INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Antigen-induced inflammatory responses have been implicated in the pathogenesis of a variety of allergic diseases, including asthma. When allergic asthmatics are exposed to an appropriate antigen, they develop an acute IgE-dependent early response, which is often followed by an inflammatory late airway response (LAR) 4 to 12 h later (1). Several investigators have studied the relationship between the LAR and airway inflammation after antigen challenge in humans and in animals. A prominent feature of the LAR was infiltration of the airways by CD4⁺ cells, basophils, and eosinophils (2, 3). The presence of inflammatory cells and their inflammatory products in the lung often correlated with disease severity and the degree of the LAR (4). Furthermore, activated CD4⁺ T lymphocytes expressed elevated levels of messenger RNA (mRNA) for T helper cell, type 2 (Th2) cytokines interleukin-4 (IL-4) and IL-5, that promoted IgE production by B cells as well as eosinophil chemotaxis, activation, and survival (5).

In contrast to the well-defined role of the CD4⁺ T-cell subset, the role of CD8⁺ T cells in the development of the LAR is less clear. Recent studies have established that CD8⁺ T cells can have a regulatory role in certain immune responses (8). These effects are not due to cell cytotoxicity but to the release of certain mediators that may be responsible for such responses. CD8⁺ T cells, like CD4⁺ T cells, can be differentiated into either T helper type 1 (Th1) cells that release interferon gamma (IFN-) and IL-2, or Th2 cells that release IL-4 and IL-5 (9). The differential development of these cell subsets is a major determinant of the outcome of physiological and pathological immune responses in autoimmune, allergic, and infectious diseases. A relative deficiency of cytotoxic/suppressor T lymphocytes has been reported in patients with allergic rhinitis (10) and in atopic dermatitis (11). Furthermore, cytotoxic/suppressor T-cell dysfunction may be associated with pathogenesis of atopic and nonatopic asthma through the production of cytokines (12).

T cell-produced cytokines are not the only mediators that are involved in inflammatory cell recruitment. Eosinophils and other leukocytes are directly drawn into tissues by local production of chemokines that act primarily as chemoattractants and leukocyte activators (13, 14). Chemokines are divided into four groups, CXC, CC, C, and CX3C, based upon spacing of conserved cysteine residues. CC group is the most active on eosinophils, monocytes, lymphocytes, natural killer (NK) cells, and basophils (13). This group includes RANTES (regulated upon activation, normal T-cell expressed and secreted) and eotaxin, which have been shown to induce migration of eosinophils in vitro and to be involved in lung eosinophilia in humans, guinea pigs, and mice (15). Furthermore, eotaxin seems to be a link between T-cell activation and recruitment of eosinophils into airways (18). RANTES also affects eosinophils and is a strong chemotactic factor for T lymphocytes in vivo and in vitro (19). Monocyte chemoattractant protein-1 (MCP-1) is active on monocytes, T cells, NK cells, basophils, mast cells, and dendritic cells and is also strongly expressed in the airways during inflammation (13, 20).

The present study was performed to better delineate the immunoregulatory role of CD8⁺ cells in the pathophysiology of allergic asthma. Sprague-Dawley (SD) rats were studied as they are low IgE producers that do not develop LAR after sensitization and antigen challenge unless their CD8⁺ cells have been depleted (21). The LAR was enhanced by prior in vivo depletion of CD8⁺ cells, and the effects of this intervention on cellular inflammation and in vivo chemokine and cytokine mRNA production were assessed. The results suggested that CD8⁺ cells are involved in LAR and pulmonary inflammation through direct or indirect effects on chemokine and cytokine production.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animals

The study was approved by the animal ethics committees of McGill University and the CHUM Research Center (Montreal, PQ, Canada). Twenty-two male SD rats 6 to 8 wk of age and weighing 250 to 325 g were obtained from Charles River Canada Inc. (Montreal, PQ, Canada). Rats were maintained in animal facilities with filtered air.

