Pathologic and Protective Roles for Microglial Subsets and Bone Marrow- and Blood-Derived Myeloid Cells in Central Nervous System Inflammation

Inflammation is a series of processes designed for eventual clearance of pathogens and repair of damaged tissue. In the context of autoimmune recognition, inflammatory processes are usually considered to be pathological. This is also true for inflammatory responses in the central nervous system (CNS). However, as in other tissues, neuroinflammation can have beneficial as well as pathological outcomes. The complex role of encephalitogenic T cells in multiple sclerosis and its animal model experimental autoimmune encephalomyelitis (EAE) may derive from heterogeneity of the myeloid cells with which these T cells interact within the CNS. Myeloid cells, including resident microglia and infiltrating bone marrow-derived cells, such as dendritic cells (DC) and monocytes/macrophages [bone marrow-derived macrophages (BMDM)], are highly heterogeneous populations that may be involved in neurotoxicity and also immunoregulation and regenerative processes. Better understanding and characterization of myeloid cell heterogeneity is essential for future development of treatments controlling inflammation and inducing neuroprotection and neuroregeneration in diseased CNS. Here, we describe and compare three populations of myeloid cells: CD11c+ microglia, CD11c− microglia, and CD11c+ blood-derived cells in terms of their pathological versus protective functions in the CNS of mice with EAE. Our data show that CNS-resident microglia include functionally distinct subsets that can be distinguished by their expression of CD11c. These subsets differ in their expression of Arg-1, YM1, iNOS, IL-10, and IGF-1. Moreover, in contrast to BMDM/DC, both subsets of microglia express protective interferon-beta (IFNβ), high levels of colony-stimulating factor-1 receptor, and do not express the Th1-associated transcription factor T-bet. Taken together, our data suggest that CD11c+ microglia, CD11c− microglia, and infiltrating BMDM/DC represent separate and distinct populations and illustrate the heterogeneity of the CNS inflammatory environment.

