IL-4-Secreting NKT Cells Prevent Hypersensitivity Pneumonitis by Suppressing IFN-γ-Producing Neutrophils¹

Hypersensitivity pneumonitis (HP)³ represents a group of related inflammatory interstitial lung diseases caused by the repeated inhalation of and consequent sensitization to a wide variety of environmental organic dusts, fungi, and molds (1, 2). The most common Ags are thermophilic actinomycetes, which cause farmer’s lung disease (1). Although the pathogenesis of HP is complex and not completely understood, it is currently accepted that a combination of humoral and cell-mediated immune responses contributes to the development of HP in humans and animals (1). Several immune cell types, including neutrophils, lymphocytes, monocytes and macrophages, are involved in the development of HP (1, 2). After Ag exposure, the number of alveolar macrophages is increased, and these cells secrete cytokines and chemokines such as TNF-α, IL-1, IL-8, and MIP-1α, which attract neutrophils and lymphocytes (3, 4, 5). IL-8 acts as a chemotactic factor for neutrophils and T cells, and its levels in bronchoalveolar lavage fluid (BALF) from patients with HP are correlated with the number of neutrophils and response to treatment (4). Therefore, immediately after Ag challenge, neutrophils infiltrate the alveoli and are followed by an influx of activated T cells with a preponderance of CD8⁺ T cells, which are considered critical cell components for the development of HP (6). Several studies have suggested that animal models of experimental hypersensitivity pneumonitis appear to be mediated by Th1 immune responses. The differences in the expression of Th2 cytokines, such as IL-4, between C57BL/6 and DBA/2 mice establishes a bias for Th1 vs Th2 adaptive immune responses in these mice, and contribute to determining their susceptibility and resistance to HP (7). IL-12, which is produced by activated alveolar macrophages in HP, induces the shift from Th0 to Th1 cells, which in turn aggravates HP (6, 8). In addition, when the Th1/Th2 balance shift toward a Th1 response, the HP is more severe, while it is attenuated by Th2-biased immune responses (9, 10, 11, 12). Moreover, in a murine model, the adoptive transfer of Th1 clones induced reversible HP (13, 14). These observations suggest that the development of HP in animal models is determined by the modulation of Th1/Th2 immune responses in vivo.

NKT cells are a distinct subset of conventional αβ T cells and are characterized by the coexpression of surface markers of both αβ T cells and NK cells (15). NKT cells express intermediate levels of a semi-invariant Vα14-Jα18 TCR in mice or an invariant Vα24-Jα15 TCR in humans (16), which recognize the glycolipid Ags presented by nonpolymorphic MHC class I-like protein CD1d (17). NKT cells contribute to regulating various immune responses in vivo, including the maintenance of self-tolerance, autoimmune diseases (18, 19, 20), tumor rejection (21), pulmonary fibrosis (22), and the response to various infectious agents (23, 24). Upon activation, NKT cells rapidly produce large amounts of IL-4 and IFN-γ (25), which are involved in the regulation of innate and adaptive immune responses by NKT cells (15). Therefore, it has been established that NKT cells regulate immune responses by modulating the Th1/Th2 balance in vivo by producing IL-4 and IFN-γ. Based on these regulatory functions of NKT cells in vivo, we postulated that NKT cells regulate hypersensitivity pneumonia by modulating Th1/Th2 immune responses in lung tissues.

Therefore, in this study, we investigated whether NKT cells play functional roles in the development of murine hypersensitivity pneumonitis by modulating Th1/Th2 immune responses.

C57BL/6 mice were purchased from Orient Company. CD1d^−/− mice (C57BL/6 background) were a gift from National Institute of Allergy and Infectious Diseases-Taconic facility (Dr. H. Gu, Columbia University, New York, NY). Jα18^−/− (B6 background) were a gift from Dr. M. Taniguchi (Chiba University, Chiba, Japan). IFN-γ^−/− and IL-4^−/− mice were purchased from The Jackson Laboratory. These mice were bred and maintained under specific pathogen-free conditions at the Clinical Research Institute of Seoul National University Hospital. All animal experiments were approved from the Institutional Animal Care and Use Committee of the Clinical Research Institute of Seoul National University Hospital.

Saccharopolyspora rectivirgula (SR) Ag was prepared from a strain of SR obtained from the American Type Culture Collection (catalog no. 29034), which was grown in a trypticase soy broth in a 55°C shaking incubator for 4 days, centrifuged, and rinsed with distilled water three times. Next, the SR Ag was homogenized and lyophilized. The SR Ag was resuspended in pyrogen-free saline. The SR Ag contained <20 ng/mg endotoxin, estimated using a limulus amebocyte lysate assay (Sigma-Aldrich).

