Sialoadhesin-Positive Macrophages Bind Regulatory T Cells, Negatively Controlling Their Expansion and Autoimmune Disease Progression¹

Murine experimental autoimmune encephalomyelitis (EAE),³ a widely used T cell-mediated animal model for multiple sclerosis, is characterized by self-reactivity directed against several myelin-derived Ags, including myelin basic protein and myelin oligodendrocyte glycoprotein (MOG). During EAE, autoreactive myelin-specific CD4⁺ T cells penetrate into the CNS and initiate destruction and demyelination processes executed by activated macrophages (1, 2).

An important role in the regulation of autoimmune and other inflammatory processes has been ascribed to CD4⁺CD25⁺ regulatory T cells (Tregs). Abnormalities in the generation and function of these cells and resulting immune dysregulation are considered the primary cause of autoimmune diseases and other immunological disorders (3). Tregs are mainly characterized by expression of the transcription factor Foxp3 and the IL-2 receptor CD25 and by a distinct cytokine profile which includes IL-10 and TGF-β. Their generation and function, which is the suppression of the activation and proliferation of the CD4⁺CD25⁻ effector T cells (Teff) during immune responses, is based on their surface markers like CD127, CD62L, CD103, CD122, CTLA-4, and GITR (3). Ab-blocking studies indicate that CTLA-4 (4, 5), GITR (6, 7, 8), and IL-2 (9, 10) are required for the generation of Tregs and are thereby involved in the maintenance of the balance between Teffs and Tregs. The suppressive activity of Tregs depends on direct cell-cell communication between Tregs, Teffs, and APCs, although the molecular basis for this suppression has not yet been clarified (11). An involvement of carbohydrate structures in Treg functions is evident from recent implications of galactose-binding molecules of the galectin family in these activities (12, 13).

Sialoadhesin (Sn) is a macrophage-restricted prototypic member of the siglec family of sialic acid-binding molecules (14) found on a subpopulation of macrophages within the subcapsular sinus and medulla of lymph nodes and on metallophilic macrophages in spleen (15, 16). During inflammatory disorders such as rheumatoid arthritis and experimental autoimmune uveoretinitis (EAU), where macrophages are considered to play decisive roles, Sn is expressed by activated macrophages within the inflamed organs (16, 17). In vitro studies have demonstrated that Sn binds several membrane proteins, mainly expressed on leukocytes, via both sialic acid-dependent (CD43, PGSL, MUC1) and independent (MGL1) mechanisms (18, 19, 20). However, the biologically relevant interactions of Sn and its ligands are still not clear, and counterreceptors for Sn within macrophage-infiltrated tissues remain to be identified (18). The structural features of Sn and its high conservation on activated macrophages are suggestive of a role in mediating cell-cell interactions.

We show here that Sn⁺ macrophages represent a significant proportion of leukocytes infiltrating into the CNS upon EAE induction by immunization with MOG_35–55. Sn knockout (KO) mice show significantly reduced EAE severity and incidence compared with wild-type (WT) littermates, which is associated with a significant increase in Treg numbers and reduction in Teff numbers within the CNS, but no differences in total macrophages. Through the use of a Sn-Fc fusion protein, a subpopulation of Tregs expressing Sn ligands was identified in the inflamed CNS, which was elevated in Sn KO mice. In vitro experiments confirmed that the Sn KO macrophages induce higher rates of proliferation of Tregs than WT macrophages. Thus, Sn-positive macrophages negatively regulate the expansion of Tregs through Sn-dependent direct cell-cell communication during the immune response. Upon loss of Sn-positive macrophages, the proliferation of Tregs is up-regulated, specifically in the CNS, resulting in a milder disease course.

Sn KO mice (Sn^−/−) and WT littermates (Sn^+/+) (backcrossed to C57BL/6 at least 10 generations) were used in active EAE experiments (21). C57BL/6-Ly5.1 (CD45.1) mice were used in passive transfer experiments. Experiments were conducted according to German Animal Welfare guidelines.

