Myelin-phagocytosing macrophages modulate autoreactive T cell proliferation

Female Lewis rats, 6-8 weeks of age, were purchased from Harlan Netherlands B.V. (Horst, The Netherlands). Animals were housed in the animal facility of the Biomedical Research Institute of Hasselt University. Experiments were conducted in accordance with institutional guidelines and approved by the local Ethical Committee for Animal Experiments of Hasselt University.

Three days prior to macrophage isolation, rats were injected intraperitoneally with 3 ml 3% thioglycolate (Sigma-Aldrich, Bornem, Belgium). Resident peritoneal macrophages were obtained by peritoneal lavage using 10 ml of ice-cold PBS (Lonza, Vervier, Belgium) supplemented with 5 mM ethylenediamine tetraacetic acid (EDTA; VWR, Leuven, Belgium). Peritoneal exudate cells (PECs) were cultured for 2 hours in RPMI 1640 medium. After 2 hours incubation at 37°C with 5% CO₂, non-adherent cells were washed away. Remaining cells were >95% macrophages [38].

Myelin was purified from rat brain tissue by means of density-gradient centrifugation, as described previously [39]. Myelin protein concentration was determined by using the BCA protein assay kit (Thermo Fisher Scientific, Erembodegem, Belgium). LPS content was determined using the Chromogenic Limulus Amebocyte Lysate assay kit (Genscript Incorperation, Aachen, Germany). Isolated myelin contained a neglectable amount of endotoxin (1.8 × 10^-3 pg/μg myelin).

Isolated myelin was fluorescently labelled, according to the method of Van der Laan et al. [11]. In short, 10 mg/ml myelin was incubated with 12.5 μg/ml 1,1"-diotadecyl-3,3,3',3',-tetramethylindocarbocyanide perchlorate (DiI; Sigma-Aldrich) for 30 min at 37°C. Next, a myelin phagocytosis assay was performed as previously described [39].

Lewis rats were injected subcutaneously with a 0.1 ml suspension containing 250 μg/ml guinea pig myelin basic protein (MBP) or ovalbumin (OVA), 2.5 mg/ml H37RA heat-killed mycobacterium tuberculosis (Difco, Detroit, USA) and 60 μl Complete Freunds adjuvant (Sigma-Alldrich) in both hind paws. Subsequently animals were injected subcutaneously with PBS, 2.6 × 10⁶ latex beads (0.8 μm mean particle size, Sigma-Alldrich), 75 μg/animal of isolated myelin or OVA (d-4, 0, 4 and 8 pre- and post-immunization). MBP-immunized rats were weighted and scored daily according to the following neurological scale: 0 = no neurological abnormalities, 0.5 = partial loss of tail tonus, 1 = complete loss of tail tonus, 2 = hind limb paresis, 3 = hind limb paralysis, 4 = moribund, 5 = death.

MBP and OVA-specific lymphocytes were obtained 9 days post-immunization by bilateral isolation of the inguinal and popliteal lymph nodes. Single-cell suspensions of harvested lymph nodes were obtained by grinding with a syringe plunger against a 70 μm cell strainer (Bellco Glass Inc., Vineland, USA). To enrich for antigen specific lymphocytes, lymph node cells were restimulated, as described previously [40]. Briefly, lymph node cells were resuspended in stimulation medium: RPMI 1640 medium (Invitrogen, Merelbeke, Belgium) supplemented with 50 U/ml penicillin (Invitrogen), 50 U/ml streptomycin (Invitrogen), 20 μM 2-mercapto-ethanol (Sigma-Alldrich), 1% sodium pyruvate (Invitrogen), 1% MEM non-essential amino acids (Invitrogen), 2% deactivated autologous serum and 33 μg/ml MBP. After 2 days, cells were washed and resuspended in RPMI 1640 medium supplemented with 50 U/ml penicillin, 50 U/ml streptomycin, 20 μM 2-mercapto-ethanol, 10% fetal calf serum (FCS, Hyclone, Erembodegem, Belgium) and 6,5% supernatants of Concanavalin (ConA, Sigma-Alldrich) stimulated spleen cells. Following 2 days, cells were washed and resuspended in RPMI 1640 medium supplemented with 50 U/ml penicillin, 50 U/ml streptomycin, 20 μM 2-mercapto-ethanol and 10% fetal calf serum for 3 days.

