INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Asthma is a chronic respiratory disease characterized by the presence of reversible airway obstruction due to inflammation and bronchoconstriction along with airway hyperresponsiveness (AHR) to a variety of stimuli. Acute exacerbations of asthma appear to be related to environmental allergen exposures (1) or viral respiratory infections (4); however, there is a significant genetic predisposition to the development of asthma in humans (5). Children who have one parent with asthma have a threefold increase in their likelihood of having asthma and children whose parents both have asthma have a sevenfold increase in their likelihood of having asthma compared with children whose parents do not have asthma (6). Twin studies have revealed that monozygotic twins of people with asthma have a 65% likelihood (fourfold that of the general population in their study) and dizygotic twins a 25% (twofold) likelihood of having asthma (7). Although studies such as these have established that there is a strong genetic basis for the predisposition to develop asthma, additional studies to localize and ultimately identify the gene or genes coding for this predisposition are difficult, especially in genetically heterogeneous human populations. Clinical, physiological, and genetic information must be obtained from very large numbers of people and large multicenter or sequential studies may be confounded by inconsistency of clinical definitions and lack of uniformity of testing methods.

Murine studies searching for genes that predispose to asthma have advantages offered by the genetic homogeneity, ready availability, and low cost of standard inbred strains of mice for experimental use. Recent studies have established the reproducibility of physiological, bronchoalveolar lavage (BAL), and histological assessment of the murine respiratory system (8). Some of those studies have also displayed applicable murine models of human asthma (11, 12). Mice can be characterized prior to the application of standardized experimental antigen challenge protocols. The short reproductive interval of mice, the ability to perform selective breeding, and the increasingly detailed analysis of the murine genome (13) promise rapid progress in localizing relevant disease genes. Murine studies have begun to identify genetic regions associated with immunoglobulin (Ig)E production (14), airway hyperresponsiveness (15), and eosinophilic bronchitis (16).

The usefulness of inbred mouse strains for genetic studies of the predisposition to develop asthmalike pulmonary changes depends on the presence of wide differences among strains in this trait. Such variability would also permit analysis of possible links between antigen-induced enhancement of airway responsiveness and traits thought to predispose to asthma, including lung eosinophilia and/or production of high IgE levels.

We present here a characterization of the pulmonary physiological and cellular alterations and serum IgE levels observed in 12 inbred mouse strains after controlled experimental antigen sensitization and exposure. We found marked variability among strains in the development of enhanced cholinergic pulmonary responsiveness, pulmonary eosinophilia, and of increased antigen-specific serum IgE levels, confirming a large degree of genetic control for each of these traits in mice. Mice of different strains exhibited a significant association between the magnitudes of antigen-induced shifts in airway responsiveness and the degree of induced eosinophilic bronchitis; the correlations between shifts in responsiveness and serum antigen-specific IgE levels were of borderline significance.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Mice

Male mice of each of 12 strains were obtained from Jackson Laboratories (Bar Harbor, ME) at 6 to 8 wk of age. Mice were housed in microisolator cages under BL1 viral antibody-free conditions, and were fed standard rodent chow, which is known to be free of chicken albumin. Our study protocol was approved by the Children's Hospital Animal Care and Use Committee.

Sensitization

Mice were sensitized intraperitoneally with 10 µg chicken ovalbumin (OVA, Grade III; Sigma, St. Louis, MO) and 1 mg Al(OH)₃ in 0.2 ml of 0.9% NaCl. Sham sensitizations consisted of 1 mg Al(OH)₃ in 0.2 ml of 0.9% NaCl. These injections were repeated 14 d later. Depending on the protocol, 7 or 10 d after the second injection, mice were exposed to either 0.5 N phosphate-buffered saline (PBS) or OVA diluted in 0.5 N PBS using an ultrasonic nebulizer (Model 5000D; DeVilbiss, Somerset, PA). A bias flow of 3 L/min was used to deliver aerosolized solutions (mean particle diameter, 1 to 3 µm) into an enclosed 22 × 22 × 12 cm Plexiglas chamber with a 6-mm hole in the opposite wall allowing for air flow.

