Multiple studies have demonstrated that 4-1BB (CD137), a member of the TNF receptor superfamily, is expressed on several immune cells including activated T cells. However, the expression and the role of 4-1BB on natural killer T (NKT) cells have not been fully characterized. In this study, it was shown that 4-1BB was not expressed on naive NKT cells but was rapidly induced on activated NKT cells by TCR engagement with α-galactosylceramide (α-GalCer). Also, 4-1BB signaling provided by 3H3, an agonistic anti-4-1BB mAb, promoted NKT cell activation resulting in enhanced cytokine production of NKT cells driven by α-GalCer. When NKT cell-driven airway immune responses were evaluated by intranasal administration of α-GalCer, airway hyperresponsiveness (AHR) and lung inflammation were significantly more aggravated in mice treated with 3H3 and α-GalCer than in mice treated with α-GalCer alone. These aggravations were accompanied by up-regulation of IL-4, IL-13, and IFN-γ production. Interestingly, AHR was not developed in IL-4Rα-deficient mice treated with α-GalCer with or without 3H3 but was exacerbated in IFN-γ-deficient mice. Our study suggests that 4-1BB on NKT cells functions as a costimulatory molecule and exacerbates the induction of NKT cell-mediated AHR, which is dependent on the IL-4Rα-mediated pathway.
Natural killer T (NKT)3 cells represent a subpopulation of T cells that express NK receptors such as NK1.1 and Ly-49 in addition to markers of conventional T cells such as the TCR/CD3 complex (1). A major subset of NKT cells is restricted by the MHC class I-like molecule CD1d. A marine sponge glycolipid, α-galactosylceramide (α-GalCer), is well known to activate these CD1d-restricted NKT cells through CD1d-mediated Ag presentation and to induce rapidly large quantities of the Th1 cytokine IFN-γ and the Th2 cytokine IL-4, as well as TNF-α and IL-2 (2, 3). This rapid production of cytokines regulates the activity of conventional T cells, NK cells, dendritic cells, and B cells (4, 5, 6).
Airway inflammatory cell accumulation is thought to play a critical role in the clinical expression and pathogenesis of asthma, including airway hyperresponsiveness (AHR). Although many studies have focused on the roles played by Th2 cells in AHR and lung inflammation in asthma, we and others have demonstrated that NKT cells also play a critical role in the development of allergen-induced AHR and airway inflammation (7, 8, 9, 10). To date, it has been reported that glycolipid activation of NKT cells is capable of inducing AHR and airway inflammation independently of conventional CD4+ T cells (11). In addition, NKT cell-mediated airway responses are similar to conventional CD4+ T cell-mediated responses in terms of Th2 cytokine production, inflammatory cell infiltration, and mucus secretion (11).
A member of the TNF receptor superfamily, 4-1BB (CD137), functions mainly as a costimulatory molecule for T cells (12). It engages with an agonistic anti-4-1BB mAb, thereby promoting the clonal expansion and survival of CD8+ T cells (13, 14), the eradication of large established tumors (15, 16), the abrogation of T cell-dependent Ab production (17, 18), and the prevention of autoimmune diseases (19, 20). Interestingly, an agonistic anti-4-1BB mAb ameliorates conventional CD4+ T cell-driven airway responses such as AHR and lung inflammation by inducing an anergy of conventional MHC class II-restricted CD4+ T cells (21, 22, 23).
Although a study using 4-1BB-deficient mice revealed the importance of 4-1BB signaling to the development of NKT cells (24), little is known about the expression of 4-1BB by naive or activated NKT cells. The effects of 4-1BB engagement on NKT cells are also wholly unknown. Therefore, we questioned whether 4-1BB is related to NKT cell activation. In this study, we evaluated whether 4-1BB engagement on NKT cells can regulate NKT cell-mediated airway responses and whether an agonistic anti-4-1BB mAb can be used for the treatment of NKT cell-mediated airway diseases similarly as in the treatment of conventional Th2 cell-mediated airway diseases.
