CD40 Expression by Microglial Cells Is Required for Their Completion of a Two-Step Activation Process during Central Nervous System Autoimmune Inflammation¹

Microglial cells are components of a CNS “fixed” macrophage system, which is thought to play a key role in the regulation of autoimmune inflammation in the CNS (1, 2). In the normal CNS, microglial cells exist in the resting state and express the macrophage cell marker CD11b, a low level of CD45, and virtually undetectable levels of MHC class II and the costimulatory molecules CD40 and CD86 (3, 4). The exact function of microglial cells is not known, but it has been suggested that microglial cells become activated and participate in propagation of inflammation in the CNS through Ag presentation and the production of proinflammatory cytokines/chemokines (5, 6, 7, 8). In vitro, the process of microglial cell activation has been characterized by the up-regulation of CD45, MHC class II, CD40, and CD86 (4, 9, 10), but this has not been studied extensively in vivo. One of the important molecules that is likely involved in microglial cell activation in vivo is CD40 (11), which is a member of the TNFR family and is a costimulatory molecule important for activation of B cells, macrophages, and dendritic cells (12). Stimulation of monocytic lineage cells through the CD40 pathway induces the expression of proinflammatory cytokines/chemokines and enhances the expression of costimulatory molecules (13). Interaction of CD40 with its ligand CD40L (CD154), which is primarily expressed by activated T cells, promotes the activation of microglial cells in vitro (9); however, this has not been definitively demonstrated in the CNS.

Experimental autoimmune encephalomyelitis (EAE)³ is broadly used as an animal model of the CNS autoimmune inflammatory disease multiple sclerosis (MS). Both MS and EAE are thought to be mediated by autoimmune CD4 T cells (encephalitogenic T cells) that recognize self-Ag within the myelin sheath such as myelin basic protein (MBP), proteolipid protein, and myelin oligodendrocyte glycoprotein (14). Activated encephalitogenic T cells have the ability to penetrate the blood-brain barrier and initiate inflammation in the CNS, which is characterized by the presence of multiple inflammatory lesions containing various populations of infiltrating leukocytes including macrophages and T cells (15). CD40-CD40L interactions were reported to be important for EAE/MS pathogenesis (16). It was demonstrated that CD40^−/− and CD40L^−/− animals were resistant to EAE (17, 18), and that myelin self-Ag-pulsed CD40^−/− dendritic cells were not able to induce EAE, while wild-type (WT) dendritic cells were (19). In addition, Abs that blocked either CD40 or CD40L ameliorated EAE disease symptoms in both mice and marmoset monkeys (20, 21, 22). In contrast, administration of an activating CD40-specific Ab overcame tolerance to myelin self-Ag and triggered EAE (23).

Although CD40 is implicated in EAE susceptibility, only one study has investigated the role of CD40 expression in the CNS during EAE (17). This study demonstrated that bone marrow (BM) chimera mice with a deficiency in CD40 expression in the CNS, and not in the periphery, exhibited a less severe EAE disease course with reduced numbers of infiltrating macrophages and lymphocytes in the CNS. Although these data suggested that CD40 expression on CNS resident cells was important for EAE development, neither the phenotype of the CNS resident cells nor their activation kinetics was examined.

In this study, we examined the role of CD40 in the activation process of microglial cells in the CNS during EAE using BM chimeras that allowed us to distinguish between CNS resident microglial cells and infiltrating macrophages and to render only the microglial cells CD40-deficient. We found that the activation of microglial cells was a multistep process with both CD40-independent and -dependent stages and that CD40 expression by microglial cells facilitated EAE disease progression. Microglial cell activation at the onset of EAE, characterized by an increase in both CD45 and MHC class II expression, was a CD40-independent process. In contrast, during peak of disease, full activation of microglial cells was dependent upon their expression of CD40. These data provide evidence that microglial cells must become activated in the CNS to facilitate the progression of EAE clinical disease. Thus, therapeutics that specifically target and prevent microglial cell activation could be efficacious in controlling MS disease progression.

