INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

CD23, the low-affinity receptor for IgE, has been implicated in a number of immune and inflammatory functions, and is thought to be important in the regulation of IgE production by B cells (1). Although structurally related to the human form, murine CD23 is different in its cellular expression and function. Mouse CD23 is found on B cells and follicular dendritic cells only and many of the effects ascribed to human soluble CD23 cannot be confirmed in mice (4,5). Mice, however, do retain one of the most important functions of CD23, that is engagement of IgE on the B cell surface.

In various models of allergic sensitization we have previously identified the predominance of a helper T cell type 2 (Th2)-like response manifested by production of specific IgE and an interleukin 5 (IL-5)-related eosinophil accumulation in the lungs (6). The T cell dependence of these events has been demonstrated in adoptive transfer and depletion experiments as well as in studies of nude mice (8,11-13). By facilitating antigen presentation via IgE-antigen complexes, CD23 was shown to be capable of altering T cell function (14) and of modulating the immune response by affecting T cell cytokine synthesis in vitro (15). Whether this function bears any significance to development of allergic responses in vivo remains to be clarified. However, by using CD23-deficient mice in a model of allergic sensitization we found that in spite of impaired IgE-mediated immune responses, these animals were capable of developing significant airway eosinophilia and airway hyperresponsiveness (AHR). In fact, the allergic responses were enhanced in CD23-deficient mice. Although these data suggest that the presence of CD23 in this model may negatively affect development of allergic airway inflammation, we could not exclude the possibility that compensatory mechanisms induced by genetic absence of this molecule were responsible for the enhanced changes (16). Here we extend our findings by evaluating the consequences of treating the mice with an antibody to CD23, an approach that has been shown to induce negative signaling events in B cell function (17). We used two different sensitization protocols that were previously shown to differ in their dependency on IgE-mediated events (18, 19). Since there are no other works published to date in which the negative regulatory effects of CD23 are clearly demonstrated in murine models of allergic inflammation and airway hyperresponsiveness, in the present study we compared the effects of targeting the CD23 gene negatively and positively. If CD23 indeed has negative regulatory effects on these events, we hypothesized that CD23 deficiency would lead to enhancement while overexpression would result in suppression of allergic inflammation.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Mice

Female BALB/c mice between the ages of 8 and 12 wk were obtained from Jackson Laboratories (Bar Harbor, ME). CD23-deficient (CD23^-/-) and wild-type (CD23^+/+) mice were generated by Fujiwara and coworkers (20). Briefly, the gene-targeted 129/Ola strain was crossed with C57BL/6, and then the CD23-deficient F₁ heterozygous mice were backcrossed to C57BL/6 mice twice. These heterozygous mice were then intercrossed to produce homozygous wild-type and CD23-deficient mice. Confirmation of the functional disruption of the CD23 gene was performed by Southern blot and Northern blot analysis and has been published in detail. A detailed FACScan analysis of splenocytes and thymocytes as well as an immunohistochemical analysis of germinal centers of spleen sections from homozygous CD23-deficient and wild-type mice were published by Fujiwara and colleagues (20).

CD23 transgenic mice were generated as described previously (21) and were obtained from the laboratory of D. Conrad. These mice were considered heterozygous for the transgene because of the variable expression of CD23, although some of the animals were likely homozygous for the transgene. To assess CD23 expression we performed FACScan (Becton Dickinson, San Jose, CA) analysis on populations of splenic lymphocytes.

All mice were housed under pathogen-free conditions and were maintained on an ovalbumin (OA)-free diet. All experimental animals used in this study were under a protocol approved by the Institutional Animal Care and Use Committee of the National Jewish Medical and Research Center (Denver, CO).

Antibodies

To prepare it for in vivo treatment of mice the anti-FcRII rat monoclonal antibody (MAb) B3B4 (IgG_2a) was purified from hybridoma supernatant as described previously (15). The supernatant was compared for CD23 binding on splenic B cells with commercially available B3B4 (PharMingen, San Diego, CA). The antibody used for identifying CD23 in studies of the CD23 transgenic mice was a fluorescein isothiocyanate (FITC)-conjugated B3B4 (PharMingen). Rat IgG (serum) was used as control antibody and was obtained from Sigma (St. Louis, MO). Use of the rabbit polyclonal anti-mouse major basic protein (MBP) antibody for immunological labeling of eosinophil granulocytes was previously described (16).

Sensitization and Airway Challenge

In the first protocol, mice were exposed to a 1% OA (or phosphate-buffered saline [PBS]) solution by aerosolization for 20 min each day over 10 consecutive days (11). In the second protocol, mice were actively immunized by intraperitoneal injection of 20 µg of OA (grade V, Sigma) together with 2.0 mg of alum (Inject Alum; Pierce, Rockford, IL) in 100 µl of PBS, or with PBS alone, on Day 1 and Day 14. On Days 24, 25, and 26 mice received an aerosol challenge for 20 min with a 1% OA-PBS solution. All mice were sacrificed 48 h after their last OA exposure.

