 1 min in the dark. After a final wash for three times for 5 min with PBS, the sections were mounted with coverslips using aqueous mounting media Mowiol.
Real-time quantitative RT-PCR
Total mRNA from polarized microglia was extracted using the mRNA Capture Kit (Roche), and cDNA was synthesized using the Reverse Transcription System kit (Promega) following the manufacturer’s guidelines. Reactions were performed using the SYBR Green method in an ABI 7900HT sequence detection system (Applied Biosystems), with GAPDH as the endogenous reference gene (ERG). Samples were analyzed in duplicate and normalized to GAPDH. Oligonucleotides were designed by using the computer software Primer Express 2.0 (Applied Biosystems) and synthesized by Invitrogen. Primer specificity was computer tested by homology search with the human genome (BLAST, National Center for Biotechnology Information) and later confirmed by dissociation curve analysis. The difference between the Ct of the target gene and the Ct of the ERG ΔCt = CtTarget − CtERG is used to obtain the normalized amount of target (Nt), which corresponds to 2−ΔCt. The Nt reflects the relative amount of target transcripts with respect to the expression of the ERG. Primer sequences: CLEC10A, Fw, 5′-TACACCTGGATGGGCCTCAG-3′, Rev., 5′-TGTTCCATCCACCCACTTCC-3′; GAPDH, Fw 5′-CCTTCCGTGTCCCCACTG-3′, rev 5′-GACGCCTGCTTCACCACC-3′.
Dataset selection and analysis
Expression of MGL (CLEC10A) in a RNA-Seq data previously published by Hendrickx et al. (GSE108000) [
20] was analyzed using the R-based software RStudio. Normalized data was downloaded using the package
GEOquery [
21] and converted to
z-score using the function
scale.
Statistical analysis
Prism7 software (GraphPad) was used for statistical analysis. For clinical EAE and weight data, the area under the curve (AUC) was calculated for each mouse over the time period assessed, and these values were used for comparison of groups by an unpaired, one-tailed Student’s t test. Statistical differences in the mean maximum scores were analyzed by one-tailed Mann-Whitney U test. Differences with P values of less than 0.05 were considered significant.
Discussion
Human MGL is expressed on immature and tolerogenic DCs, dermal CD1a
+ DCs, and blood CD1c
+ myeloid DCs [
5]. Macrophages express MGL; however, it is especially upregulated on alternatively activated macrophages [
5,
37]. Human MGL and murine MGL1 and MGL2 have immunoregulatory properties when expressed on tolerogenic APCs and have been shown to play a role in ameliorating autoimmune diseases [
5,
7,
10‐
13,
15,
16]. It has been shown that MGL is expressed by APCs at sites of chronic inflammation in rheumatoid arthritis [
8]. Here, we found that MGL is markedly expressed at sites of neuroinflammation. By microscopy and tissue microarray analysis, we observed that MGL is highly expressed in active and chronic active MS lesions, but not in chronic inactive MS lesions or NAWM (Fig.
1a, b), indicating an attempt to prevent lesion expansion and progression, as has been suggested for other anti-inflammatory genes upregulated in chronic active MS lesions [
20]. In line with these findings, we demonstrate for the first time that
Clec10a−/− mice show an exacerbated EAE disease severity (Fig.
4a), indicating a role for MGL1 in limiting CNS inflammation.
Expression patterns of immune checkpoint inhibitors in EAE indicate that their levels rise after disease onset, reach a plateau at the peak of the disease, and remain stable or diminish during the resolution phase of the inflammation [
23,
24,
31]. Moreover, the pace of the EAE in the absence of immune checkpoint inhibitors shows a similar disease onset but higher scoring when compared to WT controls [
31,
38‐
41]. In line with this, we have shown that human MGL and rat MGL are increased in the CNS during MS or EAE inflammation. In rat EAE, MGL is upregulated at the peak of the disease in both cerebellum and brainstem and then diminishes when inflammation is resolved (Additional file
1: Figure S2). Moreover, EAE in
Clec10a−/− mice exhibits a similar disease onset, but higher disease severity [
31,
38‐
41], suggesting that MGL1 might act as an immune checkpoint inhibitor in CNS autoimmune inflammation. Future research will address whether, similarly to human MS lesions and acute EAE in rats, MGL1 is upregulated at the peak and/or resolution phase of the EAE in mice. Furthermore, given the above-indicated link between MGL1 and IL-10 and the pivotal role of IL-10 in EAE resolution (ref. [
42,
43]) more work is needed to determine whether, as in human M2 microglia (Fig.
2b), MGL1 binding in murine CNS APCs induces the secretion of IL-10.
