|Home | About | Journals | Submit | Contact Us | Français|
Asthma is caused by memory Th2 cells that often arise early in life and persist after repeated encounters with allergen. Although much is known regarding how Th2 cells develop, there is little information about the molecules that regulate memory Th2 cells after they have formed. Here we show that the costimulatory molecule OX40 is expressed on memory CD4 cells. In already sensitized animals, blocking OX40–OX40L interactions at the time of inhalation of aerosolized antigen suppressed memory effector accumulation in lung draining lymph nodes and lung, and prevented eosinophilia, airway hyperreactivity, mucus secretion, and Th2 cyto-kine production. Demonstrating that OX40 signals directly regulate memory T cells, antigen-experienced OX40-deficient T cells were found to divide initially but could not survive and accumulate in large numbers after antigen rechallenge. Thus, OX40–OX40L interactions are pivotal to the efficiency of recall responses regulated by memory Th2 cells.
Memory Th2 cells secreting the cytokines IL-4, IL-5, IL-9, and IL-13 are the major driving force behind allergic asthma and these CD4 cells are thought in many cases to arise early in life during infancy (1, 2). Although the immunological mechanisms that result in the development of Th2 cells are fairly well characterized, there is a lack of knowledge regarding how an already formed memory Th2 cell is regulated. Hence, it is not clear if there are certain molecular interactions that could be targeted to either suppress ongoing lung inflammation or to prevent the reoccurrence of asthmatic symptoms when airborne allergens are repeatedly encountered.
Apart from TCR signals that arise from peptide/MHC recognition, the most likely sources of positive signals that could affect longevity and/or reactivity of memory T cells are membrane bound costimulatory molecules of the Ig and TNFR superfamilies. Although there is abundant data on the requirement of members of these families in controlling priming of naive T cells and hence dictating the development of memory Th2 cells, there is little data on whether they can influence reactivity or persistence of a memory T cell once it has been generated. In fact several groups postulated that memory T cells are less dependent, or independent, of costimulation, largely based on negative data obtained from analyses of the Ig family member CD28 (3–8).
In contrast to CD28, it has become clear in the past few years that a number of additional receptors exist that could be crucial to a T cell response. Most significantly, recent data on OX40 (CD134), a member of the TNFR family, has shown that it is a major regulator of antiapoptotic proteins such as Bcl-xL and Bcl-2 and strongly promotes the survival of antigen-activated primary CD4 T cells (9–11). OX40 is not constitutively expressed on naive T cells but is induced in a primary immune response 24–48 h after recognition of antigen (9, 12–14). OX40L, a member of the TNF family, is also inducible (9, 14, 15) being expressed on activated B cells (15, 16), dendritic cells (17), and macrophage-like cells (18). The importance of OX40 costimulation in priming naive T cells has been demonstrated in a variety of basic and disease-related systems using OX40- and OX40L-deficient animals, and in blocking studies (10, 11, 18–27). However, to date no studies have addressed whether OX40–OX40L interactions contribute to the maintenance and functionality of antigen-specific memory T cells.
OX40 is down-regulated to low levels after the effector phase of primary T cell responses, and antigen-primed CD4 T cells can rapidly up-regulate its expression upon reencounter with antigen (9). Similarly, an anergic T cell, which also represents an antigen-experienced cell, albeit functionally hyporesponsive, can also express OX40 at low levels and be receptive to OX40 engagement resulting in enhanced functionality (28). These observations suggest that, similar to naive T cells, OX40 may play an important role in a range of memory T cell activities, including survival and effector function. Hence, OX40 may be an attractive target for suppressing already formed memory Th2 cells that mediate allergic asthma.
In the present study we demonstrate that some CD4 cells in an already sensitized animal express OX40 and that costimulation through this molecule is critical for all aspects of lung inflammation driven by these memory Th2 cells. These studies directly show that the recall response of memory T cells is not costimulation independent, and can be controlled by OX40 signals. They also suggest that targeting OX40–OX40L interactions may be useful therapeutically for the modulation of ongoing chronic inflammation in allergic disorders.
