Programmed cell death 1 inhibits inflammatory helper T-cell development through controlling the innate immune response

Significance

Programmed cell death 1 (PD-1) is a membrane receptor that transmits inhibitory signals on leucocytes. Lack of PD-1 in mice results in various autoimmune diseases, which have been believed to be caused by the activation of T lymphocytes without inhibition. The current study reveals that PD-1 inhibits T-lymphocyte–mediated autoimmunity through regulating macrophage function. PD-1 deficiency only in innate immune cells caused dysregulated macrophage responses against mycobacterial adjuvant and enhanced production of IL-6, a proinflammatory cytokine, leading to the development of inflammatory autoreactive helper T cells and the exacerbation of experimental autoimmune encephalomyelitis. The lymphocyte extrinsic regulation of PD-1 provides a unique perspective on the maintenance of the immune self-tolerance and the understanding of the development of autoimmune diseases.

Abstract

Programmed cell death 1 (PD-1) is an inhibitory coreceptor on immune cells and is essential for self-tolerance because mice genetically lacking PD-1 (PD-1^−/−) develop spontaneous autoimmune diseases. PD-1^−/− mice are also susceptible to severe experimental autoimmune encephalomyelitis (EAE), characterized by a massive production of effector/memory T cells against myelin autoantigen, the mechanism of which is not fully understood. We found that an increased primary response of PD-1^−/− mice to heat-killed mycobacteria (HKMTB), an adjuvant for EAE, contributed to the enhanced production of T-helper 17 (Th17) cells. Splenocytes from HKMTB-immunized, lymphocyte-deficient PD-1^−/− recombination activating gene (RAG)2^−/− mice were found to drive antigen-specific Th17 cell differentiation more efficiently than splenocytes from HKMTB-immunized PD-1^+/+ RAG2^−/− mice. This result suggested PD-1’s involvement in the regulation of innate immune responses. Mice reconstituted with PD-1^−/− RAG2^−/− bone marrow and PD-1^+/+ CD4⁺ T cells developed more severe EAE compared with the ones reconstituted with PD-1^+/+ RAG2^−/− bone marrow and PD-1^+/+ CD4⁺ T cells. We found that upon recognition of HKMTB, CD11b⁺ macrophages from PD-1^−/− mice produced very high levels of IL-6, which helped promote naive CD4⁺ T-cell differentiation into IL-17–producing cells. We propose a model in which PD-1 negatively regulates antimycobacterial responses by suppressing innate immune cells, which in turn prevents autoreactive T-cell priming and differentiation to inflammatory effector T cells.

Autoimmune disease development is impacted by both genetic and environmental factors. Programmed cell death 1 (PD-1) is a type I membrane protein that delivers inhibitory signals to immune cells upon the binding of its ligand, PD-L1 or PD-L2 (1). PD-1 has been shown to be important for self-tolerance because spontaneous autoimmune diseases develop in PD-1^−/− mice (2–4). A single-nucleotide polymorphism that affects PD-1 expression is associated with autoimmune diseases in humans, such as systemic lupus erythematosus (5), type I diabetes (6), rheumatoid arthritis (7), and multiple sclerosis (MS) (8), suggesting that PD-1 deficiency may be a genetic factor involved in the development of autoimmunity.

Experimental autoimmune encephalomyelitis (EAE) is a rodent model of T-cell–mediated inflammatory disease in the central nervous system (CNS), causing demyelination, axonal damage, and paralysis, and is a commonly used model for human MS. Previous reports suggested that PD-1 functions to attenuate EAE. PD-1 and its ligands were found to be strongly expressed on immune infiltrates in the CNS during the peak phase of EAE (9–11). In EAE studies, PD-1–deficient mice or the use of blocking antibodies that inhibit PD-1 engagement by ligands resulted in earlier disease onset, increased inflammatory infiltrates, and increased severity of clinical symptoms compared with normal disease progression (10–16). It has been demonstrated that ligand engagement of PD-1 inhibits T-cell activation, expansion, and cytokine production (17–19). Similarly, in EAE, PD-1 signaling in CNS-specific helper T cells may inhibit their expansion and secretion of inflammatory cytokines (10–12). Recently, T-helper 17 (Th17) cells were shown to be involved in EAE by producing IL-17 and GM-CSF (20, 21). Two reports showed that PD-1^−/− mice mount an augmented Th17 response to EAE induction (14, 16). However, the fundamental mechanisms by which PD-1 regulates antigen-specific Th17 cell differentiation, expansion, and effector function in EAE remain to be understood.

