Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

The respiratory tract is one of the representative mucosal tissues, and is repeatedly exposed to a broad array of airborne foreign antigens (Ags) (1). Inhalation of excessive amounts of Ags would evoke deleterious immune and inflammatory responses in the airway, as in bronchial asthma (2). Evaluation of intrinsic autoregulatory mechanisms by which harmful responses are avoided is therefore important with respect to the management of allergic disorders.

Another mucosal tissue, the gastrointestinal tract, is equipped with a characteristic immune system, which is devised to protect the tract from exaggerated immune responses against a wide variety of dietary proteins (3). Circumvention of deleterious immune responses to ingested food Ags is known as oral tolerance, which is separated into two distinct types on the basis of the amount of the fed Ag (4, 5). Feeding with low doses of Ags favors active suppression with the increased secretion of transforming growth factor (TGF)- and minimal anergy, whereas high doses induce anergy with little or no active suppression (4). The application of oral tolerance to the amelioration of experimental autoimmune and allergic diseases has yielded successful results, with highlighting of TGF- as a key regulatory cytokine elaborated from T lymphocytes (6, 8).

The regulatory system in the airway was most extensively examined by Holt and coworkers (11), who demonstrated that the repeated inhalation of Ag triggered Ag-specific and immunoglobulin (Ig)E-isotype-specific immunologic tolerance, and that this tolerance was mediated by interferon (IFN)--secreting CD8⁺ T cells. The inhibitory effects of Ags given by inhalation or intranasal instillation were also observed in other experiments, in which the Ags downregulated the experimental immune diseases by the induction of T helper (Th)2-like responses or other mechanisms (14). However, little is known about the regulation of Ag-induced eosinophilic inflammation by the induction of tolerance as an intrinsic self-regulatory system in the airway.

In the present study, we examined the mechanisms by which the trachea exposed to a large amount of Ag evaded the provocation of excessive eosinophilic inflammation. We found that high doses of Ag in the trachea inhibited the eosinophilic inflammation by downregulating interleukin (IL)-4 production from T cells. We also found that the high-dose tolerance in the trachea was mediated by TGF- secreted from CD4⁺ T cells in an Ag-specific manner. Thus, TGF- was a key cytokine mediating an intrinsic autoregulatory mechanism in the airway.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Animals and Immunization

BALB/c mice were bred in our animal facility and were used at 5 to 10 wk of age. These animals were primed intraperitoneally with 10 µg of ovalbumin (OVA) (Sigma Chemical Co., St. Louis, MO) or keyhole limpet hemocyanin (KLH) (Calbiochem Co., La Jolla, CA) precipitated with 4 mg of aluminum hydroxide (alum) in 200 µl of phosphate-buffered saline (PBS) three times at weekly intervals. Seven days after the last immunization, the mice were challenged with intratracheally administered Ag, as described later. BALB/c mice transgenic for T-cell receptor (TCR) specific for OVA_323-339 and I-A^d were established as described previously (21). They were primed intraperitoneally with 10 µg of OVA in alum twice at weekly intervals. After 7 d, they were challenged intratracheally with 10 or 500 µg of OVA.

Adoptive Cell Transfer

BALB/c mice or anti-OVA TCR transgenic (tg) mice received an intratracheal administration of 500 µg of OVA in 50 µl of PBS. After 7 d, 10⁶ or the indicated numbers of mediastinal lymph node (LN) cells were adoptively transferred intravenously to the OVA/alum- or KLH/alum-sensitized BALB/c mice. After 1 to 2 h, these mice were challenged by an intratracheal administration of the relevant Ag to evoke eosinophilic inflammation in the trachea.

