Abstract
The TNF superfamily ligand LIGHT (lymphotoxin-like, exhibits inducible expression and competes with HSV glycoprotein D for herpesvirus entry mediator [HVEM], a receptor expressed by T lymphocytes) has been shown to play a role in T cell costimulation and be involved in apoptosis of mononuclear cells. As both T cells and monocytes are key components in the development and progression of experimental autoimmune encephalomyelitis (EAE), we studied the role of LIGHT in EAE. Following immunization with myelin oligodendrocyte glycoprotein peptide (35–55), LIGHT-deficient mice developed severe EAE that resulted in an atypically high mortality rate. Histological examinations revealed intensive activation of microglia/macrophages in the CNS and higher numbers of apoptotic cells within the CNS parenchyma of LIGHT-deficient mice. However, myelin oligodendrocyte glycoprotein peptide–specific CD4+ T cells from LIGHT-deficient mice showed reduced IFN-γ and IL-17 production and migration. Serum levels of reactive nitrogen intermediates and CNS transcripts of several proinflammatory cytokines and chemokines were also substantially decreased in the absence of LIGHT. EAE adoptive transfer experiments and bone marrow chimeras indicated that expression of LIGHT on donor cells is not required for disease induction. However, its expression on CNS host cells is a decisive factor to limit disease progression and tissue damage. Together, these data show that LIGHT expression is crucially involved in controlling activated macrophages/microglia during autoimmune CNS inflammation.
Introduction
The TNF superfamily ligand LIGHT (TNFSF14/CD258) (lymphotoxin-like, exhibits inducible expression and competes with HSV glycoprotein D for herpesvirus entry mediator [HVEM] receptor, expressed by T lymphocytes) is a type II transmembrane glycoprotein that is expressed on activated T cells, NK cells, monocytes, granulocytes, and immature dendritic cells (DCs) (1–4). It binds three receptors: HVEM, which is expressed on resting T cells, NK cells, monocytes, immature DCs, and endothelial cells (1, 2, 5–7); lymphotoxin β receptor (LTβR), which is expressed on DCs, endothelial cells, and stromal cells but is not expressed on lymphocytes (4, 8); and, in humans, Decoy receptor DcR3/TR6 (9).
LIGHT is a costimulator of T cells and can stimulate a Th1 profile of cytokines, including IFN-γ (10, 11). LIGHT can increase T cell proliferation in vitro (5–7), and in LIGHT-deficient mice the T cell responses are reduced (12–14). Its blockade or targeted disruption can alleviate alloresponses, graft rejection, or graft-versus-host disease (2, 4, 15–18). LIGHT can also induce apoptosis in thymocytes and, through LTβR, in tumors (19–23), and LIGHT has been shown to regulate lymph node (LN) hypertrophy in the generation of the adaptive immune response (24).
LIGHT is important in immunity to the parasite Leishmania donovani and early tuberculosis infection (25–27). However, LIGHT is not required for protection from Listeria, for protection from vesicular stomatitis virus, and for CD8 T cell responses to influenza virus (13, 25, 27, 28). In contrast, LIGHT-deficient mice are able to better control parasitemia from cerebral malaria than are wild-type mice (29). Expression of LIGHT on transplanted tumors results in increased antitumor responses and tumor clearance (15, 18). However, overexpression of LIGHT in transgenic mice leads to severe inflammation and development of lymphoproliferative diseases (30–32).
The involvement of members of the TNF superfamily in neuroinflammatory and neurodegenerative human diseases is of interest for potential therapies, and several important reports have been published (8, 33–37). In this study, we show the involvement of LIGHT in the development and progression of chronic stable experimental autoimmune encephalomyelitis (EAE)—one of the models used in research of multiple sclerosis. The ablation of LIGHT led to an acute aggravation of the clinical signs of the disease. Although LIGHT does not assist significantly in the development of Th1/Th17 cells, our data reveal that it plays an important role in controlling the inflammatory response in target tissue during the effector phase of the disease and that LIGHT-deficient mice showed increased activation of macrophages/microglial cells. Therefore, in this model the regulatory role by LIGHT in the target tissue supersedes its costimulatory activity on autoreactive T cells.
Materials and Methods
Mice
Mice with a targeted disruption of the LIGHT gene (LIGHT−/−) were described previously (13). The animals used in these experiments (originally in a 129sv background) were backcrossed onto the C57BL/6 background for nine generations. Age- and sex-matched wild-type homozygous littermate controls (LIGHT+/+) were used in the studies unless otherwise stated. C57BL/6 (CD45.2+) and congenic C57BL/6 (CD45.1+) mice 6–8 wk old were bred in the Australian Phenomics Facility. All mice were housed under conventional conditions and were used at 10–14 wk of age. All animal procedures were approved by the Australian National University Animal Ethics and Experimentation Committee.
Active induction of EAE and clinical evaluation
EAE was induced by immunizing male LIGHT+/+ and LIGHT−/− mice s.c. over two sites on the flanks with either 50 μg or 150 μg myelin oligodendrocyte glycoprotein peptide (35–55) (MOG35–55) (MEVGWYRSPFSRVVHLYRNGK) emulsified in an equal volume of CFA (Life Technologies/BRL) supplemented with Mycobacterium tuberculosis H37 RA (400 μg per mouse; Difco). Pertussis toxin (PTx) (250 ng; List Biological, Campbell, CA) was injected i.v. at the time of immunization and repeated 48 h later. Animals receiving adjuvant alone served as controls. Mice were assessed daily, beginning at day 5, for clinical signs of EAE and were graded according to the following scale: 0, no disease; 1, loss of tail tone; 2, hind limb weakness or partial paralysis; 3, complete hind limb paralysis and body paresis; 4, hind and front limb paralysis. Gradations of 0.5 for intermediate scores were assigned to animals exhibiting signs of lesser severity than any of these stages. Mice were euthanized when they reached a score of 4.