To induce EAE, mice are immunized with myelin autoantigens in an emulsion of Mycobacterium tuberculosis (MTB)-derived adjuvants, causing a strong innate inflammatory response, leading to Th skewing (22). Curiously, recent studies showed that PD-1^−/− mice exhibited an altered response to infection with mycobacteria, characterized by uncontrolled bacterial burden; massive production of cytokines, termed “cytokine storm”; and early death (23–25). We wondered if this unique response of PD-1^−/− mice to mycobacteria contributed to their Th response in EAE.

In this study, we took a combination of genetic and immunological approaches in which the innate response to MTB-derived adjuvant and antigen-specific T-cell polarization were separately analyzed. The present data suggest that an enhanced innate response of PD-1^−/− mice to MTB contributes to the susceptibility of these mice to severe EAE. We propose a previously undescribed function of PD-1 in controlling the basal state of the innate immune response, the failure of which can cause the activation of adaptive immune responses, provoking autoimmunity.

Results

Augmented EAE with Suboptimal Immunization of PD-1^−/− Mice.

EAE in C57BL/6 mice is typically induced by immunizing the mice with myelin oligodendrocyte glycoprotein (MOG)_35–55 peptide in an emulsion of complete Freund's adjuvant (CFA) and 200–300 μg of additional heat-killed MTB (HKMTB), followed by two separate i.p. injections of pertussis toxin (PTX), which is assumed to induce a strong innate immune response leading to MOG-specific T-cell priming. In agreement with previous reports (12, 14), the induction of EAE with this treatment resulted in accelerated disease progression in PD-1^−/− mice, characterized by earlier disease onset and an earlier peak in disease activity (Fig. 1A).

Fig. 1.

Augmented EAE with suboptimal immunization of PD-1^−/− mice. (A) EAE scores of PD-1^+/+ (n = 12) and PD-1^−/− (n = 12) mice (C57BL/6 background) after immunization with MOG/CFA/HKMTB in the presence of PTX on day 0 and day 2. (B) EAE scores of PD-1^+/+ (n = 13) and PD-1^−/− (n = 12) mice after immunization with MOG/CFA/HKMTB. (C) EAE scores of PD-1^+/+ (n = 13) and PD-1^−/− (n = 11) mice after immunization with MOG/CFA. Data presented are combined from two independent experiments (means ± SEM). Statistical analysis was performed using a two-tailed Student t test. *P < 0.05 and **P < 0.01.

Because this treatment also caused prominent disease in PD-1^+/+ mice, we next examined the effects of suboptimal immunization with reduced adjuvants on both wild-type and mutant mice. PD-1^−/− mice appeared to be susceptible to EAE by the suboptimal treatment. First, when PTX injection was eliminated from the treatment, the PD-1^−/− mice developed a comparable disease to that developed with full immunization, whereas PD-1^+/+ mice showed an attenuated response (Fig. 1B). PTX is known to open the blood–brain barrier (26), inhibit Foxp3 expression on T cells (27), and enhance Th17 development in EAE (28). Our findings and a previous study by others (29) indicated that PTX is dispensable for the development of EAE in PD-1^−/− mice. Next, when both PTX and the additional dose of HKMTB were eliminated, many PD-1^−/− mice still developed EAE, whereas the PD-1^+/+ mice remained healthy (Fig. 1C). The severe EAE in PD-1^−/− was well-correlated with the number of immune cells infiltrating the affected brain (Fig. S1).