Enrichment of LN cells for T cells, CD4⁺ T Cells, or CD8⁺ T Cells

To determine the phenotype of inhibitory cells, 10⁶ mediastinal LN cells from the BALB/c mice immunized intratracheally with 500 µg of OVA were adoptively transferred to the OVA/alum-sensitized BALB/c mice after the following enrichment. The LN cells were preincubated with PBS, anti-CD4 (Gk1.5) (22), and anti-CD8 (3.155) (23) monoclonal antibodies (mAbs) for 1 h at 4°C. They were allowed to react with antimouse Ig (Caltag, South San Francisco, CA) immobilized on plastic plates for 1 h at 4°C, and nonadherent cells were recovered as T cells, CD8⁺ T cells, and CD4⁺ T cells, respectively. Both mAbs were used at a 1:1,000 dilution of the ascites form. Flow cytometry revealed > 90% depletion of B cells, CD4⁺ cells, or CD8⁺ cells.

Neutralization of Cytokines by mAbs

A total of 0.1 µg of anti-TGF- mAb (Genzyme Co., Cambridge, MA) or 10 µg of anti-IFN- mAb (Genzyme) was administered intratracheally or intraperitoneally to the OVA/alum-sensitized BALB/c mice at the time of OVA instillation into the trachea.

Ag-Induced Eosinophilic Infiltration in the Trachea

The detailed methods have been described previously (24). In brief, 1 wk after the last intraperitoneal immunization with OVA/alum or KLH/alum, the sensitized mice were lightly anesthetized with pentobarbital (Abbott Laboratories, North Chicago, IL), and 10 µg of the relevant Ag or graded doses of OVA in 10 µl of PBS was instilled directly into the surgically exposed trachea. After 2 d, the excised trachea was fixed in 10% formalin and frozen in OCT compound (Miles Laboratories, Naperville, IL). Cryosections (7 µm thick) from the frozen tissue were stained with Diff-Quik (International Reagents Corp., Kobe, Japan). The numbers of eosinophils infiltrating into the submucosal tissue of the trachea were determined under a light microscope. The perimeter of the basement membrane of the trachea was measured using an MCID image analyzer (Imaging Research Inc., St. Catherines, ON, Canada). For every trachea, six to eight sections were used for counting, each of which was separated from the adjacent ones by more than 70 µm. Results were converted to the number of eosinophils per 1-mm basement membrane length. Data are expressed as means ± standard error of the mean (SEM) for four to seven mice in each group. Each experiment was repeated at least twice.

TGF- Assay

The mediastinal LN cells from anti-OVA tg mice given 10 or 500 µg of OVA intratracheally were cultured in RPMI 1640 supplemented with 1% Nutridoma SP (Boehringer Mannheim, Mannheim, Germany) in the presence or absence of 100 µg/ml OVA for 72 h in 96-well plates. Culture supernatants were applied to 96-well microtiter plates coated with chicken antihuman TGF-1 antibody (5 µg/ ml) (R&D Systems, Minneapolis, MN). Bound TGF- was detected by monoclonal anti-TGF- (1 µg/ml) (Genzyme), followed by peroxidase-labeled goat antimouse IgG (1 µg/ ml) (Kierkegaard & Perry Laboratories, Gaithersburg, MD) and tetramethylbenzidine reagent (Kierkegaard & Perry). Optical densities were determined at 450 nm and converted to concentrations (ng/ml) according to the standard curve obtained with titrated concentrations of human recombinant TGF-1 (R&D Systems).

IL-4 and IL-5 Assays

The mediastinal LN cells were prepared from BALB/c or anti-OVA tg mice that were primed intraperitoneally with OVA/alum as described previously, and then challenged intratracheally with 10 or 500 µg of OVA. The LN cells were cultured in the presence or absence of 100 µg/ml OVA for 48 h in 96-well plates. Otherwise, the LN cells were enriched for CD4⁺ T cells using anti-CD4 beads (Miltenyi Biotec, Gladbach, Germany) and MACS (Miltenyi Biotec), and cultured for 48 h with anti-CD3 mAb (5 µg/ml) (PharMingen, San Diego, CA) immobilized on microwells. The concentrations of IL-4 or IL-5 in culture supernatants were determined using enzyme-linked immunosorbent assay (ELISA) with paired anti-IL-4 or anti-IL-5 mAbs (PharMingen) according to the manufacturer's recommendations. Standard recombinant mouse IL-4 and IL-5 were purchased from Genzyme.