B10.PL (H-2^u) and C57BL/6 mice were purchased from The Jackson Laboratory or bred locally, and CD40-deficient mice were provided by Dr. R. Geha (Boston Children’s Hospital, Boston, MA) (24). B10.PL-CD40^−/− (CD40^−/−) mice were produced in our breeding colony by backcrossing onto B10.PL for four generations, then intercrossing to generate homozygous knockout mice. The MBP-TCR transgenic mice expressing a TCR transgene specific for the acetylated NH₂-terminal peptide of MBP (Ac_1–11) have been described (25). GFP-B10.PL mice were generated as previously described (4). Animals were housed at the Biomedical Research Center of the Medical College of Wisconsin. All animal protocols were approved by the Medical College of Wisconsin Institutional Animal Care and Use Committee.

The MBP Ac_1–11 peptide (Ac-ASQKRPSQRSK) was generated by the peptide core of the Blood Research Institute (BloodCenter of Wisconsin, Milwaukee, WI). Anti-mouse CD45-PE was purchased from eBioscience. Anti-mouse H-2K^b-FITC, anti-TCRβ-PE, anti-CD11b-biotin, and Vβ8.2-biotin were purchased from BD Pharmingen. Anti-mouse CD11b-PE-Cy5, anti-H-2D^b-biotin, and streptavidin (SA)-allophycocyanin-Cy7 were purchased from Biolegend. SA-Alexa 350 was purchased from Molecular Probes. The 2.4G2 hybridoma was purchased from American Tissue Culture Collection.

BM chimeras were generated as described previously (4). Briefly, 7 × 10⁶ total BM cells from WT (B10.PL/C57BL/6)F₁ (H-2^uxb) or GFP-B10.PL mice were transferred into lethally irradiated (950 rad) WT (WT→WT or GFP→WT) or CD40^−/− (WT→CD40) B10.PL mice and were allowed to reconstitute for 8–9 wk. BM-derived cells of donor origin were differentiated from recipient cells by the expression of MHC class I molecule H-2K^b or H-2D^b. EAE was induced in chimera mice 8–9 wk post-BM transplantation.

EAE was induced by the adoptive transfer of MBP-specific encephalitogenic T cells generated as previously described (25). Briefly, 1 × 10⁶ activated MBP-TCR T cells were i.v. injected into sublethally irradiated (360 rad) 14- to 15-wk-old WT→WT or WT→CD40 chimeric mice. Individual animals were assessed daily for symptoms of EAE and scored using a scale from 1 to 5 as follows: 0, no disease; 1, limp tail and/or hind limb ataxia; 2, hind limb paresis; 3, hind limb paralysis; 4, hind and fore limb paralysis; and 5, death.

Mononuclear cells were isolated from the CNS of WT→WT or WT→CD40^−/− chimera mice with EAE on days 7 and 10 from four to five mice perfused with 25–30 ml of cold PBS as described (4). The brains and spinal cords were homogenized, and mononuclear cells were isolated using 40/70% discontinuous Percoll gradients (Sigma-Aldrich). Total cell numbers were determined by counting on a hemocytometer, and viability was assessed by trypan blue exclusion. Four-color flow cytometry using anti-H-2K^b-FITC, anti-CD45-PE, anti-CD11b-PE-Cy5 and anti-MHC class II-biotin, or anti-CD40-biotin or anti-CD86-biotin or anti-Vβ8.2-biotin combined with SA-allophycocyanin-Cy7 was conducted using total mononuclear cell preparations. For Ab stainings, FcR were first blocked with anti-mouse FcR (2.4G2). Ab incubations were conducted on ice for 20 min, and the cells were fixed in 1% paraformaldehyde and analyzed using the LSR II (BD Biosciences).

For labeling of proliferating cells in vivo, groups of four or five WT→WT or WT→CD40^−/− mice on days 7 and 10 after EAE induction by adoptive transfer were injected i.p. with 1 mg of BrdU (Sigma-Aldrich) 14 h before isolation of brain mononuclear cells (4, 26). Freshly isolated mononuclear cells from each experimental group were pooled and stained with anti-CD45-PE, anti-CD11b-PE-Cy5, and anti-H-2D^b-biotin combined with SA-allophycocyanin-Cy7. Subsequently, BrdU incorporation into the cellular DNA was detected using the BrdU flow kit (BD Biosciences) according to the manufacturer’s instructions. Briefly, cells were fixed, permeabilized, treated with DNase, and incubated with anti-BrdU-FITC. The samples were kept on ice and immediately analyzed by four-color flow cytometry. Populations of resting microglial cells (CD11b⁺CD45^lowH-2D^b−), activated microglial cells (CD11b⁺CD45^highH-2D^b−), peripheral macrophages (CD11b⁺CD45^highH-2D^b+), and encephalitogenic T cells (CD11b⁻CD45^highH-2D^b−) were gated and analyzed for BrdU incorporation.