Abs used in immunofluorescence, FACS, and cell sorting included: pan-laminin (455) (22, 23), CD45 (30G.12; BD Pharmingen), CD45.2 (104; eBioscience) and CD45.1 (A20; BD Pharmingen), CD11c (N418; eBioscience), CD11b/MAC-1 (M1/70; BD Pharmingen), CD4 (H129.19; BD Pharmingen) and CD8 (53-6.7, eBioscience), CD25 (7D4; BD Pharmingen), CD169 (3D6.112; Serotec), Foxp3 (Abcam), F4/80 (BM8; eBioscience), IL-17 (TC11-18H10.1; BD Pharmingen), IFN-γ (XMG1.2; BD Pharmingen), IL-6 (MP5-20F3; eBioscience), IL-10 (JES5-16E3; eBioscience), TNF-α (MP6-XT22; BD Pharmingen), CD16/CD32 (2.4G2; BD Pharmingen), MHC class II (M5/114.15.2; eBioscience), CD86 (2331; BD Pharmingen), CD80 (16-10A1; BD Pharmingen), CD40 (3/23; BD Pharmingen), and CD69 (H1.2F3; BD Pharmingen). For function-blocking experiments, two Sn Abs were used, SER-4 and 1C2 (24).

EAE was actively induced in 6- to 8-wk-old female C57BL/6 mice by immunization with the 35–55 peptide of MOG as described previously (25). The mice were observed daily and scored on a scale of 0–5 for neurological defects: 1, flaccid tail; 2, hind limb weakness; 3, severe hind limb weakness; 4, hind quarter paralysis; and 5, forelimb weakness. CNS samples were collected at different EAE stages for FACS analysis or frozen in Tissue-Tek (Sakura Finetek) for immunofluorescence analysis.

Cells were isolated at day 10 after immunization with MOG_35–55 peptide from draining lymph nodes (LNs), stimulated with 20 μg/ml MOG_35–55 peptide for 2 days, and then in the presence of 20 ng/ml IL-2 and 10 ng/ml IL-1β for another 2 days. On day 4, 10 ng/ml IL-23 was added and the cells were cultured for another 5 days before transfer to recipient mice. Twenty × 10⁶ cells/mouse were transferred i.v. and mice were injected i.p. with 20 ng of pertussis toxin/mouse on the day of transfer and at day 2 after transfer.

Sn KO or WT mouse brain cryostat sections (5 μm) were analyzed by immunofluorescence as described previously (23). Primary Abs used included anti-Foxp3, anti-Sn (MOMA-1), anti-pan-laminin (455), and anti-CD45 and were visualized using Alexa Fluor 488- or Cy3-conjugated goat anti-rat or Cy3- or Cy5-conjugated anti-rabbit secondary Abs (The Jackson Laboratory and Molecular Probes). Negative controls involved incubation of sections with secondary Abs alone. Sections were examined using a Zeiss AxioImager microscope equipped with epifluorescent optics and documented using a Hamamatsu ORCA ER camera.

Mice were perfused with PBS before spleens, LNs, thymi, and brains were harvested. Spleens and brains were pretreated with 2 mg/ml collagenase D (Roche Diagnostics) and 1 mg/ml DNase I (Roche Diagnostics), and total cells were isolated using cell strainers (BD Biosciences), 70 μm for the spleens and 100 μm for the brains. Brain homogenates were separated into neuronal and leukocyte populations by discontinuous density gradient centrifugation using isotonic Percoll (Amersham Biosciences) (26). For intracellular staining, isolated leukocytes were stimulated with PMA (10 ng/ml)/ionomycin (1 μg/ml; Sigma-Aldrich) in the presence of 10 μg/ml brefeldin A (Sigma-Aldrich) at 37°C for 5 h. An intracellular staining kit (eBioscience) was used to permeabilize and fix the cells. Primary Abs used are listed above and stained cells were analyzed using a FACSCalibur (BD Biosciences). In all cases, isotype control Abs were used to determine background staining.

Cytokine concentration in culture supernatants of Tregs, Teffs, and macrophages from immunized Sn KO and WT mice was tested using a cytometric bead array kit (Bender MedSystems). Intracellular staining and FACS analyses were performed as described above.

