INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Production of helper T cell type 2 cytokines in allergic asthma is at least partly responsible for eliciting the cardinal pathogenic changes of the asthmatic phenotype. The factors in allergic diseases that govern the production of Th2 cytokines over Th1 cytokines are slowly being revealed. One likely factor is the cytokine profile of T cells, which is influenced by antigen-presenting cells, possibly through costimulatory signals. Such signals may be provided by ligation of CD28 or CTLA-4 with CD80 or CD86 (1).

The exact role of CD28/CTLA-4-mediated T cell costimulation is unclear. Blockade of CD28/CTLA-4 ligation after treatment with CTLA-4 Ig has previously been found to be effective in murine models of either Th1 (2, 3) or Th2 inflammatory conditions including allergic sensitization with either Schistosoma mansoni (4) or ovalbumin (OA) (5). Interestingly, treatment with monoclonal antibodies to CD80 and CD86 was shown to result in strikingly different outcomes in certain experimental models. In experimental autoimmune encephalomyelitis (EAE), a disease mediated by Th1 immune responses), administration of anti-CD80 reduced whereas anti-CD86 increased the incidence of disease, and transfer of Th2 clones prevented disease development (6). Conversely, in Th2-mediated autoimmune diabetes, anti-CD86 abrogated and anti-CD80 significantly accelerated the development of disease in mice (7). Allergic sensitization, however, shows some divergence from this paradigm and previous murine studies did not clarify the individual function of these costimulatory ligands (8).

We have previously characterized different animal models of allergic sensitization and found that they reflect the predominance of Th2-like responses, which are manifested by increases in allergen-specific IgE and an interleukin 5 (IL-5)- induced eosinophil accumulation in the lungs with altered airway function (11). The T cell dependence of these responses has been demonstrated in adoptive transfer (14, 15) and depletion experiments (16) as well as in studies of nude mice (17).

In the present study, we directly assessed the consequences of specific blockade of CD80 or CD86 on the development of allergic responses in two models of allergic sensitization and development of allergic airway hyperresponsiveness (AHR), one shown to be IgE dependent and the other IgE independent. We show that in both systems, only anti-CD86 treatment is effective in attenuating the allergic inflammatory response and AHR.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Mice, Sensitization and Aerosol Exposure; Treatment with Antibodies

Female, 8- to 12-wk-old BALB/cByJ mice weighing 22 ± 2.5 g were housed under pathogen-free conditions and were maintained on an OA-free diet. Experiments were performed with the approval of the Institutional Animal Care and Use Committee of the National Jewish Medical and Research Center (Denver, CO).

Mice were exposed to OA according to two different protocols: (1) Ten consecutive days of exposure to an aerosol containing 1% OA- phosphate-buffered saline (PBS), produced using a DeVilbiss Aerosonic 5000 nebulizer (DeVilbiss Health Care, Somerset, PA) as previously described (18). Either CD80 or CD86 antibody was injected intraperitoneally (100 µg in 100 µl of PBS) on Days 1, 3, and 5. Positive control sensitized mice were injected with rat IgG (Sigma, St. Louis, MO); (2) intraperitoneal sensitization by injection of 20 µg of OA (grade V; Sigma) together with 20 mg of alum (Imject Alum; Pierce, Rockford, IL) on Days 1 and 14, followed on Days 24, 25, and 26 by airway challenge for 20 min with a 1% OA-PBS solution (19). CD80, CD86, or rat IgG antibodies were injected intraperitoneally on Days 1, 7, 14, and 21 and before each aerosol challenge on Days 24, 25, and 26. In an additional protocol, mice were injected with anti-CD80 or anti-CD86 immediately before allergen challenge of sensitized mice on Days 24, 25, and 26. Each injection contained 100 µg of protein in 100 µl of PBS. All mice were sacrificed 48 h after the last OA exposure. Control, nonsensitized mice that received either anti-CD80 or anti-CD86 showed no difference from naive mice in their airway responses or OA-specific immunoglobulin levels (data not shown).