Immunofluorescent staining of coronal sections of lumbar spinal cord was preformed as described previously (4). Briefly, on day 10 after EAE induction, frozen sections 10-μm thick from spinal cords of GFP→WT chimera mice were prepared and stained with anti-TCRβ-PE and anti-CD11b-biotin combined with SA-Alexa 350.

We first investigated whether B10.PL mice with a CD40 deficiency only in the CNS exhibit a less severe EAE disease course as was reported for C57BL/6 mice (17). To create chimera mice with a CNS CD40 deficiency, we transplanted WT BM from (C57BL/6/B10.PL)F₁ mice into lethally irradiated WT (WT→WT) or CD40^−/− (WT→CD40^−/−) recipient B10.PL mice. In the WT→CD40^−/− chimeras, CNS resident cells, including microglial cells, are CD40-deficient, while the peripheral macrophages are WT. When EAE was induced by the adoptive transfer of MBP-specific encephalitogenic T cells, we found that disease symptoms in WT→CD40 chimera were reduced when compared with the WT control chimeras (Fig. 1). On day 7 after EAE induction, WT→WT chimera mice had an average disease score 1.3 ± 0.2, while WT→CD40^−/− mice exhibited no or very mild disease symptoms with an average score of 0.2 ± 0.1. Similarly, during the peak of EAE, WT→CD40^−/− chimera mice had an average disease score of 1.6 ± 0.2, while WT→WT chimeras exhibited a more severe disease with an average score of 2.5 ± 0.3 (Fig. 1 and Table I). At both time points, clinical disease was statistically significantly reduced in the WT→CD40^−/− chimera mice (p < 0.05). These data show that CD40 expression in the CNS of B10.PL mice is important for the onset and progression of EAE.

Because the most prevalent CNS resident CD40-expressing cells are the microglial cells, we next determined whether CD40 expression by microglial cells is required for their activation in vivo during EAE. To examine this, we isolated CNS mononuclear cells at EAE onset (day 7) and peak of disease (day10) from CD40→WT and WT→WT chimeras and analyzed populations of microglial cells for the expression of the activation markers CD45, MHC class II, CD40, and CD86. In mice without EAE, ∼95% of all mononuclear cells from WT→WT or WT→CD40 chimera mice were CD11b⁺H-2K^b− and exhibited a resting microglial cell phenotype characterized by a CD45^low expression pattern with little to no expression of activation markers (Ref.4 and data not shown). Because we previously confirmed that CD45 up-regulation is a marker of microglial cell activation in our EAE model (4), we first investigated the level of CD45 expression in WT and CD40-deficient microglial cells during EAE. On day 7, the day of EAE onset, 16% of WT CD11b⁺H-2K^b− microglial cells had up-regulated CD45 (Fig. 2,A, upper left quadrant), while only 8% of CD40-deficient microglial cells had a similar phenotype (Fig. 2,B, upper left quadrant), suggesting that the process of microglial cell activation was less efficient in WT→CD40 chimera mice. In addition, a statistically significant (p < 0.01) 3-fold decrease in the percentage (Fig. 2,K, onset) and 5-fold reduction in the absolute number of CD45^high-activated microglial cells in the CNS of CD40→WT chimeras was observed when compared with WT→WT animals during EAE onset (Table I). At the peak of disease, neither the percentage (Fig. 2, C, upper left quadrant, and K) nor the absolute number (Table I) of WT-activated microglial cells changed from onset. In contrast, in the CD40→WT chimeras, additional recruitment of activated microglial cells was evident at the peak of disease. This is shown by an increase in the percentage of activated microglial cells from 8% at onset (Fig. 2,B, upper left quadrant) to 13% at peak of disease (Fig. 2,D, upper left quadrant), and by a 1.8-fold increase in the absolute number of these cells from onset to peak of disease (Table I). Although the activated microglial cells in the WT→WT chimeras did not increase in number after onset, they exhibited signs of further activation as shown by an increase in the mean fluorescence intensity (MFI) of CD45 from 341 on the day of onset (Fig. 2,A) to 593 at peak of disease (Fig. 2,C). This increase did not occur in the CD40-deficient microglial cells, which had an MFI of 351 and 347 during onset and peak, respectfully (Fig. 2, B and D). During the peak of EAE, the average percentage (Fig. 2,K, peak) and absolute number (Table I) of activated microglial cells were 1.7- and 3-fold lower in CD40→WT mice compared with WT→WT animals, respectively. These differences were statistically significant (p < 0.05 and p < 0.01 for percentage and absolute number, respectfully). Thus, these cumulative data show that the first stage of microglial cell activation characterized by an initial increase in CD45 expression was attenuated but not inhibited in CD40-deficient microglial cells; whereas, the second stage of activation accompanied by a further increase in CD45 was completely blocked.