A carboxyfluorescein diacetatesuccinimidyl ester (CFSE) stock (10 mM in DMSO, Invitrogen, Merelbeke, Belgium) was diluted in PBS (Biowhittaker™). Antigen-specific lymphocytes were resuspended in PBS supplemented with 0.05% BSA and 4 μM CFSE (20 × 10⁶ cells/ml) for 7 min at 37°C with 5% CO₂. Cells were washed and diluted in 0.5 ml culture medium for 30 min at 37°C with 5% CO₂ to stabilize the CFSE-labeling. In parallel, to determine macrophage viability following a coculture with lymphocytes, macrophages were labeled with CFSE to distinguish them from unlabeled lymphocytes.

Prior to coculture with CFSE labeled lymphocytes, isolated macrophages were seeded in flat-bottem 96-well plates (15 × 10³ cells/well) in RPMI 1640 medium supplemented with 50 U/ml, 50 U/ml streptomycin and 10% FCS, and treated with 100 μg/ml of isolated myelin for three hours. Excess myelin was removed by washing twice with RPMI 1640 medium at 37°C. Subsequently, stimulation medium containing irradiated thymocytes (15 × 10⁴, 3000 rad), CFSE-labeled MBP- or OVA-specific lymphocytes (15 × 10⁴) and respectively 10 μg/ml MBP or 10 μg/ml OVA were added. Untreated macrophages were used as a control. To evaluate the involvement of respectively NO, arginase, indoleamine 2,3-dioxygenase (IDO) the phagocytosis process itself, direct cell-cell contact and IFNγ, 1.5 mM N^G-Monomethyl-L-arginine (L-NMMA; VWR), 0.5 mM N^G-Hydroxy-L-arginine (NOHA; VWR), 0.2 mM 1-Methyl-L-tryptophan (1-MT; Sigma Alldrich), latex beads (1:100), 100 μg/ml zymosan A (Sigma-Alldrich), transwell inserts (0.4 μm pore size, Sigma-Aldrich) or 10 μg/ml anti-rat IFNγ (Preprotech, London, UK) were tested in the coculture model.

Flow cytometry was used to assess proliferation and cell death of lymphocytes and macrophages after a 4 day coculture. Here, cells were stained with PE-conjugated mouse-anti-rat CD3 (Immunosource, Erembodegem, Belgium) or CD11b (AbD Serotec, Düsseldorf, Germany) and 7 aminoactinomycin D (7AAD, BD Biosciences).

Isolated lymph node cells (20 × 10⁴) were cultured with MBP (10 μg/ml), OVA (10 μg/ml) or myelin-oligodendrocyte glycoprotein (MOG, 20 μg/ml). Additionally, 100 μg/ml of isolated myelin was added in some experiments. Following 48 hr, 1 μCi [³H]thymidine (Amersham, Buckinghamshire, UK) was added to the culture. Next, cells were harvested with an automatic cell harvester (Pharmacia, Uppsala, Sweden) and uptake of radioactivity was measured in a β-plate liquid scintillation counter (Wallac, Turku, Finland).

Coculture supernatants were collected and release of NO was determined using the griess reagent system (Promega, Leuven, Belgium), following the manufacturer's instructions. Absorbance was determined by using a microplate reader at 550 nm (Biorad Benchmark).

Snap-frozen brain and spinal cord material was cut in respectively the coronal and sagittal plane with a Leica CM1900UV cryostat (Leica Microsystems, Wetzlar, Germany) to obtain 10 μm sections. The extent of demyelination and infiltration was determined by staining with Luxol Fast Blue (LFB; Gurr BDH, Poole, England). Briefly, aceton-fixed slides were incubated with LFB for 16 hr at 56°C, destained with 0.05% lithium carbonate, and counterstained with cresyl violet (VWR). Analysis was carried out using a Nikon eclipse 80i microscope and NIS Elements BR 3.10 software (Nikon, Tokyo, Japan).

DiI-labeled myelin migration to popliteal and inguinal lymph nodes was determined by immunohistochemistry. Popliteal and inguinal lymph nodes were snap-frozen directly following isolation and cut into 10 μm sections. Following fixation in respectively aceton for 10 min, sections and cells were blocked using 10% goat serum (Millipore, Brussels, Belgium) in PBS. Subsequently, sections and cells were stained with mouse-anti-rat CD169 (1/250 in PBS; Abd Serotec), a marker for macrophages in lymph nodes. As a secondary antibody Alexa fluor 488 F(ab')₂ fragment of goat-anti mouse was used (1/500 in PBS; Invitrogen). Control staining was performed by omitting the primary antibody. Nuclear staining was performed using 4,6'-diamidino-2-phenylindole (DAPI; Invitrogen) for 10 min. Autofluorescence was minimalized by using 0.1% Sudan Black in 70% ethanol.