Sensitization and Ovalbumin Challenge

Active sensitization was performed by subcutaneous injection of 1 ml of saline containing 1 mg of chicken egg ovalbumin (OA) (Sigma, St. Louis, MO) and 3.5 mg of aluminum hydroxide gel (BDH Chemicals, Poole, UK). Bordetella pertussis vaccine (0.5 ml) (IAF BioVac, Laval, PQ, Canada) containing 2 × 10¹⁰ heat-killed organisms was administered intraperitoneally as an adjuvant. Rats underwent aerosol challenge with OA (5% wt/vol) 14 d after sensitization. The equipment and methodology for measuring pulmonary resistance was as previously described (22).

Monoclonal Antibody Administration

On Days 8, 10, and 12 after sensitization, animals (n = 22) received either 1 mg of anti-CD8 (OX-8) monoclonal antibody (mAb) (n = 11) or 1 mg of BALB/c mouse ascites IgG₁ mAb (n = 11) obtained from pristaned mice injected with SP2/0 myeloma cells (Cedarlane Laboratories, Hornby, ON, Canada). The OX-8 antibody was purified from a hybridoma cell line as previously described (21).

Measurement of Airway Responses

Animals were anesthetized by intraperitoneal injection of urethane (1.1 g/kg) and were orotracheally intubated. Lung resistance (R_L) was determined as described (22) by fitting the equation of motion of the lung using multiple linear regression analysis and a commercial software package (RHT-InfoDat Inc., Montreal, PQ, Canada). To determine the early airway response (EAR) and the LAR, lung resistance was measured at the following times: at baseline, after a 5-min saline aerosol administration, immediately after the administration of a 5% (wt/vol) aerosol of OA, at 5, 10, and 15 min after OA, and subsequently at 15-min intervals, for a total period of 8 h. The EAR was derived from the highest R_L measurement between t = 0 and t = 60 min, and expressed as a percentage of the baseline. The LAR was calculated as the area under the curve for R_L versus time, between t = 3 and t = 8 h. The LAR was considered significant if its calculated value exceeded 10.

Isolation and Staining of Blood Lymphocyte Subsets

Mononuclear cells were obtained 8 h after antigen challenge from freshly heparinized peripheral blood of rats after centrifugation over Ficoll-Hypaque, and were prepared for flow cytometry as described (21). Indirect staining was performed by incubation with the mAb W3/25, OX-19, and OX-8 (Dimension Laboratories, Mississauga, ON, Canada), which recognize CD4, CD5, and CD8 surface antigens, respectively. Controls consisted of cells stained with mouse ascites IgG₁ antibody. After washing, a second incubation was performed with fluoroescein isothiocyanate (FITC)-conjugated rabbit anti-mouse IgG antibody. Stained cells were fixed in 1% paraformaldehyde and analyzed using FACScan and PC-LYSYS software (both from Becton Dickinson Canada Inc., Montreal, PQ, Canada).

Preparation of the Lungs

All animals were exsanguinated 8 h after OA challenge and the pulmonary vasculature was flushed with balanced salt solution until the lungs were white (± 10 ml). For in situ hybridization, the left lung was fixed by perfusion of freshly prepared 4% paraformaldehyde in phosphate-buffered saline (PBS) (pH 7.4) for 30 min at a pressure of 25 cm H₂O and then incubated overnight in 15% (wt/vol) sucrose in PBS. The right lung was frozen in liquid nitrogen and reserved for the preparation of total RNA.

Histologic Analysis

Lung sections (7 µm) were stained with a RAL kit (Fisher Scientific Inc., Montreal, PQ, Canada) according to the manufacturer's instructions.