HP was induced by intranasally instilling 150 μg of SR Ag in saline into mice under light anesthesia. This procedure was performed on 3 consecutive days per week for 3 wk. The mice were sacrificed with a pentobarbital injection 4 days after the final treatment. For histological examination, paraffin-embed entire lungs were cut and stained with H&E. Pathologic scores were defined as follows: 0, no lung abnormality; 1–5, the presence of inflammation and granulomas involving 10, 10–30, 30–50, 50–80, and >80% of the lung, respectively.

The trachea was cannulated, and the lung was lavaged five times with 0.7 ml of cold PBS. The BALF was centrifuged at 1500 rpm for 10 min at 4°C, and the supernatant was removed. The total BALF cells were counted using a hemocytometer, and incubated for 15 min on ice with FcγRII/III blockade. After washing, the cells were stained in a 200-μl total volume with a 1-μg combination of the following mAbs: anti-CD8, -CD4, -CD69, -NK1.1, and -TCR-β mAbs, which were purchased from BD Pharmingen.

The total hydroxyproline level of the lung was measured 3 wk after the intranasal SR administration as described previously (22).

The CFSE (Molecular Probes)-labeled BALF cells (1 × 10⁶/well) were stimulated with SR (10 μg/ml) in 12-well plates for the indicated periods. The amounts of CFSE were determined on gated CD4⁺ or CD8⁺ cells using flow cytometry.

The liver mononuclear cells were isolated (20) and were stained with PE-conjugated anti-NK1.1 (BD Pharmingen) and Cy-conjugated anti-TCR-β (BD Pharmingen). The stained cells were then sorted with a FACStar^Plus using CellQuest software, and the purity of the sorted cells was >95%. Then 3 × 10⁵ NKT cells per mouse were adoptively transferred via i.v. injection.

For quantitative real-time PCR, the lungs were homogenized using a Polytron and Tri-reagent, and total RNA was isolated. Five micrograms RNA was reverse transcribed with Moloney murine leukemia virus reverse transcriptase (Koschem), and PCR was performed. The following primers and probe were synthesized by Applied Biosystems: GAPDH (TagMan predeveloped Assay Reagent: 4352339E), IFN-γ (TaqMan predeveloped Assay Reagent: 4339850F): TGF-β1, GCAACATGTGGAACTCTACCAGAA (forward), GACGTCAAAAGACAGCC ACTCA (reverse), and ACCTTGGTAACCGGCTGCTGACCCTAMRA; IL-4, FAM-CTCCGTGCATGGGGTCCCTTC-Black Hole Quencher; and IL-12, 5′-FAM-TTCAACATCAAGAGCAGTAGCAGTTCCCCT-Black Hole Quencher (BioSource International). The results for each cytokine were normalized with respect to GAPDH expression.

BALF cells were taken from the lungs of B6 and CD1d^−/− mice administered SR Ag 3 wk after the SR Ag administration, and cultured with filtered SR Ag. The amounts of TNF-α (BioSource International) in the culture supernatant were determined using ELISA. Blood collected from the orbital sinus was centrifuged, and the serum was analyzed for SR-specific IgG. SR-specific IgG in serum (dilution 1/100) and BALF was measured using ELISA. In addition, spleen cells were taken from B6 and CD1d^−/− mice administered SR Ag 3 wk after the SR Ag administration, and cultured with filtered SR Ag. SR-specific immune cell proliferation was determined using MTS (3-(4.5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium, inner salt) assay kits (Promega).

For intracellular cytokine staining in NK, CD4⁺, and CD8⁺ cells, BALF cells were isolated from mice administered SR Ag in the first week after Ag exposure, and incubated overnight with lyophilized SR (10 μg) and Con A (10 μg). The cells were surface stained with mAb specific for NK1.1, TCR-β, CD4, and CD8, and fixed and permeabilized with Cytofix/Cytoperm according to the manufacturer’s instructions. These cells were stained using PE-conjugated anti-IFN-γ mAb. For intracellular staining of IFN-γ in neutrophils, freshly isolated BALF cells were incubated for 2 h with GolgiPlug and then stained with mAb against Gr-1 or Ly-6G, and intracellular IFN-γ. To deplete Gr-1⁺ granulocytes, the mice were given 200 μg of anti-Gr-1 (RB6-8C5; BD Pharmingen) or control mAb (BD Pharmingen) i.v. 1 day before SR Ag exposure (150 μg).

The hybridomas for anti-IFN-γ mAb (clone: R4-6A2) were purchased from the American Tissue Cell Culture. The mAbs were purified using a protein G column (Amersham Biosciences). Five hundred micrograms of each mAb was given i.v. 1 day before the administering the SR Ag. Recombinant murine IFN-γ was purchased from R&D Systems and diluted in PBS with 1% bovine albumin. Mice were given 1000 U i.p. in 0.2 ml of diluent 24 h before the first SR Ag exposure each week.

