OCH, a synthetic analog of α-galactosylceramide with a truncated sphingosine chain, stimulates natural killer T (NKT) cells to produce predominantly Th2 cytokines. Thus, OCH may be a potential agent for the treatment of Th1-mediated autoimmune diseases. This study was designed to evaluate the protective effects of OCH on collagen-induced arthritis (CIA) in mice.
Mice were immunized with type II collagen (CII) and injected intraperitoneally twice per week with OCH, before or after the onset of CIA. They were monitored to assess the effect of OCH treatment on the severity of disease. Anti-CII antibodies and cytokine production were measured by enzyme-linked immunosorbent assay. Expression of cytokine genes was determined by quantitative reverse transcriptase–polymerase chain reaction.
OCH inhibited CIA in wild-type C57BL/6 (B6) mice but not in NKT-deficient mice. OCH suppressed CIA in SJL mice, which are prone to autoimmune diseases and have a deficiency in the number and function of NKT cells which is similar to that in patients with autoimmune diseases, even after disease has already developed. Disease protection conferred by OCH correlated with its ability to selectively induce Th2 cytokine production mediated by NKT cells and to promote collagen-specific Th2 responses. Neutralization of interleukin-4 (IL-4) or IL-10 with monoclonal antibodies abolished disease protection by OCH, indicating a critical role for these cytokines.
Taken together, our findings suggest that OCH holds possibilities as a therapeutic agent for autoimmune diseases such as rheumatoid arthritis.
Rheumatoid arthritis (RA) is a common autoimmune disease characterized by persistent inflammation of joints resulting in progressive destruction of cartilage and bone. Although its precise etiology is not clearly understood, cumulative evidence suggests that Th1 cells secreting interferon-γ (IFNγ) and tumor necrosis factor α (TNFα) exacerbate disease, whereas Th2 cells producing interleukin-4 (IL-4) and IL-10 suppress arthritis (1). Studies with animal models have demonstrated that systemic or locally administered IL-4 and IL-10 can effectively protect against arthritis in mice (2–11).
Natural killer T (NKT) cells are a unique subset of T cells that coexpress T cell receptor α/β (TCRα/β) and receptors from the NK lineage. NKT cells express an invariant TCRα chain (encoded by a Vα14–Jα281 rearrangement in mice and a homologous Vα24–JαQ rearrangement in humans). Unlike conventional T cells that recognize peptides in association with major histocompatibility complex (MHC), NKT cells recognize glycolipid antigens bound to the nonpolymorphic class I MHC–like protein, CD1d. NKT cells have been implicated in a variety of immune responses such as infection and tumor immunity. One striking feature of NKT cells is their capacity to secrete a large amount of cytokines, including IL-4 and IFNγ, in response to TCR stimulation (12–14). Recently, a number of reports have indicated that NKT cells play a critical role in the regulation of autoimmune responses. Abnormalities in the numbers and functions of NKT cells have been observed in patients with autoimmune diseases (15–18) as well as in a variety of mouse strains that are genetically predisposed to the development of autoimmune diseases (19–23).
While the natural ligand for NKT cells remains to be determined, α-galactosylceramide (α-GC) (Figure 1A), a derivative of marine sponge, has been shown to bind to CD1d and strongly stimulate NKT cells to produce IFNγ and IL-4, both in humans and in mice (24–26). Previously, we have shown that OCH (Figure 1A), an analog of α-GC with a truncated sphingosine chain, efficiently inhibits induced experimental autoimmune encephalomyelitis (EAE) in C57BL/6 (B6) mice, due to its ability to stimulate NKT cells to selectively produce Th2 cytokines; in contrast, α-GC had little effect on EAE (27, 28).