We next determined whether a deficiency in CD40 expression on microglial cells affected the recruitment of macrophages from the periphery into the CNS during EAE by analyzing the percentage and absolute number of peripheral macrophages in WT→CD40 vs WT→WT chimeras. Peripheral macrophages (CD11b⁺H-2K^b+CD45^high cells) were detected in the CNS of both groups of chimera mice during EAE onset and peak of disease. However, the percentage of these cells was 4-fold lower in the CNS of WT→CD40 chimeras at the day of EAE onset when compared with that of WT→WT mice and was constituted of 11 and 37% of all CD11b⁺ cells, respectfully (Fig. 2, A and B, upper right quadrant). During peak of EAE, the percentage of peripheral macrophages in WT→WT chimera was 43%; whereas, in the WT→CD40 chimeras, this population represented only 27% of all CD11b-positive cells (Fig. 2, C and D, upper right quadrant). Fig. 2 shows one representative experiment; the average peripheral macrophage populations from three experiments are: WT→WT, onset = 33 ± 5% and peak = 45 ± 5%; WT→CD40, onset = 11 ± 2% and peak = 22 ± 3%. Similarly, the average absolute number of peripheral macrophages was statistically significantly reduced by 7- and 3-fold (p < 0.01) for WT→CD40 chimeras in comparison to control WT→WT chimeras during EAE onset and peak of disease, respectfully (Table I). Thus, deficiency in CD40 expression in the CNS did not prevent the recruitment of macrophages into the CNS, but did result in a reduction in their percentage or absolute number.

Because lower numbers of activated microglial cells and peripheral macrophages as potential APCs could result in a decrease in the expansion of encephalitogenic T cells in the CNS, we also examined the contribution that encephalitogenic T cells made to the total lymphocyte pool in the CNS of chimera mice. We found that the percentage of encephalitogenic T cells (H-2K^b−Vβ8.2⁺ cells) of the total lymphocyte population (CD11b⁻CD45^high gated cells) was ∼2-fold lower in the CNS of WT→CD40 mice when compared with WT→WT mice during EAE onset (16 and 36% of all lymphocytes, respectfully; Fig. 2, E and F, upper left quadrant) and at the peak of disease (24 and 49% of all lymphocytes, respectfully; Fig. 2, G and H, upper left quadrant). In Fig. 2,L, the average of three experiments is shown with the percentage of encephalitogenic T cells being statistically significantly reduced in the WT→CD40 chimeras at disease onset (p < 0.05) and peak (p < 0.01) as compared with WT→WT chimeras. Similarly, during onset and peak of disease, the absolute number of encephalitogenic T cells in the CNS of WT→CD40 mice was ∼3-fold lower than that of WT→WT chimeras and these differences were statistically significant for both EAE onset and the peak of disease (p < 0.001; Table I). Taken together, these data demonstrate that CD40-deficient microglial cells exhibit an intermediate level of activation during EAE, but were unable to achieve full activation, which was correlated with lower numbers of infiltrating peripheral macrophages and encephalitogenic T cells in the CNS, suggesting that microglial cells play an important role in the progression of EAE clinical disease.