Histological analysis
At the indicated time points, mice were perfused intracardially with ice-cold PBS followed by 4% paraformaldehyde in buffered PBS. The brain and spinal cord were dissected out and fixed in 4% paraformaldehyde for 7 d before being embedded in paraffin. Longitudinal sections of the brain and the entire spinal cord (5 μm thick) were prepared and stained with Luxol fast blue with periodic acid–Schiff to assess the degree of demyelination. Immunohistochemical analyses were performed in adjacent serial sections with the following Abs: rat monoclonal anti-human/mouse CD3 (clone CD3-12, dilution 1:200; Serotec) for T cells; rat monoclonal anti-mouse F4/80 (clone CI:A3-1, dilution 1:200; Abcam), a marker for activated microglia/macrophages; and glial fibrillary acidic protein (GFAP), a marker for astrocytes. Appropriated biotin-conjugated secondary Abs were used. To identify apoptosis in the spinal cord and brain, the ApopTag Peroxidase In Situ Oligo Ligation Apoptosis Detection Kit (Chemicon) was used on paraffin sections. Briefly, 5-μm paraffin sections were incubated with 3% H2O2 in PBS to quench endogenous peroxidase activity. Ag retrieval was performed by boiling sections in citrate buffer, pH 6.0, for 20 min. Sections were incubated with ApopTag Equilibration Buffer (Chemicon) for 1 min at room temperature, followed by a mixture of T4 DNA ligase and a blunt-ended biotinylated oligo (Chemicon) for 12 h at 4°C. Sections were further exposed to streptavidin–peroxidase for 1 h at room temperature, and the positive reaction was visualized with diaminobenzidine as substrate (Chemicon).
IL-17 and IFN-γ induction by T cells
Draining LNs were harvested from individual mice 10 d after immunization with MOG35–55/CFA without injection of PTx. Spleens were obtained from naive mice and CD11b+ cells were purified by positive selection using CD11b MicroBeads on an autoMACS magnetic separator (Miltenyi Biotec). Isolated CD11b+ cells (typically ∼ 97% pure) from LIGHT−/− or LIGHT+/+ mice were cocultured with corresponding LN cells at a ratio of 1:5 in the presence of IL-23 (5 ng/ml). Cells were incubated with MOG35–55 (10 μg/ml) or medium alone. At 5 d, samples were stimulated with 5 ng/ml PMA, 1 μM ionomycin (Sigma-Aldrich, St. Louis, MO), and GolgiStop for 5 h before fixation and permeabilization with Cytofix/Cytoperm. Cells were membrane labeled with Ab against CD4 and intracellularly with anti–IL-17 and anti–IFN-γ. Experiments were repeated up to four times with cells prepared from different groups of animals; n = 4.
Ag-specific proliferation assay and cytokine production
To quantify the in vitro proliferative response to Ag, LNs (inguinal, popliteal, axillary, and mesenteric) and spleens were obtained 10 d after EAE induction with MOG35–55/CFA. Tissues were mechanically homogenized to make a single-cell suspension, and cells were seeded in triplicate in flat-bottom 96-well plates at a concentration of 2 × 106 cells per milliliter (for splenocytes) or 1 × 106 cells per milliliter (for LN cells) in MLC medium supplemented with 2 mM l-glutamine, 100 mg/ ml streptomycin, 100 U/ml penicillin, 10 mM HEPES, 1 mM sodium pyruvate, 2 × 10−5 M 2-ME, and 5% heat-inactivated FCS. Cells were cultured with MOG35–55 (10 μg/ml), anti-TCR (1 μg/ml), or medium alone. After 48 h of incubation at 37°C with 5% CO2 and humidified atmosphere, plates were pulsed with [3H] thymidine (0.5 μCi per well; Amersham Pharmacia Biotech) for an additional 16 h before harvesting the cells. Results are expressed as mean thymidine uptake (cpm) of triplicate cultures. Production of IL-2 and IL-5 was determined by quantitative capture ELISA according to the supplier's guidelines (BD Pharmingen) from 2-ml cultures containing 4 × 106 spleen or LN cells. Cytokine concentrations were calculated using standard curves generated with known amounts of recombinant proteins. The concentration of IL-12, IL-6, TNF, and IFN-γ was analyzed using the murine Th1/Th2 cytokine bead array (BD Biosciences) according to the manufacturer’s instructions. Flow cytometry was conducted on a FACScan (Becton Dickinson), and all samples were analyzed at least in duplicate. For the staining of intracellular IFN-γ, lymphocytes were incubated as indicated above in the presence of GolgiStop (BD Biosciences) for the last 6 h of the 72-h culture. Cytokine staining was performed using a Cytofix/Cytoperm kit (BD Biosciences), and samples were analyzed by flow cytometry.
CFSE labeling
LN cells were resuspended at a density of 1 × 107 cells per milliliter in serum-free MLC medium containing 2 μM CFSE (Molecular Probes). The cells were incubated at 37°C for 10 min and washed in MLC medium supplemented with 5% FCS. Labeling was terminated by the addition of FCS to a final concentration of 10%, and cells were washed twice in culture medium.
IL-17 coculture assay
Draining LNs were harvested from individual mice 10 d after immunization with MOG35–55/CFA without injection of PTx. Spleens were obtained from naive mice, and CD11b+ cells were purified by positive selection, using CD11b MicroBeads on an autoMACS magnetic separator (Miltenyi Biotec). Isolated CD11b+ cells (typically ∼ 97% pure) from LIGHT−/− or LIGHT+/+ mice were cocultured with LN cells from reciprocal mice at a ratio of 1:5 in the presence of IL-23 (5 ng/ml). Cells were incubated with MOG35–55 (10 μg/ml) or medium alone for 5 d. Samples were stimulated with 5 ng/ml PMA, 1 μM ionomycin (Sigma-Aldrich), and GolgiStop for 5 h before fixation and permeabilization with Cytofix/Cytoperm. Cells were surface stained with Ab against CD4 and intracellularly with anti–IL-17 and anti–IFN-γ. Experiments were repeated up to four times with cells prepared from different groups of animals; n = 4.