Materials and Methods
Six- to 8-wk-old female BALB/c mice were purchased from Orient Bio and maintained in pathogen-free conditions in the animal-housing facilities of the College of Pharmacy, Seoul National University (Seoul, Republic of Korea) for the duration of the experiments. CD1d-deficient mice, IFN-γ-deficient mice, and IL-4Rα-deficient mice (on a BALB/c background) were bred under aseptic conditions in facilities at Seoul National University and Pohang University (Pohang, Republic of Korea). The studies reported here conformed to the principles for laboratory animal research outlined by Seoul National University.
Preparation of α-GalCer and anti-4-1BB mAb
α-GalCer was provided by Dr. S. Kim (Seoul National University). α-GalCer was dissolved in 0.5% Tween 20 in saline as a vehicle. Anti-4-1BB mAb (clone 3H3) was purified from the ascites of nude mice using a protein G column.
In vivo treatment
Mice were i.p. injected with 200 μg of anti-4-1BB Ab (clone 3H3) or rat IgG control Ab (Rockland) 24 h before α-GalCer administration. In addition, 20 μg of this mAb were intranasally administered along with α-GalCer to evaluate the effects of 4-1BB engagement on the NKT cell-mediated airway responses.
To induce NKT cell-mediated airway responses, mice were anesthetized with ketamine (Yuhan) and xylazine (Bayer Korea) and then intranasally treated with 0.2 μg of α-GalCer in 20 μl of PBS on day 0. The control group was treated with vehicle (0.5% Tween 20 in saline) for all experiments.
Flow cytometric analysis of NKT cells
The NKT cell ligand α-GalCer was incubated with CD1d dimer (BD Pharmingen, catalog No. 557599) at a molar ratio of 10:1. α-GalCer-loaded CD1d dimer complex was labeled with anti-mouse IgG1-PE (Southern Biotech) and incubated overnight at 4°C at a weight ratio of 1:2. The mixture was then used with FITC-conjugated anti-mouse TCR-β mAb (BD Pharmingen) to stain the isolated single-cell suspension of hepatic mononuclear cells, splenocytes, or thymocytes. CD1d dimers that had not been loaded with α-GalCer were used as a control. The cells were washed, fixed with 4% paraformaldehyde, and analyzed using a FACSCalibur flow cytometer (BD Pharmingen). FACS analysis of intracellular cytokines in NKT cells was performed with BD Cytofix/Cytoperm Plus with Golgi-Plug (BD Pharmingen) according to the manufacturer’s protocol.
NKT cell purification
Isolated splenocytes were incubated with α-GalCer, CD1d dimer, and anti-mIgG1-PE, followed by anti-PE microbeads (Miltenyi Biotec). CD1d-restricted NKT cells were then purified by using a VarioMACS sorter (Miltenyi Biotec).
Measurement of methacholine AHR
Pulmonary function testing was assessed in conscious, unrestrained mice using noninvasive whole-body plethysmography (Allmedicus). One day after α-GalCer treatment, methacholine AHR was assessed as previously described (9). Briefly, mice were placed in the plethysmograph chamber and exposed to an aerosol of PBS (basal readings) and then aerosol doses of methacholine at 10, 20, and 40 mg/ml. The aerosol was generated with an ultrasonic nebulizer and drawn through the chamber for 3 min. Enhanced pause (Penh) readings were taken for 3 min and averaged.
Analysis of bronchoalveolar lavage (BAL) fluids (BALF)
Mice were sacrificed on day 3 and BAL cells were obtained and analyzed as previously described (9). Differential cell counts were performed by counting at least 300 cells on cytocentrifuged preparations followed by staining with Diff-Quick (Dade Behring).
Results are expressed as the means ± SEM. Each experiment was repeated at least twice (n = 3–7). Differences between groups were analyzed using the Student’s t test. The level of significance was set at p < 0.05.
