INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

T cells are proposed to play a key role in the development of eosinophilic airway inflammation through the production of cytokines. A number of T-cell products, including interleukin-4 (IL-4), IL-5, and granulocyte-macrophage colony-stimulating factor (GM-CSF), are probably involved in the late-phase inflammatory response; however, IL-5 has unique specificity for eosinophils (1). IL-5 is involved in the migration, maturation, activation, and survival of eosinophils (2). A direct cause and effect relationship has been demonstrated in animal models of allergic airway inflammation, in which genetic IL-5 deficiency (3) or administration of neutralizing anti-IL-5 antibody (4) inhibits allergen-induced eosinophil recruitment to the lung. Increased levels of IL-5 can be measured in the bronchoalveolar lavage fluid (BALF) of allergic subjects (7), and IL-5 messenger RNA (mRNA) can be detected within the mucosa (12) and BAL cells (13) following allergen bronchoprovocation. Although other factors, such as regulated on activation, normal T-cell expressed and secreted (RANTES), have eosinophil chemotactic activity, the level of IL-5 correlates most strongly with eosinophil numbers in the airway following allergen challenge (14).

Recently, we have shown that in allergic asthmatic subjects, segmental bronchoprovocation (SBP) with allergen leads to a dose-dependent increase in IL-5 that correlates with airway eosinophilia (11). However, we did not address in that study the mechanisms for increased IL-5 in the airway following allergen challenge. A requirement for T cells has been demonstrated in animal models of allergic airway inflammation. Hence, depletion of CD4⁺ T cells inhibits eosinophilic airway inflammation (4, 15) and decreases steady-state levels of IL-5 mRNA in the lung tissue (16) of allergen-sensitized and -challenged mice. In the present study, we used SBP with saline or antigen (Ag) to evaluate the effect of Ag on cytokine production by airway lymphocytes, and the relationship of the cytokines produced to airway eosinophilia in allergic subjects. We hypothesized that local instillation of allergen into the airway would result in the activation or recruitment of T cells with an increased capacity to produce IL-5 and thus recruit eosinophils to the airway.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

Twelve allergic subjects were recruited for the study. Subjects' characteristics are shown in Table 1. The subjects included seven males and five females between the ages 21 and 44 yr, who were in good health with the exception of allergic rhinitis. Subjects had a positive skin-prick test to one or more aeroallergens, and a history of allergies with nasal congestion. Each subject gave a medical history and underwent a physical examination and pulmonary function testing with spirometry. Informed consent was obtained from each subject prior to participation. The study was approved by the University of Wisconsin-Madison Center for Health Sciences Human Subjects Committee.

SBP and BAL

One month prior to bronchoscopy, a graded inhaled Ag challenge was performed in each subject to determine the antigen dose that caused a 20% decrease in FEV₁ (Ag PD₂₀) (17). An Ag PD₂₀ of 1,500 was assigned to three subjects whose decrease in FEV₁ following an inhaled ragweed challenge was less than 20%. Allergen doses are given in Table 1. Allergens used in SBP included ragweed (GS Ragweed mix; Greer Labs, Lenoir, NC) and house dust mite (Dermatophagoides farinae; Miles Allergy Products, Spokane, WA). Bronchoscopy and SBP were done as described previously (18). Briefly, one bronchopulmonary segment was identified, and a wedge position was achieved with the fiberoptic bronchoscope. Sham (10 ml of 0.9% NaCl) SBP was done, followed by BAL 5 min later. In a separate segment, Ag-SBP was done with an Ag dose equal to 10% of the Ag PD₂₀. The Ag dose was diluted in 10 ml 0.9% NaCl and instilled through the wedged bronchoscope. This was followed by injecting 5 ml of air to clear the bronchoscope channel. BAL was performed 5 min later. After 48 h, bronchoscopy was repeated and BAL performed from each of the two previously challenged segments.