CFA itself contains a low dose of HKMTB (50 μg per mouse). This small dose of HKMTB appeared to be required for EAE in PD-1^−/− mice because PD-1^−/− mice that received MOG in incomplete Freund’s adjuvant did not develop EAE with or without PTX. These data indicated that PD-1^−/− mice were susceptible to EAE by immunization with much lower doses of adjuvants.

Enhanced Inflammatory Cytokine Production Associates with EAE in PD-1^−/− Mice.

Th1 and Th17 cells, which typically produce IFN-γ and IL-17, respectively, have been shown to be involved in EAE (30). To study the development of the two Th cell types during disease progression in PD-1^−/− mice, we analyzed the MOG-specific effector T-cell response at early and later phases of the disease. On day 8, purified CD4⁺ T cells from the spleen of PD-1^−/− mice produced significantly more IL-17 and IFN-γ than the cells from PD-1^+/+ mice upon restimulation with MOG_35–55 in vitro (Fig. 2A). In response to suboptimal immunization conditions, the cytokine levels were lower, but the difference between PD-1^−/− and PD-1^+/+ mice was still significant. These findings were consistent with the earlier onset and the severe disease in PD-1^−/− mice.

Fig. 2.

EAE in PD-1^−/− mice associates with enhanced inflammatory cytokine production. (A) PD-1^+/+ and PD-1^−/− mice (C57BL/6 background) were immunized with MOG/CFA/HKMTB/PTX (n = 12), MOG/CFA/HKMTB (n = 11), or MOG/CFA (n = 6) and analyzed 8 d later. (B) PD-1^+/+ and PD-1^−/− mice were immunized with MOG/CFA/HKMTB/PTX (n = 7), MOG/CFA/HKMTB (n = 7), or MOG/CFA (n = 13) and analyzed 30 d later. (A and B) Sorted CD4⁺ T cells were restimulated with MOG_35–55, and the cytokine concentrations were determined by ELISA. Statistical significance was determined by Student t test. Not significant (NS) P > 0.05, *P < 0.05, and **P < 0.01.

Next, we examined the T-cell responses to MOG_35–55 on day 30, when the clinical symptoms have mostly receded. As shown in Fig. 2B, the PD-1^−/− CD4⁺ cells still produced significantly higher levels of IL-17 than the PD-1^+/+ control cells; however, the IFN-γ levels were indistinguishable. These results suggested that in comparison with PD-1^+/+ mice, PD-1^−/− mice generated a stronger memory response to both the standard and suboptimal EAE-inducing immunizations, and this response was associated with enhanced IL-17 production.

The Innate Response to HKMTB in PD-1^−/− Mice Influences Antigen-Specific Th17 Development.

Because PD-1^−/− mice developed a MOG_35–55-specific Th17 response to EAE immunizations with low doses of HKMTB, PD-1^−/− mice may have intrinsic properties to exhibit augmented response to HKMTB, which results in enhanced induction of Th17. As PD-1 is expressed on innate cells (31, 32) as well as T and B cells, PD-1 may regulate innate response that influences T-cell activation. To address this point, we established a system in which we can separately evaluate the function of PD-1 in the innate and the acquired immune responses. First, we bred PD-1^−/− mice onto the recombination activating gene (RAG)2^−/− background to exclude T- and B-cell responses (Materials and Methods). Then, PD-1^−/− RAG2^−/−or control PD-1^+/+ RAG2^−/− mice were immunized with HKMTB alone in CFA to induce a genuine anti-HKMTB response in vivo. Spleen cells from the immunized mice were prepared and used to stimulate naive OVA-specific TCR transgenic T cells (PD-1^+/+) under the various Th skewing conditions (Materials and Methods).