Statistical significance was analyzed using the program Prism3.0. Student’s t tests were used to determine the p value for comparing two groups. Values of p < 0.05 were considered significant. For the hydroxyproline content in the lungs, one-way ANOVA was performed.

To investigate whether NKT cells are functionally involved in the development of HP, we administered SR Ag intranasally to B6 and NKT cell-deficient, CD1d^−/− mice (26) on 3 consecutive days per wk for 3 wk. The mice were sacrificed 4 days after the last Ag exposure to evaluate the histological alterations in the lung, and measure the levels of hydroxyproline. SR Ag-induced multifocal pulmonary lesions with peribronchial hyperplasia of lymphoid follicles, granuloma, and fibrosis in B6 and CD1d^−/− mice (Fig. 1,A). However, CD1d^−/− mice showed more extensive histological alterations in the lung parenchyma in SR-induced HP compared with B6 mice. The histological scorings of the pulmonary lesions revealed significant aggravation of HP in CD1d^−/− mice compared with B6 mice (Fig. 1,B). Moreover, the amount of hydroxyproline (a collagen component) in the lungs was significantly higher in CD1d^−/− mice than in B6 mice (Fig. 1 C). These findings indicate that the SR Ag-induced HP is significantly aggravated in NKT cell-deficient mice compared with B6 mice.

To explore infiltrating immune cells in the lung tissues of B6 and CD1d^−/− mice during SR-induced HP, the numbers of total cells, B cells, T cells, macrophages and polymorphonuclear leukocytes were counted using BALF 7 days after the first SR Ag administration. The numbers of total cells and each subset in BALF were greater in CD1d^−/− mice than in B6 mice in the SR-induced HP model (Fig. 1,D), suggesting that more immune cells were recruited into the lung tissues in CD1d^−/− mice during SR-induced HP. Clinically, the lung involvement in HP is characterized by a lymphocytic alveolitis with an increase in both the percentage and absolute number of CD8⁺ T cells in the BALF (27). Therefore, the percentages of CD4⁺ and CD8⁺ T cells in the BALF from B6 and CD1d^−/− mice were estimated during SR-induced HP. The percentage of CD8⁺ T cells in the BALF of CD1d^−/− mice was higher than for B6 mice in SR-induced HP, whereas the percentage of CD4⁺ T cells in BALF of CD1d^−/− mice was similar to that of B6 mice. Therefore, the CD4:CD8 T cell ratio was significantly reduced in the BALF of CD1d^−/− mice compared with B6 mice in SR-induced HP (Fig. 1,E). These findings suggest that NKT cells contribute to regulating the influx or proliferation of CD8⁺ T cells in BALF during SR-induced HP. The immunological responses to SR Ag in SR-induced HP were assessed by measuring SR-specific immune cell proliferation, and IgG in serum and BALF in CD1d^−/−, Jα18^−/−, and B6 mice. CD1d^−/− and Jα18^−/− mice given SR Ag showed more SR-specific immune cell proliferation, and IgG production in sera and BALF than B6 mice given SR-Ag (Fig. 1, F–H). CD1d^−/− mice lack CD1d-dependent NKT cells, and Jα18^−/− mice lack NKT cells (iNKT) expressing invariant Vα14Jα18 TCR in vivo (21, 26, 28). Combined, these findings indicate that CD1d-dependent NKT and iNKT cell-deficient mice develop more severe clinical processes, and higher SR-specific immune responses in the SR-induced HP model than B6 mice.

To estimate the kinetics of infiltrating NKT cells in the lung tissues during SR-induced HP, the numbers of total and NKT cells in BALF were counted in B6 mice at various times. The kinetic study revealed that the total numbers of immune cells were increased 2 days after the nasal administration of SR Ag, and peaked on day 5 in B6 mice in the SR-induced HP model (Fig. 2,A). The kinetics of α-Galactosylceramide/CD1d-positive NKT cell infiltration in the lungs of B6 mice was similar to those of total cells in BALF during SR-induced HP (Fig. 2,B), which showed similar kinetics of NK1.1⁺TCR-β⁺ NKT cells in the lung (data not shown). These findings suggest that NKT cells infiltrate the lung tissues and play critical roles in an early phase rather than in a later phase of SR-induced HP. To address this suggestion, sorted NKT cells from B6 mice were adoptively transferred into CD1d^−/− mice in the first (−1 day), second (8 days), or third (15 days after the first Ag administration) week, and the SR-specific IgG in sera and BALF, SR-specific immune cell proliferation, and hydroxyproline content in the lung tissues were measured in these mice. Consistent with the kinetics of NKT cell infiltration in the lung, the adoptive transfer of NKT cells in the first week into CD1d^−/− mice significantly reduced the SR-specific IgG and immune cell proliferation, and hydroxyproline content in the lungs of CD1d^−/− mice administered SR Ag (Fig. 2, C–F). In contrast, the adoptive transfer of NKT cells in the second or third week into CD1d^−/− mice caused a less significant reduction of the SR-specific IgG levels and immune cell proliferation, and hydroxyproline content in the lungs of CD1d^−/− mice administered SR Ag nasally. The production of TNF-α in the lung tissues is correlated with the severity of HP and recruitment of inflammatory cells into the lung (29). Therefore, we measured the amount of TNF-α produced by SR-stimulated BALF cells from CD1d^−/− mice given NKT cells at various time points in the SR-induced HP model. Consistent with the results of the SR-specific immune responses, the production of TNF-α was significantly reduced in the BALF of CD1d^−/− mice given NKT cells in the first week, whereas the adoptive transfer of NKT cells in the second or third week did not significantly reduce the production of TNF-α from the BALF cells of CD1d^−/− mice (Fig. 2 G). These findings indicate that adoptively transferred NKT cells in CD1d^−/− mice attenuate the SR-induced HP at an early time point, rather than during the late phase of the SR-induced HP model.