Analysis of Blood and BALF

Peripheral blood was drawn immediately prior to the lavage. Total cells counts, differential counts, and fluorescence-activated cell sorter (FACS) analyses were performed on ethylenediamine tetraacetic acid (EDTA)-treated whole blood. Total cell numbers were determined with a hemocytometer, using Turk's counting solution containing acetic acid and methylene blue. Eosinophils were also enumerated with a hemocytometer, using a Unopette pipet test with 1% Phloxine B (Becton Dickinson, purchased through Fisher Scientific, Itasca, IL). For differential cell counts, blood smears and cytospin preparations of BAL cells were stained with the Giemsa-based Diff-Quik stain (Baxter Scientific Products, McGaw Park, IL). Peripheral blood mononuclear cells (PBMC) were obtained from 20 ml of heparinized blood by centrifugation over Ficoll-Paque (Ficoll 400-diatrizoate; Pharmacia Bioethic Inc., Piscataway, NJ). Cells were collected from the interface and washed twice. BAL cells were recovered from the lavage fluid by centrifugation at 400 × g for 10 min at 4° C. BAL cells were washed twice with Hank's balanced salt solution (HBSS) containing 2% newborn calf serum. BALFs were stored at 70° C until analyzed.

Flow-cytometric Analysis

BAL cells (1 × 10⁵ cells) and 100-µl aliquots of whole blood were stained through the simultaneous addition of fluorescence isothiocyanate (FITC)- and phycoerythrin (PE)-conjugated antibodies specific for cell-surface markers (Becton Dickinson Immunocytometry Systems, San Jose, CA). Red blood cells (RBCs) were then lysed in both BAL and blood samples (BD Lyse; Becton Dickinson Immunocytometry Systems), and cells were fixed by overnight incubation in buffered formaldehyde (1% paraformaldehyde, 1% cacodylic acid in 0.78% NaCl) at 4° C. For analysis, 5,000 or 10,000 photoelectric events, for blood and BAL cells, respectively, were recorded with a Becton Dickinson FACScan II, and data analyses were performed with the Lysys II software package (Becton Dickinson Immunocytometry Systems). To normalize for the large amount of noncellular debris that was present in some BAL preparations, the relative percentage of CD19, CD3, CD4, CD8, and natural killer (NK) (CD3/CD16⁺CD56⁺) cells was calculated on the basis of the total number of CD45⁺ leukocytes present within the lymphocyte gate (19).

Flow-cytometric Analysis of Intracellular IL-5

Unseparated BAL cells were cultured in 17 × 100-mm polypropylene tubes for 24 h with phorbol myristyl acetate (PMA) (10 ng/ml) and ionomycin (10⁶ M). Monensin (2 µg/ml) was added during the last 4 h to inhibit intracellular protein transport and secretion of cytokines (20). CD4⁺ cells were labeled with peridinin chlorophyll protein (PerCP)-conjugated anti-CD4 antibody (Becton Dickinson Immunocytometry Systems) and cells were then fixed with 1% paraformaldehyde for 10 min. A 0.1% saponin (Sigma Chemical Co., St. Louis, MO) solution was used to permeabilize the cell membrane, and intracellular IL-5 was detected with PE-conjugated anti-IL-5 antibody (PharMingen, San Diego, CA). Statistical gates were set on the basis of isotype control antibodies. For analysis, 10,000 events were recorded.

Cell Cultures

Viability of the cell populations was determined through trypan blue exclusion, using a hemocytometer. Unseparated blood or BAL cell populations were cultured at 2 × 10⁶ viable cells per ml in culture media consisting of RPMI 1640, 10 mM 4-(2-hydroxyethyl)-l-piperazine-N'-2-ethanesulfonic acid (HEPES), 2 mM L-glutamine, antibiotic and antimycotic agents (100 units/ml penicillin G, 100 µg/ml streptomycin sulfate, and 250 ng/ml amphotericin B), and 5% fetal calf serum (FCS) (all culture components purchased from GIBCO BRL, Grand Island, NY). Cells were cultured in a total volume of 0.5 ml in 48-well tissue-culture plates. Phytohemagglutinin (PHA) (Sigma L9132) was used at a final concentration of 10 µg/ml. The concentration of PHA and duration of the cultures were optimized for detection of IL-5 and IFN-. All cultures were done in triplicate. Cells were cultured for 48 h at 37° C under 5% CO₂ in a humidified incubator. Culture supernatant fluids were removed and stored at 20° C until analyzed.

