B7-H1 restricts neuroantigen-specific T cell responses and confines inflammatory CNS damage: Implications for the lesion pathogenesis of multiple sclerosis

Authors


Abstract

The co-inhibitory B7-homologue 1 (B7-H1/PD-L1) influences adaptive immune responses and has been proposed to contribute to the mechanisms maintaining peripheral tolerance and limiting inflammatory damage in parenchymal organs. To understand the B7-H1/PD1 pathway in CNS inflammation, we analyzed adaptive immune responses in myelin oligodendrocyte glycoprotein (MOG)35–55-induced EAE and assessed the expression of B7-H1 in human CNS tissue. B7-H1–/– mice exhibited an accelerated disease onset and significantly exacerbated EAE severity, although absence of B7-H1 had no influence on MOG antibody production. Peripheral MOG-specific IFN-γ/IL-17 T cell responses occurred earlier and enhanced in B7-H1–/– mice, but ceased more rapidly. In the CNS, however, significantly higher numbers of activated neuroantigen-specific T cells persisted during all stages of EAE. Experiments showing a direct inhibitory role of APC-derived B7-H1 on the activation of MOG-specific effector cells support the assumption that parenchymal B7-H1 is pivotal for delineating T cell fate in the target organ. Compatible with this concept, our data investigating human brain tissue specimens show a strong up-regulation of B7-H1 in lesions of multiple sclerosis. Our findings demonstrate the critical importance of B7-H1 as an immune-inhibitory molecule capable of down-regulating T cell responses thus contributing to the confinement of immunopathological damage.

Abbreviations:
B7-H1:

B7-homologue 1

MOG:

myelin oligodendrocyte glycoprotein

OM:

original magnification

PD-1:

programmed death receptor-1

Introduction

MS is an inflammatory demyelinating disorder of the CNS of putative autoimmune origin. Murine EAE is used as a T cell-dependent animal model for MS characterized by neurological impairment and progressive paralysis resulting from inflammatory demyelination in the CNS 1. Active immunization with myelin protein antigens (e.g., myelin oligodendrocyte glycoprotein, MOG) induces encephalitogenic CD4+ Th1 and pathogenic Th17 cells 27. While mechanisms of peripheral T cell activation and expansion have been elucidated in detail during last few years, parenchymal control of CNS inflammation is incompletely understood.

T cells express co-stimulatory and co-inhibitory accessory receptors, and the delivery of additional co-stimulatory signals by APC is a crucial process for their complete activation and generation of effector function 8. Modulation of T cell activity by attenuation of co-stimulatory or augmentation of co-inhibitory signaling molecules therefore represents an attractive approach for the treatment of autoimmune and inflammatory diseases 911. B7-homologue 1 (B7-H1), a new member of the B7-CD28 family, can be expressed by a variety of cells, including T and B cells, monocytes, dendritic cells, as well as non-hematopoietic cells like keratinocytes, muscle cells, pancreatic islets, endothelial and tumor cells 1214. B7-H1 regulates immune responses by interaction with programmed death receptor-1 (PD-1), an inhibitory co-stimulatory receptor induced on activated T, B, and myeloid cells and possibly another yet unidentified receptor on activated T cells 10, 15. Previous studies using B7-H1-Fc fusion protein demonstrated a co-stimulatory role for B7-H1 16. However, more recent data suggest that the B7-H1/PD-1 pathway has considerable negative regulatory properties by inhibiting T cell proliferation, cytokine production and CTL activity 13, 1719. PD-1-deficient mice spontaneously develop various autoimmune diseases depending on their genetic background 20, 21.

Recent data from human and murine studies suggest, that the B7-H1/PD-1 pathway seems to be critically involved in limiting parenchymal inflammation 2224. Using an animal model of diabetes, Martin-Orozco et al.25 elegantly demonstrated that parenchymal B7-H1 contributes to the limitation of insulitis and the resolution of inflammation.

Approaches interfering with the B7-H1/PD-1 pathway have been shown to alter the susceptibility and progression of EAE in different mouse strains 26, 27. The absence of B7-H1 renders resistant mice susceptible to EAE and increases encephalitogenic T cell responses 28. In the murine CNS, B7-H1 expression is very limited under normal conditions. During inflammation, B7-H1 is significantly up-regulated by various CNS cells, including microglial cells 22, 2831. Microglial B7-H1 is inhibitory for encephalitogenic T cells thus fueling the hypothesis that inducible expression of inhibitory B7-H1 may counterbalance parenchymal inflammation 22.

This report characterizes the role of B7-H1 in modulating adaptive immune responses during EAE and in human neuroinflammation. Our data show that B7-H1 serves as an important negative regulator of primary T cell activation and contributes to the maintenance of antigen-specific T cell survival/responsiveness in the periphery. Absence of B7-H1 allows persistence and sustained functionality of neuroantigen-specific T cells in the CNS, while B7-H1-mediated attenuation of reactivation of neuroantigen-specific T cells by APC is missing. The concept that B7-H1 is critically involved in determining T cell fate in the CNS is corroborated by the demonstration of strong B7-H1 up-regulation in human brain MS lesions.