Immunocytochemistry for Eosinophils and Eotaxin

Cryosections (7 µm) of lung tissue were immunostained using the alkaline phosphatase (AP)-anti-AP method, as reported (23), with some modifications. Sections were fixed with methanol-acetone at 20° C for 10 min. After treatment with universal blocking solution (DAKO Diagnostics Canada Inc., Mississauga, ON, Canada) for 15 min, slides were incubated with 10 µg/ml of anti-human major basic protein (MBP) (BMK-13, kindly provided by Dr. Redwan Moqbel, University of Alberta) or anti-eotaxin mAb (R&D Systems, Minneapolis, MN) for 45 min at room temperature. Slides were then incubated for 30 min at room temperature with a rabbit anti-mouse IgG and positive staining was visualized with AP/anti-AP complex and Fast Red-TR substrate containing levamisole (Sigma) to inhibit endogenous AP. Sections were counterstained with hematoxylin and mounted in Crystal Blue (Biomeda Corp., Foster City, CA). Negative controls (no primary antibodies) were included in all studies.

In Situ Hybridization

Cryostat blocks containing embedded lung tissue were cut into 7-µm sections on superfrost slides. Sections were incubated at 37° C for 12 h and either processed immediately or stored at 80° C until used. The sections were hybridized with ³⁵S- or digoxigenin-labeled antisense and sense probes coding for rat IL-4, IL-5, and IFN-, as previously described (24). Antisense and sense riboprobes were synthesized using sulfur-35-uridine triphosphate (S³⁵-UTP) or digoxigenin-11-UTP and SP6, T3, or T7 polymerases (Promega, Montreal, PQ, Canada). Labeled probes were digested by alkaline hydrolysis to an average length of 100 to 200 bases before precipitation. Preparations were acetylated in 0.1 M triethanolamine for 5 min and then in acetic anhydride, 0.25% triethanolamine for 10 min. Sections were permeabilized with proteinase K and prehybridized with 50% formamide and 5× Denhardt's solution before application of the probes. Nonspecific binding was removed by posthybridization washing under high-stringency conditions and subsequent treatment with ribonuclease A (RNase A) (20 µg/ml; Boehringer Mannheim, Montreal, PQ, Canada). For slides with radiolabeled probes, hybridization signals were visualized by dipping slides into photographic emulsion (Amersham LM-2) and exposing for a period of 10 d. Slides were developed in Kodak D-19 developer and fixed and counterstained with hematoxylin. Slides with digoxigenin-UTP-labeled probes were treated with anti-digoxigenin AP conjugated antibody (Boehringer Mannheim) and processed using NBT-X-Phosphate substrate (Sigma, Oakville, ON, Canada). Antisense signal specificity was confirmed by the absence of sense probe signals.

Quantification of Cytokine and Chemokine Positive Cells

For immunostaining and in situ hybridization, slides were coded and positive cells were counted without knowing from which group they were obtained. Sections of the airways were cut in such a way that the axial pathway could be defined and only such sections were read. Both small and large airways (with a perimeter of basement membrane [BM] between 0.1 mm and 2 mm) were studied. Positive cells were counted only in the subepithelial area beneath the BM to a depth of 0.45 mm by using an eyepiece graticule (0.01-mm graticule calibration). Results were expressed as the mean number of positive cells per mm of BM.

RNA Preparation and Reverse Transcription

TRIzol reagent (Gibco BRL, Montreal, PQ, Canada) was used as a monophasic solution to homogenize tissue and to isolate total RNA from frozen biopsies according to the manufacturer's instructions. Reverse transcription was performed on 5 µg of total RNA with Moloney murine leukemia virus (M-MLV) reverse transcriptase (Gibco BRL) in the presence of RNase inhibitor (RNasine) (Pharmacia, Montreal, PQ, Canada). Polymerase chain reaction (PCR) was performed using an automatic thermal cycler (MJ Research Inc., Ottawa, ON, Canada). Cycle conditions were 94° C for 1 min, 60° C for 2 min, 72° C for 3 min. Complementary DNA (cDNA) (2 µl) was amplified in a 25-µl reaction volume containing 0.5 µM (each) deoxyribonucleoside triphosphates (dNTPs), 0.5 µM primers, and 1 U Taq polymerase (Gibco BRL).