Bronchoalveolar Lavage

BALB/c mice primed and challenged with OVA as described previously were used for bronchoalveolar lavage (BAL). The lungs were lavaged twice with PBS (0.25, then 0.20 ml each time) and approximately 0.4 ml of the instilled BAL fluid (BALF) was consistently recovered. After centrifugation, supernatants were assayed for IL-4 by ELISA.

Statistics

Student's t test was used in the analysis of results.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Ag Dose-Dependent Responses of Eosinophilic Inflammation in the Trachea

We started with the establishment of an optimal condition for tracheal eosinophilia evoked by intratracheally administered Ag. BALB/c mice were primed intraperitoneally with OVA/alum and given titrated doses of OVA directly into the trachea. The animals showed few eosinophilic responses without intratracheal OVA challenge, and exhibited increasing eosinophilic infiltration in response to an increase in the amount of instilled OVA (Figure 1). The optimal amount of Ag for eosinophilic responses was 10 µg of OVA. Surprisingly, a further increase in the amount of administered OVA led to a decline in eosinophilic responses.

View larger version (16K): [in this window] [in a new window]   Figure 1.   Ag dose-dependent responses of eosinophilic inflammation in the trachea. BALB/c mice were injected intraperitoneally with 10 µg of OVA in alum three times at weekly intervals. Seven days after the last immunization, these mice were given an intratracheal administration of titrated doses of OVA in PBS (4 to 6 mice/group). After 2 d, the numbers of eosinophils infiltrating the trachea were determined as described in MATERIALS AND METHODS. It is noted that a high dose of Ag failed to elicit eosinophilia in the trachea.

Inhibition of IL-4 Production from CD4⁺ T Cells in High-Dose Tolerance

It is known that eosinophilic inflammation in the trachea is induced in a Th2-dependent manner (25). We examined whether low eosinophilic responses to high doses of Ag reflected the downregulation of Th2 cells. The BALB/ c mice sensitized with OVA/alum were challenged intratracheally with either high (500 µg) or optimal (10 µg) doses of OVA, and in vitro IL-4 production from mediastinal LN cells was examined. IL-4 production showed a striking relationship with the extent of eosinophilic inflammation: IL-4 production induced by in vitro Ag stimulation was inhibited in animals instilled intratracheally with high doses of Ag, in comparison to those with low doses (Figure 2A, P < 0.001). The level of IL-4 in the BALF was also decreased in the BALB/c mice given high doses of Ag (Figure 2B, P < 0.05).

View larger version (49K): [in this window] [in a new window]   Figure 2.   Inhibition of IL-4 production from CD4+ T cells in high-dose tolerance. (A) BALB/c mice were primed intraperitoneally with OVA/ alum three times, and then challenged intratracheally (i.t.) with 10 or 500 µg of OVA. Mediastinal LN cells were cultured in the presence or absence of OVA (100 µg/ml) for 48 h, and culture supernatants were assayed for IL-4 by ELISA. Antigen-induced production of IL-4 was inhibited in animals instilled intratracheally with high doses of Ag in comparison to those with low doses (P < 0.001). (B) BALF was collected from the BALB/c mice primed and challenged in this manner. The IL-4 level in BALF was again reduced in the former compared with the latter (P < 0.05). (C ) Anti-OVA TCR tg mice were primed intraperitoneally with OVA/alum twice and then challenged intratracheally with 10 or 500 µg of OVA. The number of eosinophils in the trachea was lower in animals instilled intratracheally with high doses of Ag than in those with low doses (P < 0.05). (D) Mediastinal LN cells were prepared from the same mice used in C and cultured in vitro in the presence or absence of OVA. IL-4 production induced by in vitro Ag stimulation was lower in the former than in the latter (P < 0.0005). (E ) To verify that increased IL-4 production upon Ag stimulation derived from CD4+ T cells, CD4+ T cells were purified from the mediastinal LN and stimulated by immobilized anti-CD3 mAb for 48 h. The CD4+ LN T cells from the former produced lower amounts of IL-4 than those from the latter in response to in vitro Ag stimulation (P < 0.001). (F ) IL-5 production from anti-CD3 mAb-stimulated CD4+ LN T cells was also reduced in high-dose-tolerant mice (P < 0.01). The results of in vitro cultures are given as the means ± SEM of six wells. The data shown are representative of two to three independent experiments.