Results

Absence of B7-H1 results in accelerated and more severe EAE

To evaluate the role of B7-H1 in EAE, we immunized age and sex-matched C57BL/6 WT and B7-H1–/– mice with MOG35–55 peptide (Fig. 1 and Table 1). In the absence of B7-H1, mice had a more rapid onset of EAE followed by exacerbated progression of the disease. Interestingly, accelerated onset and peak of the disease were associated with a statistically significant increase in disease score in B7-H1–/– mice compared with WT animals (mean maximal score 2.3 ± 0.5 versus 3.3 ± 0.4, p=0.01). Unlike WT controls, EAE-affected B7-H1–/– mice showed no recovery of their disease (observation up to 30 days post immunization). Due to earlier EAE onset in B7-H1–/– mice, disease incidence at day 13 after immunization was only 15% in the WT group, whereas 88% of B7-H1–/– mice already showed clinical signs of EAE. Overall, disease incidence in both groups was 100%. Additionally, pooled data from four independent experiments revealed a higher mortality rate for B7-H1–/– mice: 4 out of 32 mice in the B7-H1–/– group died, while none out of 31 mice in the control group died.

Figure 1.

B7-H1 deficiency results in accelerated and enhanced EAE. Daily mean clinical score for C57BL/6 WT and B7-H1–/– mice are plotted. Mice were immunized with MOG35–55 and graded for disease daily. The mean daily grade for each group is shown. These data are representative of four independent experiments with identical group sizes (n=5–8) and similar results.

Table 1. EAE induction by immunization with MOG35–55 in B7-H1-deficient mice
GroupIncidenceMean day of onset a)Mean maximal score b)Median c)
  1. a) Mean day of onset was calculated for the mice which developed EAE. Data represent mean ± SE.

  2. b) Mean maximal score was calculated for all mice in the group. Data represent mean ± SE.

  3. c) Median of the maximum score for all the mice in the group.

  4. d)p values are compared with WT (Student′s t-test).

WT7/7 (100%)14.6 ± 1.02.3 ± 0.52.5
B7-H1–/–8/8 (100%)11.9 ± 1.93.3 ± 0.43
  p = 0.008 d)p = 0.01 d) 

B7-H1–/– mice generate an anti-MOG35–55 peptide antibody response after immunization similar to wild-type counterparts

Published data suggest a critical role of anti-MOG antibodies in augmentation of demyelination and worsening of EAE 32, 33. B cells constitutively express B7-H1 and PD-1 upon activation, suggesting an involvement of this pathway in the induction of an antibody response to MOG 14. Therefore, we investigated whether anti-MOG antibody production was altered in B7-H1–/– mice following EAE immunization. Measuring the relative antibody titers from immunized mice sera by ELISA we observed similar levels of antibody responses for WT and B7-H1–/– mice on days 10, 17 and 24 after immunization (data not shown). This indicates that antibody responses against MOG are not affected by the absence of B7-H1.

B7-H1 influences kinetics and maintenance of primary neuroantigen-specific T cell responses in the periphery

Next we analyzed peripheral T cell responses during EAE. Splenocytes from immunized mice were obtained and analyzed at day 7, 10, 13 and 17 after EAE induction. Splenocytes were challenged with 10 µg/mL MOG35–55 peptide to circumvent the inhibitory effect of APC-derived B7-H1 on T cell activation, which was shown to be most prominent at lower antigen concentrations 28. Ex vivo responses of MOG35–55 peptide-restimulated T cells were assessed by ELISA and ELISPOT to address the amount and frequency of MOG-specific cytokine-producing cells, respectively. In the early phases of EAE (preclinical, up to day 10 post immunization), cells from B7-H1–/– mice produced higher amounts of IFN-γ in response to MOG restimulation compared to WT controls (Fig. 2A and B). ELISPOT analysis revealed that numbers of MOG-specific IFN-γ-producing cells were significantly higher in the preclinical phase (Fig. 2C), indicating an enhanced priming of antigen-specific T effector cells in the periphery of B7-H1–/– mice in the early phase of EAE. In contrast, at later time points (clinical phase, days 13 and 17 after EAE immunization) peripheral MOG35–55-induced T cell response was diminished in B7-H1–/– mice, while it peaked in WT mice. This was demonstrated by increased IFN-γ production and higher frequencies of MOG-specific IFN-γ-producing cells (Fig. 2A–C). Due to accumulating evidence of the role of IL-17 during autoimmunity and EAE 5, 34, we also assessed the kinetics of IL-17 production. We observed very similar results as for IFN-γ: in the absence of B7-H1, MOG-specific IL-17-producing T cells expressed moderately elevated levels of IL-17 in the preclinical phase (up to day 10 after EAE immunization) with a subsequent decline (13 and 17 days post immunization) (Fig. 2D). However, quantitative differences in cytokine production between B7-H1–/– and WT mice were not statistically significant here. IL-4 production was also assessed but found undetectable in either group of mice at any investigated time points (data not shown).

Figure 2.