The studies reported here conform to the principles outlined by the animal Welfare Act and the National Institutes of Health guidelines for the care and use of animals in biomedical research. 6–8-wk-old female C57BL/6 mice were purchased from The Jackson Laboratory. OT-II TCR transgenic mice, originally generated by Barnden et al. (29), were a gift from Dr. W. Heath, and used as a source of Vβ5/Vα2 CD4+ T cells responsive to peptide 323–339 of OVA. OX40-deficient OT-II TCR transgenic mice were generated in house by crossing OT-II mice with OX40−/− mice.
The protocol for induction of pulmonary inflammation via antigen sensitization and aerosol challenge has been described previously (22). Briefly, groups of 4 C57BL/6 mice were sensitized by i.p. injection of 20 μg OVA protein (chicken egg albumin; Sigma-Aldrich), adsorbed to 2 mg aluminum hydroxide (Alum; Pierce Chemical Co.) in PBS on day 0. Unsensitized (naive) mice received 2 mg Alum in PBS. On day 25 or later, mice were challenged via the airways with OVA (5 mg/ml in 15 ml of PBS) for 30 min, once a day for four consecutive days, by ultrasonic nebulization (22). To block OX40–OX40L interaction, mice were injected i.p. with 200 μg of rat anti–mouse-OX40L blocking mAb (RM134L, rat IgG2bκ; reference 30) or isotype control antibody IgG2b; given on the indicated days in PBS, 30 min before OVA challenge.
Airway hyperreactivity (AHR)*was measured in vivo 1–3 h after the last aerosolized OVA exposure by recording respiratory curves by whole-body plethysmography (Buxco Technologies) in response to inhaled methacholine (MCh; 1.25–10 mg/ml; Aldrich Chemical) as described (22). Bronchoalveolar lavage (BAL) cytology, lung histopathology, OVA-specific and total IgE and lung cytokine profiles obtained by ELISA were determined as described (22). All data unless otherwise stated were collected 24 h after the final antigen challenge.
Peribronchial lymph nodes (PBLNs) and spleens were collected at the time of lung harvest. Lavaged lungs were digested with HBSS (GIBCO BRL) supplemented with 3 mg/ml collagenase (Type V; Boehringer), 0.1 mg/ml Dnase (Sigma-Aldrich), 100 μg/ml streptomycin (Invitrogen), and 100 U/ml penicillin (Invitrogen) for 60 min at 37°C. After lysing RBCs with ACK lysis buffer, PBLN, spleen, and lung cells were resuspended in RPMI-1640 medium (GIBCO BRL) supplemented with 10% FCS (Omega Scientific), 1% l-glutamine (Invitrogen), 100 μg/ml streptomycin, 100 U/ml penicillin, and 50 μM 2-mercaptoethanol (Sigma-Aldrich). Splenocytes (2 × 105 cells/well), lung cells (8 × 105 cells/well), or PBLN cells (2 × 105 cells/well) were plated in round-bottomed 96-well microtiter plates in 200 μl with increasing concentrations of OVA (10–100 μg/ml) for 72 h at 37°C. After 56 h, 1 μCi of 3H-thymidine (ICN Biomedicals) was added to each well. The cells were harvested 16 h later, and thymidine incorporation measured using a Betaplate scintillation counter. Each in vitro stimulation was performed in quadruplicates. Supernatants were harvested after 40 h for cytokine analysis.
Lymph node and spleen cells from wild-type (OX40+/+) OT-II mice or OT-II OX40-deficient (OX40−/−) mice were pooled and CD4 cells isolated as described previously (9). The purity of CD4+ T cells was confirmed to be >98% by FACS® analysis. To generate memory Th2 cells, naive T cells (106 cells/ml) were cultured in two ways: with plate-bound αCD3 (3 μg/ml) and soluble αCD28 (10 μg/ml), or with 2 × 106/ml syngeneic splenic APCs and 0.1 μM OVA peptide, plus IL-2 (5 ng/ml), IL-4 (20 ng/ml), αIFN-γ (10 μg/ml), and αIL-12 (10 μg/ml), for 3 d at 37°C. At the end of this culture, cells were removed, washed, and then cultured for another 3–6 d without further stimulation.
In some cases, primed T cells were labeled with CFSE (5- and 6-carboxyfluorescein diacetate succinimidyl ester C-1157; Molecular Probes) and 2 × 106 were injected into the tail vain of naive C57BL/6 mice in 200 μl. 1 d after transfer of cells, mice were challenged with inhaled OVA (5 mg/ml in 15 ml PBS) for 30 min daily for 2 consecutive days. Animals were killed for analyses 1 d later. Control mice received inhaled PBS only. In vivo cell division and T cell accumulation were assessed by tracking transferred T cells by flow cytometry based on coexpression of CFSE and Vα2.