Flow cytometry analysis
Single-cell suspensions of spleen and LN were prepared, followed by lysis of RBCs in 0.83% ammonium chloride. Cells were incubated with FITC-, PE-, PerCP-, and APC-labeled Abs against CD4, CD8, CD44, CD62L, CD25, and CD69 (Pharmingen). The bone marrow (BM)–derived macrophages were stained for CD11b, CD80, CD86, and MHC class II (I-A/I-E) (Pharmingen and BD Biosciences). The fluorescence was determined by flow cytometry on a FACScan (Becton Dickinson), and data were analyzed with Weasel v2.2.3 software.
Measurement of NO Production
The levels of nitrate and nitrite in serum samples were determined as an indirect measurement of NO production (38). Briefly, the nitrite was measured colorimetrically after addition of Griess reagent. Nitrate was first converted to nitrite by nitrate reductase in the presence of NADPH (Boehringer Mannheim, Ingelheim, Germany). The results were expressed as micromolar (μM) concentrations of reactive nitrogen intermediates (RNI), that is, the sum of nitrate and nitrite concentrations. The detection limit of this assay was ≥ 4 μM.
Adoptive transfer of EAE
Male LIGHT+/+ and LIGHT−/− donor mice were immunized s.c. with 50 μg MOG35–55 in CFA supplemented with 400 μg M. tuberculosis. At 10 d after immunization, their spleens were collected and stimulated in vitro with MOG35–55
6 blasts were injected i.v. into naive LIGHT−/− or LIGHT+/+ recipient mice (n = 6 mice per group). EAE was scored daily as described above.Quantitative real-time RT-PCR
Total RNA was isolated from whole spinal cord using TRIzol technology (Invitrogen, Melbourne, VIC, Australia) according to the manufacturer's instructions. RNA (5 μg) for each sample was reverse transcribed into cDNA, using the Omniscript RT Kit (Qiagen, Doncaster, VIC, Australia). SYBR Green–based real-time PCR (Applied Biosystems, Foster City, CA) was used to measure relative gene expression of IFN-γ, TNF, inducible NO synthase (iNOS), IL-6, MCP-1, and IFN-γ–induced protein 10 (IP-10) in each sample. Gene expression of TNFSF14 (LIGHT) was performed using the TaqMan MGB probe and primer pairs from Applied Biosystems. Each experimental sample was assayed using three replicates for each primer, including the β-actin–specific primer that was used as an internal standard. Negative controls lacking the cDNA template were run with every assay to assess specificity. Primers used in this study are listed below and were designed using the Primer Express software (Applied Biosystems). PCR amplification was carried on an ABI Prism 7700 Sequence Detection System (Applied Biosystems), and analysis was performed with the accompanying software. At the end of the PCR cycle, a dissociation curve was generated to ensure the amplification of a single product, and the threshold cycle time (Ct values) for each gene was normalized to housekeeping gene β-actin. The results were expressed as relative fold change over the values for naive mice.
BM chimeras
BM chimeric animals were generated using LIGHT−/− (CD45.2), C57BL/6 (CD45.1), or C57BL/6 (CD45.2) mice as donors and/or recipients. Eight-week-old female recipient mice were lethally irradiated (1200 rad) and were reconstituted 5 h later with 3 × 106 BM cells harvested from femurs and tibias of age-matched mice. Experimental transfers were as follows: LIGHT−/− donors into CD45.1 recipients (LIGHT−/− > B6.CD45.1), B6.CD45.1 donors into LIGHT−/− recipients (B6.CD45.1 > LIGHT−/−). Control chimeras were generated by reconstituting lethally irradiated B6.CD45.1 mice with congenic B6.CD45.2 BM (B6.CD45.2 into B6.CD45.1). Chimeras were maintained on sterile water containing 0.2% Bactrim for 3 wk. At week 8 after BM transplantation, mice were bled from the retro-orbital sinus, and the level of donor chimerism was analyzed by flow cytometry using the following Abs: FITC-conjugated anti-CD45.2, PE-conjugated anti-CD45.1 with PerCP- or APC-labeled anti-CD11b, or anti-CD4 (Pharmingen). The chimerism for CD4+ and CD11b+ cells was > 95% (data not shown). Mice were used for experiments 10 wk after reconstitution.
Isolation of mononuclear cells from CNS
BM chimeras were sacrificed at the peak of disease, and the brain and spinal cord were removed after perfusion with ice-cold PBS. Tissue was dissociated using the Neural Tissue Dissociation Kit according to the manufacturer’s instructions (Miltenyi Biotec), and mononuclear cells were isolated by centrifugation over 30%/70% discontinuous Percoll gradient (Amersham Biosciences). Cells were stained as indicated above.
BM-derived macrophages
Mice were killed by cervical dislocation, and their femurs and tibias were removed. The BM was flushed in cold MLC containing 5% heat-inactivated FCS (Life Technologies); cells released from BM were washed twice and resuspended in MLC and 30% L929 cell–conditioned medium. A total of 5 × 106 BM cells per 5 ml medium were plated into petri dishes and incubated at 37°C for 7 d. On day 3, 50% of the medium was replaced with fresh 30% L929-conditioned medium (39). For priming, the 30% L929-conditioned medium was removed from the macrophage monolayer and replaced with MLC medium without L929-conditioned medium and the indicated concentration of LPS (Escherichia coli
Preparation of L929-conditioned medium
Confluent plates of L929 cells (a kind gift from Dr. Lauren Wilson) were split 1–10 and cultured with MLC supplemented with 5% heat-inactivated FCS, 100 U/ml penicillin, 100 μg/ml streptomycin, 1 mM sodium pyruvate, and 0.1 mM nonessential amino acids for 7–8 d. The medium was removed, filtered through a 0.2-μm membrane, and stored at −20°C.