Selection of CD4⁺ Cells

In selected experiments, CD4⁺ cells were purified through immunobead selection. BAL cells were centrifuged over a Percoll (Pharmacia Bioethic Inc.) density gradient (1.060 g/ml), and the mononuclear cells were collected. Cells were incubated with a mouse monoclonal antihuman CD4 antibody (Ortho Diagnostic Systems, Raritan Park, NJ), followed by goat antimouse IgG-coated paramagnetic beads (Miltenyl Bioethic, purchased through Becton Dickinson Immunocytometry Systems, San Jose, CA), and passaged through a magnetic cell separator (MACS; Miltenyl Bioethic, purchased through Becton Dickinson Immunocytometry Systems, CA). The eluate was used as the CD4-enriched population and the effluent as the CD4-depleted population.

Histamine Quantification

Endogenous histamine was measured in EDTA-treated 1× BAL fluids with a radioenzymetric assay (21). This assay utilizes a histamine-specific n-methyl transferase to transfer a tritiated methyl group from S-adenosyl-L-methionine to histamine, forming an N-telemethyl histamine. A series of solvent extractions is performed to isolate the radiolabeled product. A standard curve is run with each assay to allow direct comparison of unknowns with a standard. The assay has a linear detection range of 30 to 50,000 pg/ml.

Cytokine ELISAs

A sensitive two-step sandwich-type enzyme-linked immunosorbent assay (ELISA) was established to measure cytokines in culture supernatant fluids. ELISA plates (Easy Wash, Catalogue No. 25805-96; Corning, Costar Corp., Cambridge, MA) were coated overnight with a predetermined optimal concentration of purified monoclonal anticytokine antibody. Nonspecific binding sites on the plate were blocked with 10% dialyzed newborn calf serum. BALFs were concentrated by 20× with a low-protein-binding concentrator (Centriprep; Amicon, Beverly, MA) with a molecular-weight cutoff limit of 3 kDa. Cell-culture supernatant fluids were diluted to optimal concentrations. Test samples were incubated on antibody-coated plates for 2 h, and cytokines were detected with biotinylated anticytokine antibodies followed by addition of a streptavidin polymer (POLY-HRP-40; Research Diagnostics Inc., Flanders, NJ). A one-component substrate, 3,3',5,5'-tetramethylbenzidine (TMB) (Kirkegaard and Perry Laboratories, Inc., Gaithersburg, MD), was used for color development, and the reaction was stopped by addition of 0.18 M sulfuric acid. Optical density (O.D.) at 450 nm was determined with a Dynatech MR5000 microplate reader, and data were analyzed with Biolink Software (Dynatech Laboratories, Inc., Chantilly, VA). The concentration of cytokines in supernatant fluids was calculated by comparison with a standard curve generated with known amounts of recombinant human cytokines. The sensitivity for each cytokine assay was < 3 pg/ml.

Antibodies

The following antibodies were used in the ELISA assay: Coating antibodies (all from Pharmingen, Inc., unless otherwise indicated): mouse monoclonal antiHuman (Hu)-IFN- (Clone 2G1; Endogen, M-700A, Cambridge, MA), mouse monoclonal anti-Hu-IL-4 antibody (Clone 8D4-8), monoclonal antimouse-IL-5 antibody (Clone TRFK5), rat monoclonal anti-Hu-IL-10 antibody (Clone JES-9D7), rat monoclonal anti-Hu-GM-CSF antibody (Clone BVD2-23B6), and rat monoclonal anti-Hu-IL-2 antibody (Clone MQ1; 17H12).

Detecting antibodies (all from Pharmingen, Inc., unless otherwise indicated): mouse monoclonal anti-Hu-IFN- antibody (Clone B133.5; Endogen, M-701A, Cambridge, MA), biotinylated rat monoclonal anti-Hu-IL-4 antibody (Clone MP-25D2), biotinylated rat monoclonal anti-Hu-IL-5 antibody (Clone JES1-5A10), biotinylated rat monoclonal anti-Hu-IL-10 antibody (Clone JES3-12G8), biotinylated rat monoclonal anti-Hu-GM-CSF antibody (Clone BVD2-21C11), and biotinylated polyclonal rabbit anti-Hu-IL-2 antibody. Biotinylation of the mouse monoclonal anti-Hu-IFN- antibody (Clone B133.5) was done according to the manufacturer's instructions, using an N-hydroxysuccinimide ester with a long side chain (NHS-LC biotinylation kit; Pierce Chemical Company, Rockford, IL).