At the end of the primary culture, an aliquot of primed T cells were retained to determine proliferation, survival, and cytokine production in recall responses in vitro. 5 × 105 CD4 T cells/ml were recultured with 2 × 106/ml syngeneic splenic APCs and OVA peptide. IL-4, IL-5, IL-13, and IFN-γ levels from cell supernatants were determined by ELISA at 40 h. Proliferation was measured in triplicate by the incorporation of 3H-thymidine (1 μCi/well; ICN Pharmaceuticals) during the last 12 h of culture. T cell survival was determined by trypan blue exclusion.
Cells were stained with FITC-conjugated anti-CD4, PE-conjugated anti-OX40, Cychrome-conjugated anti-CD44, or PE-conjugated anti-Vα2 at 4°C for 30 min. Immunostained cells were analyzed on a FACScan™ flow cytometer (Becton Dickinson) using the CELLQuest™ software.
Characteristic features of allergic asthma are produced when mice are sensitized with OVA, and subsequently challenged several weeks later by inhalation of OVA (1, 22). Using OX40-deficient mice, we previously showed the importance of OX40 in controlling initial development of asthma symptoms (22). However, as these mice are defective in generating a primary T cell response (10, 11), no conclusions could be drawn regarding any requirement for OX40 in the secondary response of the memory T cell that ultimately induces lung inflammation.
We initially looked to see if memory CD4 cells express OX40. By staining T cells from mice primed 4 wk prior with OVA in alum, we visualized a significant number of CD44hi memory CD4 cells in lymph nodes that expressed OX40 at low/moderate levels. In contrast, as shown previously (9), CD44lo naive CD4 cells did not express OX40 without antigen exposure (Fig. 1 A). Unimmunized mice also contained a proportion of CD44hi CD4 cells that expressed OX40 at low levels (unpublished data), suggesting that OX40 can be readily available to some memory T cells. OX40 levels were up-regulated on responding CD44hi memory T cells in lung draining lymph nodes after challenge with aerosolized antigen (Fig. 1 A), and the absolute number of OX40 positive CD44hi CD4 cells increased markedly (Fig. 1 C, and see below). As this large increase in the number of OX40-expressing CD44hi cells was not seen in unprimed mice challenged with aerosolized antigen, the latter directly reflects the response of memory T cells and accumulation of memory effector cells.
Few CD4 cells were present in the lungs of primed but unchallenged mice, and OX40 was not detected on either CD44hi or CD44lo cells (Fig. 1 B). However, after antigen challenge, a large number of OX40-expressing CD4/CD44hi cells were seen in the lung (Fig. 1 B, and see below). Analysis of lung sections revealed the presence of OX40+ cells in peribronchial and perivascular areas, sites previously shown to contain the majority of antigen-responding memory effector T cells (Fig. 1 D). These data show that OX40 is expressed on memory/memory effector T cells and suggest that it is available to play a role during the secondary response that occurs after reencounter with antigen.
To examine the contribution of OX40/OX40L interactions to lung inflammation, we administered a blocking anti-OX40L mAb to OVA-sensitized mice at the time of rechallenge with aerosolized antigen (Fig. 2 A). OVA-immunized mice treated with the isotype control developed increases in AHR whereas administration of anti-OX40L dramatically reduced the degree of AHR (Fig. 2 B). Sensitized mice treated with control Ab during aerosol challenge (Alum-OVA/IgG-OVA) responded with an increase in the total number of cells in BAL (Fig. 2 C), which was evident as early as 48 h and was mostly eosinophils (Fig. 2 D). In striking contrast, administration of anti-OX40L during the challenge period (Alum-OVA/RM134L-OVA) virtually eliminated the increase in total leukocytes (Fig. 2 C) and eosinophils (Fig. 2 D). As no cell infiltration was seen in unprimed but challenged animals, this directly shows that OX40/OX40L interactions were essential to the recall response.