Analysis of lymphocyte migration
The assay was performed with collagen (type I; Sigma-Aldrich)–coated Transwells, 6.5 mm in diameter with 5-μm pore size. MOG-specific LN cells (2 × 104) from LIGHT+/+ and LIGHT−/− mice were added to the upper chamber of the Transwells, and 600 μl assay medium or supernatant from these cells (previously filtered through a 0.22-μm pore size filter and diluted 1:5 in assay medium) was added to the lower compartment. Chambers were incubated at 37°C for 2 h, and cell viability was assessed by trypan blue exclusion. The cells that had migrated to the lower chamber were collected, stained, and counted by FACS. Percent migration was calculated by subtracting the number of cells that migrated to the medium from the number of cells that migrated to the corresponding supernatant divided by the total number of input cells added to the upper chamber. For the horizontal chemotaxis assay using the EZ-TAXIScan (Effector Cell Institute, Tokyo, Japan), 5 μl MOG-specific LIGHT+/+ or LIGHT−/− lymphocytes (1 × 106 cells per milliliter) were microinjected into the bottom holes of the TAXIScan chamber, and chemotaxis was initiated by the addition of 5 μl filtered supernatant into the matching holes. Cell migration was monitored for up to 90 min and recorded with a charge-coupled device camera. The number of migrating cells was quantified as a function of time from a time-lapse series of individual images taken at 1-min intervals. Only the cells that reached the upper chamber were considered.
Statistical analysis
Statistical analysis was performed using the GraphPrism program package, version 4 (GraphPad Software, San Diego, CA). Data were evaluated by performing two-way ANOVA and the Bonferroni tests, as appropriate. Unpaired Student t test was used to analyze two-variable comparisons. A p < 0.05 was considered statistically significant. Each dot in graphs corresponds to one biological replicate.
Primers
Gene-specific primer sets were designed to span intron–exon junctions to discriminate between cDNA and genomic DNA. The primer sets used in these studies were as follows: β-actin sense, 5′-CGT GAA AAG ATG ACC CAG AT CA-3′; antisense, 5′-CAC AGC CTG GAT GGC TAC GT-3; iNOS sense, 5′-AGA GAG ATC CGA TTT AGA GTC TTG GT-3′; antisense, 5′-TGA CCC GTG AAG CCA TGA C-3; IFN-γ sense, 5′-TGA ACG CTA CAC ACT GCA TCT TGG-3′; antisense, 5′-CGA CTC CTT TTC CGC TTC CTG AG-3′; MCP-1 sense, 5′-GCT GGA GAG CTA CAA GAG GAT CA-3′; antisense, 5′-CTC TTG AGC TTG GTG ACA AAA ACT AC-3′; IP-10 sense, 5′-AAA TCA TCC CTG CGA GCC TAT-3′; antisense, 5′-CTG CTC ATC ATT CTT TTT CAT CGT-3′.
Results
High mortality in the absence of LIGHT following immunization with MOG35–55 in CFA
To assess the requirement for LIGHT during EAE, we used mice with a null mutation in the LIGHT gene (LIGHT−/−) (13). LIGHT−/− mice and wild-type littermates were immunized with a standard dose (150 μg) of MOG35–55 in CFA, as outlined in Materials and Methods, and the course and severity of disease were evaluated (Fig. 1A). Control and LIGHT−/− mice had 100% disease incidence with similar disease onset (day 12.4 ± 1.5 versus 10.5 ± 0.8 for LIGHT+/+ and LIGHT−/− mice, respectively). However, whereas LIGHT+/+ mice developed an episode of ascending paralysis followed by a phase of partial recovery, in the absence of LIGHT the disease progressed rapidly to complete paralysis and all mice required euthanasia within 5 d after disease onset (Fig. 1A). Decreasing the dose of peptide in the inoculum from 150 μg to 50 μg resulted in longer survival time without remission in LIGHT-deficient mice (Fig. 1B). Immunization of LIGHT−/− mice with MOG35–55/CFA in the absence of PTx also induced clinical signs of disease in 60% of the mice, whereas the wild-type littermates did not show any clinical signs during the entire period of observation (Fig. 1C). Mice that developed EAE exhibited the clinical signs later than when PTx was coadministered. The disease score was lower, with paralysis limited to the hind limbs (average maximum disease score of 3). Thus, LIGHT expression is not required for disease development but is necessary for restraining the progression of EAE.
EAE disease in LIGHT−/− mice showed nonremitting severe disease with high mortality. LIGHT−/− (●) and LIGHT+/+ (○) mice were immunized with 150 μg (A and C) or 50 μg (B) of MOG35–55 emulsified in CFA containing 400 μg of M. tuberculosis. PTx was given on day 0 and day 2 in (A) and (B). Mice were followed daily for clinical symptoms and scored accordingly. Results are plotted as mean clinical score ± SD of all animals in each group. (A), n = 6; (B), n = 8; and (C), n = 6–7; †, mice sacrificed.
Pathological features of EAE in LIGHT-deficient mice
We next examined the pathological consequences of the absence of LIGHT signaling in the CNS. Brains and spinal cords were obtained during the acute (between 15 and 18 d postimmunization) and chronic phases of disease (day 30 postimmunization) following immunization with 50 μg MOG35–55 emulsified in CFA and with PTx. Brains were serially sectioned in a sagittal plane and spinal cords in a longitudinal plane and were analyzed for the presence of inflammatory cells and myelin loss. At peak disease, in both groups the infiltrates consisted mostly of F4/80+ activated macrophages/microglial cells and CD3+ lymphocytes (Fig. 2A–I). Intense immunoreactivity for the F4/80 Ag was detected in the cerebral meninges and brainstem of LIGHT−/− mice (Fig. 2B, 2D), but not of LIGHT+/+ mice (Fig. 2A, 2C). In the spinal cord, the F4/80+ and the CD3+ populations were predominantly in the meninges and at the edges of the perivascular cuffs (Fig. 2E–I). As disease progressed to day 30, in LIGHT+/+ mice a marked decrease was noted in the number of activated macrophages/microglia and, to a lesser extent, of lymphocytes in the affected areas (Fig. 2K, 2L, 2O). This finding contrasted with the spinal cord of LIGHT−/− mice, in which the density of F4/80+ cells augmented dramatically (Fig. 2N), accompanied by demyelination (Fig. 2M, 2P).