In the present study, we found that OCH inhibits collagen-induced arthritis (CIA), a murine experimental model for RA, in wild-type B6 but not NKT-deficient Jα281-knockout mice. We also demonstrated that OCH inhibits CIA in SJL mice even after arthritis has already developed. Experiments with anti–IL-4 or anti–IL-10 administration revealed that IL-4 and IL-10 are critical for OCH-mediated suppression of CIA. These results suggest that stimulation of NKT cells with OCH could be an attractive means of intervention in autoimmune diseases such as RA.
MATERIALS AND METHODS
B6 mice were purchased from Clea Laboratory Animal Corp. (Tokyo, Japan). SJL mice were obtained from Charles River Japan (Yokohama, Japan). Jα281-knockout mice were kindly provided by Dr. Masaru Taniguchi (Chiba University Graduate School of Medicine, Chiba, Japan). The animals were kept under specific pathogen-free conditions and studied at 7–10 weeks of age.
Induction of CIA.
Mice were immunized intradermally at the base of the tail with 100 μg of either chicken type II collagen (CII) (for B6 mice) or bovine CII (for SJL mice) (Collagen Research Center, Tokyo, Japan) emulsified with an equal volume of Freund's complete adjuvant (CFA), containing 250 μg of H37RA Mycobacterium tuberculosis (Difco, Detroit, MI). The animals were boosted by intradermal injection with the same antigen preparation on day 21. Mice were examined for signs of joint inflammation 3 times per week, and joint involvement was scored as follows: 0 = no change, 1 = focal redness of the limb or swelling and redness of 1 digit, 2 = mild swelling and erythema of the limb or swelling of >2 digits, 3 = marked swelling and erythema of the limb, 4 = maximal swelling and redness of the limb and later, ankylosis. The average of the macroscopic score was expressed as a cumulative value for all paws, with a maximum possible score of 16 per mouse. The in vivo experiments were performed with 10 mice per group and repeated twice to ensure reproducibility.
In vivo glycolipids treatment and antibody treatment.
Synthetic glycolipids were used to treat CIA. Starting from the indicated day, mice were injected intraperitoneally twice per week with either OCH or α-GC at a dose of 500 μg/kg. The control mice were injected with vehicle alone (10% DMSO in phosphate buffered saline [PBS]). To neutralize IL-4 or IL-10, anti–IL-4 monoclonal antibody (mAb) (11B11) or anti–IL-10 mAb (JES-2A5) (500 μg per mouse) was injected intraperitoneally 2 hours before glycolipid administration.
Forepaws were removed from mice killed 50 days after the first immunization of CII, then fixed in buffered formalin, decalcified, embedded in paraffin, sectioned, and stained with hematoxylin and eosin for histopathologic analysis.
NKT cell preparation.
NK1.1-positive T cells were purified from the liver of 6–8-week-old B6 mice. Liver mononuclear cells were prepared by Percoll density-gradient centrifugation. Prepared cells were then incubated with phycoerythrin-conjugated NK1.1 mAb and fluorescein isothiocyanate–conjugated CD3 mAb (BD PharMingen, San Jose, CA). The stained cells were sorted into NK1.1+,CD3+ cells. The purity of the sorted cells was >95%.
Total RNA was extracted with an RNeasy kit (Qiagen KK, Tokyo, Japan) from purified NK1.1+ T cells. Random hexamer-primed complementary DNA was prepared with the First-Strand cDNA Synthesis Kit (Invitrogen, Carlsbad, CA). For quantitative analysis of cytokines, we used the LightCycler quantitative PCR system (Roche Molecular Biochemicals, Mannheim, Germany) and performed quantitative PCR with a commercial kit (LightCycler-DNA Master SYBR Green I; Roche Molecular Biochemicals). The PCR amplification was repeated 40 times (for 15 seconds at 95°C, 5 seconds at 60°C, and 10 seconds at 72°C). All PCR reactions were normalized by GAPDH expression.
Enzyme-linked immunosorbent assay (ELISA).