To support the findings from lung lavages we conducted histological evaluations on lung sections. Mice receiving control Ab developed inflammatory lesions, characterized by a predominance of eosinophils and lymphocytes (Fig. 3, A and C) , together with hyperplasia of the mucus-secreting bronchial epithelial cells (Fig. 3 A). In contrast, lungs from anti-OX40L treated animals had almost normal bronchial epithelium, and a minimum of infiltrating cells around the bronchioles and blood vessels (Fig. 3, B and C). Many PAS+ (mucus-secreting) cells were detected in the airway of mice that were sensitized and challenged and received the control Ab (Fig. 3, D and F), whereas treatment with anti-OX40L markedly reduced the number of PAS+ cells (Fig. 3, E and F).
As an indirect test of whether OX40 signals were controlling the memory Th2 response driving the asthmatic reaction, we measured cytokine levels in the BAL and production of the Th2-associated antibody IgE in the serum. Treatment with anti-OX40L prevented the increase in IgE that resulted from the recall response (Fig. 3 G) and all Th2 cytokines (Fig. 3, H–K) associated with the secondary response. IFN-γ was either not present or at only low levels in all groups (unpublished data), suggesting that there was not a switch to a Th1 response.
To determine whether functional antigen-reactive T cells were present in mice where OX40 signals were blocked, OVA-specific proliferation and production of Th2 cytokines was measured in vitro. Strong responses were seen in lung cell cultures from OVA-primed and challenged mice but not in unprimed challenged mice (Fig. 4, A and B , compare grp B and grp A), again reflecting the memory effector response and that a functional primary response did not result from exposure to airborne antigen alone. Those animals treated with anti-OX40L during the recall response showed little or no OVA-specific T cell reactivity in the lungs (Fig. 4, A and B, grp C). To test whether recall responses were absent at other sites, spleen and lung draining lymph nodes were examined. Greatly reduced OVA-reactivity was found in secondary lymphoid organs after anti-OX40L treatment (Fig. 4, C and D). These results suggest that the effect of OX40/OX40L blockade is not tissue specific, but, rather, that OX40 signaling is critical for the development of functional memory effector cells after memory T cells reencounter antigen.
As OX40 is only present on antigen-responding or antigen-experienced T cells, we tracked the accumulation of OX40-expressing CD4 cells over time. As shown in Fig. 4 (and Fig. 1), before antigen challenge, only a low number of T cells expressed OX40. After OVA challenge, an large increase was observed in the number of CD4+OX40+ cells in the bronchial LN (Fig. 4 E), lung (Fig. 4 F), and BAL (Fig. 4 G) in primed and challenged mice, far greater than in unprimed challenged mice, reflecting the population of responding memory effector T cells. The number of OX40+ cells in the LN peaked at day 1, whereas the number in the BAL rose progressively from day 2 over the 4-d aerosol exposure. This suggests that the initial memory T cell response developed in the secondary lymphoid organs, and within 24 h the T cells started to migrate to the lung. This correlates with other published data (31). Significantly, treatment of mice with anti-OX40L reduced the number of CD4+OX40+ cells visualized in the LN at the peak of response at day 1, and subsequently. Consequently, few activated OX40+ T cells were then observed in the lung and BAL. These results suggest that OX40/OX40L interactions are required early in the response of a memory T cell and control the ability to expand in numbers and survive in order to form a large population of memory effector T cells.
As further evidence for this, we performed kinetic blocking experiments where mice received anti-OX40L 1, 2, or 3 d after the initial aerosol (Fig. 4, H and I). Although delaying anti-OX40L treatment by 1 or 2 d reduced the response by 50% and 30%, the most dramatic effect was seen when OX40/OX40L interactions were inhibited at the time of initial encounter with recall antigen.
To further investigate the requirement for OX40 in recall responses of antigen-primed Th2 cells, OVA-specific OX40-deficient CD4 cells were produced by crossing OX40-knockout mice to OT-II TCR transgenic mice. Naive CD4 cells from wild-type (OX40+/+) or OX40-deficient (OX40−/−) OT-II mice were cultured in vitro with peptide and APCs under Th2 (IL-4, anti-IL-12, and anti-IFN-γ) polarizing conditions for a time period that allows primary T cell expansion and contraction to proceed normally. Previous data have shown that such in vitro–stimulated cells mimic in vivo generated memory cells (32, 33) and can also be used in adoptive transfer experiments to directly induce lung inflammation (34, 35).