Semiquantitative Reverse Transcriptase/Polymerase Chain Reaction (RT-PCR)

For semiquantitative experiments, PCR was set up as described previously, except that the reaction mixture contained 5 µCi/ml of radioactive phosphorus-deoxyadenosine triphosphate ([³²P]dATP) as a tracer. Specific primers were used to amplify selected cytokine and chemokine messages (24, 25). Preliminary experiments determined the optimal number of cycles for each primer, which were as follows: cyclophilin (a housekeeping gene), 19 cycles; IL-4, IL-5, and IFN-, 30 cycles; eotaxin and RANTES, 23 cycles; MCP-1, 27 cycles. These cycle numbers were selected as midpoints of their respective linear ranges for amplification of cDNA (2 µl), and there was a linear correlation between input cDNA and the yield of PCR products. Quantities of cDNA were standardized to yield equivalent amounts of PCR products for cyclophilin and compared with each other. To determine the relative mRNA expression of each cytokine and chemokine present in different samples, 20 µl of the amplified product was electrophoresed through a polyacrylamide gel, containing 5% urea, in Tris-acetate/ethylenediaminetetraacetic acid (EDTA) (TAE) buffer. The gels were dried and exposed overnight at 80° C using autoradiography film (Kodak, Rochester, NY). The radioactive signal-specific bands were quantified by an Instant Imager System 2000 (Alpha Innotech Corporation, San Leandro, CA). The relative amount of mRNA synthesized for each cytokine and chemokine was calculated using the intensity of radioactive signal (RS) of specific bands for each cytokine, chemokine, and cyclophilin, as follows: relative amount of cytokine or chemokine mRNA = (RS for cytokine or chemokine in sample)/(RS for cyclophilin in sample).

Statistical Analysis

All results are expressed as mean values for the group of animals analyzed ± 1 SEM. Groups were compared using the Student t test when there was a normal distribution or the Mann-Whitney U test (for the late airway response), with a value of p < 0.05 taken as an indicator of statistical significance.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Depletion of CD8⁺ Cells In Vivo

Administration of OX-8 mAb specifically depleted CD8⁺ cells in rats. The status of depletion of CD8⁺ cells was determined by fluorescent-activated cell sorter (FACS) analysis of mononuclear cells obtained from control and test groups 8 h after OA challenge (Figure 1). The results of these analyses revealed that blood helper lymphocytes (CD4⁺) did not differ significantly in the OX-8-treated animals when compared with control animals (34.5% ± 4.3% versus 33.6% ± 2.6%). In contrast, blood CD8⁺ lymphocytes were significantly reduced in OX-8-treated versus control rats (1% ± 2.7% versus 27.7% ± 3.2%; p < 0.001). To ascertain the adequacy of CD8⁺ cell depletion, immunofluorescent staining was also performed with anti-CD5 antibody (OX-19) which recognizes T- and B-cell subsets. As shown in Figure 1, similar immunostaining profiles with anti-CD5 antibody were obtained in both groups (41.5% ± 2.5% versus 47.4% ± 1.8%), showing that CD8 depletion was specific. No significant difference was observed in the immunostaining using control mAb in both groups of animals.

View larger version (20K): [in this window] [in a new window]   Figure 1.   Flow cytometry analysis of immunofluorescent staining of blood lymphocytes obtained from OX-8 (A) and control IgG1-treated (B) SD rats. Representative histograms of cells stained with W3/25, OX-19, and OX-8 mAb, which recognize CD4, CD5, and CD8 surface antigens, respectively, are shown. Immunofluorescent staining with a control anti-mouse IgG1 mAb is included as well. Peripheral blood mononuclear cells were collected from rats treated with anti-CD8 or IgG1 control mAb 8 h after antigen challenge (see METHODS). Numbers refer to the percentage of positive cells for each cell surface marker for the individual experiment shown.