To detect CII-specific IgG1 and IgG2a, chicken CII or bovine CII (5 μg/ml) was coated onto ELISA plates (Sumitomo Bakelite, Tokyo, Japan) at 4°C overnight. After blocking with 3% bovine serum albumin in PBS, serially diluted serum samples were added onto CII-coated wells. The plates were incubated with biotin-labeled anti-IgG1 and anti-IgG2a (Southern Biotechnology, Birmingham, AL) or anti-IgG antibody (CN/Cappel, Aurora, OH) for 1 hour and then incubated with streptavidin–peroxidase. After addition of substrate, the reaction was evaluated. The levels of IL-4, IL-10, and IFNγ in serum were measured by standard sandwich ELISA, using purified and biotinylated mAb pairs and standards (BD PharMingen).
Suppression of CIA development by OCH.
In order to determine whether stimulation of NKT cells modulates arthritis, we first examined the effect of α-GC, a prototypic ligand for NKT cells, on the development of CIA. Early studies showed that CIA is restricted to mouse strains bearing the H-2q, H-2r, or H-2s haplotype (29) and is generally induced in DBA/1 mice. More recently, modified immunization conditions that are sufficient to induce CIA in B6 mice have been developed; this allowed us to study CIA in knockout mice with a B6 background, which eliminated the need to backcross the knockout mice onto a DBA/1 background (30, 31). The histologic and immunologic characteristics of the disease induced in B6 mice have been shown to be similar to those in DBA/1 mice, even though B6 mice have a slightly delayed onset and less uniformly severe disease (31). We immunized B6 mice with chicken CII in CFA to elicit CIA as described previously (30) and then injected mice intraperitoneally with either α-GC or vehicle alone twice per week starting from the day of the second immunization. As shown in Figure 1B, α-GC treatment did not improve the arthritis score significantly. We next examined the effect of OCH on CIA. The mean maximum clinical CIA score was profoundly reduced in OCH-treated B6 mice (Figure 1B and Table 1). The incidence and the time of onset of disease were not significantly different between the OCH-treated group and the control group.
Table 1. Clinical scores of collagen-induced arthritis in C57BL/6 and Jα281-knockout mice*
Maximum score, mean ± SEM
Days to onset, mean ± SEM
C57BL/6 mice or Jα281-knockout mice were sensitized with chicken type II collagen for induction of arthritis.
Vehicle or 500 μg/kg of α-galactosylceramide (α-GC) or OCH was injected intraperitoneally twice per week from day 21. Data are from 5 mice per group.
P < 0.05 versus control vehicle, by Mann-Whitney U test.
To investigate the role of Vα14 NKT cells in the suppression of CIA by OCH, we examined the ability of OCH to modulate disease in Jα281-knockout mice, in which Vα14 NKT cells are absent (32). Administration of OCH did not modulate the clinical course of CIA in Jα281-knockout mice compared with mice treated with vehicle alone (Figure 1C and Table 1). These results indicate that OCH-mediated suppression of CIA requires NKT cells.
In addition to visual scoring, on day 50 after disease induction we analyzed the histologic features in the joints of forepaws of wild-type B6 mice treated with OCH or α-GC. As shown in Figure 2, in the control and α-GC–treated groups there was severe arthritis in the joints, associated with massive cell infiltration, cartilage erosion, and bone destruction. These histologic features were significantly less apparent in the group of mice treated with OCH. These results demonstrated that administration of OCH ameliorated CIA, whereas α-GC had little effect on CIA in B6 mice.
Selective induction by OCH of NKT cell–mediated IL-4 and IL-10 production.