Restimulation of primed OX40+/+ Th2 cells in vitro with OX40L-sufficient APCs resulted in proliferation, expansion, survival, and production of Th2 cytokines (Fig. 5) . Primed OX40−/− T cells secreted Th2 cytokines normally, suggesting that there is not a role for OX40 in this T cell activity, and OX40−/− T cells also initially proliferated comparably with wt T cells. However, far fewer T cells survived in the secondary response in the absence of OX40 signals, which was also reflected in weak proliferation late in culture (Fig. 5, A and B).
To determine if a similar requirement for OX40 was apparent in vivo during a recall response to aerosolized antigen, in vitro–primed T cells were labeled with CFSE, adoptively transferred into naive recipients, and then these mice were challenged intranasally with OVA (Fig. 6) . Challenge with PBS did not result in division (Fig. 6, A and C) or expansion (Fig. 6, B and D) of either OX40+/+ or OX40−/− T cells in draining lymph nodes or lung. Challenge with OVA resulted in pronounced division of all OX40+/+ T cells and accumulation in numbers was observed in lymph nodes and particularly evident in the lung. Correlating with the in vitro data, recovered OX40−/− T cells displayed the same division profile as their wt counterparts (Fig. 6, A and C), but in total approximately fourfold fewer accumulated at the end of the antigen challenge period (Fig. 6, B and D). This directly mimicked the data obtained by tracking OX40-expressing T cells after OX40L blockade (Fig. 4, E–G). This shows that OX40 signals regulate the number of memory effector T cells that are generated after memory T cells reencounter antigen.
To correlate a lack of expansion/survival of memory effector cells with reduced lung inflammation, eosinophilia, lung histology, and Th2 cytokines were analyzed after adoptive transfer of primed OX40−/− OT-II Th2 cells (Fig. 7) . Transfer of wild-type cells followed by repeated intranasal OVA challenge induced profound inflammation accompanied by production of Th2 cytokines in BAL (Fig. 7, A, C, and F). In contrast, mice receiving OX40-deficient cells showed only a small increase in the total numbers of eosinophils recovered from the BAL (Fig. 7 A), relatively normal lung histology (compare Fig. 7 E to 7 C), and reduced Th2 cytokines (Fig. 7 F). These results clearly show that OX40 expressed on a memory Th2 cell is required for lung inflammation.
As our initial studies addressed the role of OX40 signals in the recall response of what can been termed early memory T cells, persisting 4 wk after primary immunization, we determined whether OX40/OX40L interactions were also critical to the response of memory T cells that persist for longer periods of time and survive following a secondary response. In one set of experiments, mice were primed for 8 wk and anti-OX40L administered during the secondary response to aerosolized antigen (Fig. 8 A). Again, blocking OX40 signals strongly inhibited all aspects of lung inflammation, including total leukocyte (Fig. 8 B) and eosinophil (Fig. 8 C) infiltration, and production of Th2 cytokines (unpublished data). In further experiments, mice were primed for 4 wk, challenged with aerosolized antigen in a secondary response without blocking OX40L, rested for 13 wk, and then challenged again but with anti-OX40L blockade (Fig. 8 D). Inhibiting OX40 signals during this tertiary response again strongly inhibited lung inflammation (Fig. 8, E and F), with a greater than 50% reduction in cellular infiltration. As no lung inflammation was observed in unsensitized control mice that were repeatedly exposed to aerosolized antigen, ruling out any primary response (Fig. 8, E and F, grp A), these data demonstrate that blocking OX40/OX40L interactions severely limits the recall response of memory Th2 populations that can persist for extended periods of time.
Previous work has demonstrated that OX40 and OX40L control the development of a number of primary T cell responses (9–11, 18–27) but no data has been presented regarding a role for these interactions in the recall response of memory T cells. We now show the absolute importance of OX40–OX40L interactions in antigen-specific memory that mediates allergic lung inflammation. We demonstrate that exposure of sensitized animals to antigen via the airways is accompanied by large increases in the number of OX40-expressing memory/memory effector Th2 cells, both in the lungs and in secondary lymphoid organs, which was closely associated with eosinophilia and the development of airway hyperresponsiveness. Strikingly, preventing OX40–OX40L interactions during the recall response inhibited the ensuing Th2 response and the associated asthmatic symptoms. These results provide the first in vivo evidence that memory T cells, like naive T cells, can require costimulation through OX40, and indicate that OX40–OX40L interactions are potential targets for modulating peripheral acute phase inflammatory responses.