For the initial OVA exposure regimen, mice were exposed to a 6% (wt/vol) OVA solution administered for 60 min daily for 7 d (Protocol 1). A second OVA aerosol exposure regimen, in which mice were exposed to 1% OVA or 0.5 N PBS for 20 min daily for 4 d (Protocol 2), was applied to six strains whose responses to OVA administered in accordance with Protocol 1 spanned the observed range of responses. Third, sham-sensitized groups of mice of each of the 12 strains studied with Protocol 1 were exposed to 6% aerosolized OVA for 60 min daily for 7 d and compared with control sham-sensitized mice exposed to 0.5 N PBS.

Measurement of Airway Narrowing and AHR

According to our previously published method (9), 20 to 28 h after the final aerosol challenge, mice were anesthetized with sodium pentobarbital (70 to 90 mg/kg, intraperitoneally), a 19-gauge tubing adapter (Becton Dickinson, Rutherford, NJ) was inserted into the trachea, and ventilation was instituted via the tracheostomy tube using a tidal volume of 5 to 7 ml/kg at a rate of 150 breaths/min. After a thoracotomy was performed, lung conductance (GL) values were obtained prior to agonist challenge to assess whether airway narrowing had developed in OVA-sensitized mice. GL, in ml · s¹ · cm H₂O, was calculated as the inverse of the difference between total respiratory resistance at midlung volumes and the resistance of the tracheostomy tube (0.75 cm H₂O · ml¹ · s). For each value, measurements from 10 consecutive breaths were averaged.

Logarithmically increasing doses of methacholine (MCh) from 3.3 to 3,300 µg/ml were dissolved in 0.9% NaCl and administered in a volume of 1 µl per gram of body weight via a jugular venous catheter over a 3- to 4-s interval. Maximally reduced GL values observed after each dose were recorded as percentages of the values obtained immediately before that dose. Change in airway responsiveness was measured as the ED₅₀GL, which represents the differences between the logarithms of the MCh doses required to obtain a 50% reduction in GL, comparing sham-sensitized, PBS-exposed mice with OVA-sensitized, OVA-exposed mice. We also compared sensitized versus control mice of each strain in terms of maximal responses. Because a 1 mg/kg dose of MCh induced maximal responses in each strain, we measured the differences between the magnitudes of the reductions in GL observed after administration of that dose and refer to that difference as 1 mg/kg GL.

Measurement of Eosinophil Influx

Immediately after pulmonary measurements, BAL was performed by instilling and retrieving through the tracheostomy tube three 1.0-ml aliquots of 1× PBS containing 0.6 mM ethylenediaminetetraacetic acid (EDTA). After cytocentrifugation of recovered fluid at 800 rpm for 10 min, slides were stained with Diff-Quik Stain Set (Baxter, McGaw Park, IL) and differential cell counts were performed.

Quantification of Peribronchial Inflammation

After completion of BAL, a portion of the left lung was fixed in 10% buffered formalin, embedded in paraffin, cut into 5-µm-thick sections, and stained with hematoxylin and eosin for histological examination. Semiquantitative histological scoring was performed according to the scheme shown in Table 1. The extent score and the severity score were multiplied to obtain an inflammation score, and this value was multiplied by the % eosinophils score and the product divided by 3 to obtain an eosinophilic inflammation score; the maximal possible eosinophilic inflammation score for a specimen would thus equal its inflammation score.