Kinetics of MOG-specific peripheral T cell responses in WT and B7-H1–/– mice during EAE. (A) ELISA of IFN-γ production by splenocytes from WT and B7-H1–/– mice challenged ex vivo with 10 µg/mL MOG35–55 peptide for 48 h. Data shown are the mean + SE of four wells from one representative experiment of each time point (n=3–5 mice per group). (B) Comparison of MOG-specific IFN-γ-production in WT and B7-H1–/– mice presented as relative change. WT mice were set as 100% and the kinetics of IFN-γ-production in B7-H1–/– mice relative to WT is depicted. The data are combined from three independent experiments. (C) ELISPOT analysis of MOG-reactive IFN-γ-producing cells was performed in cultures of splenocytes from WT or B7-H1–/– mice harvested at days 7, 10, 13 or 17 after EAE immunization. The y axis represents the number of positive cells per 2 × 105 cells plated after stimulation for 24 h with 10 µg/mL MOG35–55. (D) ELISA of IL-17 production by splenocytes from WT and B7-H1–/– mice stimulated ex vivo with 10 µg/mL MOG35–55 peptide for 48 h. Data shown are the relative change of IL-17 production in B7-H1–/– mice compared to WT mice (100%) and are combined from two experiments for each time point. *p<0.05 compared to WT mice.

In summary, the proinflammatory Th1/Th-17-prone peripheral immune response induced in EAE is accelerated and enhanced in the absence of B7-H1, correlating with the earlier disease onset in B7-H1–/– mice. However, peripheral Th1/Th17 responses are more sustained in WT animals, suggesting an involvement of B7-H1 in maintaining the functionality of antigen-specific T cells in the periphery.

In B7-H1–/– mice numbers of neuroantigen-specific T cells are significantly increased in the CNS

Following systemic immunization with MOG, encephalitogenic T cells primed in the periphery migrate to the CNS where they reencounter their cognate antigen and develop inflammatory infiltrates (reviewed in 35). To address the amount of immune cell infiltration into the CNS at the peak of disease, we used flow cytometry to quantify the abundance of total CNS infiltrating cells in the inflamed brains and spinal cords. We found a higher percentage of both CD4+ and CD8+ T cell populations among all isolated lymphocytes in the absence of B7-H1 (Fig. 3A, left panel). Interestingly, the effect of B7-H1 deficiency was even more pronounced on CD8+ T cell population. Additionally, absolute numbers of both CD4+ and CD8+ T cells were significantly elevated in the CNS of B7-H1–/– mice in comparison to WT animals (Fig. 3A, right panel). However, no significant changes were observed for CD11b+ cells, mostly depicting both resident microglial cells (CD45+CD11blow) and infiltrated/activated macrophages (CD45+CD11bhigh) (data not shown). Thus, the amount of T cell infiltration in the target organ was found to be in a direct correlation with the severity of clinical EAE course in B7-H1–/– compared to WT mice.

Figure 3.

Analysis of immune cells in the CNS of WT and B7-H1–/– mice during EAE. WT and B7-H1–/– mice were immunized with MOG35–55 and CNS cells were analyzed at indicated time points. (A) Dot plots are gated for lymphocytes and inserted numbers represent the percentage of CD4+ and CD8+ CNS infiltrating T cells (left). Data are representative for three separate experiments (n=3). Absolute cell numbers of CNS infiltrating immune cells in brain and spinal cord homogenates at the peak of EAE (right) are shown (n=4). (B) Flow cytometry of spleen and CNS lymphocytes from EAE-affected mice. Inserted numbers in dot plots represent the percentage of naïve, memory and effector T cells. Data presented are representative of three separate experiments. (C) ELISPOT analysis of IFN-γ production by infiltrating CNS T cells after 10 µg/mL MOG35–55 peptide restimulation. Results are representative of two independent experiments (n=3). (D) ELISPOT analysis of IL-17 production by infiltrating CNS T cells after 10 µg/mL MOG35–55 peptide restimulation. Data shown are representative of two separate experiments (n=3–5). (E) Histological examination of apoptotic cells in spinal cord sections by in situ TUNEL staining. One representative three-color overlay of DAPI (blue), CD4 (green) and TUNEL (red) from a B7-H1–/– mouse is shown. White arrows indicate apoptotic CD4+ T cells. Original magnification (OM), 63×. (F) Quantification of apoptotic cells in the CNS. Percentage of TUNEL+ CD4+ cells (circles) in the investigated areas of spinal cord sections. *p<0.05 compared to WT mice.

Encephalitogenic T cell subsets were further analyzed by comparative flow cytometry of CNS tissue and splenic lymphocytes obtained from mice at peak of the disease. Approximately 88% of CD4+ cells in the CNS of both WT and B7-H1–/– mice were CD44+CD62L effector T lymphocytes, whereas this cell type was found in only 20% of splenic lymphocytes (Fig. 3B). In contrast, the major immune cell population (∼65%) in the spleen was naïve T cells (CD44CD62L+) in both groups (Fig. 3B). Only few naïve CD62L+ cells were detectable in the CNS, with no detectable differences between WT and B7-H1–/– mice. Approximately 7% of CD44+CD62L+ memory T cells were found in spleens and around 5% in the CNS (Fig. 3B). Interestingly, CD44+CD62L+ memory T cells were markedly increased in the CNS in the late recovery phase of EAE (day 29 after EAE immunization) accompanied by a decline of activated CD44+CD62L effector lymphocytes (from ∼88% to ∼79%) (Fig. 3B). In the periphery, the number of CD44+CD62L effector T cells continuously increased (from ∼20% to ∼44%), while naïve T cells diminished (from ∼65% to ∼30%).