4-1BB is expressed on activated NKT cells, but not on naive NKT cells
We first analyzed the expression of 4-1BB on the cell surface of NKT cells in hepatic MNC using flow cytometric analysis. In naive mice, 4-1BB was not expressed on conventional T cells and CD1d-restricted NKT cells (Fig. 1 A). However, CD1d-restricted NKT cells, but not conventional T cells, expressed 4-1BB on the cell surface within 2 h after α-GalCer injection. This expression of 4-1BB returned to nearly basal levels by day 3. Similar expression profiles were seen on splenic NKT cells, but not on thymic NKT cells (data not shown).
To confirm these results, we also analyzed the 4-1BB expression pattern of the NKT cell line DN32-D3, finding that it did not express 4-1BB without α-GalCer stimulation (Fig. 1,B). Previous reports had shown that DN32-D3 constitutively expressed CD1d molecules on the cell surface and was activated by the addition of α-GalCer without additional APCs in the culture medium. We confirmed that CD69 expression on DN32-D3 was increased when DN32-D3 was cultured with α-GalCer (Fig. 2,A). TCR stimulation with α-GalCer induced the expression of 4-1BB by the DN32-D3 cells (Fig. 1,B). Stimulation using anti-CD3 Ab showed similar results (Fig. 2 B).
To determine the strength of the TCR signal required to induce 4-1BB on NKT cells, DN32-D3 cells were stimulated with varying concentrations of α-GalCer or immobilized anti-CD3 Ab. When the concentration of α-GalCer or of immobilized anti-CD3 Ab was increased, so too was 4-1BB expression by DN32-D3 cells (Fig. 2 B). These results indicate that the degree of 4-1BB expression on NKT cells induced by TCR stimulation depends on the strength of the TCR signal.
Agonistic anti-4-1BB mAb treatment enhances cytokine production from α-GalCer-stimulated NKT cells
We hypothesized that 4-1BB might function as a costimulatory molecule on NKT cells. To verify this hypothesis, BALB/c mice were i.p. administered with 200 μg of 3H3 or control Ig 1 day before α-GalCer challenge. Then, the mice were i.p. challenged with 0.2 μg of α-GalCer. After the treatment of α-GalCer or vehicle, sera were collected from the mice at the indicated time points and IFN-γ and IL-4 levels were measured. Administration of 3H3 further enhanced IFN-γ and IL-4 levels in sera induced by α-GalCer in vivo, whereas administration of control Ig did not (Fig. 3, A and B). However, 4-1BB engagement alone did not induce cytokines (Fig. 3. A and B), suggesting that 4-1BB engagement may costimulate NKT cells upon TCR stimulation.
To investigate which cell population produces these cytokines, we next conducted intracellular cytokine staining for splenocytes from BALB/c mice treated with α-GalCer and 3H3 or control Ig. The majority of cytokine-producing cells were CD1d-restricted NKT cells and their cytokine production was increased by 3H3 administration (Fig. 3 C). These data indicate that 4-1BB engagement increases the cytokine production of CD1d-restricted NKT cells in vivo.
Additionally, we performed assays for IL-4 and IFN-γ in the sera of CD1d-deficient mice. The sera from α-GalCer/3H3-treated CD1d-deficient mice showed significantly lower levels of IL-4 and IFN-γ than sera from α-GalCer/3H3-treated wild-type BALB/c mice (Fig. 3 D). The lack of cytokine increase by α-GalCer/3H3 in CD1d-deficient mice suggests that α-GalCer/3H3-activated NKT cells are required for increased cytokine levels in sera.