To evaluate the mechanism by which NKT cells contribute to attenuating SR-induced HP, we used real-time PCR to measure the levels of cytokines IL-4, IL-12, IFN-γ, and TGF-β in lung tissues of B6 and CD1d^−/− mice on days 7, 14, and 21. The IL-4 level in the lungs was lower in CD1d^−/− mice than that in B6 mice on day 7 in the SR-induced HP model, whereas the levels of IFN-γ, IL-12, and TGF-β in the lungs were higher in CD1d^−/− mice than in the lungs of B6 mice (Fig. 3,A). The pattern of cytokine production in the lungs of B6 and CD1d^−/− mice on day 7 was most prominent, but was similar to those from these mice on days 14 and 21 (data not shown). Furthermore, the adoptive transfer of NKT cells into CD1d^−/− mice restored the levels of IL-4, IFN-γ, IL-12, and TGF-β in the lung tissues of CD1d^−/− mice, much like B6 mice in the SR-induced HP model (Fig. 3,A). Consistent with the high histological scores and amount of hydroxyproline in the lungs, CD1d^−/− mice produced high levels of TGF-β in the lungs during SR-induced HP, and this appeared to be related to the enhanced fibrosis seen in these mice. Based on the cytokine pattern of IL-4, IFN-γ, and IL-12, it has been suggested that CD1d^−/− mice produce high levels of Th1 cytokines and low levels of Th2 cytokines in the lungs in SR-induced HP, and that the Th1 immune response in the lungs aggravates HP compared with B6 mice. Moreover, it has been reported that the response of the lung to intranasal SR Ag is dependent on IFN-γ, which is a potent cytokine for deriving Th1 immune responses in vivo (9). Therefore, we postulated that IFN-γ is a cytokine that regulates SR-induced HP in NKT cell-deficient mice. To establish a functional link between IFN-γ and SR-induced HP in NKT cell-deficient mice, neutralizing anti-IFN-γ Ab or rIFN-γ was injected into B6 and CD1d^−/− mice. IFN-γ blockade in B6 and CD1d^−/− mice significantly reduced SR-specific IgG production in BALF and immune cell proliferation in the SR-induced HP model, which was more prominent in CD1d^−/− mice than in B6 mice (Fig. 3, B and C). The rIFN-γ administered to B6 and CD1d^−/− mice enhanced SR-specific IgG production and immune cell proliferation compared with untreated B6 and CD1d^−/− mice in the SR-induced HP model (Fig. 3, B and C). These findings revealed that the regulatory effects of NKT cells on SR-induced HP depend on the amount of IFN-γ produced in lung tissues.

To analyze which subset of immune cells in BALF contributes to producing the high level of IFN-γ in the lung tissues of CD1d^−/− mice in SR-induced HP, we estimated the levels of IFN-γ in CD4⁺ and CD8⁺ T cells, NK cells, and Gr-1⁺ granulocytes in BALF from B6 and CD1d^−/− mice using intracellular staining. Flow cytometric analysis revealed that CD4⁺ and CD8⁺ T cells, and NK cells in BALF from CD1d^−/− mice produced low levels of IFN-γ in the SR-induced HP model, which was similar to that in B6 mice. In contrast, a large proportion (62.7%) of the Gr-1⁺ granulocytes in BALF produced IFN-γ in B6 mice with SR-induced HP, which was less than that in CD1d^−/− mice (83.5%) in the SR-induced HP model (Fig. 3,D). Gr-1 is expressed on the surface of both granulocytes (neutrophils and eosinophils) and monocytes (30). Therefore, to characterize the IFN-γ-producing cells in SR-induced HP model in detail, we analyzed the expression pattern of Ly-6G, a specific neutrophils marker, on IFN-γ-producing cells in BALF. The pattern of Ly-6G expression was similar to that of Gr-1 on IFN-γ-producing cells in BALF from B6 and CD1d^−/− mice in SR-induced HP, suggesting that most of the Gr-1⁺ and IFN-γ-producing cells were Ly-6G-expressing neutrophils in BALF (Fig. 3,D). Moreover, the adoptive transfer of NKT cells reduced the expression of IFN-γ in Gr-1⁺ granulocytes and Ly-6G⁺ neutrophils in BALF from CD1d^−/− mice, much like B6 mice in the SR-induced HP model (Fig. 3 D). These findings suggest that CD1d^−/− mice enhance IFN-γ production in Gr-1⁺ neutrophils in BALF in the SR-induced HP model, which contributes to aggravating HP in CD1d^−/− mice.