Brain and spinal cord histopathology during EAE in LIGHT+/+ and LIGHT−/− mice. Sagittal sections of brain (A–D) and longitudinal lumbar spinal cord sections (E–P) were prepared at days 15–18 (A–D, E–J) and at day 30 (K–P) after immunization with 50 μg MOG35–55/CFA and PTx. Sections were analyzed after immunohistochemistry for F4/80 (A, B, E, H, K, N) and CD3 (C, D, F, I, J, O) and histochemical staining with Luxol fast blue (G, J, M, P). Data shown are representative of five mice at days 15–18 and six mice at day 30 after immunization. Original magnification ×40.
The level of apoptosis was also tested. More TUNEL+ cells were seen in the lumbar spinal cord of LIGHT−/− mice than in their wild-type littermates during the peak (10–15 apoptotic cells versus 4–5 apoptotic cells per 500 cells, respectively) and the chronic phase of EAE (30–36 apoptotic cells versus 5–7 apoptotic cells per 500 cells) (Fig. 3). In LIGHT+/+ mice, apoptotic cells were present predominantly in the leptomeninges and perivascular spaces. By contrast, in LIGHT−/− mice most of the TUNEL+ cells were in the white matter parenchyma, particularly in areas of normal-appearing myelin. Apoptotic cells were also observed in the brainstem and cerebellum of LIGHT−/− mice, whereas fewer numbers were detected in the cerebellum of LIGHT+/+ mice (data not shown). As a single GFAP+ apoptotic cell was found in all sections, it was reasonable to accept that significantly more activated microglial cells/macrophages and/or CD3+ T cells are eliminated via apoptosis in the CNS of LIGHT-deficient mice.
Apoptosis in lumbar spinal cord sections of LIGHT+/+ (+/+) and LIGHT−/− (−/−) mice. The images, on which apoptotic cells are indicated with green arrows, are representative of the peak (days 15–18, five mice) and the chronic phase (day 30, six mice) of actively induced EAE. Original magnification ×40.
LIGHT is upregulated in CNS during the progression of EAE
We next sought to determine the kinetics of LIGHT expression during the course of EAE. LIGHT+/+ mice were immunized with 50 μg MOG35/CFA with PTx, and LIGHT mRNA was measured in the lumbar spinal cord. LIGHT expression was slightly augmented within the CNS during development of disease (day 15), and then mRNA levels increased dramatically during the recovery phase (Fig. 4A).
Kinetics of LIGHT, IFN-γ, iNOS, MCP-1, and IP-10 mRNA transcripts in the CNS during EAE. (A) Quantitative real-time PCR analysis of LIGHT mRNA expression in the spinal cord of LIGHT+/+ mice before and after immunization with 50 μg MOG35–55 and PTx. Sample sizes are as follows: naive, n = 6; day 15, n = 6; day 30, n = 5. Data are expressed in relative fold change to control week 0 ± SEM. (B–E) Quantitative RT-PCR analysis of IFN-γ, iNOS, MCP-1, and IP-10 expression in the spinal cord at 10 and 30 d in LIGHT−/− mice (black bars) compared with LIGHT+/+ (empty bars) after immunization with 50 μg of MOG35–55/CFA and PTx. Spinal cord basal mRNA levels for the chemokines studied were similar for LIGHT+/+ and LIGHT−/− mice. Transcript levels were normalized to the housekeeping gene β-actin and were calculated relative to the LIGHT+/+ group. Results are shown for n = 5–6 per group. Error bars represent SEM. **p < 0.01, ***p < 0.001.
LIGHT-deficient mice show lower levels of IFN-γ and NO in CNS during EAE
Further, we assessed the gene expression for IFN-γ, iNOS, MCP-1, and IP-10 mRNA in the lumbar spinal cord of LIGHT+/+ and LIGHT−/− mice during the disease. Expression of IFN-γ in the perfused CNS of naive mice was negligible, but it was strongly upregulated after the peak of disease (data not shown). In contrast, mRNA levels for IFN-γ were markedly lower in samples from LIGHT−/− mice compared with those from the wild-type controls before (p = 0.031, day 10) and during clinical disease (p = 0.0018, day 30) (Fig. 4B). Transcripts of iNOS, of which the major regulator is IFN-γ, were also considerably reduced in the CNS of LIGHT−/− mice compared with levels in LIGHT+/+ controls (p = 0.0041, day 30; Fig. 4C). The expression levels of IP-10 (IFN-γ–inducible protein 10) and MCP-1, both known to contribute to the recruitment of Th1 cells and macrophages in EAE (40, 41), were decreased within the inflamed tissues of LIGHT−/− mice compared with LIGHT+/+ mice, reaching significant levels at day 30 (p = 0.0034 and p = 0.0002, respectively; Fig.4D, 4E).
Lymphocytes from LIGHT−/− mice have increased T cell proliferative responses
Data gained from pathological changes in the target tissue—the CNS—raised the question about the state of T lymphocytes and monocytes in the periphery. EAE is an autoimmune condition primarily mediated by T (Th1/Th17) cells, and to identify the mechanisms that drive LIGHT-deficient mice to acute fulminant EAE, we studied the effect of the LIGHT gene disruption on some T cell functions. At 10 d after MOG35–55/CFA immunization, single-cell suspensions from LNs and spleens were analyzed for recall responses to Ag. The proliferation to MOG35–55 of LIGHT−/− LN cells and splenocytes significantly exceeded that of LIGHT+/+ lymphocytes (p < 0.002; Fig. 5A). The CFSE dilution assay showed that more LIGHT−/− CD4+ LN cells enter cell division in response to Ag (MOG35–55, p < 0.03, Fig. 5B) or TCR stimulation (p < 0.001, data not shown). The number of CD4+ T cells expressing CD44hi in LIGHT-/- mice was almost double that in wild-type controls (data not shown). The percentages of CD4+ T cells expressing the activation markers CD69, CD25, and CD62L were comparable (data not shown).