Activation of NKT cells leads to the rapid production of a variety of cytokines, including IL-4, which promotes Th2 differentiation, and IFNγ, which promotes Th1 differentiation. Previously we demonstrated that OCH stimulates NKT cells to produce predominantly IL-4, whereas α-GC stimulates NKT cells to produce both IL-4 and IFNγ (27). IL-10, as well as IL-4, has also been reported to suppress CIA (1) and to be involved in α-GC–mediated inhibition of diabetes in the NOD mouse (33). These data led us to examine whether IL-10 was induced by OCH stimulation. We measured serum levels of IL-10 in addition to IL-4 and IFNγ, 2 hours and 12 hours after intraperitoneal injection of either OCH or α-GC into B6 mice. As shown in Figure 3A, OCH, as well as α-GC, caused an elevation in IL-10 levels. Consistent with previous results, OCH injection induced a rapid rise in IL-4 levels along with a much less marked increase in levels of IFNγ. Injection of α-GC induced the production of IL-4 and IFNγ (Figure 3A). NKT cell–deficient Jα281-knockout mice did not exhibit a response to either α-GC or OCH (Figure 3A), indicating that the increase in cytokine levels in B6 mice was mediated by NKT cells.
To further clarify the difference in gene expression by NKT cells stimulated with OCH or α-GC, we performed quantitative RT-PCR to detect the levels of expression of major contributors to joint destruction, such as TNFα and receptor activator of NF-κB ligand (RANKL) in stimulated NKT cells in vivo. We sorted NK1.1+ T cells from liver mononuclear cells of B6 mice 1.5 hours after administration of either OCH or α-GC. As shown in Figure 3B, α-GC stimulation induced expression of TNFα and RANKL genes. In contrast, OCH stimulation induced much lower levels of TNFα and RANKL expression.
Efficient inhibition by OCH of CIA development in SJL mice.
In a screen of laboratory mouse strains, NKT cells in SJL mice, which exhibit a marked propensity to the Th1-mediated autoimmune diseases, were found to be reduced in number and to have a profound defect in IL-4 secretion (19–23). Furthermore, recent studies of human autoimmune diseases demonstrated that patients with these diseases exhibited a decreased frequency of NKT cells in the periphery (15–18). These data led us to investigate whether OCH could ameliorate CIA in the autoimmune-prone mice with reduced numbers of and functional defects in NKT cells. Administration of OCH in SJL mice resulted in a rapid appearance of IL-4 and IL-10, although the levels of these cytokines were lower than those in B6 mice (data not shown). In contrast, IFNγ was barely detectable in the serum of SJL mice treated with OCH (data not shown), indicating that the cytokine profile induced by OCH stimulation was similar to that seen in B6 mice.
Next, we immunized SJL mice with bovine CII to induce CIA and then treated the mice with either OCH, α-GC, or vehicle alone. As shown in Figure 4A and Table 2, OCH administration efficiently inhibited the clinical course of CIA, whereas α-GC treatment had little effect on CIA in SJL mice. To examine the potential therapeutic effect of OCH on established CIA, we injected OCH beginning on day 28 after the first immunization, when arthritis had already developed (Figure 4B and Table 2). The severity of arthritis gradually decreased after OCH treatment, and the disease was efficiently suppressed within 1 week. These results suggest that OCH has a therapeutic effect on established CIA in autoimmune-prone mice.
Table 2. Clinical scores of collagen-induced arthritis in SJL mice*
Time of injection, treatment
Maximum score, mean ± SEM
Days to onset, mean ± SEM
Mice were sensitized with chicken type II collagen for induction of arthritis. Vehicle or 500 μg/kg of α-galactosylceramide (α-GC) or OCH was injected intraperitoneally twice per week. Data are from 6 mice per group.
P < 0.05 versus control vehicle, by Mann-Whitney U test.
OCH has been demonstrated to alter the cytokine profile of autoantigen-specific T cells in vivo (27). We therefore speculated that the OCH-mediated inhibition of arthritis might be due to a modulation of the Th1/Th2 balance, resulting from a Th2 bias of CII-reactive T cells. To explore this possibility, we measured CII-specific immunoglobulin isotype levels 50 days after induction of CIA. It is generally accepted that elevation of antigen-specific IgG2a antibody results from augmentation of a Th1 immune response to the antigen, whereas a higher level of IgG1 antibody would reflect a stronger Th2 response to the antigen. In OCH-treated mice there was a greater reduction in the level of IgG2a antibody specific to CII versus IgG1 specific to CII (Figure 5A). Consequently, the IgG1:IgG2a ratio was elevated in mice treated with OCH (Figure 5B), indicating that the suppression of CIA by OCH is associated with a Th2 bias of CII-reactive T cells.