OVA-specific Serum IgE Level Measurement

OVA-specific IgE concentrations were determined before sensitization and again just before and after the 7 d of aerosolized OVA exposure in a separate group of mice of 11 of the same strains subjected to the same sensitization and exposure protocol. For OVA-specific IgE measurements, a modification of the protocol provided by Pharmingen was followed. We coated 96-well plates with purified anti-IgE (R35-72; Pharmingen, San Diego, CA) at 3 µg/ml. The next day, the plates were washed and blocked with 3% bovine serum albumin (BSA)/PBS for 2 h at room temperature, and washed again before serum diluted at 1:10 to 1:15 and standard IgE samples (IgE-3; Pharmingen) at concentrations of 6.25 to 200 ng/ml were added and allowed to stand overnight. Plates were washed and incubated for 1 h at room temperature with biotinylated OVA at 1:5,000 dilution. Plates were coated with avidin-conjugated horseradish peroxidase (Zymed, San Francisco, CA) and incubated for 30 min at room temperature. Plates were then washed again and substrate, azino-bis(3-ethylbenzthiazoline-6 sulfonic acid) (Zymed) added. Optical densities of wells were read 10 to 20 min later at 405 nm by a plate reader (Diagnostics Pasteur; Kalledstad Diagnostics, Chaska, MN). Concentrations of IgE in sera were extrapolated from a graph of standard optical densities versus concentration.

Statistical Analyses

Comparison of baseline pulmonary function among groups of mice was performed using unpaired two-tailed Student's t tests with the Bonferroni correction. The MCh dose required to induce a 50% decrease in GL was obtained by graphical interpolation as previously described (9). ED₅₀GL values were obtained by subtracting the mean ED₅₀GL values of the OVA-sensitized and exposed group of mice of each strain from that of the sham-sensitized and exposed group; hence only a single difference between mean values, without an estimate of standard deviation, can be obtained for each strain. Significance of relationships between ED₅₀GL values and eosinophil percentages in BAL cells or severities of eosinophilic infiltration were assessed using Pearson correlation. Significance of the relationship between ED₅₀GL values and serum OVA-specific IgE levels was tested with Spearman correlation because IgE concentrations did not fit a Gaussian distribution.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Baseline GL values measured 20 to 28 h after the last dose of aerosolized OVA but before MCh administration in OVA-sensitized, OVA-exposed mice ranged from 72% to 117% of those from sham-sensitized, PBS-exposed mice of the same strains (Table 1). None of these differences between OVA- and sham-sensitized groups was significant.

The ED₅₀GL values (Figure 1a) reflect differences in MCh responsiveness between the sham-sensitized, PBS-exposed mice and the OVA-sensitized, OVA-exposed mice for each strain. They ranged from 0 to 0.43, indicating that OVA sensitization and exposure caused no change in airway responsiveness in some strains (e.g., A/J) and up to a 2.7-fold reduction in the MCh dose required to induce a 50% decrease in GL in more susceptible strains (e.g., SWR). 1 mg/kg GL values also exhibited substantial variability among strains, ranging from 0 in A/J mice to 23% in FVB mice (Figure 1b).

View larger version (15K): [in this window] [in a new window]   View larger version (13K): [in this window] [in a new window]   Figure 1.   Magnitudes of shifts in airway responsiveness, expressed as (a) ED50GL or as (b) 1 mg/kg GL induced by intraperitoneal sensitization to OVA followed by exposure to aerosolized OVA according to Protocol 1 for mice of 12 inbred strains (n = 4 to 6 mice per strain). Enhancement of ED50GL due to OVA exposure ranged from zero in A/J mice to as much as log 0.43, i.e., by 3-fold, in SWR mice. Increases in effect of 1 mg/kg dose of MCh (1 mg/kg GL) ranged from 9% in AKR mice to 23% in FVB mice. In both figures the strains are portrayed in the same order that reveals that the two parameters did not correlate with one another. Unless otherwise noted, in all figures the absence of data indicates a value of 0.

Eosinophil influx varied markedly among strains studied with Protocol 1, with BAL eosinophils comprising from 3.3 ± 3.1% (SD) to 91.2 ± 5.0% of the BAL cells obtained from OVA-sensitized, OVA-exposed mice (Figure 2a). None of the sham-sensitized, PBS-exposed mice exhibited > 2% eosinophils. Lymphocytes comprised less than 7% of the BAL cells from mice of each strain except sensitized CBA mice, which exhibited 17% lymphocytes. Polymorphonuclear cells did not comprise more than 5.7% of BAL cells in mice of any strain.