Electrical Field Stimulation of Tracheal Smooth Muscle in Vitro

Airway responsiveness to electrical field stimulation was determined 48 h after the last aerosol challenge of mice as described previously (11). 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 contractions measured. Frequencies resulting in 50% of the maximal contractions (ES₅₀) were calculated from linear plots for each individual animal and were compared between the different groups.

In Vivo Measurement of Bronchial Responsiveness to Methacholine

Bronchial responsiveness was assessed as a change in airway function after challenge with aerosolized methacholine (MCh), using a modification of methods previously described in mice (22). Mice were anesthetized with an intraperitoneal injection of pentobarbital sodium (70 to 90 mg/kg). A stainless steel 18-gauge tube was inserted as a tracheostomy cannula and was passed through a hole in the Plexiglas chamber containing the mouse. A four-way connector was attached to the tracheostomy tube, with two ports connected to the inspiratory and expiratory sides of a ventilator (model 683; Harvard Apparatus, South Natick, MA). Ventilation was achieved at 160 breaths/min and a tidal volume of 0.15 ml with a positive end-expiratory pressure of 2- 4 cm H₂O. The Plexiglas chamber was continuous with a 1.0-L glass bottle filled with copper gauze to stabilize the volume signal for thermal drift. Transrespiratory pressure was detected by a pressure transducer. Changes in lung volume were measured by detecting pressure changes in the plethysmographic chamber. Flow was measured by digital differentiation of the volume signal. Lung resistance (RL) was continuously computed by Macintosh computer software (Labview, National Instruments, TX) by fitting flow, volume, and pressure to an equation of motion.

The aerosolized bronchoconstrictor agents were administered through bypass tubing via an ultrasonic nebulizer placed between the expiratory port of the ventilator and the four-way connector. Aerosolized agents were administered for 10 s with a tidal volume of 0.5 ml. After a dose of inhaled PBS was given, the subsequent values of RL 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 0.4 mg/ml. Increasing concentrations were given at 5 to 7 min intervals. Hyperinflations of twice the tidal volume were applied between each MCh concentration and performed by manually blocking the outflow of the ventilator in order to reverse any residual atelectasis and ensure a constant volume history prior to challenge. From 20 s up to 3 min after each aerosol challenge, the data for RL were continuously collected and maximum values of RL were taken to express changes in murine airway function.

ELISA for Immunoglobulins and Cytokines

Venous blood was collected from the tail vein before and at various time points during the sensitization period into serum separator tubes (Microtainer; Becton Dickinson). Serum samples were stored at 20° C pending analysis. Serum antibody levels were determined as previously described (16). Standards containing OA-specific IgE and IgG were generated as described (23). For total serum immunoglobulins, commercial standards were used (PharMingen). Enzyme-linked immunosorbent assay (ELISA) data were analyzed with the Microplate Manager software program for the Macintosh (Bio-Rad, Hercules, CA).

The levels of cytokine secreted into the supernatants of mononuclear cell cultures and bronchoalveolar lavage (BAL) fluid samples were determined by ELISA. Briefly, 96-well plates (Immulon; Dynex, Chantilly, VA) were coated overnight (4° C) with primary anti-cytokine capture antibody (1 µg/ml). Purified rat anti-mouse IL-4, IL-5, and interferon  (IFN-) were all from PharMingen. The plates were then washed three times with PBS-Tween 20 (Fisher, Pittsburgh, PA) and blocked overnight with PBS-10% fetal calf serum (FCS). After washing, 100 µl of the undiluted BAL fluid and cell culture supernatant samples were added to the wells. Serial dilutions of standards were prepared at a dilution factor of 0.33. After overnight incubation (4° C), the plates were washed and anti-cytokine antibodies conjugated to biotin (PharMingen) were added (1 µg/ml). The plates were incubated overnight and, after washing six times, avidin-peroxidase complex (Sigma) and substrate were added and incubated at room temperature. A green color developed and was read at 410-nm wavelength in a spectrophotometer (2550; Bio-Rad). The cytokine amounts were calculated by applying a standard curve to each plate. The limits of detection were 5 pg/ml for IL-4 and IL-5 and 3 pg/ml for IFN-. As standards, we used recombinant mouse IL-4 and IL-5 (PharMingen) and recombinant murine IFN- (Genentech, San Francisco, CA).

BAL Collection and Differential Cell Count

After measurement of lung function parameters, lungs were lavaged with 1-ml aliquots of 0.9% (w/v) sterile NaCl (room temperature) through a polyethylene syringe attached to the tracheal cannula. Lavage fluid was centrifuged (500 × g for 10 min at 4° C) and the cell pellet was resuspended in 0.5 ml of RPMI tissue culture medium. The cell-free supernatant of each BAL sample was stored at 20° C until cytokine assay.