Enhanced proliferation of LIGHT−/− lymphocytes. Male LIGHT+/+ and LIGHT−/− mice were immunized with 50 μg of MOG35–55 emulsified in CFA without further injection of PTx. At 10 d after immunization, their LN (2 × 106 cells per milliliter), spleen (1 × 106 cells per milliliter) and cells were incubated with 10 μg/ml MOG35–55, 1 μg/ml α-TCR, or medium alone. (A) Proliferation was determined 48 h later by measuring [3H] thymidine uptake in LIGHT+/+ (○) and LIGHT−/− (●). Each dot represents a value for an individual mouse, and the bar represents the mean for each group (n = 8–11 mice). (B) Cell cycle analysis of CD4+ LN cells by CFSE dilution after 5 d stimulation with 10 μg/ml MOG35–55 (open histograms) or in the absence of stimulus (shaded histograms). Percentages of CD4+ cells that are CFSElo after stimulation are shown as mean ± SD. n = 5 mice per group, three independent experiments. (C) IL-2 levels in culture supernatants were determined from LIGHT+/+ (empty bar) and LIGHT−/− (black bar) samples by ELISA 24 h after stimulation with MOG35–55. (D) Secretion of IFN-γ was examined by a cytometric bead array in. LIGHT+/+ (empty bar) and LIGHT−/− (black bar) samples after 72 h of stimulation. Spontaneous secretion of cytokines in the absence of Ag was subtracted. Data are expressed as mean ± SEM. n = 5–6 mice per group; three independent experiments were performed. (E) Detection of IFN-γ–expressing CD4+ T cells by intracellular cytokine staining assay after 72 h of stimulation with MOG35–55. (F) RNI levels in serum from LIGHT+/+ (○) and LIGHT−/− mice (●) were measured at days 10 and 30 following immunization with MOG35–55/CFA. In this assay, 4.5 μM was the lower limit of detection. The experiments were performed twice, yielding similar results. *p < 0.05, **p < 0.001, ***p < 0.0001.
MOG-primed LIGHT−/− lymphocytes show reduced IFN-γ, RNI, and IL-17 production
We next examined whether the difference in proliferation correlated with the release of proinflammatory cytokines in splenocyte culture supernatants. The enhanced proliferation of LIGHT−/− lymphocytes was accompanied by an increase in IL-2 production in response to MOG35–55 (p < 0.05, Fig. 5C) or TCR stimulation (p < 0.003, data not shown). However, no significant differences were found between the two groups in the levels of secreted TNF, IL-6, IL-12p40, IL-5, and IL-10, and none of the groups produced detectable levels of IL-4 (data not shown). In contrast, the levels of IFN-γ were significantly lower (p < 0.003) in culture supernatants from MOG35–55 stimulated LIGHT−/− splenocytes than in supernatants from littermate controls (Fig. 5D), and the same was true for the frequency of IFN-γ+ CD4+ T cells: 2.5 ± 0.1 for the LIGHT−/− mice versus 6.9 ± 0.1 for the LIGHT+/+ mice (p < 0.001, Fig. 5E).
As IFN-γ is able to modulate lymphocyte proliferation by inducing the release of immune mediators such as NO, the serum levels of the RNI were measured in samples taken on days 15 and 30 after MOG35–55/CFA priming (for these assays, PTx was not included in the immunization protocol, as injection of PTx alone raises the RNI levels). At both time points, the serum RNI levels in LIGHT−/− mice were significantly lower than those in littermate controls, with p < 0.01 for day 15 and p < 0.001 for day 30 (Fig. 5F).
It is already well documented that the Th17 lineage plays a major role in the pathogenesis of EAE. Different mechanisms have been proposed to drive the differentiation of this T cell subset, but a defect in IFN-γ signaling appears to facilitate the process. Therefore, the ability of CD4+ T cells from LIGHT+/+ and LIGHT−/− mice to produce IL-17 in response to MOG35–55 was examined. Lymphocytes from MOG35–55/CFA–immunized LIGHT+/+ and LIGHT−/− mice were cultured in vitro with MOG35–55 in the presence of CD11b+ cells isolated from spleens of naive LIGHT+/+ or LIGHT−/− mice (Table I). Compared with LIGHT+/+ cells, CD4+ T cells from LIGHT−/− mice produced significantly less IL-17 in response to MOG35–55 (p = 0.0002), with or without the presence of TGFβ plus IL-6 or IL-23. When wild-type primed CD4+ T cells were cocultured with LIGHT-deficient CD11b+ cells, the levels of IL-17 decreased by 2-fold. In contrast, wild-type CD11b+ cells were able to partially restore the IL-17 levels in LIGHT−/− CD4+ cells in response to stimulation with MOG35–55, particularly in the presence of IL-23. No significant change in IFN-γ was noted in any of the conditions. Taken together, these results indicate that LIGHT is involved in the expansion of autoreactive IL-17–producing T cells at Ag restimulation and that the monocytes/macrophages are involved.
LIGHT-deficient macrophages exhibited reduced inflammatory cytokine production
We assessed how the myeloid cell response (BM-derived macrophages) is affected by the absence of LIGHT. In response to LPS or to the combination of LPS with IFN-γ, the secreted levels of MCP-1, IL-6, and TNF by LIGHT-deficient macrophages (Fig. 6) were significantly reduced. The production of IL-12 and IL-10 was not altered by the lack of LIGHT (data not shown). In addition, the levels of expression of B7.1/B7.2 and MHC class II were comparable in LIGHT+/+ and LIGHT−/− cells after IFN-γ stimulation (data not shown).
Secretion of IL-6, TNF, and MCP-1 by BM-derived macrophages from LIGHT+/+ (empty bars) or LIGHT−/− (black bars) mice after stimulation with IFN-γ (10 U/ml), LPS (1 μ/ml), or a combination of IFN-γ and LPS. Protein levels were determined by BD Cytometric Bead Array in supernatants collected after 24 h culture. Error bars represent SEM for four mice per group. *p < 0.05, **p < 0.001, ***p < 0.0001.