To test Treg function, T cells were isolated from draining LNs and CD4⁺CD25⁻ and CD4⁺CD25⁺ cells were sorted by MACS (Miltenyi Biotec) and placed in RPMI 1640 plus 5% FCS and 2-ME. These Teffs and Tregs were mixed in different ratios and 2 × 10⁴ of the mixed T cells were cultured at 37°C for 3 days along with irradiated splenic dendritic cells (DCs; from a nonimmunized mouse) in the presence of different concentrations of MOG_35–55. Cell proliferation was determined by [³H]thymidine incorporation.

To test Ag presentation function of WT vs Sn KO macrophages, 2 × 10⁴ CD4⁺ T lymphocytes from LNs of immunized mice were cultured at 37°C for 3 days along with 1 × 10⁴ macrophages (from the CNS of immunized mice) and different concentrations of MOG_35–55. Cell proliferation was determined by [³H]thymidine incorporation.

To measure proliferation of Teff vs Treg lymphocytes, 2 × 10⁶ Teff or Treg lymphocytes isolated from immunized mice (by MACS sorting as described above) were incubated with 1 μM CFSE and cocultured with 10⁶ macrophages (from EAE mice) plus 20 μg/ml MOG_35–55 for 4 days before FACS analysis.

Sn-Fc protein (15 μg/ml) (24) was preincubated on ice for 1 h with 10 ng/ml Cy5-conjugated anti-human IgG Fc fragment Ab. Cells were then stained using the Sn-Fc protein-Cy5-anti-human-IgG complex or the Cy5-conjugated anti-human-IgG Fc Ab alone or in combination with other cell marker Abs and analyzed by FACS. Sialidase treatment involved adding 50 μl (60 mU) of neuraminidase from Vibrio cholerae (Calbiochem) in 1 ml of DMEM to 5 × 10⁶ cells and subsequent incubation at 37°C followed by extensive washing with PBS.

Five × 10⁶ donor Ly5.1 naive CD4⁺CD25⁺ T cells were transferred i.v. to Ly5.2 WT and KO recipient mice, and 3 days after transfer the recipients were immunized with MOG_35–55. On days 7, 9, and 11 after immunization, the mice were injected i.p. with 1.5 mg of BrdU. Brains and spinal cords were removed 12 h after the last injection. Lymphocytes were isolated as described above and analyzed by FACS using a BrdU kit (BD Biosciences) and cell markers listed above.

Quantitative data are expressed as means ± SEM. Statistical significance of EAE results was analyzed with an unpaired Student’s t test and by Mann-Whitney U test, whereas proliferation rates, cytokine production, and cell numbers were assessed by Student’s t test. Values of p < 0.05 were considered significant.

Active EAE induction in Sn KO mice and WT littermates revealed reduced disease severity throughout the disease course in KO mice but no significant alteration in the day of onset of disease symptoms (Fig. 1,A). Although 92% of WT mice developed disease symptoms, first appearing on average between days 10 and 12 after immunization, ∼80% of the KO mice developed symptoms throughout the experimental period (Fig. 1 A), indicating reduced disease susceptibility.

Despite reduced disease scores in KO animals, extensive leukocyte infiltration of the CNS parenchyma was apparent by immunofluorescence staining at peak disease severity (day 17) in both KO and WT brain sections (Fig. 1,B). However, quantitative microscopic examination of CNS sections revealed slightly fewer inflammatory cuffs per cm² in Sn KO mice compared with WT mice (Fig. 1,B). FACS analysis of perfused CNS (brain and spinal cord) at peak disease severity confirmed the absence of the Sn⁺CD11b⁺ macrophage population in KO mice, which represented 2% of total leukocytes in WT CNS (Fig. 1,C) and 20% of total macrophages, but no significant difference in total CD11b⁺ cells or in CD8⁺ T cells (Fig. 2,A) and CD11c DCs (data not shown) in the CNS of WT and KO mice. However, there was a moderate reduction in CD4⁺ T cells in the KO CNS (Fig. 2,A). No differences were observed in total CD45⁺ cells, CD4⁺ T cells, CD8⁺ T cells, CD11b⁺ cells (Fig. 2 A), and CD11c⁺ DCs (data not shown) in draining lymph nodes or in the circulation (data not shown) of KO mice vs WT littermates.