Cells from BAL fluid were resuspended in RPMI and counted with a hemocytometer. Differential cell counts were made from cytospin preparations as described (16). Cells were identified as macrophages, eosinophils, neutrophils, and lymphocytes by standard morphology and at least 300 cells were counted under ×400 magnification. The percentage and absolute numbers of each cell type were then calculated.

Immunolabeling of Eosinophils

Lung tissue was removed and fixed in 10% formalin solution. Sections (4 µm thick) were cut, deparaffinized, and treated with porcine trypsin for 30 min at 37° C. After washing three times, 10% goat serum was applied for 30 min, at room temperature. Primary antibodies (rabbit polyclonal anti-mouse MBP) were diluted in 3% goat serum and applied at 4° C overnight. Slides were then washed and stained with 1% Chromotrope 2R (Harlesco, Gibbstown, NJ) for 30 min at room temperature. After washing, FITC-conjugated goat anti-rabbit secondary antibody was used and the slides were incubated for 1 h at room temperature in the dark. Coverslips were applied with Fluoromount (Fisher Scientific, Pittsburgh, PA). Slides were kept at 20° C until they were examined under a Zeiss (Thornwood, NY) 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 as described previously (16).

Data Analysis

Data were expressed as means ± SEM. Analysis of variance (ANOVA) was used to determine significant variance among the groups. If a significant variance was found, the t test was used to analyze the differences between individual groups. In case of multiple comparisons, the Bonferroni correction was applied. A p value of < 0.05 was considered significant. Regression analysis was performed in order to establish correlation between variables. Data were analyzed with the Minitab standard statistical package (Minitab, State College, PA).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Anti-CD23 Treatment Decreases Serum IgE and IgG₁, Tracheal Smooth Muscle Reactivity to Electrical Field Stimulation, and Eosinophil Numbers in the Lung After 10-d OA Exposure

To determine the effects of anti-CD23 antibody on allergic changes after 10-d OA exposure exclusively via the airways, various groups of BALB/c mice were treated with either anti-CD23 or rat IgG (control). The animals were injected with the antibody intraperitoneally on Days 1, 3, and 5 during the 10-d sensitization period and were studied 48 h after the last OA exposure.

Airway responsiveness was monitored by measuring responses of tracheal smooth muscle preparations to electrical field stimulation. ES₅₀ values from individual dose-response curves were calculated and the ratios (%) relative to naive controls are depicted in Figure 1A. A decrease in ES₅₀ represents an increase in responsiveness (11). Active sensitization of BALB/c mice after 10-d OA exposure resulted in significant decreases in ES₅₀ when compared with PBS-exposed controls, as illustrated in Figure 1A. Mice exposed to PBS for 10 d demonstrated no difference in responsiveness compared with naive controls (data not shown). Anti-CD23 treatment significantly reduced tracheal smooth muscle hyperresponsiveness when compared with control rat IgG-treated mice.

View larger version (24K): [in this window] [in a new window]   Figure 1.   Anti-CD23 inhibits allergic responses after 10 d of OA exposure. (A) Electrical field stimulation (EFS) results are expressed as the percent change from the mean ES50 value (50% of the electrical stimulus [Hz] that results in maximal contraction) in control mice. The mean ES50 value of naive control mice was 3.5 ± 0.32 Hz (n = 10). (B) OA-specific IgE and IgG1 levels (ELISA U/ml). (C ) Lung digest cellular composition. Total = total cell number in whole lung sample; Mp = macrophage; Ne = neutrophil; Ly = lymphocyte; Eo = eosinophil. (D) BAL cytokine content (pg/ml). Data represent means ± SEM; *p < 0.05; **p < 0.01. Open bars: 10 d PBS (n = 8). Closed bars: 10 d OA/control IgG (n = 8). Hatched bars: 10 d OA/anti-CD23 (n = 8).

Assay of the serum collected from these mice demonstrated that OA aerosol exposure for 10 d resulted in significant increases in OA-specific IgE and IgG₁ levels (OA-specific IgG_2a was not detectable; data not shown). Anti-CD23 treatment significantly reduced IgE and IgG₁ levels when compared with rat IgG-treated controls (Figure 1B).

To investigate whether the suppressive effect of anti-CD23 on airway function was associated with a decrease in inflammatory cell accumulation, the cellular content of the BAL and of the lungs was examined. Bronchoalveolar lavage (BAL) and lung digestion were performed 48 h after the last exposure to allergen. Using this 10-d exposure protocol, we found only a few eosinophils in the BAL fluid of mice (not shown). Nevertheless, after lung digestion and analyses of the cellular content of the lungs, we found significant increases in eosinophil counts that were significantly reduced in anti-CD23-treated mice (Figure 1C).