Antibodies and Cell Lines

The monoclonal antibodies (IgG_2a and IgG_2b, respectively) were purified from culture supernatants of the 1G.10 and 2D.10 cell lines (American Type Culture Collection [ATCC], Rockville, MD) on protein G columns. The binding capacity of the purified antibodies was tested by flow cytometric analysis (FACScalibur flow cytometer; Becton Dickinson, Mountain View, CA) on splenic B cells activated by a combination of phorbol 12,13-dibutyrate (10 nM) and ionomycin (0.5 µM) (PI) as well as on lymphoblastoma cell lines transfected with either the CD80 or CD86 gene (generous gift of J. Hagman, National Jewish Medical and Research Center).

Airway Function Measurements

Airway responsiveness to electric field stimulation was determined 48 h after the last aerosol challenge of mice as described previously (18). Briefly, tracheas were removed and 0.5-cm-long preparations were placed in Krebs-Henseleit solution, suspended by triangular supports transducing the force of contractions. Electrical field stimulation with an increasing frequency from 0.5 to 40 Hz was applied and the contractions measured. Frequencies resulting in 50% of the maximal contractions (ES₅₀) were calculated from linear plots for each animal and were compared between the different groups.

Bronchial responsiveness was assessed after challenge with aerosolized methacholine (MCh) via the airways using a method previously described (19). Mice were anesthetized, tracheostomized, and ventilated at 160 breaths per minute and a tidal volume of 0.15 ml with a positive end-expiratory pressure of 2-4 cm H₂O. Tidal volume was adjusted to weight if the weight of the animal exceeded or fell below the average weight of 22 ± 2.5 g by 10%. Changes in lung volume were measured by detecting pressure changes in a plethysmographic chamber. Aerosolized MCh was administered through bypass tubing via an ultrasonic nebulizer. After a dose of inhaled PBS was given, the subsequent values of pulmonary resistance (RL) or dynamic pulmonary compliance (Cdyn) were used as baseline. Starting 3 min after saline exposure, increasing concentrations of MCh were given by inhalation (10 breaths), with the initial concentration set at 0.8 mg/ml. From 20 s up to 3 min after each aerosol challenge the data of (RL and Cdyn) were continuously collected by fitting flow, volume, and pressure to an equation of motion. Maximum values of RL and minimum values of Cdyn were taken to express changes in murine airway function. Provocative concentrations of MCh that cause 100 to 200% increases in lung resistance above baseline (PC100 and PC200, respectively) were calculated by log-linear transformation of the dose-response curves.

Serum Collection and ELISA for Immunoglobulins

Venous blood was collected from the tail vein before and at different time points during the sensitization period into serum separator tubes (Microtainers; Becton Dickinson). Serum samples were centrifuged in a microcentrifuge for 10 min and stored at 20° C pending analysis. OA-specific antibody levels in the serum were determined by enzyme-linked immunosorbent assay (ELISA) (19). Plates (Dynatech, Chantilly, VA) were coated with OA (20 µg/ml NaHCO₃ buffer, pH 9.6) and incubated overnight at 4° C. Plates were blocked with 0.2% gelatin buffer (pH 8.2) for 2 h at 37° C before adding serum samples, which were diluted 1:10. The antibody titers were related to a pooled standard, generated in our laboratory. ELISA data were analyzed with the Microplate Manager software program for the Macintosh (Bio-Rad, Hercules, CA).

BAL and Whole Lung Digest Differential Cell Count and Analysis of BAL Cytokine Content

After functional measurements were completed, lungs were lavaged with 1-ml aliquots of sterile saline through the tracheal cannula. After centrifuging (500 g for 10 min at 4° C) the cell pellet was resuspended and counted. The supernatants were harvested and stored at 20° C for further analysis. Lung digestion was performed after exsanguination and perfusion of the lungs, following the protocol previously described from our laboratory, and differential cell counts were made from cytospin preparations as described (16, 17, 19). Cells were identified as macrophages, eosinophils, neutrophils, and lymphocytes by standard morphology and at least 200 cells were counted under ×400 magnification. The percentage and absolute numbers of each cell type were then calculated.