To define whether the reduced numbers of CD4⁺ T cells in Sn KO brains were due to defects in T cell migration into the CNS, passive transfer experiments were performed using CD45.2⁺ WT or KO MOG_35–55-specific T cell blasts transferred to CD45.1⁺ WT recipients, or CD45.1⁺ WT MOG_35–55-specific T cell blasts transferred to CD45.2⁺ WT or KO recipients, and analysis of brains at day 3 after transfer, a time point when T cell proliferation in the brain has not yet occurred (27). FACS revealed no differences in the number of donor WT or KO T cell blasts in the brains of WT or KO recipient mice (Fig. 2 B), indicating the absence of significant T cell migration defects in KO mice.

Because Sn KO mice showed reduced EAE severity and incidence without major deficiencies in total leukocyte infiltration and because published data suggest that the balance between Teffs and Tregs is important in the development of EAE symptoms (28, 29), analyses were extended to T cell subpopulations: Th1 and Th17 cells were analyzed in the CNS of KO and WT mice at peak disease severity (day 17), revealing no differences in the proportion of Th17 cells but a reduced proportion of Th1 cells in Sn KO mice. Absolute numbers of both Th1 and Th17 cells, however, were lower in the CNS of KO mice compared with their WT littermates (Fig. 2^,C). Tregs as defined by CD4⁺CD25⁺Foxp3⁺ were also elevated in various organs of Sn KO mice as compared with WT littermates, in particular the CNS (Fig. 2 D). FACS revealed that naive WT and KO mice showed identical numbers of Tregs in multiple organs, indicating that there is no defect in Treg development (supplemental Fig. 1⁴). The elevated Treg and reduced Th1 and Th17 cell numbers in the CNS of KO mice suggest increased T cell suppressor function, implicating Sn in the regulation of the balance between Tregs and Teffs during the immune response. Other regulatory immune cell populations were also examined during EAE, including NKT cells, γ/δ T cells, and B cells, revealing no significant differences between Sn KO mice and WT littermates (data not show).

To determine how Sn⁺ macrophages up-regulate Treg numbers, expression of Sn ligands was investigated on activated immune cells during EAE by flow cytometry using a fusion protein that combines a human IgG-Fc fragment with the first three Ig domains of Sn (Sn-Fc). This chimeric protein specifically detects sialic acid-containing structures recognized by the Sn molecule and, therefore, recognizes potential Sn counterreceptors (24). In combination with specific cell markers, this Sn-Fc therefore permits identification of the cell population/s that potentially interacts with Sn expressed naturally by macrophages. Furthermore, it permitted quantification of Sn ligands on different cells during the immune response.

In accordance with previously published observations (15, 16), CD45⁺CD11b⁺ cells were identified as the major population of Sn ligand-expressing cells in nonimmunized animals (supplemental Fig. 2). Especially in the bone marrow, >40% of the CD45⁺CD11b⁺ cells expressed high levels of Sn-accessible sialic acid residues, which were removed by sialidase treatment (data not shown). These Sn ligand-expressing cells were both Sn⁺ and Sn⁻. In the spleen and LNs, ∼35 and 10% of the CD45⁺CD11b⁺ cells were Sn ligand positive, respectively, while CD4⁺ and CD8⁺ T cells and B220⁺ B cells showed little or no expression (supplemental Fig. 3A).

During EAE, Sn ligand levels on CD45⁺CD11b⁺ cells were reduced (supplemental Fig. 2), and an additional CD4⁺Foxp3⁺ Treg population was detected that expressed Sn ligand and was particularly enriched in the CNS of both WT and KO mice (Fig. 3,A). By contrast, naive Tregs from spleens and LNs showed low levels of the Sn ligand (supplemental Fig. 3B), while CD4⁺Foxp3⁻ Teffs do not express the Sn ligand (Fig. 3 A).

In Sn KO mice, which showed elevated numbers of CD4⁺Foxp3⁺ Tregs, the proportion of Sn ligand-positive Tregs was also significantly higher (32.6%) than in WT littermates (19.3%) (Fig. 3,B). CD4⁺Foxp3⁻ Teffs, in contrast, did not show any up-regulation of the Sn ligand in KO mice (Fig. 3 B). The elevated numbers of Sn ligand-positive Tregs in KO mice during EAE suggest that direct interaction of Tregs with Sn⁺ macrophages down-regulates Treg numbers, in particular in the CNS.