To further characterize the altered CNS milieu in the absence of B7-H1, we compared numbers of neuroantigen-specific T cells during different phases of EAE between B7-H1–/– and WT mice. ELISPOT analysis for MOG-reactive IFN-γ-/IL-17-producing cells were performed from cells isolated from the CNS. Interestingly, numbers of MOG35–55-specific IFN-γ-producing cells were markedly increased in the CNS of B7-H1–/– mice compared to WT mice (Fig. 3C), both at peak of the disease (days 14–17 after EAE immunization) and during recovery phase (day 25–29 after EAE immunization). Similarly, an increase of cytokine-producing MOG35–55-specific T cells was also observed for IL-17, which was consistently secreted at higher levels in B7-H1–/– mice during both investigated time points of EAE, although the differences were not statistically significant in two performed experiments (Fig. 3D).

To determine whether an increased T cell accumulation in the CNS of B7-H1–/– mice was due to enhanced infiltration of T cells from the periphery or result of decreased cell apoptosis, we analyzed apoptotic cells in the CNS by in situ TUNEL staining. Co-stained for CD4, double-positive apoptotic cells were histologically analyzed in specimens of murine spinal cords obtained at the peak of disease. The cells were counted in different areas of inflammatory foci and expressed as percentage of TUNEL+ cells among infiltrating CD4+ T cells. Here, we observed no differences between WT and B7-H1–/– mice in terms of increased apoptosis or protection against apoptosis, respectively (Fig. 3E and F). Thus, apoptosis does not appear to be a relevant mechanism responsible for the increased T cell numbers in the CNS.

Taken together, these data indicate that T cells more abundantly infiltrate the CNS in the absence of B7-H1 and, here, are able to produce higher amounts of pro-inflammatory cytokines. In contrast to the situation in the periphery, the inflammatory response in the CNS is more sustained in the absence of B7-H1.

IFN-γ production by neuroantigen-specific effector cells is controlled by B7-H1 on APC

Encephalitogenic T cells need to be (re)activated by cognate antigen-specific APC. Recent data support the assumption that presentation of neuroantigens is most likely mediated by CNS-associated dendritic cells, such as perivascular dendritic cells 36. To analyze the functional consequences of B7-H1 deficiency on APC, we studied cytokine responses of MOG-reactive encephalitogenic cells in co-culture with splenocytes or bone marrow-derived dendritic cells from WT or B7-H1–/– mice. Splenocytes isolated from MOG35–55-immunized WT mice were used as responder cells and co-cultured with irradiated, MOG-pulsed APC. Absence of B7-H1 as well as neutralization of the B7-H1/PD-1 pathway by a specific anti-B7-H1 blocking antibody (clone 10B5) markedly increased the levels of IFN-γ secretion by MOG-reactive T cells (splenocytes as APC: Fig. 4A; dendritic cells as APC: Fig. 4B). Inhibitory function of APC-derived B7-H1 was most significant at low antigen concentrations. Whereas no differences were observed between B7-H1–/– and WT APC at higher peptide concentrations (10 μg/mL), IFN-γ secretion was significantly increased at 1 µg/mL and 0.2 µg/mL of MOG pulse (Fig. 4C). In these experiments IL-17 production was also assessed, but no significant differences were found when B7-H1 was either blocked or absent on APC.

Figure 4.

B7-H1 on APC inhibits activation of MOG-reactive effector T cell IFN-γ production in a concentration dependent manner. MOG35–55-specific T cells were co-cultured with irradiated peptide-pulsed splenocytes (1 µg/mL) (A, C) or dendritic cells (2 µg/mL) (B) generated from WT or B7-H1–/– mice. Supernatants were harvested after 24 h and IFN-γ concentration was analyzed by ELISA. Using stimulator cells from WT mice, neutralizing anti-B7-H1 (10B5) or isotype-matched control antibodies (10 µg/mL) were added to the co-culture. Bars show the mean + SE from one of three representative experiments, each performed in triplicate. (C) Antigen dose-dependent inhibitory effect of B7-H1 on splenocytes as APC. Irradiated splenocytes were pulsed with 10, 1 and 0.2 µg/mL MOG35–55 peptide and used as APC in a co-culture with MOG35–55-specific T cells harvested from EAE-immunized donor. * Significant differences (p<0.05) from WT APC.

These data demonstrate the direct inhibitory function of APC-derived B7-H1 for encephalitogenic T cell reactivation and IFN-γ production. Consistent with previous reports our data confirm that the inhibitory effect of the B7-H1/PD-1 pathway is most prominent at low antigen densities 28, 37.