Inflammation is a series of processes designed for eventual clearance of pathogens and repair of damaged tissue. In the context of autoimmune recognition, inflammatory processes are usually considered to be pathological. This is also true for inflammatory responses in the central nervous system (CNS). However, as in other tissues, neuroinflammation can have beneficial as well as pathological outcomes. The complex role of encephalitogenic T cells in multiple sclerosis and its animal model experimental autoimmune encephalomyelitis (EAE) may derive from heterogeneity of the myeloid cells with which these T cells interact within the CNS. Myeloid cells, including resident microglia and infiltrating bone marrow-derived cells, such as dendritic cells (DC) and monocytes/ macrophages [bone marrow-derived macrophages (BMDM)], are highly heterogeneous populations that may be involved in neurotoxicity and also immunoregulation and regenerative processes. Better understanding and characterization of myeloid cell heterogeneity is essential for future development of treatments controlling inflammation and inducing neuroprotection and neuroregeneration in diseased CNS. Here, we describe and compare three populations of myeloid cells: CD11c + microglia, CD11c − microglia, and CD11c + blood-derived cells in terms of their pathological versus protective functions in the CNS of mice with EAE. Our data show that CNS-resident microglia include functionally distinct subsets that can be distinguished by their expression of CD11c. These subsets differ in their expression of Arg-1, YM1, iNOS, IL-10, and IGF-1. Moreover, in contrast to BMDM/DC, both subsets of microglia express protective interferon-beta (IFNβ), high levels of colony-stimulating factor-1 receptor, and do not express the Th1associated transcription factor T-bet. Taken together, our data suggest that CD11c + microglia, CD11c − microglia, and infiltrating BMDM/DC represent separate and distinct populations and illustrate the heterogeneity of the CNS inflammatory environment.
Keywords: microglia, eae, cD11c, myeloid cells, cns, M1/M2 Myeloid cell heterogeneity in EAE Frontiers in Immunology | www.frontiersin.org introduction The contribution of myeloid cells to central nervous system (CNS) function is increasingly appreciated. This is especially relevant in CNS inflammation, where they can act not only as participants in the immune attack but also as regulators of the immune response and promoters of neuronal and synaptic protection and regeneration (1). A key aspect of the inflammatory response within the CNS is the recruitment and reactivation of infiltrating CD4 + T cells and direction of their effector function within the tissue (2). Their role in multiple sclerosis (MS) and its animal model experimental autoimmune encephalomyelitis (EAE) is complex. Although there is a consensus that inhibition or blockade of such responses has therapeutic benefit in MS and EAE [reviewed in Ref. (3)], there are numerous reports of beneficial effect of the activation of such T cells for myelin clearance, remyelination, and repair (4,5). The cells implicated in interaction with CD4 + T cells within the inflamed CNS are myeloid cells, such as infiltrating dendritic cells (DC) and monocytes/macrophages, and also resident microglia (6)(7)(8)(9). These are very heterogeneous populations of immune cells that can dampen inflammation or contribute to pathological events. Better characterization of the heterogeneity of blood-derived as well as CNS-resident myeloid cells and their interaction with T cells is crucial for understanding how to effectively control inflammation and induce neuroprotection and neuroregeneration in diseases like MS.
Microglia are cells of myeloid lineage considered to be CNS macrophages [reviewed in Ref. (10)]. However, they differ in terms of origin: while macrophages originate from fetal liver, microglia emerge from yolk sac and colonize the CNS early during neonatal development (11). Microglia are autonomously maintained through proliferation and in normal circumstances, they are not replaced by bone marrow (BM)-derived cells (12). Based on many phenotypic similarities between microglia and macrophages, the discrimination between these two populations as well as other BM-or blood-derived cells during neuroinflammation is challenging. The difference in the expression level of cell membrane tyrosine phosphatase CD45 can be used to discriminate CD45 low microglia from CD45 high blood-derived cells by flow cytometry (13,14). In addition, other common markers used for microglial identification include CD11b, CD11c, Iba1, and CX3CR1 (10,(15)(16)(17)(18)(19)(20). However, these can also be expressed by macrophages as well as monocytes and DC. Recent development of a double knock-in red (CCR2)/green (CX3CR1) mouse has allowed the discrimination of CNS-resident microglia from infiltrating leukocytes, microglia being CX3CR1-positive but, in contrast to infiltrating immune cells, CCR2-negative (21).
In this study, we aimed to examine the heterogeneity amongst microglia as well as to compare them to infiltrating blood-derived myeloid cells in EAE. We show that CD11c + microglia, CD11c − microglia, and CD11c + blood-derived cells are functionally distinct populations of myeloid cells within the inflamed CNS. Our findings illustrate the heterogeneity of the CNS inflammatory environment, reflecting the complexity of immunological processes that lead to pathology, tissue protection, or immunoregulation.

Materials and Methods
Mice Female C57BL/6j bom (B6) mice aged 6-8 weeks were obtained from Taconic Europe A/S (Lille Skensved, Denmark) and CX3CR1 gfp/gfp were obtained from The Jackson Laboratory (Bar Harbor, ME, USA) and maintained in the Biomedical Laboratory, University of Southern Denmark (Odense). All experiments were approved by the Danish Animal Experiments Inspectorate (approval number 2014-15-0201-00369).

NMO-like disease
To induce NMO-like lesions, 7-10-week-old female B6 mice received IgG antibodies isolated from NMO patients and human complement by injection in the striatum, as described in Ref. (22).

EAE-model
Seven to ten weeks old female mice were immunized by injecting subcutaneously 100 μl of an emulsion containing 100 μg of myelin oligodendrocyte glycoprotein (MOG)p35-55 (TAG Copenhagen A/S, Frederiksberg, Denmark) in incomplete freunds adjuvant (DIFCO, Alberstslund, Denmark) supplemented with 400 μg H37Ra Mycobacterium tuberculosis (DIFCO). Bordetella pertussis toxin (300 ng; Sigma-Aldrich, Brøndby, Denmark) in 200 μl of PBS was injected intraperitoneally at day 0 and day 2. Animals were monitored daily from day 5 and scored on a 6-point scale as follows: 0, no symptoms; 0.5, partial loss of tail tonus; 1, complete loss of tail tonus; 2, difficulty to right, 3, paresis in one or both hind legs; 4, paralysis in one or both hind legs; 5, front limb paresis; and 6, moribund. Severe EAE usually developed 14-18 days after immunization and was defined as a score of 3-5.