We found that when splenocytes from HKMTB-immunized PD-1^−/− RAG2^−/−mice were incubated with PD-1^+/+ naive OVA-reactive T cells under neutral condition, IL-17⁺ cells were produced more efficiently (average ± SD of 22.5 ± 3.627%) than when splenocytes from HKMTB-immunized PD-1^+/+ RAG2^−/− mice (13.4 ± 4.68%) were used (Fig. 3A, second column; P = 0.021 from five combined experiments). When exogenous IL-6 and TGF-β are added to the neutral condition (Th17 condition; Fig. 3A, fourth column), splenocytes from both backgrounds showed efficient induction of IL-17⁺ cells (PD-1^+/+ RAG2^−/− vs. PD-1^−/− RAG2^−/− : 34.33 ± 3.99 vs. 37.43 ± 5.84, P = 0.54). In contrast, splenocytes from both PD-1^+/+ RAG2^−/− and PD-1^−/− RAG2^−/− mice were similarly efficient in promoting the differentiation of IFN-γ–producing cells under Th1 skewing conditions (P = 0.48 in nonimmunized and P = 0.65 in immunized mice; five experiments) (Fig. 3B). We got essentially the same data when we performed the same experiments with the mice in different genetic backgrounds (using PD-1^−/− RAG2^−/− in BALB/c background and naive CD4⁺ cells from DO11.10 TCR Tg), suggesting that the finding is general to the PD-1^−/− strain (Fig. S2). Also, splenocytes from HKMTB-immunized PD-1^−/− mice on RAG2-sufficient background were more potent in inducing Th17 than their PD-1^+/+ counterparts (Fig. S3). It should be noted, however, that exogenous IL-6 and TGF-β were required for efficient Th17 induction (Th17 condition) when RAG2-sufficient splenocytes were used. It suggested that T-cell and B-cell responses in RAG2-sufficient mice somehow attenuate the primary innate anti-HKMTB response. In any case, the current results indicated that the innate immune responses of PD-1^−/− mice, when exposed to mycobacteria, were more efficient at promoting a Th17 response compared with PD-1^+/+ mice.

Fig. 3.

Splenocytes from HKMTB-immunized PD-1^−/− RAG2^−/− mice facilitate Th17 induction in vitro. Splenocytes from nonimmunized or HKMTB/CFA-immunized PD-1^+/+ RAG2^−/− and PD-1^−/− RAG2^−/− mice (C57BL/6 background) were cultured with naive CD4⁺ T cells from OTII TCR Tg mice in the presence of OVA_323–339 (5 μg/mL) under conditions indicated. Development of the (A) IL-17–producing or (B) IFN-γ–producing CD4⁺ T cells was examined by intracellular cytokine staining. Essentially the same data were obtained from experiments immunizing PD-1^−/− RAG2^−/− mice (BALB/c background) and naive CD4⁺ cells from DO11.10 TCR Tg mice (BALB/c background) (Fig. S2).

Innate Immune Cells of PD-1^−/− Mice Causes Enhanced Th17 and Exacerbation of EAE.

To directly determine contribution of enhanced innate response of PD-1^−/− mice to HKMTB in the development of EAE, we reconstituted lethally irradiated wild-type C57BL/6 mice with bone marrow cells from PD-1^+/+ RAG2^−/− and PD-1^−/− RAG2^−/− mice. Then, recipients received adoptive transfer of CD4⁺ T cells from PD-1^+/+ and were immunized with MOG_35–55 in the emulsion with CFA and HKMTB. We found that the mice reconstituted with PD-1^−/− RAG2^−/− bone marrow developed EAE with earlier onset and more severe clinical scores than the mice reconstituted with PD-1^+/+ RAG2^−/− bone marrow (Fig. 4A). The severe EAE in the PD-1^−/− RAG2^−/− bone marrow recipients were associated with significantly higher production of IL-17 in the recall response compared with PD-1^+/+ RAG2^−/− recipients (Fig. 4B). The data strongly suggested that PD-1 deficiency in the innate cells caused priming of CD4⁺ T cells from PD-1^+/+ into pathogenic Th17, leading to severe EAE.

Fig. 4.