View larger version (19K): [in this window] [in a new window]   View larger version (30K): [in this window] [in a new window]   Figure 2.   (a) Eosinophilia in cells obtained by BAL and (b) inflammation and eosinophilic inflammation in lung tissue induced by intraperitoneal sensitization to OVA followed by exposure to aerosolized OVA according to Protocol 1 for mice of 12 strains (n = 8 to 12 mice per strain). Histologically evident periluminal inflammation and eosinophilic inflammation, scored according to the scheme shown in Table 1, was barely detectable in lungs from A/J mice, whereas severe and extensive inflammation was found in lung tissue of FVB and SWR mice. The strains are portrayed in the same order as in Figure 1.

The degree of eosinophilic bronchial inflammation assessed histologically also varied dramatically among the groups of sensitized mice of different strains exposed to aerosolized OVA according to Protocol 1 (Figure 2b). Eosinophilic inflammation scores ranged from 0.12 for the 129/Sv mice to 6.20 for mice of the SWR strain. As reported by others (11, 12), the inflammation was not localized solely around bronchioles; substantial perivascular eosinophilic inflammation was also present in all affected mice.

Mice of the different strains also varied dramatically in their serum OVA-specific IgE concentrations measured after OVA sensitization and aerosol exposure, ranging from none detectable (< 3 ng/ml) to 455 ng/ml (Figure 3a). OVA-specific IgE was not detected in any serum samples obtained prior to sensitization and OVA-specific IgE levels measured after intraperitoneal sensitization but just before OVA aerosol exposures also exhibited considerable variation among strains (Figure 3b). Notably, mice of four strains, C3H/He, C57BL/6, BALB/c, and DBA/2, exhibited minimal levels at that point but higher levels after the aerosol exposures.

View larger version (15K): [in this window] [in a new window]   View larger version (11K): [in this window] [in a new window]   Figure 3.   (a) Serum OVA-specific IgE concentrations after intraperitoneal OVA sensitization and aerosol exposure to OVA for mice of each of 11 strains (n = 3 to 4 mice per strain). (b) OVA-specific IgE concentrations in serum obtained after intraperitoneal sensitization but immediately prior to beginning OVA aerosol exposures. The strains are portrayed in the same order as in Figure 1.

The effects of OVA sensitization and exposure according to Protocol 2, which were generally of lesser magnitude and detected in a smaller proportion of strains than those obtained with Protocol 1, also varied among strains (Figure 4). Moreover, there were interstrain differences in the relative effects of Protocol 2 as compared with Protocol 1; the less intensive exposure had the same effect on airway responsiveness as the more intensive exposure in A/J, 129, and BALB/c mice, whereas much smaller effects were seen after the less intensive exposure in DBA/1 and FVB mice.

View larger version (9K): [in this window] [in a new window]   View larger version (15K): [in this window] [in a new window]   Figure 4.   (a) Shifts in airway responsiveness, expressed as ED50GL, and (b) eosinophilia in cells obtained by BAL and eosinophilic inflammation in lung tissue induced by intraperitoneal sensitization to OVA followed by exposure to aerosolized OVA via Protocol 2. The mice strains are displayed in the same order as in Figure 1. The lower concentration of OVA in solution to be aerosolized and the reduction in duration and number of OVA exposures compared with Protocol 1 led to reductions in BAL eosinophilia and eosinophilic inflammation only for those strains most severely affected by OVA exposure according to Protocol 1. (An eosinophilic inflammation score was unavailable for C57BL/6 mice exposed according to Protocol 2.)

Effects of exposure to 6% OVA for 60 min/d for 7 d in sham-sensitized mice, while never greater than those observed in OVA-sensitized mice, also varied markedly among the different strains (Figure 5). Significantly enhanced MCh responsiveness was observed only in AKR, FVB, and SWR mice, the three strains that developed the greatest pathophysiological alterations after sensitization and exposure in accordance with Protocol 1. In these sham-sensitized, OVA-exposed mice, eosinophils comprised less than 7% of BAL cells and eosinophilic inflammation scores were less than 1.0 (data not shown).