The levels of cytokine secreted into the supernatants of mononuclear cell cultures and bronchoalveolar lavage (BAL) fluid samples were determined by ELISA according to standard protocols. The cytokine amounts were calculated from the standard curves in each plate. The limits of detection were 5 pg/ml for IL-4 and IL-5 and 10 pg/ml for interferon  (IFN-). As standards, we used recombinant mouse IL-4 and IL-5 (PharMingen, San Diego, CA) and recombinant murine IFN- (Genetech, San Francisco, CA).

Immunolabeling of Eosinophils

Immunocytochemistry was performed as described previously (19). Briefly, lung tissue was fixed in 10% formalin solution. Sections (4 µm thick) were prepared for immunolabeling. The primary antibodies (rabbit polyclonal anti-mouse major basic protein [MBP]) were diluted in 3% goat serum and applied at 4° C overnight. Slides were stained with 1% Chromotrope 2R (Halesco, Gibbstown, NJ) and then fluorescein isothiocyanate (FITC)-conjugated goat anti-rabbit secondary antibody was applied. Slides were examined using a Zeiss (Oberkochen, Germany) microscope equipped with a fluorescein filter system at ×200 magnification. For counting, a computer software program was used (IP Lab Spectrum; Signal Analytics, Vienna, VA) and results were expressed as number of positive cells per unit area.

Data Analysis

Data were expressed as means ± SEM. Analysis of variance (ANOVA) was used to determine significant variance among the groups. If significant variance was found, the t test was used to analyze the differences between individual groups. A p value of < 0.05 was considered significant. Data were analyzed with the Minitab standard statistical package (Minitab, State College, PA).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Anti-CD86 Treatment Decreases serum IgE and IgG₁, Lung Eosinophilia, Tracheal Smooth Muscle Reactivity to Electrical Field Stimulation, and Cytokine Levels in the BAL Fluid After 10 d of Ovalbumin Exposure

To investigate the effects of anti-CD80 and anti-CD86 antibodies on allergic changes, mice were exposed for 10 d to OA exclusively, via the airways. This protocol was previously shown to result in AHR that is dependent on T cells (17) and IgE production (21). Animals were treated with either anti-CD80, anti-CD86, or control rat IgG. We monitored airway responsiveness by measuring responses of tracheal smooth muscle preparations to electrical field stimulation. ES₅₀ values from individual dose-response curves were calculated and the ratio (%) relative to naive controls is depicted in Figure 1A. A decrease in ES₅₀ represents an increase in tracheal smooth muscle responsiveness (18). Ten days of OA exposure of BALB/c mice resulted in significant decreases in ES₅₀. Mice receiving 10 d of PBS exposure demonstrated no difference in responsiveness relative to naive controls (data not shown). Mice injected with anti-CD86 demonstrated significantly reduced airway reactivity whereas treatment with anti-CD80 was ineffective.