Bromodeoxyuridine (BrdU) Proliferation assay
B6 mice received 100 μl of 1 mg/ml BrdU by daily intraperitoneal injection, starting from day 10 after immunization. Mice with severe EAE were sacrificed and analyzed by flow cytometry using BrdU flow kit (BD Pharmingen) according to the manufacturer's protocol. Data were acquired on an LSRII™ flow cytometer and analyzed using FACSDiva™ 7 software (BD Biosciences). rna extraction, reverse Transcription, and Quantitative real-Time Pcr Sorted splenic CD11c + cells from individual mice as well as CD11c + microglia, CD11c − microglia, and CD45 high CD11c + cells from pools of 2-3 mice were placed in RLT buffer (Qiagen, Copenhagen, Denmark). Total RNA was extracted using RNeasy columns as per the manufacturer's protocol (Qiagen). Reverse transcription was performed with M-MLV reverse transcriptase (Invitrogen) according to the manufacturer's protocol.
histology Fifty-micrometer sections (coronal and sagittal) from 4% PFAfixed, frozen brains of PBS-perfused B6 mice were cut on a cryostat and stored in de Olmos cryoprotectant solution (24) at −14°C. Sections were washed in PBS and incubated for 30 min with 10% methanol and 10% H2O2 in PBS to block endogenous peroxidase. Alternatively 12 μm spinal cord sections were fixed with 4% PFA. After repeated rinses with 0.2% Triton X100 in PBS (PBST), sections were incubated for 1 h in 3% BSA in PBS to block unspecific binding. Next, sections were incubated overnight at 4°C with corresponding primary antibodies: rabbit anti-Iba1 (Wako, Neuss, Germany), hamster anti-CD11c (25) (kindly provided by Dr. Diego Gomez-Nicola; Centre for Biological Sciences, University of Southampton, SO16 6YD, Southampton, UK), rat anti-mouse Mac-1/CD11b (clone MCA711, AbD Serotec, Kidlington, UK), and rabbit anti-T-bet (H-210, Santa Cruz Biotechnology). Following primary antibody incubation, the sections were washed with PBST and incubated for 1 h with the appropriate secondary antibody: goat anti-Armenian hamster IgG-HRP (Santa Cruz Biotechnology, Heidelberg, Germany), goat anti-rabbit Alexa 488 (Invitrogen, Taastrup, Denmark), goat anti-rat Alexa 488 or biotinylated goat anti-rabbit (Abcam, Cambridge, UK) and PE-streptavidin (Biolegend). To visualize CD11c following secondary antibody incubation, the sections were washed with PBS and incubated for 4 min with TSA™ Plus Cy3 System (PerkinElmer, Skovlunde, Denmark). After immunofluorescent labeling, the sections were counterstained with DAPI and mounted with Fluorescence Mounting Medium (DAKO, Glostrup, Denmark). The sections were visualized on an Olympus FV1000MPE Confocal microscope and analyzed by FV10-ASW 4.06 software (Olympus, Denmark).

statistical analysis
All experiments were repeated at least three times and data are presented as means ± SEM. Statistical significance was assessed using the two-tailed Mann-Whitney U-test (GraphPad Prism 4). P-values <0.05 were considered significant. results and Discussion cD11c + Microglia emerge During neuroinflammation CD11b and CD11c form integrin heterodimers with CD18 and function as complement receptors. Both can be expressed by microglia [reviewed in Ref. (10)]. However, CD11c expression is usually attributed to activation of these cells. In humans, CD11c was shown to be constitutively expressed at a low level in microglia and upregulated in Alzheimer's disease (AD) (26). CD11c + Iba1 + cells were found in developing mouse brain as early as E16, they were also evident during postnatal neurodevelopment P2 (27), while in adult brain they were hardly visible (20,27). We have confirmed prominence of CD11c-positive cells in the developing brain, most of them being CCR2 − CX3CR1 + Iba1 + microglia (unpublished data). In adult mice, only very few CD11c + cells are present in the steady state, averaging 2-3% of total microglia (Figure 2A). However, their proportions increase dramatically during active neuroinflammation, such as in EAE (19), cuprizone-induced demyelination (23), and in a mouse model for neuromyelitis optica (Figure 2A), and also in an animal model for AD (28). Due to many phenotypic overlaps, bone marrow-derived macrophages (BMDM)/DC and microglia within the inflamed CNS are difficult to distinguish from each other by histology. Nevertheless, we assessed the localization of CD11c + cells within the inflamed spinal cord. We identified three different populations of myeloid cells within the lesion: Iba1 + CD11c + and Iba1 + CD11c − cells, that are likely to include microglia/BMDM, as well as CD11c + Iba1 − cells, most likely DC. Beside differential expression of these markers, they show different morphology, ranging from round, as well as dendritic-like CD11c + Iba1 − cells to amoeboid double-positive and Iba1-single positive cells. All of them were uniformly distributed within the lesion, with obvious possibility for interaction with T cells and with each other (Figure 2B). These data emphasize the high morphological and phenotypic heterogeneity of myeloid cell populations in neuroinflammation.