Effect of CD8 Depletion on Airway Responses after Antigen Challenge

Lung resistance measurements were performed for 8 h on 11 control and 11 CD8-depleted rats after OA challenge. The baseline R_L for OX-8 treated animals was 0.186 ± 0.03 cm H₂O · ml¹ · s, which was not significantly different from control animals (0.175 ± 0.02 cm H₂O · ml¹ · s). The EAR to OA challenge was observed in four of 11 animals in both groups. Overall, the mean maximal percent of baseline R_L during the first 60 min after OA challenge was not significantly different from controls (174.6% ± 5.6% in CD8-depleted rats versus 161.7% ± 6.3% in control animals). The LAR as measured by the area under the R_L versus time curve was significantly increased in the OX-8-treated group (Figure 2, 16.65 ± 6.6 in CD8-depleted rats as compared with 5.39 ± 2 in controls, p < 0.036). The LAR to OA challenge was observed in seven of 11 OX-8-treated animals, compared with two of 11 IgG₁-treated control rats.

View larger version (12K): [in this window] [in a new window]   Figure 2.   Effect of treatment with OX-8 (anti-CD8 mAb; n = 11) or with IgG1 (control mAb; n = 11) on the LAR to antigen challenge in SD rats. The LAR was measured as the area under the curve of RL (cm H2O · ml1 · s) versus time (min) between 3 and 8 h after OA challenge. The values of RL were corrected by subtracting the baseline value. Means for each group are indicated with horizontal lines. p < 0.04 in control versus OX-8-treated animals.

Antigen-induced Pulmonary Eosinophilia in CD8-depleted Rats

Histological analyses of lung sections revealed distinct differences between OX-8-treated rats and control rats. Lungs from control rats contained significantly fewer eosinophils around the small and large airways (with a perimeter of BM longer than 0.1 mm and less than 2 mm) than in CD8-depleted rats (Figure 3).

View larger version (117K): [in this window] [in a new window]   Figure 3.   Effect of treatment with anti-CD8 (a, c) or control IgG1 (b, d ) mAb on eosinophil influx 8 h after antigen challenge. Sections were stained with a RAL kit (see METHODS). Note significantly more eosinophils around the airways after OA challenge in CD8-depleted rats compared with control animals. Sections shown are from a representative experiment viewed at ×20 (a, b) and ×100 magnification (c, d ).

Immunostaining with a mAb against eosinophilic granule MBP was simultaneously performed to obtain more sensitive and specific staining of rat eosinophils versus RAL kit staining. MBP-positive cells were counted in the airway submucosa and expressed as the number of positive cells per millimeter of airway BM, as shown in Figure 4. CD8-depleted rats had significantly more MBP-positive cells in airways after antigen challenge than did control rats (CD8-depleted rats: 15.2 ± 4.2 positive cells/mm, n = 6, versus controls: 4.6 ± 1.2 positive cells/ mm BM, n = 6; p < 0.004).

View larger version (12K): [in this window] [in a new window]   Figure 4.   Immunohistochemical analysis of MBP-positive cells in lung tissue sections. Results are expressed as the number of positive cells per mm of BM. Means for each group are indicated with horizontal lines. The differences in the numbers of MBP-positive cells between CD8-depleted and control animals were statistically significant (p < 0.004).

Antigen-induced Chemokine Production in CD8-depleted Rats

Using semiquantitative RT-PCR, we assessed the expression of mRNA for eotaxin, RANTES, and MCP-1 in RNA extracted from the lungs of OX-8-treated and control rats. Eotaxin, RANTES, and MCP-1 mRNA were constitutively expressed in the lungs of animals from both groups (Figure 5A). However, eotaxin mRNA was expressed significantly more in the lungs of CD8-depleted rats (Figure 5B; p < 0.01). There was no significant difference in the expression of mRNA for RANTES and MCP-1 in the RNA extracted from the lungs of rats from both groups (Figure 5B).