In CD1d^−/− mice with SR-induced HP, high IFN-γ production in neutrophils and increased numbers of CD8⁺ T cells in the lung tissues were prominent compared with B6 mice. Therefore, we postulated that the proliferation of CD8⁺ T cells in BALF might be functionally linked to IFN-γ produced by neutrophils in the lungs in the SR-induced HP model, which is tightly regulated by NKT cells. To address this issue, we administered a neutralizing mAb against IFN-γ or rIFN-γ to B6 or CD1d^−/− mice in the SR-induced HP model and analyzed the numbers and percentages of CD4⁺ and CD8⁺ T cells in BALF. The numbers of total cells in BALF were not significantly altered in B6 or CD1d^−/− mice by IFN-γ blockade, whereas it was slightly increased by the administration of rIFN-γ in these mice on day 7 during the SR-induced HP model (data not shown). IFN-γ blockade in B6 and CD1d^−/− mice increased the CD4:CD8 T cell ratio in BALF and reduced the expression of CD69 on CD8⁺ T cells in the SR-induced HP model, whereas the rIFN-γ administered to B6 and CD1d^−/− mice significantly reduced the CD4:CD8 T cell ratio and enhanced the expression of CD69 on CD8⁺ T cells (Fig. 4, A and B). In contrast, IFN-γ blockade or the administration of rIFN-γ did not alter the percentages or expression levels of CD69 on CD4⁺ T cells in BALF from B6 or CD1d^−/− mice during SR-induced HP. These findings suggest that IFN-γ activates and increases the percentages of CD8⁺ T cells in the lungs during SR-induced HP. To determine the contribution of Gr-1⁺ neutrophils to increasing CD8⁺ T cells in BALF during SR-induced HP, we depleted Gr-1⁺ granulocytes in B6 and CD1d^−/− mice using anti-Gr-1 mAb. The Gr-1⁺ granulocyte-depleted B6 and CD1d^−/− mice had fewer CD8⁺ T cells in BALF during SR-induced HP compared with B6 and CD1d^−/− mice treated with isotype-matched control mAb (Fig. 4 A), which was more prominent in CD1d^−/− mice than in B6 mice. These findings suggest that IFN-γ-producing Gr-1⁺ neutrophils increase the activation and percentage of CD8⁺ T cells in BALF from CD1d^−/− and B6 mice in the SR-induced HP model, which contributes to the low CD4:CD8 ratio in the BALF of CD1d^−/− mice.

Upon activation, NKT cells rapidly produce large amounts of IL-4 and IFN-γ (25), which play critical roles in regulating immune responses by modulating Th1/Th2 balance in vivo (15). Therefore, to explore which cytokine produced from NKT cells contributes to the development of SR-induced HP, we adoptively transferred NKT cells sorted from B6, IFN-γ^−/− or IL-4^−/− mice into CD1d^−/− mice, and measured SR-specific immune responses and the levels of cytokines in the lung tissues. The adoptive transfer of NKT cells from IFN-γ^−/− mice reduced SR-specific IgG in the sera and immune cell proliferation in CD1d^−/− mice, as shown by NKT cells from B6 mice (Fig. 5, A and B). Moreover, the levels of hydroxyproline, IFN-γ, IL-12, and TGF-β were reduced in the lung tissues from CD1d^−/− mice in the SR-induced HP model by the adoptive transfer of B6 or IFN-γ^−/− mouse NKT cells, whereas the level of IL-4 increased in the lung tissues of these mice (Fig. 5, C–G). In contrast, the adoptive transfer of IL-4^−/− mouse NKT cells did not alter the SR-specific IgG in sera or immune cell proliferation, and the levels of hydroxyproline, IFN-γ, IL-4, IL-12, and TGF-β production in the lung tissues of CD1d^−/− mice (Fig. 5, A–G). These findings suggest that IL-4 secreted by NKT cells contributes to attenuating SR-induced HP by suppressing Th1 immune responses in the lung tissues. Next, we investigated whether IL-4 produced by NKT cells regulates IFN-γ-producing neutrophils, and the activation and proliferation of CD8⁺ T cells in the SR-specific HP model. The adoptive transfer of B6 or IFN-γ^−/− mouse NKT cells reduced IFN-γ-producing Gr-1⁺ neutrophils in BALF from CD1d^−/− mice in the SR-induced HP model (Fig. 6 A). In contrast, the adoptive transfer of IL-4^−/− mouse NKT cells did not reduce IFN-γ-producing Gr-1⁺ neutrophils in BALF from CD1d^−/− mice in the SR-induced HP model. These findings indicate that IL-4-producing NKT cells suppress IFN-γ-producing neutrophils in the lung tissues during SR-induced HP.