View larger version (30K): [in this window] [in a new window]   Figure 1.   Anti-CD86 treatment decreases tracheal smooth muscle reactivity to electrical field stimulation, serum IgE and IgG1, lung eosinophilia, and cytokine levels in BAL fluid after 10 d of OA exposure. Mice were exposed to 1% OA aerosol for 10 d and injected intraperitoneally with 100 µg of rat IgG, anti-CD80, or anti-CD86 on Days 1, 3, and 5. (A) Electrical field stimulation (EFS) was performed on tracheal preparation of mice 48 h after the last OA exposure. Results are expressed as the percent change in ES50 (50% of the electrical stimulus [Hz] that results in maximal contraction). The mean ES50 of naive control mice was 3.7 ± 0.3 Hz (n = 12). (B) Serum samples were taken 48 h after the last OA exposure and were assayed for OA-specific IgE and IgG1 by ELISA (C) Lung digests were obtained 48 h after the last allergen challenge. Total = total cell number in lung extract; Mp = macrophage; Ne = neutrophil; Ly = lymphocyte; Eo = eosinophil. (D) Mice were treated as described above and BAL samples were collected 48 h after the last aerosol exposure. BAL supernatants were assayed for cytokine content by ELISA. Open bars, 10 d of PBS exposure (n = 14); solid bars, 10 d of OA exposure plus rat IgG injections (n = 14); shaded bars; 10 d of OA exposure plus anti-CD80 injections (n = 14); and hatched bars, 10 d of OA exposure plus anti-CD86 injections (n = 14). *p < 0.05, **p < 0.01, rat IgG versus anti-CD86; # p < 0.05, ## p < 0.01, 10 d of PBS versus 10 d of OA. Results are expressed as means ± SEM. A total of 14 mice was used in three different experiments.

Allergen-exposed mice had significant increases in OA-specific IgE and IgG₁ levels in the serum but OA-specific IgG_2a levels were not detectable (not shown). Mice exposed to PBS had no OA-specific immunoglobulin in their serum (not shown). Anti-CD86 treatment significantly reduced IgE and IgG₁ levels when compared with rat IgG-treated controls (Figure 1B); anti-CD80 had no significant effect on specific antibody production.

To investigate whether the suppressive effect of anti-CD86 on airway function was associated with a decrease in airway inflammation, we studied the cellular content of the whole lung digest and cytokine levels of BAL fluid of mice treated with anti-CD86 as described. We found a significant eosinophilia in the whole lung digests of allergen-exposed mice as previously described (16, 17), with a decrease induced by anti-CD86 treatment (Figure 1C). BAL fluid contained significantly increased levels of IL-4, IL-5, and IFN- that were abolished by anti-CD86 treatment (Figure 1D).

There were no significant effects of anti-CD80 or anti-CD86 treatment on OA-specific IgE and IgG₁ levels, inflammatory cell influx, tracheal smooth muscle responsiveness to electrical field stimulation, or BAL cytokine content in nonsensitized mice (not shown).

Anti-CD86 Decreases AHR, Airway Eosinophilia, and BAL IL-4 and IL-5 Levels and Delays Development of Specific Antibody Responses in Sensitized and Challenged Mice

To delineate the effects of treatment with anti-CD80 and anti-CD86 in an inflammatory process where AHR is not dependent on IgE production (22), we used a model of systemic sensitization followed by repeated airway challenge with OA. This protocol resulted in a significant increase in airway responsiveness when compared with nonsensitized mice. Figure 2A shows the dose-response curves of pulmonary resistance (RL) plotted against concentrations of inhaled MCh in mice sensitized with OA/alum followed by airway challenge. Treatment with anti-CD86 resulted in a significant reduction in airway responsiveness with a shift to the right of the MCh dose- response curve. There were slight but significant increases in airway responses at the highest MCh dose in anti-CD86-treated mice when compared with naive animals. The reasons for this (residual) increase in RL at the high dose of MCh is not clear. The effects of anti-CD86 in normalizing dynamic compliance (Cdyn) are shown in Figure 2B and the effects are seen throughout the MCh dose-response curve. Cdyn encompasses several determinants including airway resistance to flow, tissue resistance to deformation, and elastic recoil of the lung tissue, and may be useful in detecting changes in the small airways and peripheral tissue. PC100 and PC200 values were also significantly higher in anti-CD86-treated mice than in the mice injected with anti-CD80 (Figure 3A).