To determine whether 4-1BB directly elicits a costimulatory signal on NKT cells in vivo, we developed an adoptive transfer system. Control Ig or 3H3 was administered to irradiated wild-type mice. Two groups of irradiated mice received wild-type NKT cells; the others received 4-1BB-deficient NKT cells. NKT cells were transferred together with NKT-depleted, 4-1BB-deficient splenocytes to provide APCs expressing the CD1d molecule. On the following day the recipient mice were i.p. challenged with 0.2 μg of α-GalCer, sera were collected from the mice at the indicated time points, and IFN-γ and IL-4 levels were measured. The levels of IFN-γ and IL-4 in sera increased more in wild-type NKT-transferred recipients after the administration of 3H3 than of control Ig in vivo, whereas no such increase was noted in 4-1BB-deficient NKT-transferred recipients (Fig. 3,E). In addition, we assessed the patterns of cytokine production by purified NKT cells using the in vitro APC-free system, i.e., a plate was coated with a CD1d dimer and then with α-GalCer. NKT cells were purified from the spleen by MACS using anti-PE microbeads after the staining of α-GalCer-loaded CD1d dimers with anti-mouse IgG1-PE. When purified NKT cells were incubated in plate wells with an immobilized α-GalCer-loaded CD1d dimer, IL-4 and IFN-γ in the culture supernatant were higher in 3H3-treated than in control Ab-treated groups (Fig. 4, A and B). These data provide evidence that 4-1BB functions as a direct costimulatory molecule on NKT cells.
The exacerbation of NKT cell-induced AHR and lung inflammation after agonistic anti-4-1BB mAb administration was accompanied by up-regulation of Th1 and/or Th2 cytokines
Previous studies have shown that an agonistic anti-4-1BB mAb decreases conventional CD4+ T cell-driven airway responses such as AHR and lung inflammation (21, 22, 23). We questioned whether 4-1BB engagement could also inhibit NKT cell-induced AHR and lung inflammation by modulating the cytokine production of NKT cells. To confirm this, we treated BALB/c mice by administering 3H3 or rat IgG control Ig i.p. and intranasally with intranasal α-GalCer challenge. Mice treated with α-GalCer showed increased methacholine-induced AHR than did mice treated with vehicle. The severity of AHR was greatest in mice treated with α-GalCer plus 3H3 (Fig. 5,A). The enhancing effect of 4-1BB engagement was sustained 48 h after α-GalCer treatment, but the difference in AHR between the α-GalCer plus 3H3 groups and the α-GalCer plus control Ab group was not significant 72 h after α-GalCer treatment (Fig. 5,B). In addition, the number of inflammatory cells, including granulocytes and lymphocytes in the BAL, was increased by α-GalCer administration, and lung inflammation was also more severe in mice cotreated with 3H3 than in those cotreated with control Ab (Fig. 5 C).
In terms of Th1 and Th2 cytokine production, the current study showed that BAL IL-4, IL-13, and IFN-γ levels were higher in mice treated with α-GalCer than with vehicle. Moreover, the levels of these cytokines in BAL were more enhanced in mice cotreated with the agonistic anti-4-1BB mAb than in those cotreated with control Ab (Fig. 5 C). Taken together, these results suggest that 4-1BB engagement on NKT cells contributes to NKT cell-mediated airway responses, including AHR and lung inflammation, and that these effects may be mediated by up-regulation of cytokine production.
The agonistic anti-4-1BB mAb augmented α-GalCer-induced AHR via an IL-4Rα- but not IFN-γ-mediated pathway
When stimulated by glycolipid Ags, NKT cells produced large numbers of cytokines, thereby affecting innate and adaptive immune responses (3). Because 4-1BB engagement augmented the induction of IFN-γ and IL-4/IL-13 by NKT cells, we attempted to assess the contribution of these cytokines to the α-GalCer-induced increases in AHR and lung inflammation. To do so, we tested the NKT cell-mediated AHR and lung inflammation model in IFN-γ-deficient mice or IL-4Rα-deficient mice.