cX3cr1 Deficiency influences eae susceptibility
Fractalkine receptor CX3CR1 and its ligand CX3CL1 are expressed within the CNS by microglia and neurons, respectively (29). Their interaction plays an important role in neurodevelopment and has also been implicated in neuroinflammation. Disruption of CX3CR1-CX3CL1 axis leads to microglial activation and hypersensitivity and may result in neurotoxicity [ (18) and reviewed in Ref. (30)]. For instance, mice deficient in CX3CR1 have been reported to develop more severe EAE symptoms than WT animals (31). Here, we show that mice lacking CX3CR1 are more susceptible to EAE, with a significantly earlier onset and higher incidence of the disease (Figure 3). This data support conclusions from BM chimeras (31) that CX3CR1 regulates myeloid cell and possibly microglial responses, which in this case would act to dampen neuroinflammation, as also evidenced in CX3CR1-deficient mice (18).  defense, such as tumor necrosis factor (TNFα), IL-12, or nitric oxide (NO) produced by inducible NO synthase (iNOS). By contrast, alternatively activated macrophages that are involved in the regulation of the immune response to limit inflammation and promote tissue repair express arginase-1 (ARG1) or chitinase-like proteins, such as YM1. Such cells also secrete growth factors, such as insulin-like growth factor 1 (IGF-1) or anti-inflammatory cytokine IL-10 (33). We have addressed functional correlates to CD11c phenotypic distinction for microglia sorted from the CNS of B6 and CX3CR1 GFP/GFP mice with EAE, by analysis of mRNA levels of ARG1, YM1, iNOS, IL-10, and IGF1. All of them showed clear-cut preferential expression by one or another microglial subset (Figure 4). There was no difference in TNFα expression between these microglial populations (not shown). These results, therefore, point to functional distinctions between CD11c + and CD11c − microglia. Moreover, CX3CR1 deficiency did not influence M1/M2-associated gene expression in microglia except for YM1, which was upregulated in CX3CR1-deficient microglia (Figure 4). These data suggest that the susceptibility to EAE in CX3CR1-deficient mice is not dependent on microglial M1/ M2 phenotype.

Myeloid cell heterogeneity in inflamed cns in eae
We were further interested in differences between microglia and blood-derived myeloid cells; thus, we compared mRNA levels of ARG1, YM1, iNOS, IL-10, and IGF1 between microglia populations and CD45 high CD11c + cells (BMDM/DC). Here, we show that BMDM/DC express significantly higher levels of ARG1, iNOS, and IL-10 than both populations of microglia, they also expressed significantly more transcripts of YM1 than CD11c + microglia, but less than CD11c − microglia ( Figure 5). Interestingly, while CD11c + microglia showed high expression of IGF1 transcripts, neither CD11c − microglia nor infiltrating CD45 high cells expressed significant levels of this growth factor ( Figure 5). Association of IGF1 expression with CD11c + microglia has also been reported in glatiramer acetate-treated transgenic mice with an AD-like phenotype and has been suggested as a mechanistic basis for the ability of CD11c + microglia to promote neurogenesis (34).
Since neurogenesis may be considered as a part of the anti-inflammatory or protective spectrum, we asked whether these cells additionally selectively expressed other regulatory cytokines. It was recently published that IFNβ produced by a   Figure 6A). On this basis, there was, therefore, no evidence for a selective regulatory role for CD11c + microglia. By contrast, infiltrating BMDM/DC expressed very low levels of IFNβ mRNA comparable to unmanipulated splenic CD11c + cells, consistent with lack of a regulatory role for these cells, distinct from CNS-resident microglia. However, all of these brain myeloid populations may respond to type I IFN, as evidenced by equivalent expression of IRF7 and IRF3 (Figures 6B,C).
Interestingly, IRF3 was upregulated in BMDM/DC compared to unmanipulated splenic CD11c + cells (Figure 6C), suggesting ongoing active response within the CNS. These data further suggest a beneficial role for microglia in demyelinating diseases as well as emphasize functional distinctions between them and infiltrating BMDM/DC. Peripherally administered IFNβ is used as a first-line therapy for MS (37). IFNβ, constitutively expressed by microglial cells, is increased in EAE and play a protective role in the CNS (35,36). Whether this protection might involve IFNβ-driven DC response may be speculated but would be consistent with the expression of the IFNβ response gene IRF7 by DC, which we have shown. Our findings support observations made by Yamasaki et al. who used a combination of differential reporter gene expression and morphological correlates in EAE to infer that whereas BMDM were implicated in myelin stripping at nodes of Ranvier, microglia were not intimately associated with the demyelinating axon. Their location, relative morphology, and gene expression profile pointed to microglial involvement in phagocytosis and clearance of debris (38). Together with our findings, now based on consensus phenotypic distinction between microglia and BMDM, microglia can be seen to play a beneficial role in CNS inflammation, as has been discussed elsewhere (39).