Nonlymphocyte populations of PD-1^−/− mice cause enhanced Th17 development and exacerbation of EAE. (A) Lethally irradiated C57BL/6 mice were reconstituted by bone marrow transplantation and adoptive transfer of CD4⁺ cells as indicated. Mice were immunized with MOG/CFA/HKMTB, and the EAE score was evaluated. Data presented are combined from two independent experiments (total of 18 mice included in each group) (mean ± SEM). Statistical significance was determined by Student t test (*P < 0.05 and **P < 0.01). (B) CD4⁺ T cells were harvested at day 30 and restimulated by MOG_35–55 as indicated in Materials and Methods. IL-17 and IFN-γ were measured. Each point represents an individual mouse, with the black bar indicating the mean value. Mice reconstituted with PD-1^−/− RAG2^−/− BM showed significantly higher IL-17 than their PD-1^+/+ RAG2^−/− counterparts by one-tailed Student t test. *P = 0.04.

CD11b⁺ Macrophages from HKMTB-Treated PD-1^−/− Mice Efficiently Induce Th17 Cell Polarization.

We wanted to determine the innate immune cell population responsible for the enhanced Th17 induction in PD-1^−/− mice. Because macrophages and some dendritic cells (DCs) express CD11b, we magnetically separated PD-1^−/− splenocytes into the CD11b⁺ and CD11b⁻ populations, treated each of them with HKMTB in vitro, and then assayed their ability to induce Th17. We found that the CD11b⁺ population from PD-1^−/− mice skewed naive T cells into Th17 cells more efficiently than did CD11b⁺ cells from PD-1^+/+ mice, in a manner dependent on the added bacterial dose (Fig. 5 A and B). In contrast, we detected little or no Th17 skewing by the CD11b⁻ population (Fig. 5B), suggesting that the antigen-presenting cells (APCs) in the CD11b⁻ fraction (mostly B cells) were not involved in the Th17 skewing by PD-1^−/− APCs. Further FACS sorting of CD11b⁺ population into macrophages (CD11b⁺) and DCs (CD11b⁺ CD11c⁺) revealed that HKMTB-treated macrophages but not HKMTB-treated DCs promote Th17 skewing (Fig. S4). Similar results were obtained when CD11b⁺ cells were prepared from PD-1^+/+ RAG2^−/− and PD-1^−/− RAG2^−/− mice (Fig. 5C). These data suggested that CD11b⁺ macrophages in PD-1^−/− mice influence Th17 development upon the recognition of HKMTB.

Fig. 5.

CD11b⁺ cells from PD-1^−/− mice efficiently induce Th17 cell development upon HKMTB stimulation. CD11b⁺ cells were purified from the splenocytes of (A) PD-1^+/+ and PD-1^−/− mice or (C) PD-1^+/+ RAG2^−/− and PD-1^−/− RAG2^−/− mice (all in C57BL/6 background) and were either untreated (ctrl) or stimulated by HKMTB (10 and 100 μg/mL). Three days later, the stimulated cells were cocultured with CD4⁺ T cells from OTII TCR Tg mice in the presence of OVA_323–339 (5 μg/mL) for 4 d. IL-17–producing OTII⁺ CD4⁺ T cells (PD-1^+/+) were quantified by intracellular cytokine staining. (B) Summary of the data shown in A. Note that CD11b⁻ cells from PD-1^+/+ and PD-1^−/− mice did not induce Th17 by two independent experiments. Data shown for CD11b⁺ cells are mean ± SD from six independent experiments (for ctrl and HKMTB 100-μg/mL condition) and four experiments (for HKMTB 10-μg/mL condition). Statistical significance was determined by Student t test. *P < 0.05 and **P < 0.01.

Enhanced Production of IL-6 by CD11b⁺ Cells from PD-1^−/− Mice Causes Strong Th17 Development.