Statistical Analysis

Data are expressed as medians within quartiles of 25 and 75. Wilcoxon's signed rank test (or a paired t test for normally distributed data) was used to compare cells obtained from the same segment at different time points. For comparison of cells obtained from the Ag versus the saline segments, a Mann-Whitney ranked sign test (or an unpaired t test for normally distributed data) was used. Correlations were made with Spearman's rank order correlation. A value of p < 0.05 was considered significant. Statistical analysis was done with a SigmaStat software package (Jandel Scientific Software, San Rafael, CA).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subject Characteristics and Safety

Twelve subjects completed the study. Ten were challenged with ragweed and two received house-dust-mite allergen (Table 1). At the time of SBP, all subjects had normal spirometric findings. Following bronchoscopy, FEV₁ was within 5% of the baseline value for each subject. Oxygen saturation during Ag-SBP and BAL ranged from 94 to 99%, with the exception of one subject (Subject 10) who experienced mild oxygen desaturation (O₂ saturation of 91%) after Ag-SBP. This subject's oxygen saturation returned to normal within 30 min. Bronchoscopy and BAL, which were performed 48 h after Ag-SBP, were well tolerated and produced no adverse effects.

Effect of Local Allergen Challenge on Histamine and IL-5 Release into the Airway

BALF histamine levels were markedly increased immediately (5 min) after allergen SBP, and returned to baseline by 48 h (Figure 1). When IL-5 levels were measured in 20×-concentrated BALF, a significant increase was detected at 48 h after Ag-SBP. There was also a slight but significant increase in IL-5 at 48 h after saline-SBP (Figure 1).

View larger version (17K): [in this window] [in a new window]   Figure 1.   BALF histamine and IL-5. Histamine was measured in 1× BALF obtained 5 min (imm) and 48 h (late) after SBP with saline or Ag. IL-5 was measured in 20×-concentrated BALF, and data are expressed as the calculated amount of IL-5 in 1× BALF. A cutoff level of 0.25 pg/ml was used for these calculated values. Bars represent medians within quartiles of 25 and 75; error bars represent 10th and 90th percentiles. *p < 0.05, late versus immediate; p < 0.05, late Ag versus late saline; p < 0.05; late Ag versus immediate saline.

Leukocyte Distribution Before and After Airway Allergen Challenge

In the blood, there was a significant increase in the absolute numbers of eosinophils (from 247/mm³ to 467/mm³) at 48 h after SBP (Table 2). There were no significant changes in the relative distribution of other cell populations. Forty-eight hours after SBP, eosinophils were selectively recruited into the Ag-challenged airway segment, and accounted for 35% of the total cell population. There was a small but significant increase in polymorphonuclear leukocytes (PMN) in both the saline- and Ag-challenged segments. The increase in granulocytes (eosinophils or PMN) resulted in a corresponding reduction in the percent of alveolar macrophages (AM) in both segments. The relative number of lymphocytes in the saline and Ag segments was similar and did not change following bronchoprovocation. As seen in Table 3, there were no significant variations in the relative proportions of B cells (CD19⁺), NK cells (CD3/CD16⁺CD56⁺), and T cells (CD3⁺) in the blood or BALF following SBP. In addition, there was no change in the overall distribution of CD4⁺ cells or in the relative numbers of CD4⁺ cells that expressed IL-2 receptors. There was a small but significant decrease in the proportion of CD8⁺ T cells present in the Ag-challenged segment at 48 h.

Effect of Allergen Challenge on the Functional Properties of Airway T Cells

Spontaneous secretion of cytokines from PBMC or BAL cells was not detectable. Following ex vivo stimulation with PHA, PBMC secreted IL-2, IL-4, IL-5, IL-10, GM-CSF, and IFN- (Figure 2). The cytokine profile did not change following allergen SBP. Airway cells obtained immediately after saline-SBP secreted large amounts of IFN- and IL-2, with little or no IL-4, IL-5, IL-10, or GM-CSF. At the 48-h time point, there was a small increase in IL-5 production in the saline segment. In contrast, cells obtained 48 h after allergen SBP showed a large increase in the capacity to produce IL-5, and smaller but statistically significant increases in the production of IL-4, IL-10, and GM-CSF. The ability of cells to produce IFN- or IL-2 was not significantly altered after Ag SBP.