Critical role of IL-4 and IL-10 in OCH-mediated suppression of CIA.
To confirm the involvement of IL-4 and IL-10 in the suppression of CIA, we next examined whether the inhibitory effect of OCH was abrogated after neutralization of IL-4 or IL-10 in vivo. Groups of SJL mice were injected with anti–IL-4 or anti–IL-10 mAb 2 hours before OCH or vehicle was administered. OCH-mediated suppression of CIA was partially abolished when anti–IL-4 mAb was injected (Table 3). More remarkably, in the presence of anti–IL-10 mAb, the protective effect of OCH against CIA was no longer evident at all (Table 3). Injection of neutralizing antibody to either IL-10 or IL-4 reversed the beneficial effect of administration of OCH in B6 mice also (data not shown). These results imply that IL-4 and IL-10 are critical in the OCH-mediated suppression of CIA and are consistent with our hypothesis that OCH modulates CIA by stimulating production of Th2 cytokines by NKT cells.
Table 3. Abrogation of the ability of OCH to suppress collagen-induced arthritis after neutralization of IL-10 or IL-4 in vivo*
Maximum score, mean ± SEM
Days to onset, mean ± SEM
SJL mice were sensitized with bovine type II collagen for induction of arthritis. Vehicle or 500 μg/kg of OCH was injected intraperitoneally twice per week from day 21. Anti–interleukin-10 (anti–IL-10) or anti–IL-4 monoclonal antibody (mAb) (500 μg per mouse) was injected each time vehicle or OCH was administered. Data are from 5 mice per group.
P < 0.05 versus control vehicle, by Mann-Whitney U test.
A number of studies have shown that treatment with Th2-promoting cytokines or with monoclonal antibodies directed against Th1-promoting cytokines can effectively protect against the development of CIA in mice (1–11). Here we have demonstrated that specific activation of NKT cells with their ligand OCH provides an alternative way to shift the balance from a pathogenic Th1 response toward a protective Th2 response and that disease protection is dependent on NKT cells.
We also identified a critical role of the Th2 cytokines IL-4 and IL-10 in the ability of OCH to confer protection against CIA. Recently, local delivery of Th2 cytokines, using hybridomas (8) or dendritic cells (10, 11) transfected with either IL-4 or IL-10, was found to be effective in the prevention of arthritis in animal models. In light of the fact that NKT cells are known to rapidly invade and accumulate in inflammatory lesions in a manner similar to inflammatory cells (34), stimulation of NKT cells to selectively induce Th2 cytokines might be a powerful strategy to deliver these cytokines to inflammatory lesions. It has been shown in vivo that neutralizing of IL-10, but not IL-4, increases the severity of CIA in DBA mice (3). However, we did not observe worsening of the clinical course of arthritis by neutralizing IL-4 or IL-10 in SJL mice in this study. Also in B6 mice, the clinical course of arthritis was not worsened when we neutralized IL-4 or IL-10 using the same mAb. Although the precise reason for the discrepancy with results of the earlier study is not clear, it may be because IL-4 and IL-10 levels were not high enough to modulate the severity of the disease in the natural course of arthritis in these strains.