The adoptive transfer of B6 or IFN-γ^−/− mouse NKT cells reduced the expression of CD69 on CD8⁺ T cells and increased the CD4:CD8 ratio of BALF in CD1d^−/− mice, whereas the adoptive transfer of IL-4^−/− mouse NKT cells into CD1d^−/− mice did not affect the activation of CD8⁺ T cells, or the CD4:CD8 T cell ratio (Fig. 6, B and C). Moreover, SR-induced proliferation of CD8⁺ T cells, as estimated by CFSE labeling, was enhanced in the BALF of CD1d^−/− mice compared with BALF of B6 mice, whereas the SR-induced proliferation of CD4⁺ T cells in the BALF from CD1d^−/− mice was similar to that in B6 mice (Fig. 6 D). These findings suggest that CD8⁺ T cells proliferated more in the lung tissues during SR-induced HP in CD1d^−/− mice than in B6 mice. In addition, IL-4-deficient-NKT cells could not reduce the number of neutrophils in BALF, whereas B6 or IFN-γ-deficient NKT cells inhibited the reflux of neutrophils into the lung tissues in the HP model (data not shown). Taken together, IL-4 produced by NKT cells suppresses IFN-γ-producing Gr-1⁺ neutrophils, which regulate the activation and proliferation of CD8⁺ T cells in the lung tissues, and in turn attenuate SR-induced HP in mice.

In clinical cases, the presence of serum precipitating Abs or cell-mediated hypersensitivity against presumptive Ags may be helpful for the diagnosis of patients with HP (31, 32, 33). Moreover, in several studies using the SR-induced murine HP model, SR-specific serum IgG or IL-2 production of BAL cells was measured to assess SR-induced HP (34, 35). Therefore, to determine whether mice given SR-Ag develop HP or not, SR-specific IgG and immune cell proliferation in these mice were evaluated. However, these Ab titer and cell-mediated responses do not have strong correlation with disease severity. Alternatively, we analyzed the histology and measured amounts of hydroxyproline of the lung to assess the severity of HP in these mice. In our study, the deficiency of NKT cells in B6 mice aggravates SR-induced HP, as estimated from the amounts of SR-specific IgG in serum and BALF, SR-specific immune cell proliferation, histological analysis, and the amounts of hydroxyproline in the lung tissues. These findings suggest that NKT cells contribute to attenuating SR-induced HP. The critical roles of NKT cells during the development of HP were further substantiated by the fact that the adoptive transfer of B6 mouse NKT cells attenuated HP in CD1d^−/− mice. During the development of SR-induced HP, NKT cells infiltrated the lung tissues of B6 mice 3–7 days after administering the SR Ags. The adoptive transfer of NKT cells in the first week attenuated SR-induced HP more in CD1d^−/− mice compared with CD1d^−/− mice given NKT cells in the second or third week. These findings indicate that NKT cells play a suppressive role in SR-induced HP at an early phase rather than during the later phase of HP.

Several studies have demonstrated that IFN-γ production in the lung make a critical contribution to the development of murine HP (9, 36). After exposure to SR Ag, IFN-γ^−/− mice with the BALB/c background developed minimal inflammation and no granulomas in the lung tissues, whereas wild-type (WT) mice showed marked granulomatous inflammation with increased BALF cells (9). Similar results were demonstrated in B6 mice administered neutralizing mAb against IFN-γ following SR Ag exposure (36). These findings suggest that IFN-γ-producing cells in the lung tissues are involved in the development of SR-induced HP. Of the cytokines examined in the lungs in the SR-induced HP model, IFN-γ production was significantly elevated in the lungs of CD1d^−/− mice compared with B6 mice. The adoptive transfer of NKT cells into CD1d^−/− mice reduced the level of IFN-γ in the lungs, and IFN-γ blockade in B6 or CD1d^−/− mice reduced the severity of HP, which was more prominent in CD1d^−/− mice than that in B6 mice. These findings indicate that the elevated IFN-γ production in the lung tissues contributes to aggravating HP in CD1d^−/− mice, which is tightly regulated by NKT cells in the lung.