In view of the elevated Treg numbers and increased levels of Sn ligand on Treg surfaces in Sn KO mice, in vitro proliferation assays were performed to exclude the possibility that high levels of Sn ligand on Tregs alters their suppressor effects on Teff proliferation: CD4⁺CD25⁻ T lymphocytes isolated from MOG_35–55-immunized WT mice were cultivated together with WT splenic DCs as APCs plus MOG_35–55 in the presence of increasing ratios of CD4⁺CD25⁺ Tregs isolated from immunized WT or KO mice. Ag-specific CD4⁺CD25⁻ Teff lymphocyte proliferation was inhibited by increasing ratios of Tregs regardless of whether they were isolated from WT or KO mice (Fig. 4,A). Levels of IL-10 as an indicator of Treg activity (30) was also comparable in KO and WT Tregs (Fig. 4 B), indicating that Treg function per se is not altered by the higher than usual Sn ligand levels on KO cells.

Because macrophages are potential APC that can regulate T cell proliferation (31, 32), the T cell proliferative response induced by Sn⁺ vs Sn⁻ macrophages was investigated. Since Sn could not be used to isolate macrophages from the KO mice, the CD11b and F4/80 double-positive macrophage population was isolated from immunized KO and WT mice, which in the WT mice contained the Sn⁺ population (Fig. 5,A). This macrophage population showed identical levels of surface markers, including MHC class II, B7-1, B7-2, and CD40 (33), in KO and WT animals (data not shown) but significant differences in cytokine profiles, with reduced intracellular levels of TNF-α, IFN-γ, IL-6, and increased IL-10 in KO cells (Fig. 5,B). Similarly, supernatants from cultured Sn KO macrophages also showed reduced levels of TNF-α, IFN-γ, IL-6, and IL-1α and increased IL-10 in comparison to WT macrophages (Fig. 5 C). This altered cytokine profile in the Sn KO mice provides evidence for altered macrophage function.

In vitro proliferation of MOG_35–55-specific CD4⁺ T cells using the KO or WT macrophages as APC revealed reduced T cell proliferation in the presence of Sn KO as compared with WT macrophages, suggesting that Sn KO macrophages are compromised in their ability to induce an Ag-specific T cell proliferative response (Fig. 5,D). In view of the elevated Treg numbers in KO mice, this inhibitory effect was further investigated using isolated WT Teffs or Tregs: due to the low numbers of Tregs that can be isolated even from inflamed brains and the limited sensitivity of the [³H]thymidine incorporation method, it was necessary to use enriched Teff and Treg lymphocyte samples isolated from immunized mice, which were CFSE-labeled, cultured together with Ag-loaded CD11b⁺F4/80⁺ WT or KO macrophages, and subsequently analyzed by FACS to determine the proportion of proliferating cells. Teffs exhibited comparable proliferation in the presence of either WT or KO macrophages. By contrast, Treg lymphocyte expansion was up-regulated by coculture with KO macrophages (Fig. 5 E), indicating that Sn KO macrophages specifically induce proliferation of Tregs but have no impact on the CD4⁺Foxp3⁻ Teffs. Hence, reduced proliferation of the CD4⁺ T cell population in the presence of the Sn KO macrophages is likely to be due to an expansion of Treg lymphocytes, which in turn reduced proliferation of Teffs.

To determine whether regulation of Treg proliferation requires direct interaction between the Sn molecule on the macrophage cell surface and Tregs, WT macrophage Sn was blocked by incubation with the Sn-specific Abs SER-4 and 1C2 and proliferation of Ag-specific CD4⁺ T lymphocytes was measured in the presence of increasing proportions of anti-Sn-treated or untreated macrophages. Anti-Sn-treated WT macrophages induced the same inhibition of CD4⁺ T lymphocyte proliferation as Sn KO macrophages (Fig. 6,A). Similar experiments performed with CFSE-labeled isolated Teffs or Tregs revealed enhanced proliferation of Tregs by anti-Sn-treated WT macrophages and no effect on Teff proliferation (Fig. 6 B).