Although 4-1BB signaling on NKT cells also significantly increased IFN-γ production by TCR-stimulated NKT cells (Fig. 4,B), the absence of IFN-γ rather exacerbated the effects of agonistic 4-1BB mAb administration on the NKT cell-induced AHR and inflammatory cell accumulation by α-GalCer (Fig. 6, A and C). Further, significantly higher levels of IL-4, IL-5, and IL-13 were elicited in the BALF of IFN-γ-deficient mice than in that of naive mice by the administration of α-GalCer, and these results were independent of the coadministration of an agonistic anti-4-1BB mAb (Fig. 6 D).
We next investigated the relationship between IL-4/IL-13 signaling and the effects of an agonistic anti-4-1BB mAb. Because IL-4Rα is shared by receptors for IL-4 and IL-13, IL-4Rα-deficient mice are unresponsive to both IL-4 and IL-13. When α-GalCer was intranasally administrated, IL-4Rα-deficient mice did not develop AHR (Fig. 6,B). These results are consistent with previous studies using IL-4−/−, IL-13−/−, and IL-4−/−IL-13−/− mice (11). In addition, coadministration of an agonistic anti-4-1BB mAb in IL-4Rα-deficient mice did not aggravate α-GalCer-induced AHR (Fig. 6 B). These data suggest that α-GalCer-induced AHR is mediated by IL-4/IL-13 and that 4-1BB engagement increases AHR by enhancing the production of IL-4/IL-13 from NKT cells.
The expression levels and functions of 4-1BB are well understood in conventional T cells (12) but not in NKT cells. In this study, we demonstrate that 4-1BB is expressed on TCR-stimulated NKT cells but not on naive NKT cells and that TCR-stimulated NKT cells produced more cytokines in the presence of a costimulatory signal by 4-1BB engagement. These results indicate that 4-1BB functions as a costimulatory molecule on NKT cells. Also, we show that an agonistic anti-4-1BB mAb aggravates NKT cell-mediated AHR and inflammatory cell accumulation and enhances cytokine production in BALF. In addition, our results from the study using IFN-γ- or IL-4Rα-deficient mice show that the agonistic anti-4-1BB mAb augments α-GalCer-induced AHR via an IL-4Rα- but not an IFN-γ-mediated pathway.
Several reports have suggested that NKT cells play a prominent role in AHR and the inflammatory cell accumulation characteristic of allergic asthma (25). In CD1d-deficient mice and Jα18-deficient mice, both of which lack NKT cells, allergen-induced AHR failed to develop and inflammatory cells did not accumulate (7, 8). We have previously shown that the coadministration of exogenous protein Ag with α-GalCer in mice triggered allergen-induced AHR and inflammatory cell accumulation by increasing Th2 cell-associated sensitization to the Ag (9). In addition, even in the absence of conventional T cells NKT cells were able to induce AHR and inflammatory cell accumulation in mice (11). NKT cells were shown to be a major population of CD4+CD3+ cells in the lungs of patients with moderate to severe persistent asthma (26). Many costimulatory molecules have been shown to regulate the division and function of NKT cells (27, 28, 29) and thus are also thought to be critical to NKT cell-mediated airway responses. However, it has not yet been established whether costimulatory molecules on NKT cells participate in controlling NKT cell-mediated airway responses.
Our results revealed that 4-1BB is a costimulatory molecule of NKT cells and that it affects NKT cell-mediated AHR and inflammatory cell accumulation by regulating NKT cell activation. Other costimulatory molecules on NKT cells, such as CD28, GITR, and ICOS, could play a role similar to that of 4-1BB. The signals through these costimulatory molecules may commonly enable the activation of NKT cells with smaller doses of glycolipid Ags. Although TCR stimulation through α-GalCer is sufficient to activate NKT cells, the immunogenicity of natural glycolipid Ags could be lower than that of α-GalCer. Therefore, a costimulatory signal through 4-1BB could be more important in the stimulation of NKT cells by naturally occurring NKT cell ligands.