In the SR-induced HP model, neutrophils are most prominent in BALF immediately after Ag exposure, which corresponds clinically to the phase of acute symptoms. These findings suggest that neutrophils play a critical role in the early development of HP (6, 8). The neutrophilic alveolitis may be stimulated by the formation of immune complexes, the direct activation of the complement by the alternative pathway, or the endotoxin effect of inhaled Ag (37). However, the functional roles of neutrophils in the SR-induced HP model remain poorly characterized. Recently, Nance et al. (38) suggested that the majority of IFN-γ producing cells in the lung were Gr-1^high-neutrophils in the SR-induced HP model. However, the mechanisms by which IFN-γ-producing neutrophils in the lung tissues are regulated during the development of HP remains unclear. Our study demonstrated that the level of IFN-γ produced by Gr-1⁺ neutrophils was higher in the lung tissues of CD1d^−/− mice than in B6 mice in the SR-induced HP model, whereas the expression levels of IFN-γ in CD4⁺ and CD8⁺ T cells, and NK cells of BALF from CD1d^−/− mice were low, and similar to those of BALF from B6 mice. Moreover, the adoptive transfer of WT NKT reduced the level of IFN-γ in Gr-1⁺ or Ly-6G⁺ neutrophils of BALF from CD1d^−/− mice in the SR-induced HP model. The percentages of IFN-γ-producing Gr-1⁺ granulocytes from the lungs of B6 and CD1d^−/− mice were similar to those of IFN-γ-producing granulocytes expressing Ly-6G, a specific marker for neutrophils. Based on these results, it is reasonable to consider the Gr-1⁺ granulocytes producing IFN-γ in our experiments as neutrophils rather than macrophages or monocytes. These findings demonstrated that NKT cells suppress IFN-γ-producing Gr-1⁺ neutrophils in the lung tissues of the SR-induced HP model, suggesting that NKT cells bridge innate and adaptive immunity by regulating IFN-γ-producing neutrophils in SR-induced HP. Very recently, Yasunami et al. (39) reported a bridging role for NKT cells between innate and adaptive immunity in an islet cell transplantation model. Unlike the suppressive role of NKT cells in IFN-γ-producing neutrophils in SR-induced HP, Vα14 NKT cells triggered IFN-γ production by Gr-1⁺CD11b⁺ granulocytes, which contributed to mediating immediate graft failure and the early loss of islet grafts (39). In addition, IFN-γ secreted by NKT cells appeared to contribute to triggering IFN-γ-producing Gr-1⁺CD11b⁺ granulocytes, and NKT cells did not affect the influx of neutrophils in the islet cell transplantation model, whereas IL-4 produced by NKT cells appeared to suppress the influx of neutrophils into the lungs and IFN-γ production by neutrophils in the SR-induced HP model. However, it is unclear why NKT cells have opposing effects on IFN-γ production by neutrophils in different animal models of disease. Therefore, we speculated that the regulatory effects of NKT cells on IFN-γ-producing granulocytes are diverse and depend on the cytokine profiles secreted by NKT cells or the immunological microenvironment of different murine model systems.