The above data suggest a direct interaction between Sn⁺ macrophages and Tregs. To investigate whether Treg lymphocytes and Sn⁺ macrophages colocalize in vivo, spleens and inflamed WT brains were immunofluorescently stained for Sn⁺ macrophages, Foxp3⁺ Tregs, and pan-laminin to define the perivascular space and the marginal sinus. Sn⁺ macrophages occurred in close contact to Foxp3⁺ Tregs in the splenic marginal sinus (supplemental Fig. 4) and, within the brain, both in the perivascular cuff and in the CNS parenchyma, with the majority of the interactions occurring in the perivascular space (Fig. 7 A).

To confirm that the association between Sn KO macrophage and Tregs in the CNS has the same inhibitory effect on Treg expansion as observed in vitro, in vivo expansion of Tregs was measured in the CNS of WT and Sn KO mice during EAE: WT CD45.1⁺CD4⁺ encephalitogenic T cells were passively transferred to MOG_35–55-immunized CD45.2⁺ Sn KO mice and WT littermates, and their proliferation in the CNS was measured by BrdU incorporation. At day 12 after transfer, CD45.1⁺CD4⁺ donor T cells were recovered from the CNS and analyzed by flow cytometry for the expression of Foxp3 and the incorporation of BrdU. Since passively transferred autoreactive T cells directly migrate to the CNS without extra activation and proliferation in the periphery and the majority of T cell proliferation observed in the periphery is dominated by host cells, specific measurement of BrdU incorporation into donor CD4⁺CD25⁺Foxp3⁺ Tregs in the CNS reflects mainly cells proliferating within that organ. Approximately 13% of the donor CD4⁺CD25⁺Foxp3⁺ Tregs recovered from the CNS of KO mice represented proliferating cells, while only 6% of donor CD4⁺CD25⁺Foxp3⁺ Tregs had proliferated in WT recipient mice (Fig. 7 B), supporting an in vivo immunosuppressive role for Sn KO macrophages.

Sn-positive macrophages have been previously shown to represent a population of macrophages that stem from the circulation and are present in the CNS early after the induction of EAE (34, 35). Injection of chlodronate-loaded liposomes and subsequent transfer of encephalitogenic T lymphocytes results in resistance to EAE, even in the presence of infiltrating encephalitogenic T lymphocytes and a Th1 proinflammatory cytokine milieu, and correlates with reduction in F4/80⁺ and Mac-1⁺ macrophages and complete absence of Sn⁺ macrophages. The data suggest a significant role for macrophages within the CNS at early stages of EAE that is independent of later demyelinating events. Data presented here indicate that one of these functions is regulation of Treg numbers and thereby Teff lymphocyte function by a subpopulation of Sn⁺ macrophages.

The reduced EAE incidence and severity observed in Sn KO mice and associated up-regulation of Treg and reduction in Th1 and Th17 cells suggest that Sn⁺ macrophages regulate the balance between Treg and Teff lymphocytes in EAE. Data shown here clearly indicate that Sn⁺ macrophages represent a functionally distinct macrophage population that directly interacts with Sn ligand on Tregs to inhibit Treg proliferation, but not their function. Absence of defects in Treg function in the Sn KO mice was demonstrated by the normal cytokine profiles of Tregs isolated from Sn KO mice and, in particular, their ability to inhibit proliferation of MOG_35–55-specific Teff lymphocytes to the same extent as Tregs from WT littermates. In vitro assays suggest that direct interaction between Sn⁺ macrophages and Tregs is required for these proliferation effects since Ab blocking of the Sn molecule on the macrophage surface was sufficient to demonstrate the same enhanced proliferation of Tregs as observed in the presence of Sn⁻ macrophages. The direct interaction between Sn⁺ macrophages and Tregs was supported by the in vivo localization studies which showed that these two cell populations occur in close proximity both in the periphery, in the LN and spleen, and during EAE also in the CNS. During EAE, Treg numbers were particularly elevated in the brain where Foxp3⁺ Treg and Sn⁺ macrophages colocalized in both the parenchyma and the perivascular cuffs, bordered by the endothelial and parenchymal basement membranes, which represent sites of high cell density and therefore high potential for cell-cell interactions.