View larger version (13K): [in this window] [in a new window]   Figure 5.   Magnitudes of shifts in airway responsiveness, expressed as ED50GL, induced by exposure to aerosolized OVA via Protocol 1 after sham sensitization (n = 6 to 12 mice per strain). Strains are displayed in the same order as in Figure 1. Enhancement of ED50GL due to OVA exposure was significant only in AKR, FVB, and SWR mice.

BAL eosinophil percentages and severity of eosinophilic inflammation each correlated significantly with ED₅₀GL (r = 0.710 and 0.766, respectively, p < 0.001, Figure 6). C57BL/6 and C57BL/10 mice were unusual in that they developed substantial BAL eosinophilia and eosinophilic inflammation with only minimal enhancement of airway responsiveness. Eosinophil percentages in BAL fluid correlated strongly with periluminal eosinophilic inflammation scores (r = 0.836, p < 0.001). Neither BAL lymphocyte nor polymorphonuclear leukocytes (PMN) percentages correlated significantly with the ED₅₀GL values. There was a weak correlation between OVA-specific IgE levels and ED₅₀GL values for 11 strains subjected to OVA exposure via Protocol 1 (r = 0.55, p = 0.056). There were two striking exceptions to the general correlation between OVA-specific IgE concentrations and degrees of enhancement of airway responsiveness after OVA stimulation; BALB/c mice exhibited the highest IgE concentrations but only a modest ED₅₀GL value, and AKR mice developed no detectable increase in OVA-specific IgE levels despite a large ED₅₀GL value.

View larger version (8K): [in this window] [in a new window]   View larger version (9K): [in this window] [in a new window]   Figure 6.   (a) Mean percentages of BAL cells found to be eosinophils and (b) mean eosinophilic inflammation scores of OVA-sensitized, OVA-exposed groups of mice versus mean changes in responsiveness to intravenous MCh, expressed as ED50GL, induced by OVA sensitization and exposure. Data from Protocols 1 and 2 are combined (n = 6 to 12 mice per point). Both correlations were significant (r = 0.710 and 0.766, respectively, p < 0.001). In both panels, a = C57BL/6 mice exposed to OVA according to Protocol 1; b = C57BL/10 mice exposed to OVA according to Protocol 1.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We have shown that inbred mouse strains exhibit significant genetic variability in their susceptibilities to develop asthma-like enhancement of airway responsiveness, pulmonary eosinophilia, and elevation of antigen-specific serum IgE levels due to standardized airway antigen exposure after systemic sensitization. The magnitudes of antigen-induced shifts in airway responsiveness to MCh correlated significantly with the degrees of BAL and lung tissue eosinophilia. The correlation between enhancement of responsiveness and the presence of elevated OVA-specific IgE was just below that required for statistical significance. The extensive genetic variability in susceptibilities to develop these responses to antigen thus leads to evidence supporting a close, if not causal, relationship between antigen-induced pulmonary eosinophilia and enhancement of cholinergic responsiveness.

Other investigators have demonstrated and used interstrain differences in antigen-induced, asthmalike pulmonary alterations in mice (16, 17). Kuperman and coworkers examined recombinant inbred mouse strains whose parent strains, A/J and C3H/HeJ, exhibited differing responses to airway antigen exposure. They thereby localized the genetic region predisposing to eosinophilic bronchitis in mice to chromosome 11 (16). Interestingly, in contrast to our results, those investigators found A/J mice to be susceptible to OVA-induced pathophysiological changes, perhaps related to a difference in method of airway antigen exposure. Zhang and colleagues reported that after OVA sensitization and aerosol exposure, BALB/c and C57BL/6 mice developed virtually identical antigen-specific IgE and IgG₁ levels and degrees of airway eosinophilia, but only BALB/c mice developed enhanced airway responsiveness (17). The current study provides a more complete and detailed characterization of the responses of different mouse strains to antigen.