To further characterize the process of activation of WT vs CD40-deficient microglial cells during EAE onset, we investigated the expression of the activation markers MHC class II, CD40, and CD86 on gated populations of CD11b⁺H-2K^b−CD45^low resting microglial cells, CD11b⁺H-2K^b−CD45^high activated microglial cells, and CD11b⁺H-2K^b+CD45^high peripheral macrophages isolated from the CNS of CD40→WT and WT→WT mice. Our previous study demonstrated that both populations of CD45^high-activated and “resting” CD45^low microglial cells became activated during EAE, but they exhibited two distinct patterns of activation (4). CD45^high-activated microglial cells up-regulated MHC class II, CD86, and CD40 and were colocalized with peripheral macrophages in inflammatory lesions, while the CD45^low “resting” microglial cells up-regulated MHC class II without significant up-regulation of costimulatory molecules, a process we described as bystander activation (4). In our current study, we investigated whether CD40 expression on microglial cells was required for bystander activation of CD45^low microglial cells and full activation of CD45^high microglial cells. We found that during EAE onset, populations of both WT and CD40-deficient CD45^low-resting microglial cells underwent similar levels of bystander activation, characterized by up-regulation of MHC class II (Fig. 3, A and D) and to a lesser extent CD86 (Fig. 3, B and E). CD45^low-resting microglial cells from both WT→WT and WT→CD40 chimeras were negative for CD40 during disease onset (Fig. 3, C and F). Analysis of CD40-deficient CD45^high-activated microglial cells showed that these cells expressed MHC class II and CD86 (21–24 and 16–19% positive, respectfully; Fig. 3, J and K) similar to that of WT microglial cells (21–23 and 20–23% positive, respectfully; Fig. 3, G and H). As expected, CD40-deficient microglial cells were negative for CD40 expression (Fig. 3,L), whereas 18–19% of WT CD45^high-activated microglial cells expressed this molecule (Fig. 3,I). Peripheral macrophages from both WT→WT (Fig. 3, M–O) and WT→CD40 (Fig. 3, P–R) chimeras expressed MHC class II, CD40, and CD86 at similar levels. These data demonstrate that during EAE onset, CD40 expression by microglial cells is not required for the first step of microglial cell activation characterized by increased expression of activation markers.

In the next series of experiments, we continued our analysis of activation marker expression by different myeloid populations isolated from the CNS at a later time point during EAE. At the peak of disease, bystander activation of CD45^low microglial cells reached maximum in both groups of chimeras, with similar levels of MHC class II (Fig. 4, A and D) and CD86 (Fig. 4, B and E) expression. Similar to EAE onset, CD45^low-resting microglial cells in both groups of chimera mice remained negative for CD40 (Fig. 4, C and F), further showing that bystander activation is a CD40-independent process. In contrast to bystander activation of “resting” CD45^low microglial cells, activation of CD45^high microglial cells during the peak of EAE was dependent upon CD40 expression. This was shown by a reduced level of MHC class II expression from 54–65% in WT→WT chimera (Fig. 4,G) to 15–25% in the WT→CD40^−/− chimera (Fig. 4,J). Similarly, 34–48% of CD45^high microglial cells from WT→WT chimeras expressed CD86 (Fig. 4,H) and 33–45% expressed CD40 (Fig. 4,I), while activated CD40-deficient microglial cells continued to express low levels of CD86 (4–6%; Fig. 4,K) in the absence of CD40 (Fig. 4,L). This low level of MHC class II and CD86 expression on CD40-deficient activated microglial cells during peak of EAE (Fig. 4, J and K) was similar to the level of expression of these molecules at disease onset (Fig. 3, J and K). At the same time, the level of activation marker expression on peripheral macrophages during the peak of EAE was similar in both groups of chimeras (Fig. 4, M–R). These data show that the second stage of microglial cell activation is a CD40-dependent process, and that peripheral macrophage activation is not affected by an absence of CD40 expression in the CNS.