In addition to direct interaction between Sn⁺ macrophages and Tregs, the altered cytokine profile of the macrophages, with down-regulation of proinflammatory cytokines such as TNF-α, IL-6, IFN-γ, and up-regulation of the anti-inflammatory cytokine IL-10, may enhance the immunosuppressive phenotype as documented in other inflammatory situations (36, 37, 38). It is not possible to determine whether this altered cytokine profile results from the loss of the Sn molecule from the macrophage subpopulation or changes in macrophage phenotype due to the altered Teff:Treg ratios. However, in view of the proproliferative effect of activated macrophages treated with Sn Abs on Tregs in vitro, the altered cytokine profile is unlikely to account for the up-regulation of Tregs observed in the Sn KO brain. Rather, the data suggest that this requires direct interaction between Sn⁺ macrophages and Tregs, the nature of which remains to be defined. Sn can potentially bind a wide range of glycoconjugates carrying sialic acid ligands in α2,3, α2,6, and α2,8 linkages, and although previous studies identified CD43 as a potential counterreceptor on T cells, it is not clear whether up-regulation of Sn binding to Tregs reflects an overall increased sialylation or the induction of a high avidity counterreceptor(s). Although investigations are underway to identify potential Treg Sn counterreceptors, they are hampered by the fact that they are expressed only on activated Tregs, which are present in very low numbers, and by the very low affinity of Sn for sialic acids.

The data presented here are consistent with recent studies of EAU in Sn KO mice (39). Like EAE, EAU is an organ-specific (retina) CD4⁺ T cell-mediated autoimmune disease, which when induced in the Sn KO mouse shows the same reduced disease severity and incidence as reported here. Reduced EAU symptoms were shown to be associated with reduced Teff proliferation and although a molecular mechanism for this effect was not proposed, they could be explained by elevated numbers of Tregs as described here. Similarly, in two genetic models of peripheral and CNS demyelination, in which not only macrophages but also CD8⁺ T lymphocytes play an important pathogenic role, disease symptoms are ameliorated in Sn KO mice and are associated with reduced CD8⁺ T cell and macrophage numbers (40, 41). These results are also consistent with an up-regulation of the CD4⁺CD25⁺Foxp3⁺ Treg population in the Sn KO mice. However, since glycosylation pathways in CD4 and CD8 pathways are known to be differentially regulated (42), direct interactions between Sn and activated CD8 T cells may also be important (43).

A large body of data documents the suppressive effects of CD4⁺CD25⁺Foxp3⁺ Tregs on Teff proliferation and proinflammatory cytokine production (3, 11, 44), and development and differentiation of Treg cells has been elucidated by different research groups (45, 46, 47). However, the expansion of this subpopulation of T cells during the immune response is not well understood. Existing data on the regulation of Tregs focuses on DCs, the long-known classic APCs. There are several examples of specific DC populations being required for Treg expansion, for example, epidermal RANKL expression alters DC function to maintain the number of the peripheral Tregs (48) and, in EAE, SOCS3-deficient DCs have been shown to selectively induce expansion of Foxp3⁺ Tregs in a TGF-β-dependent manner (49). We show here that in addition to DCs, Sn⁺ macrophages also regulate Treg expansion, however, in a negative manner. During EAE development, engagement of macrophage Sn with its ligand on the cell surface of Tregs inhibits Treg proliferation, permitting the expansion of the Teff population and induction of disease symptoms. These events are early in the EAE development and the in vivo proliferation study along with the absence of defects in T cell migration to the CNS in the Sn KO mice suggest that passively transferred Treg lymphocytes hyperproliferate both in the periphery and in the CNS of the Sn KO mice compared with WT mice, which correlated with the milder phenotype of the disease. Sn KO macrophages recovered from the CNS of EAE mice also showed a significantly higher capacity of inducing Treg proliferation compared with those recovered from the spleen (data not shown), indicating that after penetration into the CNS macrophages acquire additional functions that favor Treg suppression.

In conclusion, we show here that during EAE Sn⁺ macrophages represent a distinct subpopulation of macrophages that interact directly with Treg lymphocytes, principally at the site of inflammation, to inhibit their proliferation and, thereby, regulate Teff lymphocyte function. Targeting Sn may, therefore, represent a novel means of regulating autoimmune disease progression.

We thank Rupert Hallmann for advice throughout the project and for critical reading of this manuscript.

The authors have no financial conflict of interest.