The maximum score of CIA in the Jα281-knockout mouse was relatively low compared with that observed in wild-type B6 mice, suggesting that NKT cells may act as a modifier of the inflammation in the natural course of arthritis. We also observed a lower maximum score of CIA in CD1d-knockout mice (data not shown). Although the CD1d-knockout mice were backcrossed to B6 mice for only 6 generations, this observation further supports the notion that NKT cells increase the inflammation in the natural course of CIA. In contrast, the NOD and SJL strains of mice, which exhibit a marked propensity to the Th1-mediated autoimmune diseases, were found to have reduced numbers of NKT cells. Increasing the number of NKT cells in NOD mice by either transfer or transgenic expression of the invariant TCR Vα chain (Vα14–Jα281) resulted in a decrease in insulitis and diabetes, suggesting that NKT cells play a protective role in the development of diabetes (35, 36). In these strains of mice, the defect of NKT cells may contribute to disease susceptibility. Identification of a natural antigen for NKT cells would provide further insight into the precise role of NKT cells in the pathogenesis of autoimmune diseases such as arthritis.
Alpha-galactosylceramide, a prototypic ligand for NKT cells, has been reported to prevent diabetes in NOD mice (22, 33, 37). Even though we confirmed that α-GC inhibited the development of diabetes in NOD mice (data not shown), we did not observe any inhibitory effect of α-GC on CIA. In a previous study, we demonstrated the inhibitory effect of α-GC on EAE induced in IFNγ-knockout mice but not in wild-type B6 mice (28), suggesting that α-GC is not effective in B6 mice because NKT cell–derived IFNγ would mask the therapeutic effect of the IL-4 simultaneously produced by NKT cells. In fact, the serum levels of IL-4 and IL-10 after administration of OCH or α-GC were similar, whereas IFNγ production was much lower after injection of OCH compared with α-GC injection. This suggests that the balance of Th1/Th2 cytokines produced by NKT cells is important with regard to the protection against Th1-mediated disease conferred by the glycolipid ligand. Thus, OCH is a unique ligand that is beneficial in the treatment of a wide variety of Th1-mediated autoimmune diseases.
Another advantage of using OCH rather than α-GC is the reduced production of factors that are harmful in arthritis, such as TNFα and RANKL. TNFα is one of the major contributors to joint inflammation and destruction (38). It induces the production of other proinflammatory cytokines, stimulates endothelial cells to express adhesion molecules, increases the synthesis of metalloproteinases, and inhibits the synthesis of proteoglycans in cartilage (39). TNF levels are chronically elevated in the blood and, more specifically, in the joints, of patients with RA (40), and it has been proven that the blocking of TNF-related pathways is a strong therapeutic tool in RA (41). RANKL has emerged as one of the essential pathogenic factors in the destruction of cartilage and bone in RA (42–45). RANKL is part of the TNF ligand family and is an important regulator of both osteoclastogenesis and functions of the immune system, including lymph node organogenesis and lymphocyte development. Even though the roles of RANKL in the pathogenesis of RA during the chronic stage have not yet been elucidated, it is known that it is expressed both by synovial fibroblasts and by activated T lymphocytes derived from synovial tissue from patients with RA (43–45). Moreover, blocking of the RANKL pathway at the onset of adjuvant-induced arthritis prevents bone and cartilage destruction (42). Therefore, it is a noteworthy finding that OCH stimulation induces much lower levels of gene expression of TNFα and RANKL compared with stimulation by other glycolipid antigens such as α-GC.
We demonstrated in this study that OCH was effective in the treatment of established CIA in autoimmune-prone mice of the SJL strain, which have a quantitative and functional NKT cell deficiency; this suggests that OCH might be useful for the treatment of patients with various autoimmune diseases associated with reduced numbers of NKT cells. Furthermore, in contrast to classic MHC molecules, CD1d molecules are nonpolymorphic and are remarkably well conserved among the population and may become extremely valuable in the development of HLA-independent treatment approaches for autoimmune conditions. These findings highlight the potential use of OCH for therapeutic intervention in autoimmune diseases such as RA.
We thank Masaru Taniguchi (Chiba University, Graduate School of Medicine, Chiba, Japan) for providing Jα281-knockout mice, and Karen Mullane-Robinson (Boston, MA) for critical reading of the manuscript.