View larger version (27K): [in this window] [in a new window]   View larger version (53K): [in this window] [in a new window]   Figure 5.   Semiquantitative RT-PCR assessment of chemokine mRNA expression in the lungs of control IgG1 (n = 11) and anti-CD8-treated (n = 11) SD rats. (A) Representative autoradiographs taken from six rats per group chosen at random demonstrate the expression of eotaxin, RANTES, and MCP-1 mRNA in control (n = 6; lanes 1 to 6) and CD8-depleted (n = 6; lanes 7 to 12) SD rats. Cyclophilin was used for standardization. (B) Semiquantitative comparison of integrated density values of the radioactive signal of specific bands for each chemokine as a percent of the cyclophilin-specific signal from airways of all CD8-depleted (n = 11; hatched bars) and all control (n = 11; empty bars) SD rats. *p < 0.01 CD8-depleted versus control animals.

Because eotaxin is one of the most potent eosinophil chemoattractants, it was of interest also to determine whether cells in CD8-depleted rats would express eotaxin protein. Airways from CD8-depleted rats showed significantly increased numbers of eotaxin-immunoreactive cells in the subepithelial region (Figure 6, CD8-depleted rats: 14.63 ± 1 positive cells/mm BM, n = 11; controls: 3.27 ± 0.5 positive cells/mm BM, n = 11; p < 0.001).

View larger version (91K): [in this window] [in a new window]   View larger version (13K): [in this window] [in a new window]   Figure 6.   Immunohistochemical analysis of eotaxin-positive cells in lung tissue sections. (Top panels) Expression and cellular distribution of eotaxin protein in lung tissue sections from CD8-depleted (a) and control (b) rats after antigen challenge. Note significantly more eotaxin-positive cells around the airways after OA challenge in CD8-depleted rats compared with control animals. Sections shown are from a representative experiment viewed at ×20 magnification. (Bottom left panel ) The numbers of eotaxin-positive cells per mm of BM in CD8-depleted (n = 11) and control (n = 11) rats. Individual data points are shown along with group mean values (horizontal lines). The differences between two groups were significant (p < 0.001).

Antigen-induced Cytokine Production in CD8-depleted Rats

Because cytokines may also be involved in inflammation and the LAR, we assessed whether CD8 depletion affected IL-4, IL-5, and IFN- production 8 h after antigen challenge by performing semiquantitative RT-PCR on total lung RNA. Figure 7A shows the representative autoradiographs of mRNA expression for each cytokine from six randomly chosen control and CD8-depleted animals. As shown in Figure 7B, only the Th1 cytokine IFN- was significantly decreased in the lung tissue of CD8-depleted rats (n = 11) after antigen challenge when compared with control IgG₁-treated animals (n = 11, p < 0.04). There were no significant differences in IL-4 and IL-5 mRNA expression in OX-8-treated versus control IgG₁-treated rats.

View larger version (24K): [in this window] [in a new window]   View larger version (48K): [in this window] [in a new window]   Figure 7.   Semiquantitative RT-PCR assessment of IL-4, IL-5, and IFN- mRNA expression in the lungs of control IgG1-treated (n = 11) and OX-8-treated (n = 11) SD rats. (A) Representative autoradiographs taken from six rats per group chosen at random demonstrate the expression of IL-4, IL-5, and IFN- mRNA for control (n = 6) and CD8-depleted (n = 6) animals. Semiquantitative RT-PCR for these cytokines was performed using total RNA purified from the lungs of control IgG1-treated (lanes 1 to 6 ) and CD8- depleted (lanes 7 to 12) SD rats 8 h after antigen challenge. Cyclophilin was used as a housekeeping gene for standardization. (B) Quantitation of IFN-, IL-4, and IL-5 mRNA expression in the lung of all OX-8 (n = 11) or all control IgG1 (n = 11) mAb treated SD rats by semiquantitative RT-PCR. Quantitation was done using the Instant Imager System 2000, normalizing integrated density values of the radioactive signal of specific bands for each cytokine (expressed as mean ± SD) to the values obtained by quantitation of the cyclophilin products from the airways of control (empty bars) and CD8-depleted (hatched bars) SD rats. *p < 0.04 control versus CD8-depleted rats.