We predicted that changes in inflammatory cell numbers and altered airway function were associated with changes in local cytokine production. Supernatants of the BAL samples obtained 48 h after the last allergen exposure were analyzed for the presence of IL-4, IL-5, and IFN-. Naive, nonsensitized mice or mice that had been exposed to PBS for 10 d had no detectable levels of any of the cytokines studied. In contrast, 10-d OA exposure resulted in significant increases in IL-4 and IL-5 and to a lesser extent in IFN- (Figure 1D). Treatment with anti-CD23 (but not rat IgG) significantly reduced the levels of IL-4 and IL-5 while increasing IFN- levels in the BAL fluid.

Anti-CD23 Treatment of Sensitized and Challenged Mice Abolishes Airway Hyperresponsiveness to MCh, OA-specific IgE and IgG₁, Airway Inflammation and IL-4 and IL-5 Production

In the second protocol, BALB/c mice were sensitized by two intraperitoneal injections of OA/alum over a 14-d interval and then exposed to aerosolized OA challenge on Days 24, 25, and 26. Forty-eight hours after the last exposure the mice were studied for airway responsiveness, serum IgE and IgG₁ antibody levels, and cellular changes in the lungs. Anti-CD23 was administered on Days 0, 7, 14, and 21 and before each challenge. Figure 2A presents the dose-response curves for pulmonary resistance (RL) plotted against increasing concentrations of inhaled MCh. Sensitization and challenge with OA resulted in significant increases (p < 0.01) in airway responsiveness when MCh responses were compared with nonsensitized mice that received OA challenge alone. There were significant differences between the anti-CD23 and rat IgG-treated groups when compared over a wide concentration range of MCh (p < 0.01) (Figure 2A). PC₂₀₀ values (provocative concentrations of MCh that cause 200% increases in lung resistance above baseline) were calculated by log-linear transformation of the dose- response curves. These PC values were significantly higher in the anti-CD23-treated mice than in their rat IgG-treated counterparts (12.8 versus 5.1) (p < 0.01) (Table 1).

View larger version (22K): [in this window] [in a new window]   Figure 2.   Anti-CD23 inhibits allergic responses after intraperitoneal sensitization and airway challenges with OA. (A) Lung resistance (RL) results are expressed as the percent change from baseline resistance to methacholine (MCh). (B) OA-specific IgE and IgG1 levels (ELISA U/ml). (C ) BAL cellular content. Total = total cell number in BAL sample; Mp = macrophage; Ne = neutrophil; Ly = lymphocyte; Eo = eosinophil. (D) BAL cytokine content (pg/ml). Data represent means ± SEM; *p < 0.05; **p < 0.01. Open bars and open circles: Ch: three challenges on Days 24, 25, and 26 (no sensitization) (n = 12). Closed bars and closed circles: Sens/Ch/Ctr IgG: intraperitoneal sensitization + 3 challenges + intraperitoneal IgG (control) treatment; (n = 12). Hatched bars and closed squares: Sens/Ch/anti-CD23: intraperitoneal sensitization + 3 challenges + intraperitoneal anti-CD23 treatment (n = 12).

To study the kinetics of immunoglobulin production, mice were injected with anti-CD23 on Days 0, 7, 14, and 21, and before each challenge (on Days 24, 25, and 26). Serum samples were obtained from mice on Days 0, 7, 14, 21, and 28. There was no OA-specific IgE or IgG present in the serum samples before sensitization. OA-specific IgE concentrations in the mice treated with anti-CD23 antibody reached a plateau by Day 14, at a significantly lower level than in the mice receiving control rat IgG, and further declined through Day 28 (Figure 2B). IgG₁ levels in the serum were studied 48 h after the last aerosol exposure on Day 28. Anti-CD23 treatment also resulted in a significant decrease in OA-specific IgG₁ in mice sensitized and exposed to OA (Figure 2B). Aerosol exposure to OA for three consecutive days alone did not alter serum IgE or IgG₁ levels in either group. Further, control rat IgG treatment did not have any effect on the OA-specific immunoglobulin levels in nonsensitized mice.

The numbers of total leukocytes recovered from BAL of sensitized and challenged mice were significantly higher than in nonsensitized animals (Figure 2C). While there were no eosinophils and neutrophils in nonsensitized mice, sensitized and challenged animals demonstrated a marked increase in inflammatory cells, particularly eosinophils. The proportion of eosinophils reached approximately 45%, which was reduced to 5% after anti-CD23 treatment.

To examine directly the number and localization of the eosinophils in the lung, we immunolabeled eosinophil MBP in formalin-fixed, paraffin-embedded tissue sections. The tissue samples were taken 48 h after the last OA aerosol challenge. We detected only a few randomly scattered, positively stained cells in the lung parenchyma of mice receiving three OA nebulizations alone (Figure 3A). Sensitized and challenged mice developed a marked eosinophilia and the eosinophils (MBP⁺ cells) accumulated in the peribronchial and perivascular submucosal tissue, while the lung parenchyma remained relatively eosinophil free (Figure 3B). In addition to the cellular infiltrate we observed a relative widening of the submucosal tissue probably due to edematous changes and the presence of a large number of inflammatory cells. No basement membrane thickening, epithelial disruption, or airway hyperplasia were observed, probably owing to the acute nature of our model. The inflammatory changes, particularly the peribronchial/perivascular eosinophilia, were largely eliminated in the mice treated with anti-CD23 antibody (Figure 3C).