Essentially identical observations were obtained with anti-OVA/I-A^d TCR tg mice. The OVA/alum-sensitized anti-OVA tg mice were challenged intratracheally with either high (500 µg) or optimal (10 µg) doses of OVA. Animals receiving high doses of OVA showed a lower amount of eosinophilic inflammation in the trachea than did those receiving optimal doses (Figure 2C, P < 0.05). Induction of IL-4 production by Ag stimulation was lower in the high-dose group (Figure 2D, P < 0.0005). To verify that IL-4 production induced by Ag was derived from CD4⁺ T cells, purified CD4⁺ T cells were stimulated with immobilized anti-CD3 mAb. The CD4⁺ mediastinal LN T cells from tg mice receiving high doses of OVA produced less IL-4 in response to in vitro Ag stimulation than did those receiving optimal doses (Figure 2E, P < 0.001). IL-5 production from the CD4⁺ mediastinal LN T cells also decreased in the tg mice receiving high doses of Ag (Figure 2F, P < 0.01). Thus, the high-dose Ag-induced inhibition of tracheal eosinophilic inflammation paralleled the downmodulation of IL-4-producing Th2 cells. Herein we refer to this inhibition as high-dose tolerance in the trachea.

High-Dose Tolerance Is Mediated by CD4⁺ T Cells

To evaluate the mechanisms of the high-dose tolerance in the trachea, we performed adoptive transfer experiments. BALB/c mice were instilled intratracheally with 500 µg of OVA and the mediastinal-region LN cells were transferred to the syngeneic mice primed with OVA/alum. The preliminary experiment revealed that the adoptive transfer of the whole LN cells inhibited the eosinophilic inflammation in the donor mice (data not shown). We performed an additional experiment to clarify the phenotype of cells with an inhibitory effect on eosinophilic inflammation. OVA/ alum-sensitized BALB/c mice showed eosinophilic inflammation in the trachea (Figure 3, bar 1), and this control response was inhibited by adoptive transfer of 10⁶ OVA- instilled LN cells enriched for T cells (Figure 3, bar 2; P < 0.01). A similar level of suppression was observed after the transfer of CD4⁺ T cells (Figure 3, bar 3; P < 0.002), whereas T cells depleted of CD4⁺ T cells had no ability to suppress the eosinophilic inflammation (Figure 3, bar 4 ). These results indicate that high-dose tolerance in the trachea was attributable to the active suppression mediated by CD4⁺ T cells.

View larger version (30K): [in this window] [in a new window]   Figure 3.   High-dose tolerance is mediated by CD4+ T cells. BALB/c mice were instilled intratracheally with 500 µg of OVA. After 7 d, 106 of the mediastinal LN cells depleted of B cells, B + CD8+ cells, or B + CD4+ cells were adoptively transferred to OVA/alum-sensitized syngeneic mice (4 mice/group) before intratracheal challenge with 10 µg OVA. Transfer of T cells (P < 0.01) or CD4+ T cells (P < 0.002) inhibited tracheal eosinophilia. Experiments were repeated twice with similar results.

Adoptive Transfer of Titrated Numbers of LN Cells from Tolerant tg Mice

Because the active suppression was found to be mediated by CD4⁺ T cells, the suppressor cells were induced in anti-OVA/I-A^d TCR tg mice in the following experiments. We first determined the optimal cell numbers for adoptive cell transfer. BALB/c tg mice were instilled intratracheally with 500 µg of OVA, and after 7 d titrated numbers of mediastinal LN cells were injected intravenously to OVA/alum-primed BALB/c mice, which were then challenged with 10 µg of OVA. The control response was inhibited upon adoptive transfer of the LN cells from tg mice instilled with OVA in a cell number-dependent manner (Figure 4). In this experiment, the maximum suppression was obtained with transfer of as few as 10⁴ cells.