The expression of IL-4, IL-5, and IFN- mRNA in rat lungs was further evaluated by in situ hybridization using digoxigenin-labeled or [³⁵S]UTP-labeled riboprobes. The airways from CD8-depleted rats had significantly less cells that expressed IFN- mRNA (Figure 8; p < 0.001). We also found no difference in hybridization signals for IL-4 and IL-5 between both groups of rats (Figure 8). The sense probe controls were negative (data not shown).

View larger version (123K): [in this window] [in a new window]   View larger version (17K): [in this window] [in a new window]   Figure 8.   Assessment of cytokine mRNA expression by in situ hybridization in the airways of OX8- or IgG1-treated SD rats. (Left panels) Representative examples of in situ hybridization of sections obtained from CD8-depleted (a, c, e) and control (b, d, f  ) SD rats, respectively. Lung sections were hybridized with digoxigenin-labeled antisense probes coding for rat IFN- mRNA (a, b), IL-4 (c, d ), and IL-5 (e, f  ) mRNA. (Top right panel ) Quantitation of IFN-, IL-4, and IL-5 mRNA expression in the lung of OX-8 (hatched bars, n = 11) or control IgG1 (empty bars, n = 11) mAb treated SD rats by in situ hybridization. *p < 0.001.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We previously reported that depletion of CD8⁺ T cells before antigen challenge significantly increased the LAR in SD rats (21). The current study assesses factors, such as chemokines and cytokines, that may contribute to the regulation of antigen-induced airway responses in CD8-depleted SD rats. As previously reported, CD8-depletion before antigen challenge led to a significant enhancement of the LAR. It also caused an enhancement of the inflammatory response and an accumulation of a large number of eosinophils in the airways. In addition, there was an increase in eotaxin, a downregulation of IFN- mRNA, and no detectable effect on IL-4, IL-5, RANTES, and MCP-1 mRNA expression in CD8-depleted rats after OA challenge. These findings suggest a mechanism involving CD8⁺ cells that may be important in preventing the LAR and the inflammatory reaction that occurs in allergic asthma.

The pathogenesis of the LAR has been closely linked to airway inflammation, including eosinophilia. It has been shown that the LAR in asthmatic subjects is associated with an eosinophil-rich influx of inflammatory cells into the bronchial lumen (26). Eosinophils play an important role in airway inflammation by releasing toxic granular proteins and membrane-derived products that induce epithelial damage, mucous secretion, cellular recruitment, and subsequently intensify the LAR (27). Antigen challenge in sensitized guinea pigs (28), rats (29), and dogs (30) also showed marked LAR-associated eosinophilia. To elucidate the mechanisms by which CD8⁺ cells may be involved in the LAR and airway inflammation, we assessed whether they affected eosinophil influx after antigen challenge. Selective depletion of CD8⁺ cells significantly amplified the inflammatory response and promoted eosinophil accumulation within the airways of sensitized animals. This finding was consistent with a previous report that demonstrated a significant eosinophil influx into the lung tissues of CD8-depleted rats (31). The absence of CD8⁺ cells in SD rats may have resulted in a deficiency of major histocompatibility complex (MHC) class I-dependent, CD8⁺ / T-cell receptors (TCR)-expressing T cells. / T cells have been reported to suppress allergic responses in the lung (32). The absence of this regulatory T-cell subset may have affected the local production of mediators and led to the lung eosinophilia observed after antigen challenge.

The mechanisms of eosinophil recruitment and persistence in inflammatory sites include increased production and mobilization of eosinophils from bone marrow by a variety of factors, including the cytokines IL-3 and IL-5 (33), intravascular priming followed by endothelial adhesion and transmigration (34), and stimulation of eosinophil chemotaxis by chemokines (13). Because CD8 depletion led to an increase in eosinophils in the airways 8 h after antigen challenge, we assessed whether chemokine production had been affected. We first examined whether mRNA expression of eotaxin, RANTES, and MCP-1 was different in CD8-depleted rats. Eotaxin mRNA was significantly increased in the lungs of CD8-depleted rats when determined by semiquantitative PCR. These data were confirmed by the finding that more cells were positive for eotaxin protein. Eotaxin has been previously shown to play an important role in the selective recruitment of eosinophils into the airways of asthmatic subjects (16) and in animal models of allergic inflammation (35). Our results provide evidence for the involvement of CD8⁺ cells in the regulation of eotaxin production in this model of allergic asthma. Further analysis will be required to define how CD8⁺ cell can regulate eotaxin expression.