Less migration of lymphocytes from LIGHT-deficient mice
As described above, LIGHT−/− mice showed more infiltrating CD3+ cells in the CNS, despite having lower mRNA levels for several chemokines involved in the recruitment of those cells to the site of inflammation. One possibility is that LIGHT, directly or indirectly, influences the migratory properties of the effector cells. Using a standard Transwell assay, we analyzed the migratory capacity of MOG-specific LIGHT+/+ and LIGHT−/− lymphocytes. As shown in Fig. 7A, the diapedesis of LIGHT−/− lymphocytes was 50% lower than that of wild-type cells. The result was checked also with the EZ-TAXIScan chemotaxis system, in which both types of lymphocytes showed the common migratory patterns, that is, with cells exhibiting periods of higher and lower motility. However, as seen before, the diapedesis of LIGHT-deficient lymphocytes was lower (Fig. 7B), and this was true also of their migration in response to the chemoattractant IP-10 (Fig. 7C).
LIGHT-deficient lymphocytes show lower diapedesis and migration. (A) Diapedesis of MOG-specific LIGHT+/+ (empty bar) or LIGHT−/− (black bar) lymphocytes was determined in a Transwell assay. Data represent the percentage of cells from the total number used that reached the lower chambers containing culture medium alone; n = 4 per group. Results of two independent experiments were combined. (B) Diapedesis of MOG-specific lymphocytes assessed by the EZ-TAXIScan chamber. Numbers of LIGHT+/+ (gray line) and LIGHT−/− (black line) lymphocytes are given for a period of 30 min. (C) Migration of LIGHT+/+ (gray line) and LIGHT−/− (black line) lymphocytes in response to IP-10. **p < 0.001.
LIGHT expression in CNS is critical for recovery from disease
To further understand the apparently paradoxical effect of EAE exacerbation in LIGHT−/− mice with lower Th1/Th17 effector function, we performed adoptive transfer studies. Lymphocytes from MOG35–55/CFA–sensitized LIGHT+/+ or LIGHT−/− mice were differentiated in vitro with IL-23 and then transferred into LIGHT+/+ or LIGHT−/− recipients (Fig. 8A). LIGHT+/+ mice receiving IL-23 expanded MOG-specific LIGHT+/+ cells developed severe EAE without the need for PTx injection. The disease severity and clinical course paralleled the disease observed when LIGHT+/+ mice were given LIGHT−/− cells, indicating that the encephalitogenic potential of LIGHT−/− MOG-primed lymphocytes was comparable to that of LIGHT+/+. However, the transfer of LIGHT+/+ MOG-specific cells into LIGHT−/− recipients led to the death of 30% of the mice, demonstrating a critical role for LIGHT in the recipient. Although no mortality was registered in the LIGHT−/− → LIGHT−/− group, these mice did not enter the remission phase seen in the LIGHT+/+ recipient groups.
The presence of LIGHT in the CNS determines the severity of EAE. (A) Adoptive transfer of MOG-reactive lymphocytes generated by immunizing LIGHT+/+ (WT) and LIGHT−/− (KO) donor mice with MOG35–55/CFA and culturing their spleen cells in the presence of MOG peptide and IL-23 for 3 d. A total of 5 × 106 lymphocytes were injected i.v. into recipient mice. Groups were as follows: KO cells into WT recipients (open square), WT cells into WT recipients (open circle), WT cells into KO recipients (filled square), and KO cells into KO recipients (filled circle). Mice were scored daily for clinical disease (n = 5–6 group). (B) Analysis of BM chimeras. BM chimeras were constructed with reconstitution of LIGHT−/− (KO.CD45.2) mice with 100% LIGHT+/+ (WT.CD45.1) BM cells (filled circle) and LIGHT+/+ (WT.CD45.1) mice with 100% LIGHT−/− (KO.CD45.2) BM cells (filled square). As a control, LIGHT+/+ (CD45.1) mice were reconstituted with 100% LIGHT+/+(CD45.2) BM cells (open circle). BM chimeras (n = 6–8 group) were immunized with 50 μg of MOG35–55/CFA and PTx, and mice were assessed daily for clinical signs of disease. Results are shown as mean clinical score ± SD of all animals in each group; †, mice sacrificed. (C) Chimeric mice were sacrificed when they reached a clinical score of 3, and spinal cord inflammation was assessed by flow cytometry. Cells obtained from individual spinal cords were stained with Abs against CD45.1, CD45.2, and CD11b. Data are from one of two experiments with similar results.
To discern the exact role of LIGHT expressed by nonblood-derived cells in the effector phase of EAE, chimeras were constructed by reconstitution of sublethally irradiated LIGHT−/− [KO(CD45.2)] and congenic LIGHT+/+ [WT(C45.1)] mice with MHC-compatible BM cells isolated from congenic WT(CD45.1) and KO(CD45.2) mice, respectively. An additional group of chimeras with sublethally irradiated WT(CD45.1) mice reconstituted with WT(CD45.2) BM cells was included as control. After 10 wk of BM reconstitution, mice were immunized with 50 μg MOG35–55/CFA together with PTx and monitored for disease development. As shown in Fig. 8B, LIGHT deficiency in BM-derived cells [KO(CD45.2) into WT(CD45.1)] did not affect disease development when compared with control chimeras [WT(CD45.2) into WT(CD45.1)]. Both chimeric groups developed severe EAE followed by a phase of partial recovery. In sharp contrast, WT(CD45.1) into KO(CD45.2) chimeric mice developed severe EAE and were euthanized 10–15 d after disease appearance at a clinical score of 4. Disease onset was significantly accelerated compared with that in KO(CD45.2) into WT(CD45.1) chimeric mice (9.4 ± 0.6 versus 16.5 ± 1.8, p < 0.001). No mortality was recorded in any of the other groups. Together these data provide evidence that the inability of LIGHT-deficient mice to downregulate the inflammatory response resides in the radioresistant population within the CNS.