View larger version (16K): [in this window] [in a new window]   Figure 4.   Adoptive transfer of titrated numbers of LN cells from tolerant tg mice. Anti-OVA TCR tg mice were instilled intratracheally with 500 µg of OVA. After 7 d, titrated numbers of mediastinal LN cells from these tg mice were intravenously injected into OVA/alum-sensitized BALB/c mice (4 to 5 mice/group) before intratracheal OVA challenge.

Ag Specificity of Tracheal Tolerance

To examine the Ag specificity of the high-dose tolerance, the suppressor cells induced by OVA were tested for the inhibitory effects on irrelevant Ag, KLH-specific eosinophilia. Adoptive transfer of OVA-induced suppressor T cells inhibited the OVA-specific airway eosinophilia (Figure 5, bar 2; P < 0.05), whereas the same cells failed to inhibit KLH-induced tracheal eosinophilia when transferred to the KLH/alum-primed mice (Figure 5, bar 4). This failure was not attributable to the lack of activation of bystander suppressor T cells because the simultaneous intratracheal administration of KLH with OVA did not restore the inhibition (Figure 5, bar 5). Thus, the high-dose tolerance in the trachea was induced in an Ag-specific manner.

View larger version (41K): [in this window] [in a new window]   Figure 5.   Ag specificity of tracheal tolerance. Airway eosinophilic inflammation was induced in BALB/c mice primed with OVA or KLH in alum by introducing the relevant Ag into the trachea. A total of 106 mediastinal LN cells from anti-OVA TCR tg mice instilled intratracheally with 500 µg of OVA were adoptively transferred to the donor animals (4 mice/group) before intratracheal Ag challenge. Adoptive transfer of the tg LN cells from OVA-instilled animals suppressed OVA- (bar 2, P < 0.05) but not irrelevant KLH- (bar 4) specific eosinophilic responses, indicating the Ag specificity of the suppressor cells induced by tracheal tolerance. The simultaneous administration of OVA with KLH failed to restore the inhibitory activity of the transferred cells (bar 5), thereby excluding the possible bystander suppressive effects. The numbers of eosinophils induced by intratracheal instillation of OVA and KLH were 12.7 and 6.1 cells, respectively, and these control responses were expressed as 100%. The data shown are representative of two separate experiments.

Inhibition by High-Dose Ag Instillation Was Mediated by TGF-, But Not by IFN--Secreting Th1 Cells

We examined the cytokines involved in the inhibitory process by using neutralizing antibodies. Eosinophilic infiltration induced by OVA in the trachea was inhibited by transfer of the mediastinal LN cells from mice instilled with high doses of the Ag (Figure 6A, bar 2; P < 0.05). This inhibition could not be affected by 10 µg of anti-IFN- mAb given to the trachea simultaneously with OVA challenge (Figure 6A; bar 3). In sharp contrast, the inhibitory activity by the transferred cells was reversed by as little as 0.1 µg of anti-TGF- mAb administered intratracheally at the time of OVA challenge (Figure 6A, bar 4; P < 0.05). Thus, the high-dose tolerance in the trachea was found to be mediated by TGF-, but not by the dominance of Th1 cells over Th2 cells. In an additional experiment, anti- TGF- mAb given intratracheally reproducibly restored the eosinophilic responses (Figure 6B, bar 3; P < 0.05), whereas the same amount of anti-TGF- mAb failed to revert the suppression when given intraperitoneally at the time of OVA challenge (Figure 6B, bar 4), indicating that the TGF--mediated suppression of eosinophilic responses was regulated at the regional immune system.