One consequence of microglial cell activation is induction of proliferation. We previously showed that microglial cells in the normal CNS of chimera mice do not show any detectable signs of proliferation, but that they do cycle during EAE and upon progression of disease symptoms the level of proliferation decreases (4). Here, we investigated the level of proliferation in populations of resting microglial cells, activated microglial cells, and peripheral macrophages in the CNS of chimeric mice during EAE by BrdU incorporation. Resting CD45^low microglial cells and peripheral macrophages from both types of chimeras underwent little proliferation (∼5%) during EAE onset and peak of disease (Ref.4 and data not shown). During onset of EAE, WT and CD40-deficient activated CD45^high microglial cells proliferated at the same rate (14–15%; Fig. 5, A and B). However, during the peak of disease only 5% of WT-activated microglial cells incorporated BrdU (Fig. 5,E); whereas, 12% of CD40-deficient microglial cells were still cycling in the CNS (Fig. 5,F). The average of three experiments is shown in Fig. 5,I, showing a statistically significantly (p < 0.05) higher level of proliferation in the CD40-deficient microglial cell population during the peak of disease (Fig. 5 I). Thus, the data shows that during the peak of EAE, CD40-deficient microglial cells continued to proliferate at a high rate, suggesting that signaling through CD40 was required to achieve a full level of microglial cell activation and maturation.

Because we found that the absolute numbers of encephalitogenic T cells was lower in the CNS of WT→CD40 chimera as compared with control chimeras (Table I), we examined whether it was due to decreased proliferation. Indeed, we found an ∼2-fold decrease in the percentage of BrdU-positive encephalitogenic T cells in the CNS of WT→CD40 chimera mice as compared with WT→WT chimera mice during both EAE onset (12 and 19% BrdU⁺ cells, respectfully; Fig. 5, C and D) and peak of disease (14 and 25% BrdU⁺ cells, respectfully; Fig. 5, G and H). These differences in the level of proliferation of encephalitogenic T cells in the CNS of WT→WT vs WT→CD40 cells were statistically significant for onset and peak of disease (p < 0.01; Fig. 5 J). Thus, CD40 expression on microglial cells also seems to contribute to the proliferation of encephalitogenic T cells in the CNS during EAE.

Our data support a model (see Fig. 7) whereby CD40 expression by microglial cells is required for their full activation. To determine whether encephalitogenic T cells could provide the CD40 signal, we examined the expression of CD40L on encephalitogenic T cells in the CNS during EAE using flow cytometry, and found that 27% of the cells were positive (Fig. 6,A). We further analyzed whether cell-cell contacts between encephalitogenic T cells and microglial cells could be detected in the CNS during EAE. To distinguish between CNS microglial cells and immune cells from the periphery, we generated chimera mice by transplanting WT mice with BM from GFP-B10.PL mice (GFP→WT) and performing immunofluorescence examining the location of encephalitogenic T cells and activated microglial cells in CNS inflammatory lesions in mice with EAE. In these chimera mice, infiltrating peripheral macrophages and T cells express GFP (green), while the transferred encephalitogenic T cells are GFP⁻ and TCRβ⁺ (red) (Fig. 6,B). We found that encephalitogenic T cells constituted the majority of T cells in the lesions (Fig. 6, B and C), which was confirmed by flow cytometry analysis with 79–85% of all TCRβ⁺ not expressing GFP (data not shown). Peripheral macrophages and microglial cells were identified by their expression of CD11b (blue) and were found in proximity to encephalitogenic T cells (Fig. 2,C). In Fig. 6,D, an overlay of all three colors is shown with the GFP⁻TCRβ⁺ (red) encephalitogenic T cells juxtaposed to the GFP⁻CD11b⁺ microglial cells (blue) as indicated by arrows. Encephalitogenic T cells were also seen in contact with peripheral macrophages (Fig. 6 D). These data demonstrate that encephalitogenic T cells express CD40L in the CNS during EAE and colocalize with microglial cells in inflammatory lesions, demonstrating that encephalitogenic T cells are the likely source of the required microglial cell CD40 signal.

This study provides several new insights into the role of CD40 on microglial cell activation in vivo during EAE. First, we show that microglial cell activation is a multistep process characterized by an early CD40-independent stage characterized by an intermediately activated phenotype and a second CD40-dependent stage evident at the peak of EAE that is required for the full activation and maturation of microglial cells in the CNS. In addition, we show that insufficient activation of microglial cells by a lack CD40 resulted in decreased proliferation of encephalitogenic T cells in the CNS and amelioration of disease clinical symptoms. Thus, our study demonstrates that activation of microglial cells in vivo is a complex multistep process which likely involves interactions in the CNS between microglial cells and encephalitogenic T cells during EAE.