To examine the mechanism of the enhanced Th17 skewing by PD-1^−/− macrophages, we measured the cytokine production from the CD11b⁺ population from PD-1^−/− mice after HKMTB treatment. Among the cytokines tested (IL-2, IL-4, TNF-α, IL-6, IL-17, IFN-γ, IL-10, IL-12, and IL-23), the PD-1^−/− CD11b⁺ cells produced more IL-6 in the presence of HKMTB than did the PD-1^+/+ CD11b⁺ cells (Fig. 6A). Other cytokines detected, such as TNF-α (Fig. 6B) and IL-10 (Fig. 6C), were similar between CD11b⁺ cells from PD-1^+/+ and PD-1^−/− cells. In addition, this augmentation of IL-6 production by PD-1^−/− cells was specific for HKMTB because other Gram-positive bacteria, such as heat-killed listeria monocytogenes (HKLM), did not induce the production of significant levels of IL-6 from either PD-1^+/+ or PD-1^−/− CD11b⁺ cells (Fig. 6A). The difference in IL-6 production between bacterial strains may be due to differences in their ability to activate toll-like receptors and other receptors, such as dectin, on the CD11b⁺ cells.

Fig. 6.

Enhanced production of IL-6 by CD11b⁺ cells from PD-1^−/− mice causes strong Th17 development. CD11b⁺ cells purified from the splenocytes from PD-1^+/+ and PD-1^−/− mice (C57BL/6 background) were either untreated (ctrl) or stimulated with HKMTB (10 and 100 μg/mL) or HKLM (10 and 100 μg/mL). Three days later, the production of (A) IL-6, (B) TNF-α, and (C) IL-10 was evaluated. Data presented are combined from three independent experiments (mean ± SD). Significance was determined by Student t test. *P < 0.05. (D) HKMTB-stimulated CD11b⁺ cells were cocultured in the presence of OTII⁺CD4⁺ T cells and OVA_323–339 together with rat IgG (200 ng/mL) or anti–IL-6R (as shown) under neutral condition for 4 d. IL-17–producing OTII⁺ CD4⁺ T cells were quantified by intracellular cytokine staining. Data shown are representative of three independent experiments.

Because IL-6 is important for Th17 induction, we confirmed that the IL-6 from HKMTB-treated PD-1^−/− CD11b⁺ cells functioned to induce the differentiation of IL-17–producing cells. As shown in Fig. 6D, the antigen-specific Th17 differentiation by HKMTB-treated CD11b⁺ cells was inhibited in the presence of a blocking anti–IL-6 receptor antibody in a dose-dependent manner. Taken together, these data suggested that PD-1 regulates the IL-6 production from macrophages to inhibit Th17 induction.

Discussion

Here we found that the induction of EAE in PD-1^−/− mice required considerably lower amounts of bacterial adjuvants compared with the EAE induction in PD-1^+/+ mice. In addition, PD-1^−/− mice mounted an augmented Th17 effector/memory response against myelin antigen. Although PD-1 has previously been shown to function on T cells, our in vitro and in vivo assays using naive PD-1^+/+ T cells and lymphocyte-null PD-1^−/− RAG2^−/− mice suggested that PD-1 regulates Th17 differentiation indirectly by suppressing the IL-6 production from innate immune cells.

Previous studies have shown that IL-6 is necessary for Th17 induction in EAE because IL-6^−/− mice (33) or treatment with an IL-6 blocking antibody (34) abrogates EAE. Although activated T cells can produce autocrine IL-6, the receptor for IL-6 is expressed only on naive T cells and is down-modulated after activation (35, 36). Thus, T cells that are exposed to IL-6 early in their activation differentiate into pathogenic Th17 cells. Our experiments showed that macrophages but not DCs from PD-1^−/− enhanced the induction of Th17 upon HKMTB treatment in vitro (Fig. S4). DCs are necessary for EAE development but are not potent to produce IL-6 upon mycobacteria recognition, at least in vitro (37, 38). Thus, although our data do not exclude PD-1’s function on DCs in vivo, they suggest that CD11b⁺ macrophages are the primary population that produces early IL-6 in response to HKMTB.