View larger version (27K): [in this window] [in a new window]   Figure 6.   Inhibition by high-dose Ag instillation was mediated by TGF- but not by IFN--secreting Th1 cells. (A) A total of 0.1 µg of anti-TGF- mAb or 10 µg of anti-IFN- mAb was administered intratracheally (i.t.) to the OVA/alum-sensitized BALB/c mice at the time of OVA instillation into the trachea (5 mice/ group). The suppression was reversed by intratracheal administration of anti-TGF- mAb (bar 3), but not of anti-IFN- mAb (bar 4), indicating that the suppressive effect of the transferred T cells on eosinophil recruitment reflected the inhibitory effect of secreted TGF- on Th2-dependent responses, but not the dominance of the transferred Th1 cells over Th2 cells. (B) A total of 0.1 µg of anti-TGF- mAb was administered either intratracheally or intraperitoneally (i.p.) to the OVA/alum-sensitized BALB/c mice at the time of OVA instillation into the trachea (5 mice/ group). Anti-TGF- mAb given intratracheally, but not intraperitoneally restored the eosinophilic responses, indicating that the TGF--mediated suppression of eosinophilic responses was regulated at the regional immune system. The data shown are representative of two separate experiments conducted with similar results.

TGF- Production by LN Cells from High-Dose-Tolerant Mice

To verify that the suppression was mediated by TGF-, we examined the production of TGF- by the mediastinal LN cells using ELISA. The mediastinal LN cells from anti-OVA tg mice instilled intratracheally with high (500 µg) or optimal (10 µg) doses of OVA produced a minimal amount of TGF- in the absence of OVA (Figure 7, bars 1 and 3), whereas the former produced a higher amount of TGF- than did the latter upon in vitro stimulation with OVA (P < 0.005) (Figure 7, bars 2 and 4). These results add validity to the conclusion that the suppression of eosinophilic inflammation in the trachea was mediated by TGF-.

View larger version (36K): [in this window] [in a new window]   Figure 7.   TGF- production by LN cells from high-dose- tolerant mice. The mediastinal LN cells from anti-OVA tg mice given 10 or 500 µg of OVA intratracheally (i.t.) were cultured in vitro in the presence or absence of 100 µg/ml OVA for 72 h. The concentrations of TGF- in culture supernatants were determined using ELISA. The LN cells from the tg mice instilled with high doses, but not optimal doses, of Ag produced TGF- upon in vitro Ag stimulation. Results are given as means ± SEM of six wells. The data shown are representative of three separate experiments.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Downregulation of unfavorable immune responses against inhaled Ags is one of the possible approaches in immunotherapy for bronchial asthma. Two major components that play pivotal roles in the pathogenesis of bronchial asthma are elevated IgE and eosinophilic inflammation in the airways, both of which are orchestrated by cytokines elaborated by Th2 cells (32). Selective abrogation of Th2 cell- mediated responses specific for allergen with sparing of other immune functions is the most desirable immunotherapy for avoiding adverse effects, such as deterioration of host defense ability, which are inevitable as long as Ag-nonspecific immunosuppressants are used.

In this report we examined the mechanisms by which the trachea exposed to excessive amounts of Ag evaded provocation of eosinophilic inflammation. We found that the inhibition of eosinophilic inflammation paralleled the downregulation of IL-4 production from T cells in the regional LN. The high-dose tolerance in the trachea was mediated by CD4⁺ T cells in an Ag-specific manner. The inhibition of Th2-dependent responses reflected the activation of regulatory T lymphocytes secreting TGF-, but not the dominance of Th1 cells over Th2 cells.

Tolerance induction against Ags intruding through the respiratory tract has been most extensively and systemically studied by Holt and coworkers (11), who reported that the repeated inhalation of Ag triggered Ag-specific and IgE-isotype-specific immunologic tolerance, and that this tolerance was mediated by IFN--secreting CD8⁺ T cells. In a more recent experiment, McMenamin and colleagues found that the suppression of IgE responses was adoptively transferable by T cells (33). Because both IgE and eosinophilic responses are orchestrated by Th2 cells (32), the control of Th2 cell functions is the ultimate target for manipulation by the regulatory networks in both our and Holt and colleagues' experimental systems. However, the involved cytokine and the phenotype of cytokine-secreting T cells are notably different, as we showed that the key cytokine mediating tolerance was TGF- secreted from CD4⁺ T cells but not IFN-. Other investigators reported that Ags given by inhalation or intranasal instillation had inhibitory effects on experimental immune diseases (14). The cytokine involved in these processes either remained to be resolved (14) or was found to be secreted from Th2 cells, which inhibited the Th1-mediated experimental autoimmunities (18, 20). Thus, regulation of Th2-mediated immune responses through TGF- secretion from CD4⁺ T cells is a novel regulatory system of the respiratory tract.