View larger version (28K): [in this window] [in a new window]   Figure 2.   Effect of allergen challenge on the cytokine profile of ex vivo-stimulated airway T cells. Data are expressed picograms of cytokine per milliliter of culture supernatant fluid, and indicate the amount of cytokine secreted into the supernatant fluid by 2 million cells after a 48-h culture period. Bars represent medians within quartiles of 25 and 75; error bars represent 10th and 90th percentiles. *p < 0.05, versus immediate; p < 0.05, late Ag versus late saline; p < 0.05, late Ag versus immediate saline.

IL-5 Production by CD4⁺ BAL T Cells

Two methods were used to determine whether CD4⁺ BAL cells were a source of IL-5. First, IL-5 release from CD4⁺ cells was measured after enrichment or depletion of CD4⁺ cells with paramagnetic beads. Secretion of IL-5 was detected in the CD4-enriched but not CD4-depleted cultures (Figure 3A). Second, IL-5-positive cells were identified by intracellular flow-cytometric analysis. The majority of cells that expressed intracellular IL-5 were within the CD4⁺ population (Figure 3B).

View larger version (30K): [in this window] [in a new window]   Figure 3.   Identification of CD4+ BAL T cells as a source of IL-5. (A) CD4 cells were enriched (hatched bars) or depleted (dark-shaded bars) by paramagnetic bead selection. Data are expressed as picograms of cytokine per milliliter of culture supernatant fluid, and indicate the amount of cytokine secreted into by 1 million cells after a 48 h stimulation. (B) Cells expressing intracellular IL-5 were characterized by flow-cytometric analysis. Two-color dot plots were generated after gating on the lymphocyte population. Statistical gates were set on the basis of the isotype control antibody. The upper right-hand quadrant shows IL-5-positive cells within the CD4+ population. CD4 cells are represented in the left-hand quadrant.

Relationships Among Ex Vivo Production of Cytokines by BAL Cells, IL-5, and Eosinophils in BALF

There was a close correlation between the release of IL-5 into BALF and airway eosinophilia (Figure 4A). Likewise, the ex vivo generation of IL-5 in response to PHA was significantly correlated with BAL eosinophils (Figure 4B). Interestingly, the amount of IL-5 secreted in vitro was proportional to IL-5 levels present in the BALF (Figure 4C). Weaker correlations were also noted between BAL eosinophils and PHA-induced IL-4 (r = 0.64, p = 0.03) and GM-CSF (r = 0.64, p = 0.02). However, the amount of IL-4 and GM-CSF produced by ex vivo-stimulated BAL cells was smaller than that of IL-5 (Figure 2), and in some subjects, IL-4 and/or GM-CSF could not be detected in the supernatant of stimulated BAL cells. Furthermore, the levels of IL-4 and GM-CSF present in vivo in BALF were at or below the level of detection in all subjects (data not shown).

View larger version (15K): [in this window] [in a new window]   Figure 4.   Correlation of PHA-induced IL-5 with IL-5 and eosinophils in BALF. The number of BAL eosinophils (expressed as eosinophils per milliliter of BAL fluid recovered) correlated with levels of IL-5 in BALF (r = 0.76, p < 0.001) (A) and with PHA-induced IL-5 production by airway cells obtained from the antigen-challenged segment (r = 0.78, p < 0.001) (B). IL-5 levels in the BALF correlated with PHA-induced IL-5 production by airway cells obtained from the antigen-challenged segment (r = 0.84, p < 0.001) (C ). Data represent events in the allergen-challenged airway before (open circles) and after (closed circles) SBP. Note that IL-5 was measured in 20×-concentrated BALF, and data are expressed as the calculated amount of IL-5 in 1× BALF. Values  0.25 pg/ml are considered negative, based on a cutoff of 5 pg/ml measured in 20×-concentrated fluids.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Bronchoscopy with segmental bronchoprovocation was used to determine the effects of allergen on cytokine release from airway lymphocytes. We found that local deposition of allergen in the airway resulted in the release of IL-5 into the lavage fluid, influx of eosinophils, and elicitation of a lymphocyte population with the capacity to secrete IL-5 when stimulated ex vivo with T-cell mitogen. Interestingly, the amount of IL-5 secreted in response to PHA stimulation closely paralleled the level of IL-5 in the BALF and correlated with the number of eosinophils present within the airway. Our findings suggest that lymphocytes present within the airway following allergen challenge have the capacity to contribute to eosinophilic airway inflammation through the secretion of IL-5.