The massive production of IL-6 by PD-1^−/− macrophages may explain the unique response of these mice to mycobacterial infection. Although the Th1 response is ideal for controlling MTB (39), infected PD-1^−/− mice exhibit high serum IL-6 levels (23, 24), which may lead to an altered Th1 response and an augmented Th17 response. An augmented Th17 response induced by overproduced IL-6 not only inhibits efficient antimycobacterial immune responses but also contributes to autoimmune-like diseases in mice. We conclude that PD-1 functions to control innate IL-6 production, which in turn suppresses autoantigen-specific T-cell responses. Our conclusions are consistent with a previous report that immunizing PD-1^−/− mice with Bacille Calmette-Guerin (bacillus Calmette-Guérin) induces symptoms resembling Kawasaki disease, a vascular inflammatory disease associated with IL-6 production (40).

As PD-1 down-modulates the tyrosine phosphorylation of cytoplasmic proteins including Syk tyrosine kinase, through activation of nonreceptor tyrosine phosphatase, SHP-2 (41), a direct function of PD-1 in macrophage activation could explain the augmented IL-6 production by PD-1^−/− macrophages. However, we could not demonstrate the expression of PD-1 on the monocyte/macrophage population. This inability to reproduce results from a previous study in which PD-1 was detected on macrophages (32, 42) may be due to the high background binding of available anti-mouse PD-1 antibodies to this cell type. Indeed, the mAb used in previous research (42) was found to stain PD-1^−/− macrophages as efficiently as PD-1^+/+ macrophages. Thus, it is likely that the regulation of macrophage activation by PD-1 is not intrinsic to the macrophage. Because basal macrophage activation occurred in PD-1^−/− RAG2^−/− mice, this would suggest that non-T cells, possibly innate lymphoid cells, are responsible for the basal macrophage activation in these mice. For example, PD-1 is expressed on natural killer cells (43) and regulates their cytokine production (44). In the absence of PD-1, natural killer cells may produce more cytokines during immune surveillance or upon infection, possibly affecting the activation status of macrophages. How PD-1 regulates macrophage activation in a cell-extrinsic manner is currently unclear and should be addressed in future studies.

In conclusion, we propose that PD-1 can regulate a magnitude of primary innate responses to certain microbes, and this regulation in turn inhibits the differentiation of self antigen-specific inflammatory T cells. Our findings provide a unique perspective on the function of PD-1, which was previously well established as a key molecule for the maintenance of immune tolerance.

Materials and Methods

Mice.

PD-1^−/− mice were described before (45) and thoroughly backcrossed into C57BL/6 or BALB/c background (13 times). PD-1^−/− and the littermate control (PD-1^+/+) were obtained from heterozygote (PD-1^+/−) to heterozygote breeding. DO11.10 TCR transgenic (Tg) mice and OT-II TCR Tg mice were obtained from K. Murphy (Washington University, St. Louis, MO) and K. Kabashima (Kyoto University), respectively. RAG2^−/− mice were from F. Alt (Harvard Medical School, Boston, MA) and backcrossed into BALB/c or C57BL/6 background 10 times. To generate mice doubly deficient for PD-1 and RAG2, we first mated PD-1 single KO with RAG2 single KO in the same genetic background. The resulting F1 generation (PD-1^+/− RAG2^+/−) was intercrossed to obtain PD-1^−/− RAG2^−/− (double KO). Littermate PD-1^+/+ RAG2^−/− mice from the intercross breeding were used as control. All mice were maintained under specific pathogen-free conditions and treated according to protocols approved by the Animal Research Committee, Graduate School of Medicine, Kyoto University.

Reagents.

Neutralizing rat anti–IL-6 mAb (MR16-1) was kindly provided by T. Kishimoto (Osaka University, Suita, Japan) and Chugai Pharmaceutical Co. Ltd. Chrompure Rat IgG (IgG) was from Jackson Immuno Research. Information about FACS antibodies and peptide sequence is shown in SI Materials and Methods.

EAE.