View larger version (17K): [in this window] [in a new window]   Figure 2.   Effects of anti-CD80 and anti-CD86 on airway responsiveness to MCh after sensitization and challenge with OA. Mice were sensitized by intraperitoneal injection on Days 0 and 14 followed by three challenges with OA (on Days 24, 25, and 26), and were injected with anti-CD80, anti-CD86, or rat IgG on Days 0, 7, 14, 21, and before each challenge. Control mice received three aerosol challenges with OA alone. Forty-eight hours after their last exposure mice were studied for lung function. Airway responsiveness was monitored by measuring lung resistance. (RL; A) and dynamic compliance. (Cdyn; B) Results are expressed as the percent change from baseline in response to increasing concentrations of MCh. Data points represent the mean increase in RL (A) and decrease in Cdyn (B) relative to the baseline, measured in response to PBS in individual mice in each group. Open circles, challenge alone (n = 16); solid circles, sensitization + challenge + rat IgG (n = 16); hatched squares, sensitization + challenge + anti-CD80 (n = 16); open squares, sensitization + challenge + anti-CD86 (n = 16). Data represent means ± SEM; a total of 16 mice was used in each group in four different experiments. *p < 0.05, **p < 0.01, rat IgG versus anti-CD86.

View larger version (33K): [in this window] [in a new window]   Figure 3.   Anti-CD86 treatment increases methacholine PC100 and PC200, and decreases serum IgE and IgG1, BAL eosinophilia, and cytokine levels in the BAL fluid after sensitization and challenge with OA. Groups of mice were treated as described above. Forty-eight hours after their last exposure mice were studied for airway responses, serum IgE and IgG1 levels, and inflammatory changes of the airways. (A) PC100 and PC200 values of airway responsiveness to MCh. Solid bars, sensitization + challenge + rat IgG (n = 16); shaded bars, sensitization + challenge + anti-CD80 (n = 16); hatched bars, sensitization + challenge + anti-CD86 (n = 16). (B) Serum samples were obtained on Days 0, 7, 14, 21, and 28 and analyzed for production of OA-specific IgE and IgG1. Data are expressed as ELISA units/per milliliter. Solid circles, sensitization + challenge + rat IgG (n = 16); open squares, sensitization + challenge + anti-CD80 (n = 16); hatched triangles, sensitization + challenge + anti-CD86 (n = 16). (C) BAL was performed 48 h after the last allergen challenge. Total = total cell number in BAL sample; Mp = macrophage; Ne = neutrophil; Ly = lymphocyte; Eo = eosinophil. (D) BAL samples were analyzed for cytokine content by ELISA. Cytokine levels are expressed as picograms per milliliter. Open bars, challenge only (n = 16); solid bars, sensitization + challenge + rat IgG (n = 16); shaded bars, sensitization + challenge + anti-CD80 (n = 16); hatched bars, sensitization + challenge + anti-CD86 (n = 16). Data represent means ± SEM; a total of 16 mice was used in each group in four different experiments. *p < 0.05, **p < 0.01, rat IgG versus anti-CD86. # p < 0.05, ## p < 0.01, naive versus sensitized/challenged mice.

To study the kinetics of immunoglobulin production after sensitization and exposure to OA, we obtained serum samples weekly throughout the sensitization and challenge period. Sensitization and repeated airway challenge with OA resulted in a significant increase in OA-specific IgE and IgG₁ levels. Anti-CD80 did not affect these changes while anti-CD86 treatment resulted in a significant shift in the time course of OA-specific IgE and IgG₁ production (Figure 3B).

In BAL fluid there were significant increases in the numbers of macrophages, neutrophils (albeit low), and particularly eosinophils following sensitization and challenge of mice. The extent of AHR showed a statistical correlation with eosinophilia but not neutrophilia in the BAL; for BAL eosinophils the correlation with RL (lung resistance) was r = 0.765 (p = 0.04). Although neutrophils may be involved in the development of AHR, at the time of airway functional measurements their numbers were so small compared with the number of eosinophils that their importance at this time point is likely to be negligible. Treatment with anti-CD86 significantly decreased numbers of eosinophils and neutrophils while anti-CD80 resulted in no change in eosinophil numbers when compared with mice injected with rat IgG (Figure 3C).