Myeloid cells Turn-Over During eae
Microglia have been shown to be a self-renewing population within the CNS. Their proliferation in response to neurodegeneration and their maintenance in adult mice were reported to be dependent on macrophage colony-stimulating factor receptor (CSF1R) signaling (25,40). CSF1R signaling in response to its ligands, CSF1 and IL-34, induces microglial proliferation (40), suggesting tonic self-renewal as a component of normal homeostasis. Given the emergence of CD11c + microglia in neuroinflammation, we asked whether CSF1R signaling might play a selective role in expansion of these cells. We used BrdU incorporation to assess microglial proliferation. Both CD11c + and CD11c − populations of microglia proliferated equivalently during EAE (Figures 7A,B), arguing against proliferation per se as basis for relative expansion of CD11c + cells. Consistent with this, there was no significant difference in expression of CSF1R by CD11c + and CD11c − microglia ( Figure 7C). Strikingly, infiltrating BMDM/ DC (CD45 high CD11c + cells) expressed very low levels of CSF1R ( Figure 7C). This again emphasizes that differences between subsets of CNS-resident microglia are minor when both are compared to BMDM/DC.

BMDM/Dc but not Microglia express T-Bet During eae
We have shown that CD11c + microglia in EAE express MHC-II and co-stimulatory molecules CD80 and CD86, and that these cells have potent ability to induce proliferation of antigen-primed CD4 + T cells (19). This has led to speculation that they may reflect a more DC-like subset of microglia, as was previously suggested, largely on the basis of morphology and CD11c expression (41). However, they differ from DC in their relative expression of Th1and Th17-inducing cytokines and in their quantitative ability to induce such T cell responses, CD11c + microglia being, noticeably, less effective (19). We have now further explored possible overlap between CD11c + microglia and DC by comparison of their expression of the transcription factor T-bet. T-bet was first isolated as a novel Th1-specific T box transcription factor that controls the expression of the hallmark Th1 cytokine, IFNγ (42).  . T-bet is classically associated with the Th1 CD4 + T cell subset but also controls IL23R expression by Th17 (44). It has been implicated in regulation of effector responses by CD8 + T cells (45,46) and in control of Th-1 expression of granulocyte-macrophage colony-stimulating factor (GM-CSF or CSF2) (47), which has been proposed to be the link between pathogenic T cells and infiltrating myeloid cells (48). T-bet was reported to be required for Th1 and Th17 encephalitogenicity (49), although this has been contradicted by later studies (50). Earlier studies had shown that T-bet was also expressed by DC and regulated DC functions (45,46), and although reported to play a role in IFNγ production by DC, that is unlikely to be a major activity of this cell type. Expression of T-bet by DC was critical for Th1 induction and for the generation of T cell memory (43). T-bet was later shown to play a role in DC control of mast cell precursor migration (51). We compared levels of T-bet mRNA in microglia and BMDM/DC isolated from the CNS of mice with EAE. Whereas CD45 high CD11c + BMDM/DC robustly expressed high levels of T-bet mRNA, it was undetectable in most microglial samples, regardless of CD11c expression ( Figure 8A). The expression of T-bet is, therefore, a selective property of infiltrating cells, likely to be specific for DC based on previous reports, and does not occur in CNS-resident microglia.
Glimcher and colleagues showed that DC express T-bet (45). Microglial expression had not previously been reported and our findings, somewhat surprisingly given the many overlaps in myeloid cell phenotype and their involvement in Th1 induction and chemokine responses, indicate that, in fact, these brain macrophages do not express significant levels of T-bet. Thus, it was of interest that such a high proportion of T-bet-expressing cells in EAE infiltrates should turn out to be DC rather than T cells.