CD4⁺ cells appear to be a principal source of IL-5 in the airway following allergen challenge. IL-5 secretion could be stimulated in purified CD4⁺ cells, and depletion of CD4⁺ cells from cell cultures ablated IL-5 production. It has been reported that CD8⁺ cells (22), NK cells (23), mast cells (24), and eosinophils (25) can also express IL-5 under the appropriate conditions. In our study, evaluation of the lymphocyte population showed that the majority of cells that expressed intracellular IL-5 were CD4⁺ lymphocytes. Even though CD4⁺ cells were a major source of IL-5, the increased capacity of BAL cells to produce IL-5 following allergen provocation did not appear to be due to relative changes in lymphocyte populations. In ex vivo-stimulated cultures, equal numbers of lymphocytes, including CD3⁺, CD4⁺, and NK cells from the saline and Ag segments were present both immediately and 48 h after allergen challenge. Furthermore, similar numbers of CD4⁺ cells expressing CD25 (low-affinity IL-2 receptor) were cultured at both time points.

In addition to the large increase in IL-5 production, BAL cells from the allergen-challenged airway also secreted small but increased amounts of IL-4, IL-10, and GM-CSF when obtained at 48 h as compared with 5 min after challenge. The low levels of these cytokines which, like IL-5, have been associated with allergic inflammation (26), raises important points. First, IL-4 levels were much lower than those of another Th2-associated cytokine: IL-5. We suspect that the majority of secreted IL-4 was not detectable because of its binding to IL-4 receptors expressed on cells within the cultures. In a pilot experiment, stimulation of cells in the presence of anti-IL-4 receptor antibody, increased IL-4 levels from 38 pg/ml to 300 pg/ml in PBMC cultures, and from 8 pg/ml to 90 pg/ml in BAL cell cultures (data not shown). Second, although PBMC secreted high levels of GM-CSF and IL-10 in response to T-cell mitogen, production of these cytokines by BAL cells was high only when induced with lipopolysaccharide (LPS). The median levels of IL-10 and GM-CSF following LPS stimulation of BAL cells obtained at 48 h after AG-SBP were 1,000 pg/ml and 15,000 pg/ml, respectively (data not shown). These data suggest that in contrast to the blood, macrophages/monocytes, and not T cells, are a principal source of GM-CSF and IL-10 in BALF.

It has been proposed that allergic airway inflammation is a Th2-associated response, characterized by production of IL-4 and IL-5 with a reduction in Th1-associated cytokines, IFN-, and IL-2 (30). Our study shows that prior to allergen challenge, airway cells were predominantly Th1-like and had the capacity to produce IFN- and IL-2, but not IL-4 or IL-5. This cytokine profile is consistent with previous reports that airway cells from normal healthy subjects produce primarily IFN- and IL-2 (31). Following exposure to antigen, cells in our study showed cytokines associated with a Th2 profile (increased production of IL-4, IL-5, and IL-10). Interestingly the "switch" to a Th2-like phenotype was not absolute; cells that could secrete Th1-associated cytokines remained within the airway after allergen challenge. The interaction of these Th1- and Th2-like populations, and the possible contribution of Th1-type cells to eosinophilic airway inflammation, is not yet established.