Data were statistically analyzed using GraphPad Prism for windows (version 4.03) and are reported as mean ± SEM. D'Agostino and Pearson omnibus normality test was used to test normal distribution. An analysis of variances (ANOVA) or two-tailed unpaired student T-test (with Welch's correction if necessary) was used for normally distributed data sets. The Kruskal-Wallis or Mann-Whitney analysis was used for data sets which did not pass normality. *P < 0,05, **P < 0,01 and ***P < 0,001.

Initially we assessed the capacity of peritoneal macrophages to internalize myelin. By culturing macrophages with different concentrations of DiI-labeled myelin for divergent periods of time, it was demonstrated that myelin is internalized in a time- and dose-dependent manner (Figure 1a). Next, the most optimal macrophage/lymphocyte coculture ratio was determined by using untreated macrophages. At high macrophage/lymphocyte ratio's, lymphocytes demonstrated a reduced viability and proliferation (Figure 1b and 1c). The decline of lymphocyte viability was absent at low macrophage/lymphocyte ratio's (<1/10). To observe differences in lymphocyte proliferation following coculture with mye-macrophages, subsequent experiments were conducted using a macrophage/lymphocyte ratio of 1/10.

To study whether mye-macrophages affect antigen-specific lymphocyte proliferation in a different manner compared to untreated macrophages, macrophages or mye-macrophages were cocultured for 4d with MBP- or OVA-reactive lymphocytes, irradiated thymocytes and purified MBP or OVA. Here it was demonstrated that mye-macrophages inhibit TCR-triggered lymphocyte proliferation more pronounced than untreated macrophages (Figure 1d and 1e). This process was independent of antigen-specificity, since both MBP and OVA-reactive lymphocytes showed the same reduction in proliferation. Proliferation differences were not due to an altered viability of lymphocytes or mye-macrophages (data not shown).

To elucidate the mechanisms behind the increased inhibition of lymphocyte proliferation by mye-macrophages, we assessed whether the phagocytosis process as such is responsible for the observed effects on lymphocyte proliferation. For this purpose, macrophages were loaded for 3 consecutive hours with latex beads or zymosan A prior to coculture with lymphocytes. Macrophage treatment with beads or zymosan significantly affected their capacity to modulate lymphocyte proliferation compared to control and myelin treated macrophages (Figure 2a), indicating that the observed increased inhibition of lymphocyte proliferation relies on myelin-specific effects, instead of being induced by the phagocytosis process itself.

Like DCs, macrophages can act as messengers of innate and adaptive immunity by presenting antigen in context of MHC molecules. Mye-macrophages can therefore be assumed to process endogenous myelin and present it to myelin-reactive lymphocytes, possibly influencing reactivity of these lymphocytes. To determine whether the observed immune suppressive effects are dependent on antigen presentation, untreated macrophages and mye-macrophages were cocultured with MBP-reactive lymphocytes in the absence of MBP and irradiated thymocytes. Mye-macrophages did not affect lymphocyte proliferation (Figure 2b), suggesting that, at a macrophage/lymphocyte ratio of 1/10, mye-macrophages do not influence the proliferation of MBP-reactive lymphocytes. These results are in line with the observed equal inhibition of proliferation of both MBP and OVA-reactive lymphocyte by mye-macrophages.

Lipid components of myelin, like cholesterol and arachidonic acid-containing phosphatidylcholine, have been reported to directly inhibit proliferation of ConA stimulated lymphocytes [41, 42]. Accordingly, we assessed whether isolated myelin has a direct effect on lymphocyte proliferation. By culturing lymphocytes with increasing concentrations of isolated myelin, in the absence of macrophages, it was established that myelin itself, even at high concentrations, had no significant influence on lymphocyte reactivity (data not shown).

Together, these data indicate that direct effects of myelin on lymphocyte proliferation, myelin-antigen presentation and the phagocytosis process itself are not responsible for the observed effects on lymphocyte proliferation.