Mice were immunized s.c. with 0.2 mL of an emulsion containing 200 μg MOG_35–55 peptide, CFA, and 250 μg HKMTB. In addition, the mice received two i.p. injections of 200 ng PTX at the time of and 48 h after immunization. In some experiments, “suboptimal” immunization conditions were used, in which HKMTB, PTX, or both were excluded. Mice were observed daily for clinical signs of disease, and EAE scores were defined as follows: 0, no disease; 1, limp tail; 2, hind limb weakness; 3, paralysis of one hind limb; 4, paralysis of both hind limbs; 5, limited movement; and 6, moribund or death. Scores are shown as the mean clinical scores for each experimental group. For restimulation assay, spleens were harvested 8 or 30 d after immunization, and CD4⁺ T-cell–enriched samples were obtained using anti-CD4 microbeads and the autoMACS system (Miltenyi). In a 1-mL volume, 2 × 10⁶ of the isolated CD4⁺ T cells were cultured with 2 × 10⁶ mitomycin C-treated splenocytes from PD-1^+/+ mice in the presence of MOG_35–55 peptide (30 μg/mL). Supernatants were harvested 3 d later for cytokine analysis.

Antigen-Specific Th Skewing by HKMTB-Immunized Splenocytes.

Mice were immunized s.c. with 0.2 mL of an emulsion containing 250 μg HKMTB and CFA. Eight days later, 1 × 10⁶ splenocytes were cultured with 1 × 10⁵ CD4⁺ T cells purified from DO11.10 or OTII TCR Tg mice in the presence of 5 μg/mL OVA in 48-well plates. For Th skewing, the following recombinant cytokines and antibodies were added to the cultures: for Th1, IL-12 (20 ng/mL) and anti–IL-4 (10 μg/mL) were added; for Th17, human TGF-β1 (5 ng/mL), IL-6 (20 ng/mL), anti–IL-4 (10 μg/mL), and anti–IFN-γ (10 μg/mL) were added; and for neutral conditions, anti–IL-4 (10 μg/mL) and anti–IFN-γ (10 μg/mL) were added. Four days later, the cells were restimulated with 10 ng/mL phorbol 12-myristate 13-acetate and 1 μg/mL Ionomycin (Sigma-Aldrich) in the presence of 0.66 μL/mL Golgistop [Becton Dickinson (BD)] for 6 h. After stimulation, the cells were washed in FACS buffer [PBS, 2% (vol/vol) FCS, and 0.02% NaN₃] and stained for cell surface markers. The cells were then treated with Cytofix/Cytoperm buffer (BD), washed with perm/wash buffer (BD) twice, and subsequently stained with anticytokine antibodies. Samples were analyzed by FACSCanto II (BD).

Antigen-Specific Th Skewing by Mycobacteria-Activated CD11b⁺ Cells.

Splenocytes were labeled with Miltenyi CD11b microbeads and subjected to three positive selections by autoMACS (Miltenyi). The purified cells contained ∼95% CD11b⁺ cells. In a total volume of 200 μL, 3 × 10⁵ of the isolated CD11b⁺ cells were stimulated with HKMTB or HKLM at 10 and 100 μg/mL. Three days later, cell culture supernatants were evaluated for cytokine production. For antigen-specific Th skewing, 1 × 10⁶ CD11b⁺ cells were stimulated with HKMTB or HKLM at 10 and 100 μg/mL for 3 d. The stimulated CD11b⁺ cells were then cultured with 1 × 10⁵ CD4⁺ T cells purified from OTII TCR Tg mice for 4 d in the presence of 5 μg/mL OVA_323–339 under neutral conditions. In some experiments, a rat anti–IL-6R blocking antibody or control rat IgG was added to the cultures. After a 4-d incubation, the development of Th1 and Th17 cells was examined by intracellular cytokine staining, as described above.

Cytokine Analysis.

Culture supernatants were stored at −80°C until the analysis. IL-12p70 and IL-23 (p19/p40) were determined by ELISA (eBioscience). The IL-2, IL-4, TNF-α, IL-10, IL-6, IL-17, and IFN-γ levels were measured by the mouse Th1/Th2/Th17 cytometric bead array kit (BD) or ELISA (eBioscience), according to the manufacturer’s protocols.

Statistics.

Statistical significance between two groups was calculated using a two-tailed Student t test, and P < 0.05 was considered to be statistically significant.