View larger version (88K): [in this window] [in a new window]   Figure 3.   Anti-CD23 treatment abolishes eosinophilia in the peribronchial-perivascular tissue. Immunolabeling of eosinophil major basic protein (MBP) in formalin-fixed, paraffin-embedded tissue sections was performed on frozen tissue samples taken 48 h after the last OA aerosol challenge. (A) Representative tissue section from mice that received three aerosol challenges with OA. (B) Representative tissue section from mice that received intraperitoneal sensitization, three airway challenges, and treatment with control IgG. (C ) Representative tissue section from mice that received intraperitoneal sensitization, three airway challenges, and anti-CD23 treatment as described.

Supernatants of the BAL samples obtained 48 h after the last allergen challenge were analyzed for the presence of IL-4, IL-5, and IFN-. Naive, nonsensitized mice had no detectable levels of any of the cytokines studied. After sensitization and airway challenge the levels of IL-4 and IL-5 significantly increased (Figure 2D). Treatment of mice with anti-CD23 antibody significantly reduced IL-4 and IL-5 levels but, interestingly, it increased IFN- levels when compared with the effects of control antibody (Figure 2D).

Prechallenge Treatment with Anti-CD23 Reduces Eosinophilia and AHR but Does Not Affect Immunoglobulin Levels

To determine whether anti-CD23 could prevent allergic inflammatory changes when administered only before local allergen challenge, intraperitoneally sensitized mice were treated with anti-CD23 antibody before each airway allergen challenge and were compared with mice treated during the whole period of sensitization/challenge as described above. The results indicate that administration of anti-CD23 before challenge alone had a significant inhibitory effect on eosinophilic airway inflammation and AHR comparable to antibody administration throughout the sensitization and airway challenge period (Table 1). There was no significant inhibitory effect detected on OA-specific IgE and IgG₁ levels if anti-CD23 was administered only before airway challenge (data not shown). Immunoglobulin levels were already significantly elevated before allergen challenge (Table 1 and Figure 2B).

Anti-CD23 Treatment Inhibited, and Gene Targeting Enhanced Development of Allergic Sensitization

To compare the effects of anti-CD23 treatment with CD23 gene-manipulated mice, animals were sensitized by two intraperitoneal injections of OA/alum over a 14-d interval and then exposed to aerosolized OA as described over 3 d (Days 24, 25, and 26). Forty-eight hours after the last exposure the mice were studied for airway responsiveness, serum IgE and IgG₁ antibody levels, and cellular changes in the lungs. The CD23^-/- mice were developed on a C57BL/6 × 129 background. Since these mice have a slightly different profile in their allergic responses compared with BALB/c mice, as a control we used their CD23^+/+ littermates receiving the same treatment. To determine whether anti-CD23 had any effect unrelated to its capacity to bind CD23, we also treated CD23-deficient mice with the monoclonal antibody. Figure 4A presents the dose-response curves of pulmonary resistance (RL) plotted against increasing concentrations of inhaled MCh in sensitized/challenged mice. Nonsensitized, but OA-challenged animals from the CD23^-/- and CD23^+/+ groups were not significantly different and demonstrated little change in RL in response to increasing concentrations of MCh, similar to normal BALB/c mice (data for naive animals shown here only for CD23^+/+ mice). As we demonstrated previously (16), sensitization and challenge with OA resulted in significant increases in airway responsiveness in both CD23^+/+ and CD23^-/- mice, the changes being significantly higher in CD23^-/- mice. Treatment with anti-CD23 affected only the CD23^+/+ animals. In contrast, in the CD23^-/- group, anti-CD23 treatment had no effect on airway responsiveness. After anti-CD23 treatment, CD23^-/- mice remained hyperreactive while responses in the CD23^+/+ mice were significantly diminished (Figure 4A).