B7-H1 expression in human CNS tissue specimens

Having demonstrated the critical role of B7-H1 in EAE and the murine CNS, we finally investigated the expression of B7-H1 in the human CNS under non-pathological and inflammatory conditions. In non-pathological CNS tissue specimens, B7-H1 immunoreactivity was only found on a few scattered cells. A small elongated nucleus and thin processes suggest a microglial origin of these cells (Fig. 5A and B). In normal-appearing white matter of MS patients, the density of B7-H1-expressing cells was considerably higher than in the CNS control tissues (Fig. 5C). Here, mainly activated ameboid-like cells exhibited B7-H1 immunoreactivity. In chronic active MS lesions, the cell population expressing B7-H1 was strongly increased and correlated with the morphology of activated microglial/macrophage-like cells (Fig. 5D and E). However, a sharp border to the cerebral cortex with almost no B7-H1 immunoreactivity was observed (Fig. 5D). In acute MS plaques, the majority of cells were positively stained for B7-H1, and displayed a cell morphology similar to chronic active MS lesions (Fig. 5F–H). These data provide first evidence that B7-H1 is up-regulated on human resident CNS cells under inflammatory conditions. B7-H1 immunoreactivity, mostly found on cells with most likely microglial/macrophage morphology, was found to be most intense in areas of strongest inflammation. This is compatible with the hypothesis that inducible B7-H1 expression would serve to attenuate T cell activation and contribute to immune homeostasis in the CNS.

Figure 5.

B7-H1 expression in human normal CNS and MS tissue specimens. B7-H1 expression in healthy control CNS tissue (A) OM 20× (B) OM 60×. B7-H1 expression in CNS tissue of MS patients in (C) normal-appearing white matter (OM 20×), (D) chronic active MS lesions (OM 2x), (E) within the chronic active MS plaque (OM 20×) and in (F–H) acute MS plaques (F, OM 2×; G, OM 20×; H, OM 60×).

Discussion

During inflammation, T cell activation is a complex process dependent on the balance of co-stimulatory and inhibitory signals 38, 39. Due to its inhibitory properties, B7-H1 was proposed as a factor contributing to the maintenance of peripheral tolerance and confining T cell-mediated tissue damage during autoimmunity 40, 41. Several groups including ours 28, 30, 31, 42 previously demonstrated a critical negative regulatory role for the PD-1/B7-H1 pathway in vitro and in vivo. Treatment studies of MS patients using IFN-β, which strongly up-regulates B7-H1, also suggest an important regulatory function of B7-H1 during CNS autoimmunity in humans 43. This report characterizes the role of B7-H1 in modulating adaptive immune responses during EAE and in human neuroinflammation. Our data show that B7-H1 serves as an important negative regulator of primary T cell activation and contributes to the maintenance of antigen-specific T cell responsiveness in the periphery. B7-H1–/– mice exhibited a more rapid and enhanced antigen-specific pro-inflammatory (Th1/Th17) T cell response in the periphery, but showed earlier turning-off here. In striking contrast, numbers of neuroantigen-specific T cells in the CNS were significantly elevated in B7-H1–/– mice and this expansion persisted through late stages of EAE. APC-derived B7-H1 inhibited cytokine production of encephalitogenic cells, therefore contributing to the direct control of CNS-T cell activation. Compatible with our findings in the mouse system, we found up-regulation of B7-H1 in human MS lesions. This was most prominent in areas of strongest inflammation and in close proximity to microglial/macrophage-like cells. Taken together our data support the hypothesis that B7-H1 is critical for the parenchymal control of neuroantigen-specific T cell expansion and activation in the CNS.

Hematopoietic APC as well as different parenchymal cell types have been identified as a source of B7-H1 expression. PD-1, its receptor, is predominantly found on activated T and B cells 38, 40, 41. Both PD-1 and B7-H1 are up-regulated during EAE and their expression has been shown to be correlated with the course of the disease 22, 29, 30. B7-H1/PD-1 interactions are critically involved in the modulation of experimental autoimmune encephalomyelitis. Anti-PD-1 treatment in the preclinical phase exacerbates disease in MOG-induced EAE in C57BL/6 mice, neutralization of B7-H1 increases susceptibility to EAE in BALB/c mice, genetic ablation of B7-H1 renders 129Sv mice susceptible to EAE and exacerbates disease course in C57BL/6 mice, respectively 28, 30, 31. Therefore, B7-H1 displays an important role in attenuating T cell responses in EAE, with some differences depending on the investigated mouse strain. Recently, Keir and coworkers elegantly showed that tissue expression of B7-H1 is crucial for mediating peripheral tolerance 44. Using an animal model for diabetes they demonstrated that parenchymal B7-H1 is a key player delineating inflammatory responses in the pancreatic tissue. The B7-H1/PD-1 pathway therefore is not only crucial for the initial phase of T cell activation but also influences subsequent effector functions and target organ infiltration.

The CNS had long been considered an immunologically privileged site. An anti-inflammatory and, with regard to invading immune cells, pro-apoptotic environment in the brain, the limited access of brain-derived antigens to the lymphoid organs, the presence of the blood-brain barrier, and low major histocompatibility complex (MHC) expression in the brain parenchyma, were classically used to explain the lack of an effective immune response to antigens in the brain. However, numerous studies in infectious, autoimmune and tumor models have challenged this view by showing that potent immune reactions can and do occur in the CNS 45. However, limiting the local inflammatory response is crucial for an organ as vulnerable as the CNS. Critical players in the immune homeostasis of the CNS environment are APC, e.g., microglial cells. Expression of the co-inhibitory B7-H1 signal was recently proposed by our group as a mechanism to counterbalance inflammation during CNS inflammation 22.