We observed a significant positive relationship between BAL eosinophil percentages or histological eosinophilic periluminal infiltrate scores and ED₅₀GL values. This correlation could result from a strong genetic link between a gene predisposing to development of antigen-induced eosinophilic bronchitis and a gene predisposing to enhancement of airway responsiveness. An alternative explanation for this correlation is that eosinophilic bronchitis may cause AHR, as previously suggested (18). Our findings that C57BL/6 and C57BL/10 mice developed severe eosinophilic bronchitis with minimal enhancement of responsiveness and that sham-sensitized, OVA-exposed mice of three strains developed enhanced responsiveness with minimal airway eosinopilia suggest that the relationship is not causal and/or that other genetically controlled factors can modify this link.

The pulmonary effects of the different sensitization and exposure protocols in the mouse strains permit inferences about the nature of their genetic basis. Comparison of the responses of sensitized mice exposed via Protocol 1 with those sensitized and exposed via Protocol 2 or with those exposed to OVA without prior sensitization reveals three patterns. Mice of some strains (A/J, 129) did not respond regardless of the concentration of the antigen or the duration or number of exposures. Mice of other strains (BALB/c) developed pulmonary alterations after OVA exposure but the magnitude of their pulmonary alterations did not increase with intensity of antigen exposure. The third pattern was shown by strains that developed minor enhancement of airway responsiveness after OVA exposure via Protocol 2, greater pulmonary alterations after exposure via Protocol 1, and significant enhancement after OVA exposure without prior intraperitoneal sensitization. Based on these observations, it is tempting to speculate that one genetic effect confers susceptibility to develop asthmalike responses to mild antigenic exposure, whereas another genetic effect limits the magnitude of antigen-induced pulmonary responses.

The information presented here allows reexamination of associations between preexisting traits such as atopy (21) and major histocompatibility complex (MHC) type (25) and subsequent susceptibility to develop asthmalike pulmonary changes upon antigen exposure. Our results show an insignificant correlation between ED₅₀GL and antigen-specific serum IgE levels and that AKR mice developed pulmonary eosinophilia and enhanced airway responsiveness in the absence of detectable OVA-specific IgE. These data are also consistent with previous findings that eosinophilic bronchitis and enhanced airway responsiveness can develop in mice lacking the ability to produce IgE (26), IgG (27), or CD40 (28), indicating that at least in mice these responses can be generated independently of immunoglobulin production. These results support the notion that while atopy is associated with asthma, the genetic predispositions to develop atopic and asthmatic responses to antigen are distinct (29). This is not altogether surprising because approximately 20% of people with asthma have "nonatopic" asthma with normal serum IgE levels and no clinical evidence of other allergic disease (29) and half of people with clinical atopy have no symptoms or medical histories indicative of asthma (21, 22).

Our data also argue against a connection between the MHC allotypes of the different strains and their susceptibilities to develop enhanced pulmonary cholinergic responsiveness after OVA exposure (25). We found that mouse strains of a given MHC allotype do not exhibit consistent susceptibilities to these pulmonary responses to OVA. Specifically, AKR mice, which possess the k allele at all the major MHC class II loci, were among the most susceptible to development of enhanced airway responsiveness, whereas CBA and C3H/He mice, which also possess the H-2^k haplotype, were among the least susceptible to development of enhanced airway responsiveness after OVA exposure.

In conclusion, we have demonstrated striking variability in the degree to which different inbred mouse strains develop asthmalike pulmonary hyperresponsiveness and eosinophilic bronchitis after being subjected to standardized antigen sensitization and exposure regimens. We observed a significant correlation between the magnitudes of induced pulmonary hyperresponsiveness and eosinophilic bronchitis, and a lesser correlation between shift in responsiveness and antigen-specific IgE levels, and noted striking exceptions to each relationship. The mouse presents an excellent animal model in which to pursue genetic studies to identify the etiologies of the susceptibility to develop each element of the asthma syndrome.