View larger version (38K): [in this window] [in a new window]   Figure 4.   CD23 has inhibitory effects on allergic responses. (A) Lung resistance (RL) results are expressed as the percent change from baseline resistance to methacholine (MCh). (B) OA-specific IgE and IgG1 levels (ELISA U/ml). (C) BAL cellular content. Total = total cell number in BAL sample; Mp = macrophage; Ne = neutrophil; Ly = lymphocyte; Eo = eosinophil D. BAL cytokine content (pg/ml). Data represent means ± SEM; *p < 0.05; **p < 0.01. Open bars and open squares: CD23+/+ mice; Ch: three challenges on Days 24, 25, and 26 (no sensitization) (n = 8). Closed bars and closed circles: CD23+/+ mice; Sens/Ch/Ctr IgG: intraperitoneal sensitization followed by challenges and intraperitoneal treatment with IgG (control) (n = 8). Hatched bars and open circles: CD23+/+ mice; Sens/ Ch/anti-CD23: intraperitoneal sensitization followed by challenges and intraperitoneal treatment with anti-CD23 (n = 10). Dotted bars and open triangles: CD23-/- mice; Sens/ Ch/anti-CD23: intraperitoneal sensitization followed by challenges and intraperitoneal treatment with anti-CD23 (n = 8). Gray bars and closed triangles: CD23-/- mice; Sens/Ch/ctr IgG: intraperitoneal sensitization followed by challenges and intraperitoneal treatment with IgG (control) (n = 8).

Both CD23^-/- and CD23^+/+ mice showed significantly increased OA-specific antibody production after sensitization and challenge with OA. OA-specific IgE and IgG₁ concentrations in the mice treated with anti-CD23 antibody were significantly lower than in the mice receiving control rat IgG in the CD23^+/+ mice but not in the CD23^-/- mice (Figure 4B).

The numbers of total leukocytes, neutrophils, and eosinophils recovered from BAL of sensitized and challenged CD23^+/+ and CD23^-/- mice were significantly higher than in nonsensitized animals (Figure 4C). Anti-CD23 significantly reduced the inflammatory cell influx in the airways of CD23^+/+ mice but not in those of CD23^-/- mice, resulting in a significant difference between these groups (p < 0.01). Cytokine levels in the CD23^-/- mice were not affected by anti-CD23 treatment (Figure 4D) but CD23^+/+ mice showed similar changes to that of BALB/c mice as shown in Figure 2.

These data confirm that CD23^-/- mice demonstrate increased airway responsiveness to MCh, eosinophilic inflammation, and Th2-type cytokine responses when compared with CD23^+/+ mice and these responses are unaffected by anti-CD23 treatment.

Methacholine Responsiveness and Airway Eosinophilia Is Inversely Correlated with Expression of CD23 on Splenic T and B Cells of CD23 Transgenic Mice

To define further a potential negative regulatory role of CD23, heterozygote CD23 transgenic mice (21) were sensitized and then challenged with OA via the airways. Measurements of airway function, analysis of BAL cellular content and lung eosinophilia was performed 48 h after the last airway challenge. Airway responsiveness to MCh in these sensitized and challenged animals showed a wide range of variability. The range of individual airway responses to MCh is demonstrated in Figure 5A. This variability was not seen in control (challenged alone) transgenic mice (from 256 to 476% change from baseline at 50 mg/ml MCh); the range of airway responsiveness was similar to that seen in control CD23^+/+ mice (Figure 2A). When eosinophilia was quantitated in the sensitized and exposed heterozygote CD23 transgenic mice, it correlated with the extent of AHR. Regression analysis revealed a significant positive correlation between increases in lung resistance to MCh (at 50 mg/ml) and numbers of BAL eosinophils (r = 0.765, p = 0.045) (not shown) and MBP⁺ cells in the tissue (r = 0.757, p = 0.03; Figure 5B). Flow cytometric analysis was performed on splenic mononuclear cells of these transgenic mice in order to determine the level of CD23 expressed on the cell surface of T and B lymphocytes. Results of double staining of CD23 with CD4, CD8, or B220 were analyzed. We found no CD23 expression on wild-type T cells while all CD4⁺ and CD8⁺ cells were positive for CD23 in the transgenic mice. CD23 expression on B220⁺ cells increased from ~ 50 to ~ 85% of the cells (data not shown). Expression of CD23 on nonselected splenic mononuclear cell populations ranged from 40.1 to 79.9%. There was a negative correlation between expression of CD23 and increases in lung resistance to MCh (50 mg/ml) (r = 0.766, p = 0.027; Figure 5D). Numbers of eosinophils in BAL (r = 0.926; p = 0.003; Figure 5C) as well as MBP⁺ cells in the airway mucosal tissue (r = 0.785; p = 0.021; Figure 5B) also inversely correlated with CD23 expression. These data reflect the results of statistical analyses performed on a nonselected mononuclear cell population. Analysis of CD23 expression on subpopulations of CD4⁺, CD8⁺, or B220⁺ cells in these mice also negatively correlated with MCh responsiveness and eosinophil numbers (not shown).