To characterize cells in the CNS by flow cytometry, some chimeric mice in each group were sacrificed when they reached a clinical score of 3. Approximately double the number of mononuclear cells were recovered from the CNS of the LIGHT−/− recipient [WT(CD45.1) into KO(CD45.2)] group compared with the other two groups (data not shown). Flow cytometry analysis showed that the majority of CD11b+ cells (>98%) detected in host mice in which LIGHT is expressed on CNS parenchymal cells (CD45.1) consisted of BM-derived macrophages from donor mice (Fig. 8C). By contrast, in LIGHT-deficient hosts (CD45.2) reconstituted with BM wild-type cells (CD45.1), the number of activated resident microglia represented 8.5–9% of the total CD11b cells.
Thus, the enhanced disease observed in LIGHT−/− mice is consistent with an increase in macrophages and inefficient local destruction and removal of infiltrating mononuclear cells.
Discussion
LIGHT has been previously reported to be an important costimulator for T cell activation (12–14). Our work confirmed this; that is, when restimulated in vitro, the MOG-specific CD4+ T cells from LIGHT-deficient mice produced less IFN-γ and IL-17 than the MOG-specific CD4+ T cells from LIGHT+/+ mice. That finding should reflect in the generation of less Th1 and Th17 MOG-specific EAE effector cells in the periphery. How, then, could the EAE onset occur at the same time or even earlier on a LIGHT-deficient background? Moreover, our results showed a lower CNS expression of the chemoattractants MCP-1 and IP-10 and lower in vitro migration of the LIGHT−/− T cells to IP-10. An explanation of this “contradiction” could be our previous work on EAE when we reported that high RNI levels induce actin polarization of T lymphoblasts and inhibit their migration (42). In the absence of LIGHT, levels of serum RNI are low, and we speculate that this would be reflected in a low T lymphoblast polarization that would allow enough EAE effector cells to reach the target tissue.
With respect to the situation in the target tissue, one should consider the results of Wu et al. (43) concerning accumulation of greater numbers of macrophages/activated microglia late in EAE in mice deficient in NOS. These data parallel our observation of low expression of iNOS in the CNS of LIGHT−/− mice and accumulation of F4/80+ cells in the chronic phase of EAE, at day 30 after immunization. The increase of iNOS and LIGHT in the CNS correlates with the phase of recovery from an EAE episode in wild-type mice (43–47). We also reported that iNOS increases at the recovery from EAE in rats (48); that is, one comes back to the dual role of IFN-γ and the IFN-γ–regulated iNOS, NO, and LIGHT, respectively. Our experiments support the role of LIGHT in the generation of NO to the levels required to downregulate the inflammatory response, leading to a phase of partial recovery in this model of EAE.
In EAE, both tissue-resident microglia and blood-borne macrophages have been implicated in the disease owing to their role in demyelination and production of proinflammatory cytokines (49–51). Recently, the roles of these two cell types in EAE were more clearly defined (52). Microglia were shown to be important for disease initiation, along with CD4+ T cells, compared with blood-borne macrophages, which triggered EAE progression. Our experiments with BM chimeras proved that LIGHT expression on radioresistant CNS cells was required to modulate the numbers of activated macrophages/microglial cells and the associated clinical disease. More important, LIGHT-deficient mice showed a higher level of cell death in the CNS, consistent with greater demyelination and higher disease score.
The T cell apoptosis occurring in the CNS and its role in the spontaneous recovery from an EAE episode have been well researched in models of acute and chronic EAE (53–55). Of particular interest to our work are the results on T cell apoptosis in MOG peptide–induced EAE in gene knockouts (37), (56). Impairment of the TNFR 1 pathway led to a 50% reduction of T cell apoptosis in CNS lesions, whereas other genetic deletions (TNF, lymphotoxin, TNFR 2, Fas-L, iNOS, perforin, and IL-1β–converting enzyme) had no significant effect (36). Furthermore, mice with reduced apoptosis (transgenic for Bcl-xL) developed a more severe form of EAE in comparison with wild-type mice, which could reflect the increased survival of activated cells in the CNS, leading to persistent chronic disease (56).
In another disease model—experimental cerebral malaria—the disruption of LIGHT–LTβR signaling protected mice from the disease (29). Blocking LTβR with Ab led to increased numbers of splenic monocyte/macrophages, which correlated with a lower parasitemia. Soluble LIGHT has been shown to mediate hepatocyte death in an experimental hepatitis model (57). LIGHT–LTβR interaction mediated the cell death of certain tumor cell lines in the presence of IFN-γ (19, 22). In our EAE model, it is possible that LIGHT mediates removal of macrophages/microglial cells through LTβR signaling on these cells. Therefore, these cells accumulate to cause excessive disease in the absence of LIGHT–LTβR signaling.
This study began with the question of whether LIGHT is needed for the induction and disease progression of EAE. We confirm previous reports that LIGHT is important for optimal T cell activation. However, its presence or absence in the CNS was critical in modulating the regression of EAE. Thus, LIGHT expression is not required for disease development. Conversely, it appears that LIGHT expression is necessary for restraining inflammatory cells—namely, macrophages/microglial cells—that mediate the effector phase of EAE.
Disclosures
The authors have no financial conflicts of interest.
Acknowledgments
We thank David Willenborg for critical advice, Tania Watts for critical reading of the manuscript, the Australian Phenomics Facility for animal care and genotyping, Anne Prins for histology, and David Bernal for technical help.
Footnotes
This work was supported by National Health and Medical Research Council Grant 316928 (to E.M.B.).
Abbreviations used in this article:
- BM
- bone marrow
- DC
- dendritic cell
- EAE
- experimental autoimmune encephalomyelitis
- iNOS
- inducible NO synthase
- IP-10
- IFN-γ–induced protein 10
- LIGHT
- lymphotoxin-like, exhibits inducible expression and competes with HSV glycoprotein D for herpesvirus entry mediator (HVEM), a receptor expressed by T lymphocytes
- LN
- lymph node
- LTβR
- lymphotoxin β receptor
- MOG
- myelin oligodendrocyte glycoprotein
- PTx
- pertussis toxin
- RNI
- reactive nitrogen intermediate.
- Received October 31, 2012.
- Accepted April 25, 2013.
- Copyright © 2013 by The American Association of Immunologists, Inc.