We were interested whether T-bet + DC co-localized with microglial cells within EAE lesions. Using immunohistochemistry and immunofluorescence protocols, we showed that although some CD4 + cells co-stained with T-bet, as expected, the majority of T-bet + cells did not express CD4 (data not shown), but stained for Mac-1/CD11b (Figure 8B). These represent the sorted CD45 high BMDM/DC, which we showed to be the only myeloid cell source of T-bet mRNA ( Figure 8A). EAE lesions are, therefore, highly heterogeneous, containing regulatory microglia with differential ability to present antigen to and induce cytokine responses from infiltrating T cells as well as infiltrating BMDM/DC that show no regulatory phenotype, many of which express T-bet.
All together, these data suggest that T-bet may be useful as a potential marker (along with myeloid markers, such as CD11b) for infiltrating DC in neuroinflammation, given that these populations of myeloid cells are morphologically indistinguishable.
summary We, thus, show that CNS-resident microglia that are maintained at least in part by proliferation include functionally distinct subsets that can be distinguished by expression of CD11c. These subsets differ in their expression of ARG1, YM1, iNOS, IL-10, and IGF1. These markers have been used as a basis for a M1/M2 phenotype classification for macrophages and were extended to describe microglia. Although both subpopulations of microglia clearly differed from each other, they did not show a pure M1 or M2 phenotype. This further supports the detailed transcriptional analysis of LPS-activated macrophages, which has revealed limits to such a binary distinction (52) and is in line with other studies where microglia have been shown to express M1 and M2-related genes simultaneously (53,54). It is, thus, probably more useful to define phenotypic/functional correlates on a case-by-case basis as, for instance, in Ref. (55). The issue of M1/M2 microglial phenotype and its limitations have been thoroughly discussed in Ref. (39).
We further show that despite these phenotypic and functional distinctions, microglia differ from blood-derived or BMDM/ DC in that they express the regulatory cytokine IFNβ, to which BMDM/DC may respond but do not express, and that BMDM/ DC express the Th1-associated transcription factor T-bet, whereas microglia do not. The fact that DC and microglia co-localize within Myeloid cell heterogeneity in EAE Frontiers in Immunology | www.frontiersin.org inflammatory lesions in EAE emphasizes that the microenvironment in which autoreactive T cells are activated is complex and includes both regulatory and pro-pathologic myeloid cells that themselves may exert effector functions in the inflamed CNS.
conclusion Over two decades of research on CD4 + T cell subsets based on their cytokine production profile has led to major increases in understanding of the inflammatory process. Myeloid cells in neuroinflammation also show heterogeneity and may need to be equivalently thoroughly studied in order to better describe their functions and their interaction with T cell subsets and neural cells in the CNS.
author contributions AW and TO designed the work. AW, KJ, AJ, JM, RK, and OC performed the experiments. AW, OC, and TO analyzed the data and wrote the paper. All authors read and approved the final manuscript.

acknowledgments
We thank Dina Draeby and Pia Nyborg Nielsen for expert technical assistance, Inger Andersen and Lars Vitved for help with cell sorting, Zsolt Illes and Justyna Okarmus for collaboration on cuprizone experiments. We are grateful to Nasrin Asgari (Department Neurology, Vejle Hospital, Denmark) for providing us with NMO-IgG and to Diego Gomez-Nicola (University of Southampton, UK) for providing anti-CD11c antibody. The bioimaging experiments reported in this paper were performed at DaMBIC, a bioimaging research core facility, at the University of Southern Denmark. DaMBIC was established by an equipment grant from the Danish Agency for Science Technology and Innovation and by internal funding from the University of Southern Denmark.