The tracheal tolerance documented in the present study shows some similarities to oral tolerance with respect to the involvement of TGF-. Oral tolerance was separated into two distinct types on the basis of the amount of the fed Ag (4, 5). Feeding with low doses of Ags favored active suppression with the increased secretion of TGF- (5). TGF--secreting T cells induced by low-dose tolerance could ameliorate Th1-mediated immune diseases, including experimental allergic encephalomyelitis, experimental autoimmune myasthenia gravis, and experimental granulomatous colitis (6, 8). In the other type of oral tolerance, feeding with high doses was reported to induce anergy with little or no active suppression (4, 5, 7). In addition to this prevailed notion, we recently observed that TGF--secreting splenic CD4⁺ T cells induced by oral high-dose tolerance could ameliorate Th2-dependent tracheal eosinophilia (24). In the present study, we found that high-dose tolerance in the trachea could induce TGF-- secreting cells, which inhibited tracheal eosinophilia. Therefore, a key regulatory role of TGF--secreting cells in immune responses was a feature shared by both the intestinal and respiratory tract mucosae. It is of particular interest that both of the representative mucosal systems can activate TGF--secreting CD4⁺ T cells in response to high doses of Ag, and that secreted TGF- can ameliorate Th2-dependent immune and inflammatory responses.

The Ag specificity of the TGF--secreting CD4⁺ T cells was another characteristic we found. It has been reported that once the TGF--secreting CD8⁺ T cells generated following oral tolerization were activated in an Ag-specific fashion, they suppressed in an Ag-nonspecific fashion (6, 34). In contrast, the suppressor T cells in this study inhibited allergic eosinophilia in an Ag-specific manner, as the suppressor cells induced by OVA instillation inhibited the OVA-specific airway eosinophilia; whereas the same suppressor cells failed to inhibit irrelevant Ag KLH-specific airway eosinophilia even in the concomitant presence of OVA (Figure 5). The Ag specificity could be a characteristic that TGF--secreting CD4⁺ T cells possess in common, because those induced by high-dose oral tolerance also exhibited Ag specificity (24). This could be a great advantage over the bystander effects, because the targeting of TGF--secreting T cells on the allergen-specific immune response could enable immunotherapy without interfering with other essential responses such as host defense against microorganisms.

In the present study, we first observed the inhibitory effects of high doses of Ag in OVA/alum-sensitized animals and we characterized suppressor cells from nonsensitized animals. The high-dose-tolerant cells from OVA/alum-sensitized BALB/c mice failed to manifest inhibitory effects upon transfer (data not shown). We reasoned that under the conditions where eosinophilia was induced, Th2 cells dominated over high-dose OVA-induced suppressor cells. In these mice, mediastinal LN cells still produced a certain amount of IL-4, even after inhibition by high doses of Ag (Figure 2). Thus, the suppressor cells were demonstrable only in Th2-unbiased conditions, such as in naive mice.

The physiologic role of TGF- in the lung is principally highlighted as a key effector cytokine during the course of tissue repair and fibrosis (35). Another aspect of TGF- as an inhibitor of a wide variety of inflammatory responses was evidently shown by TGF--deficient mice, which developed a wasting syndrome characterized by multifocal infiltration of lymphocytes and macrophages in many organs (39, 40). The inhibitory regulation by TGF- of inflammatory and immune responses in the lung is, however, poorly understood. Here, we revealed that TGF- is also an important negative regulator in the respiratory tract.

In conclusion, we demonstrated a novel regulatory mechanism by which excessive inflammatory responses in the trachea are inhibited. The delineation of a TGF-- mediated suppression of eosinophilic inflammation might help provide a new strategy for treating Th2-dominated allergic diseases, particularly bronchial asthma.