The mechanisms underlying the anti-inflammatory activity of MGL1 remain poorly understood. Human MGL inhibits DC maturation and migration, increases their IL-10 secretion [
9‐
14], and induces CD45-dependent CD4
+ T cell apoptosis, thereby dampening the inflammatory response [
8]. Available information about the role of MGL1 in autoimmune disease models is limited but suggestive of an anti-inflammatory role. In dextran sulfate sodium salt-induced colitis, intestinal lamina propria macrophages expressing MGL1 increase IL-10 production in response to commensal
Streptococcus bacteria, playing a protective role against autoimmune colitis [
10]. In this work, we show for the first time that MGL is also upregulated on M2 microglia, cells that show neuroprotective and regenerative properties in the CNS [
27,
44], co-localizing with the M2 macrophage/microglia marker P2Y12R (Fig.
2c, d) [
19]. At the same time, we demonstrate that M2 macrophages express MGL and, like MGL-expressing DCs [
13], produce IL-10 in response to stimulation with MGL-specific ligands (Fig.
2a, b). Further research has to address whether MGL triggering on human M2 microglia cells induces IL-10.
Apoptosis induction of antigen-specific T cells is the mechanism of action shared by the checkpoint inhibitors PD-1, Galectin-1, and Tim-3 [
38,
41,
45]. Mouse MGL1 and MGL2 possess non-redundant roles; however, the functions described for each one fulfils most of the human MGL characteristics (reviewed in [
5]). Here, we show that MGL1, but not MGL2, induces apoptosis of activated CD4
+ T cells (Fig.
6b). Moreover, ex vivo experiments show that mononuclear MHC II
− DLN cells from MGL1-devoid mice with EAE bind stronger to MGL1, correlating with higher MGL1-induced apoptosis, when compared to cells from WT mice (Fig.
6a, b). Whether MGL1 is inducing apoptosis of CD4
+ T cells or of a different MHC II
− lymphocyte subpopulation in vivo still needs to be investigated. In this sense, and given that autoreactive T cells play a crucial role in the disease severity of MS and EAE, the augmented EAE severity in
Clec10a−/− mice observed could be the result of an increased number of MOG
35–55-specific effector T cells. Accordingly, splenocytes and DLN cells from MGL1-devoid mice showed higher proliferation and secreted more IL-17 than their WT counterpart (Fig.
5a–d). Intriguingly, we found no differences in IL-10 production. This was unexpected given the central role of IL-10 in the resolution of autoimmune neuroinflammation [
42,
46] and the previously described MGL1-dependent IL-10 secretion mechanisms [
5,
10]. Whether these mechanisms are also driven by MGL
+ cells in MS requires further research. Altogether, these findings indicate that MGL M2 microglia could play a role in resolving neuroinflammation by increasing IL-10 production and by inducing apoptosis in autoreactive CD4
+ T cells. Moreover, we have uncovered a novel marker for M2 microglia identification.
Human MGL or mouse MGL1 are not expressed on B cells and have not been previously associated with the alteration of the B cell response. Nevertheless, we have found that the humoral response in
Clec10a−/− mice switched to a predominant Th1 response, evidenced by the high titer of IgG2c isotype in comparison with higher levels of the more anti-inflammatory Th2 isotype IgG1 in the serum of WT mice (Fig.
5e, f) [
34‐
36]. B cells at the B cell follicle form the germinal center under the influence of specialized T follicular helper (T
FH) cells [
47]. T
FH cells collaborate with B cell proliferation and class-switch recombination by the expression of CD40L and cytokine production [
47]. Whether MGL1
+ dendritic cells or macrophages migrate to secondary lymphoid organs and affect T follicular helper cells or the cytokine expression at the germinal center microenvironment remains to be explored.
Successful treatment of MS should aim at two different processes in order to achieve a complete remission: (1) to abrogate disease progression and (2) to stimulate the resolution of inflammation. However, just a few ligand-receptor axes have been described as fulfilling one or both of these requirements. So far, CTLA-4–CD80/86, PD-1–PD-L1/2, TIGIT–CD112/155, TIM-3–galectin-9, and CD43/45–galectin-1 are the only immune inhibitory pathways involved in EAE amelioration with a strong genetic or functional correlation in MS [
4,
24,
31,
38‐
41,
48]. Here, we show that the endogenous lectin MGL tackles both targets. MGL is highly expressed at sites of neuroinflammation and triggers IL-10 production by macrophages that in turn might collaborate to the resolution of the disease. At the same time, MGL binding to its ligand CD45 on CD4
+ T cells induces apoptosis of effector T cells, thereby inhibiting disease progression. Altogether, these results provide for the first time evidence that both MGL- and MGL1-expressing APCs might play a role in limiting the inflammatory response by fostering an anti-inflammatory environment in MS or EAE by increasing IL-10 secretion and inducing apoptosis of effector T cells.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.