In humans, the lung involvement in HP is characterized by a lymphocytic alveolitis with a preponderance of CD8⁺ T cells in the BALF, these cells express activation markers, including IL-2R, VLA-1, and HLA-DR, and the dominant TCR-Vβ segments (40, 41, 42). The analysis of BAL cells in the SR-induced HP model revealed that there were more CD8⁺ T cells in BALF in CD1d^−/− mice than in B6 mice, resulting in a low CD4:CD8 T cell ratio in BALF of CD1d^−/− mice, which was restored by the adoptive transfer of WT NKT cells to CD1d^−/− mice. The levels of CD69 expression and SR-induced proliferation in CD8⁺ T cells were higher in CD1d^−/− mice than in B6 mice. These findings suggest that more CD8⁺ T cells are activated in CD1d^−/− mice, and they proliferate in the lung tissues during SR-induced HP compared with B6 mice. However, the possibility that the increased migration of CD8⁺ T cells into the lung during the development of SR-induced HP contributes to the increased CD8⁺ T cells in BALF should be ruled out. Moreover, IFN-γ blockade in B6 or CD1d^−/− mice increased the CD4:CD8 T cell ratio in BALF and reduced the expression levels of CD69 on CD8⁺ T cells in the SR-induced HP model, whereas the administration of rIFN-γ to B6 or CD1d^−/− mice significantly reduced the CD4:CD8 T cell ratio, and enhanced the levels of CD69 expression on CD8⁺ T cells. In addition, the depletion of Gr-1⁺ granulocytes in B6 and CD1d^−/− mice reduced the number of CD8⁺ T cells in BALF in the SR-induced HP model, resulting in an increase in the CD4:CD8 ratio. These findings indicate that the activation and proliferation of CD8⁺ T cells in lung tissues are regulated by IFN-γ producing Gr-1⁺ neutrophils. Combined, NKT cells suppress the activation and proliferation of CD8⁺ T cells by regulating IFN-γ-producing neutrophils in SR-induced HP. Although the precise functions of CD8⁺ T cells in the development of HP are unknown, it has been suggested that CD8⁺ T cells and activated macrophages produce MIP-1α, which induces granuloma formation by facilitating the differentiation of alveolar macrophages into epithelioid cells and multinuclear giant cells (31, 43). In addition, IL-8 and MCP-1 are chemotactic for Ag-sensitized CD8⁺ T cells and, to lesser extent, CD4⁺ T cells in the lungs (4, 44). Therefore, it is conceivable that several chemokines produced by activated macrophages recruit CD8⁺ T cells into the lung tissues and are activated and proliferate due to the effects of IFN-γ-producing Gr1⁺ neutrophils, and consequently induce granuloma formation in SR-induced HP, which is regulated by NKT cells. In tumor surveillance in vivo, CD4⁺ NKT cells suppress CD8⁺ T cell-mediated tumor rejection via the IL-4Rα-STAT6 pathway by producing IL-13 (45). In contrast, NKT cells enhanced the activation and expansion of diabetogenic CD8⁺ T cells, and their differentiation into effector cells producing IFN-γ, exacerbating the development of diabetes (46). Combined with our results, this suggests that NKT cells regulate immune responses in vivo by enhancing or suppressing the activation and proliferation of CD8⁺ T cells.

In our study, several lines of evidence suggest the critical roles of IL-4 secreted by NKT cells in suppressing the SR-induced HP. First, the level of IL-4 transcription was lower in the lungs of CD1d^−/− mice vs the B6 mice in SR-induced HP. Second, the adoptive transfer of WT or IFN-γ-deficient-NKT cells attenuated HP by suppressing SR-specific IgG production in serum and the proliferation of CD8⁺ T cells and Th1 type (IFN-γ and IL-12) cytokine production in the lung tissues, whereas IL-4-deficient NKT cells could not reduce the number of neutrophils (data not shown) and the level of IFN-γ secreted by neutrophils in BAL cells of CD1d^−/− mice, resulting in no change in the SR-induced HP. Taken together, these findings indicate that IL-4 produced by NKT cells critically contributes to attenuating HP by reducing the recruitment of neutrophils to the lungs and suppressing their IFN-γ production, which suppresses the activation and proliferation of CD8⁺ T cells, and regulates Th1/Th2 immune responses to SR Ag in the lung. Several studies have demonstrated that IL-4 suppresses the neutrophil influx into the lung tissues in Ab-induced glomerulonephritis, and into joint tissues in an arthritis model (47, 48). Moreover, IL-4 suppresses the ex vivo phagocytic activation of human neutrophils in response to IFN-γ or TNF-α (48). These findings partly support our result that IL-4 secreted by NKT cells contributes to attenuating HP by inhibiting the recruitment of neutrophils into the lung tissues and suppressing IFN-γ production by neutrophils.

The most important factor in treating patients with HP is avoiding Ag (2). However, the small percentage of patients who develop advanced pulmonary fibrosis may have continued progression of their disease, although they completely avoid re-exposure to the causative Ags (31). Therefore, it is recommended that patients with severe or progressive chronic HP be treated with a trial of corticosteroids, which are nonspecific immunosuppressive agents. To suppress Ag-specific immune responses, a novel therapeutic approach should be developed for managing HP. In an animal model, the blockade of T cell costimulation by CTLA4-Ig inhibits lung inflammation in murine HP, providing a possible approach for the treatment of HP (34). Moreover, in previous work, we demonstrate that NKT cells activated by glucocorticoid-induced TNFR engagement protect SR-induced hypersensitivity pneumonitis by reducing SR-specific humoral and T cell responses in vivo (49). Therefore, we propose that NKT cells be considered for developing therapeutic approaches based on their suppressive effects on HP.

In conclusion, IL-4-producing NKT cells attenuate hypersensitivity pneumonitis by suppressing the IFN-γ-producing neutrophils, which enhance the activation and proliferation of CD8⁺ T cells in the lung. We demonstrated that NKT cells play critical roles in attenuating HP by functionally bridging the innate and adaptive immune responses.

We thank Dr. Seong Hoe Park for his support of this project, Drs. Kyeong Cheon Jung and Eun Young Choi for helpful comments on the manuscript, and all members of the Department of Experimental Animals at the Clinical Research Institute of Seoul National University Hospital for animal management.

The authors have no financial conflict of interest.