View larger version (18K): [in this window] [in a new window]   Figure 5.   Airway responsiveness to methacholine and airway eosinophilia inversely correlated with expression of CD23 on splenic T and B cells in CD23 transgenic mice. (A) Graphs 1- 8 represent the individual MCh dose-response curves of each CD23 transgenic mouse studied. (B) Scattergram demonstrates the relationship between lung resistance (y axis) and MBP+ cells in lung tissue (x axis). Individual lung resistance values to MCh (at 50-mg/ml dose) were plotted against the numbers of MBP+ cells in the lung tissue of each mouse. MBP+ cell numbers were evaluated as described. (C ) Scattergram demonstrates the relationship between MBP+ cells in lung tissue (closed squares), BAL eosinophil cell numbers (closed circles) (y axis) and percentage of CD23+ splenocytes (x axis) in each mouse (n = 8). CD23+ cell numbers were analyzed by flow cytometry as described in METHODS. (D) Scattergram demonstrates the relationship between lung resistance to MCh (50 mg/ml) (y axis) and percentage of CD23+ splenocytes (x axis) in each mouse (n = 8).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Human allergic asthma is characterized by a Th2-dominant immune response in which IgE and eosinophils are conspicuously elevated. It is suggested that IgE has a central role in the development of disease (24), but the function and importance of IgE-mediated mechanisms are largely unknown, especially in chronic asthma. We are able to demonstrate for the first time, in a complex murine model of allergic airway hyperresponsiveness, that CD23 exerts a limiting function. In this study, we describe how CD23 has an inhibitory role in the development of allergic airway inflammation and AHR by comparing the in vivo effects of gene deficiency and overexpression in transgenic mice with those of anti-CD23 treatment. The genetic absence of CD23 was associated with increased allergic responses to OA sensitization and challenge while overexpression of CD23 was inversely correlated with these changes. Correspondingly, anti-CD23 induced negative regulatory effects on IgE and IgG₁ production as well as on airway inflammation and AHR after allergic sensitization. This suppression was associated with a shift in locally produced cytokines manifested by increases in IFN- and decreases in IL-4 and IL-5 production.

It has been suggested that IgE, via interactions with CD23, mediates augmentation of specific immune responses through increased antigen presentation, which may alter the nature of T cell help to B cells (14, 25). Antigen-IgE immune complexes may modulate T and B cell interactions and alter antigen-specific T cell functions. We previously showed that the pattern of cytokine production may be modulated by IgE-immunocomplex binding to CD23 in cell cultures after addition of IgE (15). In these cultures, the effects of exogenous IgE were absent when splenocytes from CD23-deficient mice were used or were reversed by anti-CD23 (15). The in vivo contribution of these CD23-IgE-mediated events to the development of AHR is not entirely clear. We previously studied allergic responses in a mouse model in which mice were passively sensitized with OA-specific IgE before OA challenge and found that this method elicited airway inflammation and AHR (23). It is interesting that when we repeated these experiments in CD23^+/+ and CD23^-/- mice, although the latter were impaired in their capacity to enhance OA-specific immunoglobulin production, they still developed significant airway changes (16). Thus, facilitated antigen presentation through IgE-CD23 may be of little importance in this murine model of AHR or at least under normal circumstances.

To extend our observations on the role of CD23 and to elucidate further the role of IgE-mediated events and CD23 during active sensitization, we specifically used two models of induced airway hyperresponsiveness distinguished by different IgE dependencies. Aerosolization of OA for 10 consecutive days in the absence of adjuvant was shown to induce allergen-specific IgE and T cell activation associated with immediate cutaneous hypersensitivity (26), as well as tracheal smooth muscle hyperreactivity to electrical field stimulation (11). Since B cell-deficient mice were unable to exhibit these airway changes but could be reconstituted after passive sensitization with OA-specific IgE, airway hyperreactivity evoked after 10-d OA aerosol treatment appeared to be IgE dependent (10). In contrast, after systemic sensitization and repeated airway challenge, the development of AHR has been shown to be IgE independent (18), in spite of the high IgE levels produced in these mice. In the present experiments anti-CD23 suppressed AHR in both the IgE-dependent and IgE-independent models of AHR. These data suggest that binding of antibody to CD23 may evoke negative regulatory mechanisms that are independent of IgE-mediated events.