Our observation that airway cells obtained after allergen challenge have an increased capacity for IL-5 production parallels previous reports. Studies by our group (7, 8, 11) and others (9, 10) have shown that IL-5 levels in BALF are increased after allergen challenge. In some studies, GM-CSF (8, 10, 26) and IL-4 (28) have also been detected in the BALF. In addition, mRNA for IL-4, IL-5, IL-10, and GM-CSF has been demonstrated in mucosal biopsies (12, 32) and BAL cells (13, 29) of allergic asthmatic individuals following allergen inhalation. T-cell lines that secrete IL-4, IL-5, and GM-CSF have been derived from BAL after allergen exposure (33, 34). Recently, antigen-specific production of IL-5 by CD4⁺ T cells was shown to be increased in PBMC obtained from atopic asthmatic individuals as compared with atopic nonasthmatic or normal subjects (35). Our study is unique, however, in that we used ex vivo stimulation with a T-cell mitogen to characterize cytokine secretion from freshly cultured BAL cells. To our knowledge, this is the first report of selective cytokine profiles of PHA-stimulated airway cells that relate to in vivo events observed after antigen challenge.

Nonspecific mitogenic stimulation is an important feature of our study. PHA stimulation allowed amplification of cytokine signals that could not be detected in unstimulated airway-cell cultures. Furthermore, nonspecific stimulation enabled us to determine the overall capacity of cells within the airway to produce cytokines, and to characterize the potential cytokine profile of these cells. Thus, we demonstrated that resident airway cells (obtained after saline-SBP) did not secrete IL-5 even when activated with a strong, nonspecific stimulus. This observation suggests that following allergen challenge, IL-5-positive cells are not merely expanded from a resident lymphocyte population, but rather are actively recruited into the airway or undergo antigen-specific changes that allow IL-5 to be produced.

It is of interest that the allergen-induced changes in IL-5 production seen in our study were primarily limited to the site of allergen exposure. The amount of IL-5 secreted by PBMC did not change after Ag-SBP. Moreover, within the lung, changes in lymphocyte function were mostly localized to the site of antigen challenge; only slight increases in IL-5 were seen in the saline-challenged segment. The minimal increase in eosinophils or IL-5 in the saline-challenged segment has been noted previously (7), and could be related to minimal cross contamination of the sham-challenged segment with Ag used in the Ag-SBP segment, or to a systemic effect of locally instilled Ag. The mechanisms that led to selective recruitment of IL-5-positive cells to the site of antigen deposition have not been elucidated. Involvement of chemokines such as RANTES, macrophage inhibitory protein-1 (MIP-1), and IL-16 appears likely, but this question remains to be explored.

Although the design of our study has many advantages, we recognize that it also has some limitations that need to be discussed. First, it could be argued that mucosal airway cells are more important in directing airway inflammation, and that BAL cells might not be representative of the mucosal cell population. Nonetheless, our study demonstrates that BAL cells have the potential to secrete IL-5 and may therefore contribute to cytokine release into the airway. Second, PHA stimulation is not a physiologic stimulus and thus does not allow direct inference of in vivo production of cytokines. The close correlation between the level of IL-5 secreted after ex vivo stimulation with PHA and in vivo release of IL-5 into the airway indicates that PHA stimulation does reveal events that are relevant to the airway response in vivo. Furthermore, nonspecific stimulation of airway cells in vitro may be relevant to in vivo events at the site of inflammation, where populations of memory cells may be induced to release cytokines in response to nonspecific stimuli, such as extracellular matrix proteins and adhesion molecules. Finally, it could be argued that local deposition of a relatively large dose of allergen through SBP is not equivalent to allergen exposure via the inhalation route. Although the magnitude of the inflammatory response is certainly greater after SBP, the components of the response (i.e., airway eosinophilia) and the mechanisms that lead to inflammation (i.e., IL-5 production) appear to be comparable regardless of allergen dose (11) or mode of challenge (18). SBP was chosen for our study to allow simultaneous evaluation of saline- versus allergen-induced changes within an individual subject.

In summary, we have used the combination of SBP with saline or antigen and ex vivo stimulation with mitogen to determine the capacity of airway cells to secrete cytokines. Our study demonstrated that allergen challenge induced a population of lymphocytes that, unlike resident airway cells, had the capacity to secrete IL-5. The biologic significance of our model is suggested by the close correlation between the increased capacity of airway cells to produce IL-5 in vitro, the in vivo release of IL-5 into the BALF, and the recruitment of eosinophils into the airway. Our data also suggest that CD4⁺ cells are an important source of IL-5 in the airway.