Despite broad usage in the literature, the meaning of the term “activated microglia” is still not clear. There is a discrepancy concerning what is considered “activation”. Many studies, especially those that examined microglial cells in vitro, define activated microglial cells as cells that up-regulated MHC class II and costimulatory molecules (9). Histology based studies, investigating the role of microglial cells in various CNS pathologies, defined activated microglial cells as MHC class II-positive cells or as cells that up-regulated macrophage markers such as CD11b (27, 28), while others defined microglial cell activation as a sequence of morphological changes from ramified to amoeboid forms (29), or as cells that proliferate in the CNS (30). Only a limited number of in vitro studies have suggested that microglial cell activation could be a multistep process (9), although little is known about the process of microglial cell activation in vivo during CNS pathologies. The difficulties in defining the activated state of microglial cells in vivo are mostly due to lack of current knowledge about the particular molecular mechanisms that mediate microglial cell activation. It was demonstrated that CD40 is expressed in lesions of MS patients and EAE, but the exact populations of macrophages and microglial cells that express this molecules were not characterized (11).

Based on this current study and our previous work (4), we were able to demonstrate that there are at least two distinct paths for microglial cell activation during EAE: 1) bystander activation and 2) specific activation. We defined bystander activation of microglial cells as a global up-regulation of MHC class II on the CD45^low resting population of microglial cells during EAE (4). In the current study, we demonstrated that bystander activation did not require CD40, which supports our previously stated hypothesis that bystander activation is likely due to IFN-γ production in the CNS by encephalitogenic T cells (4, 31). Because CD45^low microglial cells up-regulated MHC class II without significant up-regulation of costimulatory molecules, these cells could play a role in down-modulation of CNS inflammation by presenting Ag in the absence of costimulation to encephalitogenic T cells, which would likely be tolerogenic. In support of this possibility is a study which showed that microglial cells isolated ex vivo activated by allospecific T cells in a rat graft-vs-host disease model exhibited a CD45^lowMHC class II⁺CD80/CD86⁻ phenotype and induced apoptosis of encephalitogenic T cells (32). In contrast to bystander activation, specific activation of microglial cells resulted in up-regulation of MHC class II, but also an increase in CD45, CD86, and CD40 (Fig. 3, G–I and Fig. 4, G–I). We found that these CD45^high-activated microglial cells exhibited a morphology similar to the peripheral macrophages and were located in inflammatory lesions (4), which would facilitate their interaction with infiltrating leukocytes by both direct cell-cell contact and by soluble factors.

In summary, our cumulative work provides evidence for a multistep model of specific microglial cell activation during CNS autoimmune inflammation (Fig. 7) (4). Our model, shown in Fig. 7, is based on the hypothesis that activation of microglial cells results from interactions with CD40L⁺ encephalitogenic T cells that are present in the CNS prior to the onset of EAE (4). We propose that the first step of microglial cell-specific activation that occurs just before EAE onset is CD40-independent and is mediated by proinflammatory cytokines secreted by encephalitogenic T cells, such as IFN-γ and GM-CSF. We have previously shown that encephalitogenic T cells are the major producers of IFN-γ in the CNS during EAE (31), and in our current unpublished studies, we have shown that GM-CSF production by encephalitogenic T cells is required for microglial cell activation and EAE onset (E. D. Ponomarev, L. P. Shriver, and B. N. Dittel, submitted for publication). After completion of the first step of activation, microglial cells then exhibit an intermediately activated phenotype with moderate levels of MHC class II, CD40 and CD86 expression combined with a high level of proliferation (Fig. 7). During the CD40-dependent second step of activation, we propose that encephalitogenic T cells, which express CD40L (Fig. 6,A), interact with the intermediately activated CD45^high microglial cells, which express CD40, resulting in full activation characterized by further up-regulation of activation molecules and a decrease in proliferation, suggesting that these cells become fully activated and mature (Fig. 4, G–I; Fig. 7). Finally, we propose that the fully activated microglial cells phagocytize and efficiently process and present myelin self-Ag to encephalitogenic T cells in combination with CD28 costimulation leading to their expansion in the CNS and exacerbation of clinical symptoms (Fig. 7).