Macrophages have been reported to inhibit lymphocyte proliferation in vitro by mechanisms involving inducible nitric oxide synthase (iNOS), arginase I or IDO [43‐55]. By administrating either an inhibitor of NOS (L-NMMA), arginase (NOR-NOHA) or IDO (1-MT) to the culture, the involvement of these enzymes was evaluated. Here we demonstrated that an increased activity of both IDO and arginase did not account for the increased inhibition of lymphocyte proliferation by mye-macrophages (Figure 2c). In contrast, administration of L-NMMA completely abrogated the inhibition of lymphocyte proliferation by both macrophages and mye-macrophages. Yet again, cocultures of both OVA and MBP-reactive lymphocytes were affected equally (data not shown). The importance of NO was further delineated by assessment of NO levels in the coculture supernatant of MBP-reactive lymphocytes. In contrast to cocultures of lymphocytes with latex beads or zymosan treated macrophages, a significantly increased concentration of NO was observed in the supernatants derived from cocultures with mye-macrophages, when compared to untreated macrophages (Figure 2d). This increased NO production by macrophages following myelin internalization was absent in monocultures (data not shown), indicating that the observed increased concentration of NO in the coculture with mye-macrophages was induced by lymphocytes. Lymphocyte-derived IFNγ has previously been described to induce NO production in macrophages [56]. However, in our coculture model no increase in IFNγ in the supernatants of cocultures with mye-macrophages compared to untreated macrophages was found (data not shown). Correspondingly, IFNγ neutralization did neither abrogate the increased inhibition of proliferation by mye-macrophages or decrease the NO production in cocultures (data not shown). Nonetheless, when transwell inserts were used to restrict direct cell-cell contact, lymphocyte proliferation in cocultures of both untreated or myelin-treated approached control values (Figure 2e), indicating a role for direct cell-cell contact in the induction of NO.

Noteworthy, when lymph node cells, isolated 9 days post-immunization, were exposed directly to MBP and myelin, an even more pronounced myelin-mediated inhibition of lymphocyte proliferation was observed (Figure 2f). The latter indicates that local lymph node phagocytes show a similar immune suppressive response as peritoneal macrophages following myelin ingestion. Differences detected in proliferation were not due to an altered viability of lymphocytes (Figure 2f).

Since we established that myelin internalization by macrophages alters their capacity to modulate T cell proliferation in vitro, we assessed the in vivo suppressive capacity of mye-macrophages. First, to determine whether subcutaneous injected myelin reaches the draining lymph node and is taken up by macrophages, we injected DiI-labeled myelin in the footpad of healthy animals. A notable migration of myelin towards popliteal lymph nodes was observed (Figure 3a). Immunohistochemical analysis further revealed that myelin was contained primarily in CD169⁺ macrophages located at the border of the medulla and in the subcapsular sinus (Figure 3b).

Next, OVA-immunized animals were treated subcutaneously in the footpad with myelin (d-4, 0, 4, 8 pre/post-immunization). Recall stimulation in vitro, 9d post-immunization, revealed a reduced cognate antigen specific proliferation in animals treated with myelin (Figure 4a). Interestingly, LPS-stimulated lymph node cultures from myelin-treated animals demonstrated an increased NO production (Figure 4b). These results demonstrate that myelin is capable of suppressing antigen-specific proliferation in vivo and suggest that an increased NO production by these cells is responsible for this effect.

In contrast to OVA-immunized animals, MBP-immunized animals revealed an increased reactivity towards MBP and MOG following myelin treatment (Figure 5a). Moreover, myelin-treated animals demonstrated an earlier onset of paralysis and a more severe neurological score at the peak of disease (Figure 5b). In concordance, a LFB/cresyl violet staining of brain and spinal cord sections demonstrated increased cellular infiltrates and the presence of demyelinated areas in myelin-treated animals (Figure 5c). Noteworthy, when MBP-immunized animals were treated with latex beads no overt effect on disease score and onset was found (data not shown), indicating that the aggravated disease score following myelin treatment is myelin-specific.

To exclude the possibility that in the restricted area of the lymph node the availability of OVA in OVA-immunized animals is altered by the additional presence of myelin proteins, hereby reducing OVA-specific proliferation, MBP-immunized animals were treated subcutaneously with OVA. As expected, an increased reactivity towards OVA was observed in isolated lymph node cells (Figure 6a). However, MBP-specific proliferation was unchanged (Figure 6b), demonstrating that the observed myelin-mediated inhibition of proliferation of lymph nodes cultures from OVA-immunized is independent of a reduced availability of OVA by myelin proteins.

These results indicate that mye-macrophages have a dual influence on proliferation, depending on the nature of surrounding lymphocytes. On one hand they aggravate autoimmunity by activating myelin-reactive lymphocytes and on the other hand they have the capacity to suppress lymphocyte activation by the secretion of NO.