Development of EAE is critically influenced by priming and reactivation of auto-reactive T effector cells. Following systemic immunization with myelin antigens like MOG, naïve T cells interact with APC within secondary lymphoid organs, leading to activation and expansion of neuroantigen-specific T cells 46. Before invading the CNS parenchyma, MOG-specific T cells need to reencounter their cognate antigen in the periphery to be fully activated 35. Candidates recently proposed to mediate antigen presentation during EAE are dendritic cells 36. In the subsequent recruitment and effector phases, activated T cells migrate into the CNS parenchyma and cause tissue damage accompanied with demyelination and neurological deficits. We have followed the pathway of activated T cells during different phases of EAE and delineated the role of B7-H1 in these processes. We observed an earlier onset of EAE and a more severe EAE course accompanied by an exacerbated increase in the secretion of pro-inflammatory cytokines IFN-γ and IL-17 at earlier time points of EAE in the absence of B7-H1–/–. This observation is in accordance with previous reports documenting the inhibitory role of B7-H1/PD-1 interactions in the phase of primary immune response and the contribution of these cytokines to the EAE course 2, 7. In B7-H1–/– mice, increased IFN-γ production was accompanied by a higher frequency of MOG-specific IFN-γ-producing T cells, indicating an earlier priming of naïve T cells that resulted in different kinetics of Th1 and Th17 cell activation compared to WT mice. Interestingly, absence of B7-H1 led to a more rapid decrease of peripheral levels of neuroantigen-specific IFN-γ-production at later time points and lower frequency of MOG-specific IFN-γ-producing cells were found in B7-H1–/– mice. This observation suggests an important involvement of B7-H1 in maintaining the survival and/or functionality of antigen-specific T cells in the circulation.

B7-H1 plays an important role within the CNS. As previously shown, a variety of CNS cells including microglial cells, up-regulate B7-H1 under inflammatory conditions 22, 30. The more severe disease course in B7-H1–/– mice was accompanied by higher numbers of infiltrating immune cells (CD4+, CD8+) and significantly elevated numbers of neuroantigen-specific T cells in the CNS. The decrease in cytokine production of neuroantigen-specific T cells seen in the circulation was not observed in the CNS, which clearly emphasizes the role of B7-H1 modulating the fate of encephalitogenic effector T cells in the target organ. It is important to note that distribution of general immune subsets in the periphery or in the CNS did not differ significantly during various phases of EAE between B7-H1–/– and WT mice. However, B7-H1 had a major impact on frequency and activation of neuroantigen-specific T cells within the CNS. While absence of B7-H1 allowed persistence and sustained responsiveness to cognate antigen of MOG-specific cells in the target organ, peripheral levels significantly decreased. Our data suggest that this discrepancy cannot be explained by local protection against apoptosis in the CNS, but could be due to altered distribution of neuroantigen-specific T cells or possibly increased local T cell expansion in the CNS. Since the up-regulation of co-inhibitors in the CNS during inflammation might serve as a negative feedback loop to control inflammatory activity of T cells and to limit tissue damage, the lack of such molecules like B7-H1 could explain an extensive pathogenesis during autoimmune inflammation and the lack of recovery in B7-H1–/– mice.

This concept is further substantiated by experiments showing higher IFN-γ production by MOG-specific T cells in the absence of APC-derived B7-H1. Thus, disruption of the PD-1/B7-H1 pathway may therefore remove a critical signal for controlling encephalitogenic T cell fate determined by interaction with CNS-associated cells. This might result in increased pro-inflammatory cytokine production, amplification of CNS tissue damage and immune cell infiltration. It is interesting to note that the inhibitory effect of B7-H1 molecule was strongly dependent on the antigen concentration, indicating a predominant control of low avidity T cell clones or T cells triggered in the presence of low antigen density. Therefore, the absence of B7-H1 may facilitate immune responses against auto-antigens rather than against foreign antigens. Naturally, high-avidity self-specific T cells are deleted during thymic negative selection, but residual populations of auto-reactive T cells with low to intermediate avidity to self antigens may remain and can be released into the periphery 47, 48. To avoid cellular activation and initiation of an autoimmune response, co-inhibitory signaling pathways like the B7-H1/PD-1 may help to arrest these escaped negative selection cells under steady-state conditions. However, in an inflammatory environment or in situations with high antigen concentrations such as a viral infection, TCR triggering seems to overrule inhibitory signals provided by B7-H1/PD-1 signaling, and result in efficient T cell activation and expansion.

For the first time, we demonstrate the importance of this regulatory pathway in humans by comparing B7-H1 expression in CNS tissue of MS patients and healthy controls. A massive B7-H1 expression was correlated with a strong CNS inflammation in chronic and acute MS plaques, further substantiating the critical role of B7-H1 in down-regulating encephalitogenic T cell expansion in the CNS, thus contributing to the confinement of immunopathological damage.

In summary, our report characterizes the role of B7-H1 in modulating adaptive immune responses during EAE and in human neuroinflammation. Our data show that (i) B7-H1 serves as an important negative regulator of primary T cell activation against an auto-antigen, (ii) B7-H1-mediated maintenance of antigen-specific T cell survival/responsiveness in the periphery differs from encephalitogenic T cells in the CNS, (iii) APC-derived B7-H1 attenuates reactivation of neuroantigen-specific T cells, and (iv) B7-H1 is up-regulated in human brain MS lesions.