Our model is supported by several in vitro studies, including our own, indicating that coculture of encephalitogenic T cells with microglial cells in the absence of Ag resulted in an intermediate level of MHC class II expression and costimulatory molecule expression on microglial cells (intermediately activated microglial cells), while in the presence of Ag, the level of expression of these molecules reached a higher level (fully activated microglial cells) (10, 33). Thus, presentation of self-Ag by preactivated CD45^high microglial cells to encephalitogenic T cells through MHC class II may facilitate the engagement of CD40-CD40L interactions and further perpetuate the activation of microglial cells. In addition to in vitro studies, which support our model, there is also in vivo data showing that CD40-CD40L interactions between microglial cells and encephalitogenic T cells likely occur in the CNS during autoimmune inflammation. It was reported that CD40L was detected on CD4⁺ T cells in the lesions of MS patients and CD40L⁺ cells were colocalized with CD11b⁺CD40⁺ cells which include both microglial cells and peripheral macrophages (11).

Our current study provides support for bidirectional interactions between microglial cells and encephalitogenic T cells during autoimmune neuroinflammation. Encephalitogenic T cells likely contribute to the activation of microglial cells through cytokine secretion and CD40-CD40L interactions, while activated microglial cells may stimulate infiltration of peripheral macrophages and lymphocytes from the periphery by producing chemokines (6), and promoting encephalitogenic T cell proliferation in the CNS by presenting self-Ag (7, 34). In addition to secretion of chemokines and Ag presentation, activated microglial cells may be the main source of proinflammatory cytokines such as IL-23, which is thought to be important for expansion of encephalitogenic T cells in the CNS and development of EAE symptoms (35). Recent studies demonstrated that activation of microglial cells is an important event for development of EAE symptoms (5), proliferation of encephalitogenic T cells in the CNS (35), and epitope spreading in a model of relapsing/remitting disease (34). Thus, it is becoming widely accepted that activated microglial cells play an important role in disease initiation and progression in both EAE and MS.

Apart from activated microglial cells, other CNS resident cells such as astrocytes and neurons may also contribute to the regulation of autoimmune inflammation in the CNS through the expression of CD40. We find it unlikely that the lack of CD40 expression on these CNS resident cell types affected EAE in the WT→CD40 chimera mice. Although it was reported that neurons in normal spinal cord express CD40 (36), this molecule is mostly involved in neuronal development as well as protection of neurons from apoptosis rather than regulation of CNS inflammation (37). MHC class II and costimulatory molecules can be induced in astrocytes in vitro; however, their level of expression of CD40 and CD86 and Ag presentation is very limited (38). In contrast, microglial cells can easily be induced to express CD40 in vitro (9, 33, 38) and CD40-CD40L interactions transform these cells into efficient APCs (9). Therefore, we suggest that reduced symptoms of EAE in mice with a CD40 deficiency in the CNS are due to a blockade of the second step of microglial cell activation preventing their full activation and maturation into “macrophage-like” cells. Among CNS-infiltrating cells, peripheral macrophages represent the most important population that contribute to development of CNS autoimmune inflammation through CD40-CD40L interactions. Peripheral macrophages express high levels of CD40 in the CNS during EAE (Figs. 3 and 4) and chimera mice with a CD40 deficiency in the periphery develop mild symptoms of EAE (17). Although infiltrating macrophages express a higher level of activation markers as compared with activated microglial cells (Figs. 3 and 4) and outnumber them in the CNS during EAE (Table I), our data suggest that activation of CNS resident microglial cells is an important early event required for disease initiation in EAE and MS during the preclinical stage of disease, before infiltration of peripheral macrophages into the CNS. Thus, specific targeting of microglial cell activation at different stages of disease may be an effective therapeutic approach to treat MS patients.

We thank Shelley Morris and Monica Mann for assistance with the animal colony and Katarzyna Maresz for critical reading of the manuscript.

The authors have no financial conflict of interest.