BAL supernatant IL-4, IL-5, and IFN- levels were significantly elevated after sensitization and airway challenge but not in naive mice. Treatment of sensitized and challenged mice with anti-CD86 antibody significantly reduced IL-4 and IL-5 levels when compared with the effects of control or CD80 antibody (Figure 3D); IFN- levels were not significantly affected by antibody treatment.

Eosinophils in the lung tissue were localized by immunolabeling of eosinophil major basic protein (MBP). We detected only a few randomly scattered, positively stained cells in the lung tissue of mice receiving three OA nebulizations alone (Figure 4A). However, sensitized and challenged mice developed a marked eosinophilia (Figure 4B). The eosinophils (MBP⁺ cells) accumulated in the peribronchial and perivascular submucosal tissue, whereas the lung parenchyma remained relatively eosinophil free. The number of MBP⁺ cells was significantly correlated with changes in airway function (r = 0.757, p = 0.003). The inflammatory changes, particularly the peribronchial/perivascular eosinophilia, were unaffected in mice treated with anti-CD80 (Figure 4C) but were inhibited in the mice treated with anti-CD86 antibody (Figure 4D).

View larger version (146K): [in this window] [in a new window]   Figure 4.   Effects of anti-CD80 and anti-CD86 on eosinophil accumulation in lung tissue. Mice were sensitized and challenged as described. Tissue samples were taken 48 h after the last OA challenge. Control mice received three OA challenges alone. Immunolabeled (MBP+) eosinophils were studied by fluorescence microscopy (original magnification: ×200). (A) OA nebulization alone; (B) sensitization and challenge; (C) anti-CD80 treatment; (D) anti-CD86 treatment.

Anti-CD86 Treatment before Airway Challenge Does Not Affect AHR

To clarify whether blockade of CD86 would prevent development of inflammatory changes after local allergen challenge, we performed a series of experiments in which mice were treated with anti-CD86 antibody before airway allergen challenges but after sensitization (Table 1). Administration of anti-CD86 before challenge in sensitized mice had a limited inhibitory effect when compared with antibody administration throughout the sensitization period and airway challenges. Importantly, airway responsiveness was not altered in animals receiving the antibody before challenge alone. There was also no inhibitory effect detected in OA-specific IgE and IgG₁ levels, with the serum levels being similar to those of mice treated with the control antibody (not shown).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In the present study, we tested the hypothesis that selective interference with the important costimulatory molecules CD80 and CD86 can result in altered responses in the lung after allergen exposure. Two models of allergen exposure and AHR were assessed, one IgE dependent and the other IgE independent, in order to reveal the functional differences between CD80 and CD86.

The conditions for developing and manipulating allergic AHR may differ depending on the mode of sensitization, challenge, and treatment. These alternative approaches to sensitization and challenge we have used have proved important in analyzing the role of different components (IgE and eosinophilia) in the development of allergic AHR. When mice are exposed to allergen exclusively via the airways, i.e., 10 d of OA aerosol in the absence of adjuvant (mimicking natural conditions more closely), they develop a mild eosinophilic inflammation and AHR detectable only by electrical field stimulation of tracheal smooth muscle responsiveness (14, 16, 17). The development of altered airway responsiveness in this way is both IgE and IL-5 dependent (16, 17, 21). In contrast, when mice are sensitized by intraperitoneal injection of allergen followed by repeated airway challenge, altered airway function appears to be dependent on the significant eosinophilia that develops and is independent of the high levels of serum IgE that are also observed (22).

Using these two models, we demonstrated that anti-CD86 reduces AHR in association with significantly diminished IgE, IgG₁, and eosinophil responses after either 10 d of OA exposure or intraperitoneal sensitization and airway challenge, indicating that the effects of CD86 are not restricted to effects on IgE or mechanisms depending on IgE production. Interestingly, administration of anti-CD86 immediately before exposing sensitized mice to allergen was ineffective in inhibiting AHR, indicating that this antibody does not affect the local inflammatory cell influx that develops after allergen challenge of sensitized mice. This is in contrast to CTLA-4 Ig treatment, which was effective in suppressing inflammatory changes even when administered only before allergen exposure (4, 5).