Understanding the mechanism of B7-H1 regulation in the context of CNS autoimmunity might lead to design of targeted therapies to modulate PD-1/B7-H1 interactions and, thus, to enforce the tolerogenic role of this co-inhibitory pathway in MS.

Materials and methods

Animals

Wild-type C57BL/6 mice were purchased from Harlan Winkelmann (Borchen, Germany). B7-H1–/– mice were generated by L. Chen (Baltimore, USA 12). Mice were kept under specific pathogen-free conditions in our animal facility according to German guidelines for animal care. All experiments were conducted according to animal experimental ethics committee guidelines and were approved by the local authorities (Regierung von Unterfranken; 54–2531.01–36/06).

Induction of EAE

MOG35–55 (EVGWYRSPFSRVVHLYRNGK; synthesized and HPLC purified by R. Volkmer, Charite, Berlin, Germany) was used for active induction of EAE. Age- and sex-matched C57BL/6 WT or B7-H1-deficient mice were immunized subcutaneously with 200 µg MOG35–55 emulsified in CFA (Sigma-Aldrich, Steinheim, Germany) that was further enriched with Mycobacterium tuberculosis (5 mg/mL) (Difco, Detroit, MI, USA). In addition, mice were injected i.p. with 400 ng pertussis toxin (List Biological Laboratories, Campbell, CA, USA) at the time of immunization and 48 h later. Using this standard immunization protocol, we observed a typical, peaking chronic disease course followed by a partial recovery from EAE in the WT mice. Animals were weighed and observed for clinical signs of disease daily and scored over a maximum period of 25–30 days based on the following scale (EAE score): 0, no disease; 1, limp tail; 2, hind limp weakness; 3, hind limp paralysis; 4, hind and fore limp paralysis; 5, moribund or death. The data are plotted as the mean daily clinical score for all animals per group.

Generation of murine dendritic cells

Dendritic cells from WT or B7-H1–/– mice were prepared as described previously 49. Briefly, bone marrow cells were flushed out of femur and tibia bones with PBS. Cells were incubated for 30 s with ACK buffer (0.15 M NH4Cl, 10 mM KHC03, 0.1 mM EDTA) and filtered through a 70-µm cell strainer. The single-cell suspension was cultured in RPMI 1640 (BioWhittaker, Verviers, Belgium) supplemented with 10% FBS, 2 mM L-glutamine (PAA Laboratories, Pasching, Germany), 50 µm 2-mercaptoethanol, antibiotics (100 U/mL penicillin/10 µg/mL streptomycin; Biochrom, Berlin, Germany) and 20 ng/mL mGM-CSF (Peprotech, Hamburg, Germany). On days 3 and 6 medium containing 20 ng/mL mGM-CSF or 10 ng/mL mGM-CSF was added. Bone marrow-derived dendritic cells were harvested on day 8 and characterized as more than 85% pure CD11c+ cells.

Isolation of spleen cells and preparation of CNS mononuclear cells

Spleens of MOG35–55/CFA-immunized mice were removed at times indicated in the Results. For each individual mouse, single-cell suspensions were generated by mashing the spleens through a 70-µm strainer and lysing red blood cells with ACK buffer. Splenocytes were cultured in DMEM (BioWhittaker) supplemented with 5% FBS (PAA Laboratories), 10 mM HEPES (Gibco, Invitrogen GmbH, Germany), 2 mM L-glutamine (PAA Laboratories), 50 µM 2-mercaptoethanol (Gibco), 1% nonessential amino acids (BioWhittaker) and 25 μg/mL gentamicin (Gibco).

Brains and spinal cords from PBS-perfused mice were isolated as published previously 22. In brief, CNS material was cut into pieces and CNS cells were isolated from the interface of 30–50% Percoll (Amersham Biosciences, Freiburg, Germany) centrifuged for 30 min at 2500 rpm. Mononuclear cells were washed and resuspended in culture medium for further analysis. Cell numbers isolated from the CNS were determined by cell counting and flow cytometry based on the percentage of specific cells from the total cell population acquired.

Flow cytometry

For flow cytometric analysis of lymphocytes isolated from spleen or CNS of mice, single-cell suspensions were prepared as described above. Flow cytometry was performed using standard methods. The following antibodies were used: anti-mouse CD4-PerCP (clone RM4–5); anti-mouse CD8-FITC, CD8-PE or CD8-APC (clone 53–6.7); anti-mouse MHCII I-A/I-E-PE (clone M5); anti-mouse B7-H1-PE (clone MIH 5, eBioscience, San Diego, CA, USA); anti-mouse CD11b-PerCP (clone M1/70), anti-mouse CD11c-APC (clone HL3); anti-mouse CD44-FITC (clone IM7); anti-mouse CD62L-APC (clone MEL-14), anti-mouse CD25-PE (clone 7D4). All antibodies were purchased from BD PharMingen (Heidelberg, Germany) unless indicated differently. Flow cytometry was performed on a FACSCalibur (Becton Dickinson) and analyzed with FlowJo software (Treestar).