Our recent data show that OX40 principally functions by suppressing T cell death and does so by maintaining high levels of antiapoptotic proteins such as Bcl-xL and Bcl-2 (11) and inhibiting expression or activity of proapoptotic proteins such as Bad and Bim (unpublished data). This conclusion has been reinforced by in vivo studies where agonist antibodies directed to OX40 on a responding naive CD4 cell enhanced primary T cell expansion and survival, and in so doing promoted the development of greater numbers of memory T cells (10, 36). The data here now complement these studies and show that OX40 can act in the response of a memory Th2 cell in an analogous manner and dictate the expansion and persistence of antigen-specific memory effector T cells during recall lung inflammation.
Analysis of BAL and dispersed lung showed that anti-OX40L reduced the absolute numbers of activated (OX40+) memory effector T cells, potentially consistent with the idea that OX40 could be involved in memory T cell recruitment. However, low numbers of activated OX40+ T cells were also observed in secondary lymphoid organs, along with reduced antigen-specific activity, supporting the notion that preventing OX40/OX40L interactions did not simply block T cells from entering the lung. Other studies of lung inflammation have suggested that memory T cells do not proliferate in the lung in response to intranasal antigen challenge, but proliferate in the lymph nodes (37). Moreover, prior reports show that T cell emergence in the lung is only appreciable 24–48 h after antigen reexposure (31), consistent with the idea that expansion and survival signals are required to generate a large pathogenic population. Our kinetic blocking data showing maximal inhibition only when OX40/OX40L interactions are absent at the time of initial antigen challenge support this, along with the idea that OX40 signals are required early (0–48 h) and most likely provided in the secondary lymphoid organs. Further in line with the idea that OX40 signals dictate the subsequent generation or survival of memory effector T cells after antigen reactivation, primed OX40-deficient Th2 cells were found capable of proliferating initially in the recall response, but did not subsequently accumulate in high numbers.
Previous studies of memory T cells have largely concluded that they had a reduced requirement for costimulatory signals and hence may not be susceptible to interventions that target such membrane bound molecules (38). For example, blocking B7–CD28 interactions during secondary responses to the nematode parasites Heligmosomoides polygyrus (7) and Nippostrongylus brasiliensis (8), or immunogenic anti-mouse IgD antibody treatment (6), failed to inhibit memory Th2 responses, whereas blocking CD28 at the time of priming was suppressive (6–8). These and other observations therefore implied that activation of memory Th2 cells may be costimulation, or at least B7/CD28, independent. However, in light of our data on OX40, rather than becoming costimulation independent, T cells may shift after antigen exposure to being more reliant on other molecules. Recent studies from Gonzalo et al. (39) on an inducible Ig family member, ICOS, also support this idea. These latter data in another variant of the asthma model showed that blocking B7RP-1–ICOS interactions effectively inhibited most asthmatic symptoms at the time of allergen exposure, producing an overall result similar to our data with OX40. Furthermore, this report showed that ICOS was active at a time when CD28 was not a major factor. Unlike OX40, it has been proposed that ICOS may primarily regulate cytokine production (40) rather than T cell survival, and therefore it is an intriguing idea that both of these molecules act in concert but dictate distinct phases during the Th2 recall response.
In summary, our experiments show a critical role for OX40/OX40L interactions in the recall response to inhaled antigen, for the subsequent activation and recruitment of memory effector CD4 T cells into the airway, and for the induction of morphological changes to the airways that are reminiscent of human asthma. As CD4+ T cells have been shown to play a critical role in both initiation and maintenance of allergic pulmonary responses, our results have important ramifications for the development of therapeutic strategies in which anti-OX40L antibodies, alone or in combination with other therapies, can be used in clinical situations where memory Th2 cells have been implicated. Importantly, this approach has a distinct advantage of blocking the production of several Th2 cytokines simultaneously rather than suppressing the activity of a single cytokine, and may profoundly affect the ability to continue to mount a Th2 response.
This work was supported by an award from the Sandler Program for Asthma Research, and National Institutes of Health grant AI50498, to M. Croft. This is publication # 515 from the La Jolla Institute for Allergy and Immunology.
*Abbreviations used in this paper: AHR, airway hyperreactivity; BAL, bronchoalveolar lavage; PBLN, peribronchial LN.