Our findings in these studies demonstrate and confirm that CD86 likely exerts its inhibitor effects through suppression of Th2 cell development because the effects of anti-CD86 were associated with marked decreases in IL-4 and IL-5 production in the BAL fluid. The effects on IFN- were less obvious. Although in one study lung T cells isolated from mice sensitized with OA and treated with anti-CD86 had increased IFN- release (9), there is no evidence that anti-CD86 directly affects IFN- production by T cells in allergic inflammation. In an attempt to study the effects of these antibodies on a Th1 inflammatory response of the airways, in which increases in IFN- may play a role, we are investigating a respiratory syncytial virus (RSV)-induced model of AHR (23).

The effects of antibodies to costimulatory molecules on allergic sensitization in vivo are complex. It was previously shown, for instance, that administration of anti-CD80 resulted in upregulation of Th2 responses (6, 7). However, in our experiments together with two studies of OA-induced allergic sensitization (9, 10) and another on H. polygyrus-infected mice (24), anti-CD80 treatment failed to exhibit significant effects. In contrast, Harris and coworkers found that selective anti-CD80 treatment abolished antigen-induced accumulation of eosinophils and lymphocytes in the lungs of OA-immunized mice (8). It is not readily apparent why selective in vivo blockade of this molecule can result in enhancement of Th2-type immune responses in some studies (6, 7), no effect (9, 10, 24), or even inhibition in others (8). Establishing when, where, and to what extent CD80 and CD86 are expressed during allergen sensitization will likely be necessary in order to distinguish or explain the different effects of blocking one or the other ligand. It is possible that in order to achieve any blocking effect on a Th2-type immune response with the anti-CD80 antibody in vivo, a relatively high local concentration may be necessary, higher than that required for anti-CD86. CD80 is expressed at significantly lower levels than CD86, especially on resting cells. Further, CD86 (but not CD80) is rapidly upregulated after stimulation (25). Indeed, it was shown that after allergen challenge, murine B cells expressed predominantly CD86 (9) and peripheral blood B cells from human asthmatic patients expressed significantly higher levels of CD86 after allergen exposure than did unexposed or nonatopic healthy controls (25).

In addition to differential expression patterns of these two molecules, the effects of anti-CD80 and anti-CD86 have been suggested to reflect differential signaling after ligation of either CD28 or CTLA-4 during an immune response (6, 26, 27). We have investigated the costimulatory effects of a hybridoma cell line transfected either with CD80 or CD86 on murine T cells (data not shown). Both of these transected cell lines were previously found to be effective in rescuing thymocytes from dexamethasone-induced apoptosis, an effect that was reversed by CTLA-4 Ig treatment (20). We found that both CD80 and CD86 transfectants induced significant proliferative responses and IL-4 as well as IFN- production by T cells when compared with the nontransfected parental cell lines. Studies by Freeman and co-workers suggested that CD86 but not CD80 is capable of stimulating IL-4 synthesis by CD4⁺ T cells (26) but in other studies the costimulatory signals provided by CD80 and CD86 cells were reportedly indistinguishable (28, 29). The nature of the conditions of T cell activation, including the concentration of antigen, numbers of antigen-presenting cells, and numbers of T cells, may be important in dictating the dependence of cytokine synthesis on specific costimulatory markers (30). CD28-mediated costimulation may therefore be important in both Th1 and Th2-type immune responses under limiting conditions. When overexpressed on the antigen-presenting cell, CD80 and CD86 appear to induce similar signaling events in T cells.

In summary, we demonstrate that blocking CD86 but not CD80 costimulation is effective in inhibiting allergic responses and that these effects are associated with reduced Th2 cytokine production. We speculate that CD86 is the dominant costimulatory ligand in allergic inflammation and that its selective blockade alone can significantly influence the development of allergic responses including allergen-driven AHR.