At the present time, it is unclear whether the negative regulatory effects seen after anti-CD23 treatment are the result of activation of target cells through CD23 or secondary to blocking the interaction of CD23 on the cell surface with other effector molecules such as IgE or CD21. In a rat model, a positive feedback loop on IgE synthesis was proposed, as both whole IgG and Fab fragments of a polyclonal rabbit anti-CD23 were found to inhibit IL-4 induced IgE production. Since Fab fragments did not trigger signal transduction, the authors concluded that anti-CD23 exerted its effects exclusively by blocking IgE binding to CD23 or interactions with CD21 (27). Such positive feedback function could not be confirmed in mice (16, 21, 28, 29), indicating potential species differences in the function of CD23. These differences may be explained by the differential function of the soluble and membrane-bound forms of CD23. Human soluble CD23 increases spontaneous as well as IL-4-induced IgE production (2), while the membrane-bound form of CD23 may have an inhibitory function (30). Although murine soluble CD23 may be able to bind IgE with reduced affinity, its functional importance remains unclear (31), suggesting that in murine models, membrane-bound CD23 may be the functionally dominant form. The antibody used in our studies (B3B4) has been extensively characterized (32) and exhibits a high affinity for CD23 as well as a reciprocal inhibitory pattern with IgE, suggesting a close proximity of their binding sites. Other studies have also implicated CD23 in negative regulatory aspects. For example, CD23 may directly inhibit B cell activation and immunoglobulin production, especially IgE (17). A decrease in IgE production in CD23 transgenic mice in vivo was confirmed with purified B cells from the transgenic animals when stimulated with CD40 ligand and IL-4 (21). Thus, binding of antibody to CD23 may inhibit allergic responses by triggering signals that have negative regulatory effects on B cell function, B cell-T cell interactions, or T cell function. These conclusions appear confirmed by a study in which Cernadas and coworkers also examined the allergic response in CD23-deficient mice and in mice treated with anti-CD23 antibody (33). Administration of anti-CD23 MAb, but not anti-CD23 Fab fragments, inhibited pulmonary inflammation and AHR. On the basis of a model that the anti-CD23 MAb transduces, whereas the Fab fragment inhibits, CD23 signaling, their results also suggested that CD23 negatively regulates pulmonary inflammation and AHR.

Since eosinophilic airway inflammation is a hallmark of human asthma and it is also a characteristic feature of our mouse models of allergic AHR, we determined whether manipulation of CD23 would affect airway eosinophilia and whether these changes were related to altered airway function. Sensitized and challenged mice developed a marked increase in the number of eosinophils both in the BAL fluid and in the airway submucosal tissue when compared with nonsensitized controls. Immunolabeling of the lung tissue indicated an accumulation of MBP⁺ eosinophils largely in the peribronchial- perivascular area. Lung tissue eosinophilia was quantified in CD23 transgenic mice and showed a significant correlation with the degree of AHR. These data, revealing a significant positive correlation between the extent of AHR and number of MBP⁺ cells in the submucosal tissue, indicate that a causal relationship may exist between these parameters in this model. The exact mechanism whereby eosinophils affect airway function is not known, but release of membrane-derived and granule-associated products from activated eosinophils may contribute to airway hyperresponsiveness. Human eosinophils were shown to express IgE receptors, which participate in IgE-dependent reactions (34), suggesting that CD23 may play a role in IgE-mediated eosinophil degranulation. On the other hand, it has been shown that eosinophils in mice do not express CD23 or Mac-2 and show no binding of IgE (35), excluding CD23 as a direct modulator of murine eosinophil function. In the present studies, the numbers of eosinophils were significantly enhanced in CD23^-/- mice and inversely correlated with CD23 expression in CD23 transgenic animals. In addition, anti-CD23 reduced eosinophilia and AHR, even when administered just before allergen challenge. These data suggest that CD23 has no direct role in eosinophil activation in mice, but likely has an indirect negative effect on recruitment of eosinophils to the airways. Further, this inhibitory effect appears to be linked with suppression of AHR in our murine model.

Since development of eosinophilic inflammation is dependent on the production of Th2 cytokines, we investigated whether the effects of anti-CD23 treatment and gene targeting were associated with altered cytokine production in the airways. We previously found that T cell cytokine synthesis can be modulated by IgE-immunocomplexes in vitro, as IFN- production was decreased after addition of specific IgE to splenocyte cultures (15). This effect was reversed by anti-CD23, suggesting that binding to this molecule can influence T cell function. Since normal murine T cells do not express CD23, it is possible that antigen-presenting cells carrying CD23 on their surface can influence T cell cytokine production indirectly. Indeed, we have shown that differential inhibition of the costimulatory molecules B7.1 and B7.2 on antigen- presenting cells alters the cytokine profile of cultured T lymphocytes in an OA-specific system. Here, we analyzed cytokine levels in the BAL of mice sensitized to OA and treated with anti-CD23 and found that the inhibitory effects of the antibody on airway function and eosinophilia were associated with relative increases in IFN- and suppression of IL-4 and IL-5 levels. These data imply that anti-CD23 results in a shift in T cell cytokine production, where Th2-type cytokines no longer dominate. Such changes in the cytokine profile may be responsible for the inhibition of antigen-specific IgG₁ and IgE production, tissue eosinophilia, and AHR observed in our studies.

In summary, the role of CD23 in the development of allergic responses is complex. The studies suggest that this molecule is not essential in the development of allergic AHR, at least in murine models. On the other hand, taken together, the studies on targeting the CD23 gene, increasing or eliminating CD23 expression, as well as the results with anti-CD23, indicate that CD23 expression is associated with negative regulatory effects on allergic inflammation in mice. We propose a novel mechanism in which CD23 not only affects IgE production but significantly-even if indirectlymodifies development of Th2 responses on allergic stimulation. CD23 may therefore act not simply as a feedback molecule of the IgE circuit, but may have more general regulatory functions in allergic airway hyperresponsiveness.