Antigen restimulation assays and ELISA

Irradiated (35 Gy) splenocytes (2 × 106) or dendritic cells (5 × 105) from C57BL/6 WT or B7-H1–/– mice were pulsed with various amounts of MOG35–55 peptide and used as APC co-cultured with 2 × 106 syngeneic MOG-reactive T cells in 48-well plates. Neutralizing anti-B7-H1 antibody (10B5) or isotype control was added at 10 µg/mL where indicated. The supernatants were analyzed for IFN-γ production by ELISA according to the manufacturer's manual.

For restimulation, 3 × 106 splenocytes/well were cultured as triplicates in 24-well plates with 10 µg/mL MOG35–55 for 48 h. Control splenocytes were activated with anti-CD3/CD28 beads at a ratio of 1:1 (Dynal Biotech, Oslo, Norway) or left untreated. After 48 h, cytokines were determined in the supernatants by mouse IFN-γ, IL-4 or IL-17 ELISA according to the manufacturer's instructions (Duoset, RnD Systems, Wiesbaden, Germany).

ELISPOT

For ELISPOT assays, 2 × 105 splenocytes or 5 × 104 CNS cells per well were cultured with 10 µg/mL MOG35–55 peptide in 96-well-plates. ELISPOT assay was performed according to the manufacturer's instructions (for IFN-γ: BD Pharmingen; for IL-17: eBioscience). Spots were counted using a Wild Heerbrugg M3Z dissecting microscope or evaluated by CTL Europe GmbH (Aalen, Germany).

Anti-MOG antibody ELISA

Anti-MOG antibody ELISA was performed as described previously 50. Briefly, serum samples were obtained from blood of naïve or MOG35–55-immunized mice and stored at -20°C. Nunc-Immuno 96-well plates were coated with recombinant human MOG1–120 overnight at 4°C. Following washing (PBS/0.05% Tween-20) and blocking (10% FBS) diluted serum samples were applied and incubated for 1 h. After washing, HRP-conjugated goat anti-mouse IgG (Serotec, Düsseldorf, Germany) was added for 1 h followed by tetramethylbenzidine substrate (Sigma-Aldrich, Steinheim, Germany) incubation. Reaction was stopped by 2 N H2SO4 and the samples were analyzed by a multiplate reader (Labsystems Titertek Multiscan Plus) at 450 nm. Serum samples of immunized mice were compared with a naïve serum standard. Anti-MOG monoclonal antibody (818C5) was used as a positive control according to 51.

In situ tissue TUNEL staining

For immunohistochemical assessment of murine tissue, spinal cords were harvested from mice transcardially perfused with 0.1 M PBS. Material was immediately frozen and embedded in Tissue-Tek OCT compound (DiaTec, Hallstadt, Germany). Cryosections (10 µm thick) were fixed in 4% PFA in PBS prior to TUNEL staining using the In Situ Cell Death Detection Kit (Roche, Mannheim, Germany) following the directions from the manufacturer. Co-staining with anti-mouse CD4-FITC (clone GK1.5, BD PharMingen) was performed simultaneously. Different areas on two sections per animal (n=4 per group) were analyzed using a Zeiss Axiophot 2 microscope equipped with an image analysis system. Data are expressed as percentage of TUNEL+ cells from all CD4 positively stained cells in this area.

Immunohistochemistry

For immunohistochemical analysis of human CNS tissue, post mortem CNS tissue samples from ten MS cases (nine brain, one spinal cord) were provided by the UK Multiple Sclerosis Tissue Bank (three females, seven males; average age 54.9 years, range 38–73). In addition to MS lesions, in three cases normal-appearing white matter as well as in two cases normal-appearing spinal cord were analyzed. Information on the clinical course was given in five cases (relapsing progressive, secondary progressive, primary progressive multiple sclerosis). As controls, different regions of three non-pathological control CNS tissue specimens were investigated. The tissue specimens were snap-frozen in liquid nitrogen, cut in 10-µm slices, and fixed in acetone. Immunohistochemical tissue labeling was performed using the Benchmark immunohistochemistry system (Ventana, Strassbourg, France). Monoclonal anti-B7H1-antibody (clone 5H1) was applied at a concentration of 15 μg/mL. The tissue samples were co-stained with H&E labeling to follow cellular infiltration. The sections were photographed at different magnifications on Olympus Vanox AH-3 light microscope using digital camera.

Statistical analysis

Two-tailed Student's t-test was used to determine the statistical significance of difference. A value of p<0.05 was considered significant. Error bars in figures represent SE.

Acknowledgements

The authors are grateful to Carolin Kiesel, Barbara Reuter and Susanne Hellmig for technical assistance. Support granted by the UK Multiple Sclerosis Tissue Bank for providing CNS tissue samples is gratefully acknowledged. This work was supported by grants from the German Research Foundation (DFG), SFB 581 (to H.W.).

Appendix

The authors declare no financial or commercial conflict of interest.

Footnotes

  1. 1

    WILEY-VCH

  2. 2

    WILEY-VCH

  3. 3

    WILEY-VCH

  4. 4

    WILEY-VCH

  5. 5

    WILEY-VCH

Ancillary