Unlike the effects on eotaxin expression, CD8 depletion had no significant effect on RANTES and MCP-1 expression in the airways after antigen challenge. These results suggest that RANTES and MCP-1 are not involved in the CD8-depletion-mediated increase in the LAR and airway inflammation. This observation parallels other reports showing only minute amounts of RANTES in lesional scales of atopic patients (36) and after allergen-induced cutaneous late-phase responses (37).

CD8 depletion also altered the local cytokine profile after antigen challenge. In CD8-depleted rats, IFN- mRNA expression was found to be significantly reduced in airways 8 h after OA challenge when assessed by semiquantitative RT-PCR and by in situ hybridization. Results obtained in rats are not the only ones suggesting that IFN- may be involved in allergic airway responses. Transfer of CD8⁺ T cells from OA-sensitized mice to sensitized recipients prevented the increase in airway responses to electrical field stimulation that usually occurs in sensitized animals. It was suggested that IFN- production by CD8⁺ cells mediated the inhibitory effects (38). Our demonstration of reduced IFN- production in infants with bronchiolitis that developed asthma would support this idea (39).

The mechanism by which IFN- may be affecting the LAR is unclear; however, several possibilities exist. CD8⁺ T cells are capable of suppressing IgE production through IFN- production, which in turn can downregulate allergic responses (32). IFN- has numerous effects on T-cell differentiation (40), eosinophil migration and degranulation (41), mast cell differentiation and degranulation (42), and on basophil function (43) which may all be relevant to the LAR.

A decreased production of IFN- may be involved in the development of antigen-induced eosinophilia. Mucosal IFN- gene transfer inhibited both antigen-induced and Th2-cell-induced pulmonary eosinophilia and airway hypersensitivity (44). It has also been demonstrated that administration of recombinant IFN- (rIFN-) prevented the infiltration of antigen-induced T cells and eosinophils in the trachea of antigen-sensitized mice (41, 45). Mice lacking the IFN- receptor had an impaired ability to resolve an eosinophilic response in the lung (46). One of the mechanisms by which a decreased production of IFN- may be involved in the development of antigen-induced eosinophilia could be an increased production of eotaxin. Some studies have shown that IFN- is a potent inhibitor of eotaxin synthesis in vitro (47) and that eotaxin expression was increased in IFN-^/ thyroids with eosinophil infiltration (48). Thus, IFN- has the potential to normalize airway function by blocking both the IgE-mediated limb of the response as well as by modulating the response of T lymphocytes, cytokine and chemokine production, and inflammatory effector cells.

The present study examined whether CD8 depletion affected the expression of Th2 cytokines IL-4 and IL-5 after antigen challenge. No differences between control rats and CD8-depleted SD rats were detected in total lung mRNA expression or in the number of airway cells that were expressing mRNA for these cytokines after antigen challenge. It seems that in the sensitized SD rat, an increase in Th2 cytokines is not critical for the development of the LAR. Although this may contradict a widely accepted paradigm that the LAR and eosinophil infiltration are dependent on the presence of IL-4 and IL-5, it is possible that these cytokines are necessary for LAR to occur, but do not regulate the severity of the response. Other factors (such as low IFN- production and changes in other cytokines or chemokines) may need to act together with Th2 cytokines in order for increased LAR and airway inflammation to occur after antigen challenge.

In summary, in the SD rat CD8⁺ cells are important in the prevention of the LAR and eosinophilia that occur after sensitization and antigen challenge. These cells appear to exert their effects through decreased production of eotaxin and increased production of IFN- mRNA.