Directing agonistic mAbs to 4-1BB has been shown to regulate immune responses in several disease models, including cancers (15, 16), viral infection (30), and autoimmune diseases (19, 20). In one study, agonistic mAbs directed to 4-1BB abrogated T cell-dependent humoral immune responses through the induction of helper T cell anergy (18). Agonistic anti-4-1BB mAb treatment in mice was also demonstrated to inhibit conventional CD4+ T cell-driven airway responses such as AHR and inflammatory cell accumulation in the airway and lung (21, 22, 23). These results suggest the clinical promise of agonistic anti-4-1BB mAbs for allergic asthma therapy.
Because NKT cell activation induced AHR and inflammatory cell accumulation in the airway and lung by the same means as conventional CD4+ T cells, we questioned whether the agonistic anti-4-1BB mAb could also suppress these NKT cell-mediated airway responses. Unexpectedly, we found that the agonistic anti-4-1BB mAb exacerbated NKT cell-mediated AHR and inflammatory cell accumulation. Currently, the relative contribution of NKT cells, Th2 cells, and other T cells to the pathogenesis of asthma is not well defined in the clinical setting. Thus, our results strongly suggest that the use of agonistic anti-4-1BB mAbs for the treatment of asthma must be approached with caution.
Although our results suggest that agonistic anti-4-1BB mAb cross-linking exacerbates NKT cell-driven AHR and inflammatory cell infiltration, it is not clear whether the natural interaction between 4-1BB ligand and 4-1BB is involved in the pathogenesis of this disorder. In our α-GalCer-induced airway disorder model, the anti-4-1BB ligand blocking Ab did not reduce the severity of asthmatic symptoms (data not shown). However, this finding may be due to compensation by other costimulatory signals induced by TCR stimulation. Additional studies will be necessary to confirm this possibility.
Our study investigated only NKT cell-driven AHR and inflammatory cell infiltration. However, NKT cells may not only mediate direct induction of AHR and inflammatory cell infiltration but also induce or inhibit the development of Th2 cell-driven AHR and inflammatory cell infiltration (9, 31, 32). Because it was reported that the anti-4-1BB mAb abrogates the induction and development of OVA-induced asthma (18), we applied the model to Jα281−/− mice, which are deficient in a major population of invariant NKT cells, to see whether the effect of anti-4-1BB in the OVA-induced asthma model is due to the influence of anti-4-1BB on NKT cells. An agonistic anti-4-1BB mAb had similar effects on the inhibition of OVA-specific splenocyte proliferation, serum IgG and IgE levels, and cell infiltration into the lungs in both Jα281−/− and wild-type mice (data not shown). These data indicate that the suppressive effects of the anti-4-1BB mAb in the OVA-induced AHR model might be not due to the effects of the anti-4-1BB mAb on NKT cells. In addition, we are currently seeking to elucidate whether 4-1BB engagement on NKT cells in α-GalCer-treated mice can facilitate the modulation of Th2 cell-driven responses by NKT cells. Elucidating additional effects of 4-1BB engagement should lay the groundwork for the more effective use of agonistic anti-4-1BB mAbs as therapeutic agents in airway diseases such as allergic asthma.
In conclusion, we show here for the first time that 4-1BB engagement provides a costimulatory signal on NKT cells that aggravates NKT cell-mediated airway responses, including airway hyperreactivity and inflammatory cell infiltration. These results sound a cautionary note, suggesting that a more prudent approach is needed in the clinical application of anti-4-1BB mAbs.
We thank Dr. R. Mittler of Emory University (Atlanta, GA) for providing 3H3 hybridoma and Dr. Sanghee Kim of Seoul National University (Seoul, Korea) for providing α-GalCer.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by a Korea Research Foundation Grant from the Korean Government (Ministry of Education and Human Resource Development) (KRF-2006-312-E00093).
Abbreviations used in this paper: NKT, natural killer T; AHR, airway hyperresponsiveness; α-GalCer, α-galactosylceramide; BAL, bronchoalveolar lavage; BALF, BAL fluid.