Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Airway mucosal eosinophilia is a prominent feature of allergic airway diseases such as asthma and rhinitis (1, 2). By releasing their cytotoxic granule contents, eosinophils are believed to play a key role in causing tissue damage and symptoms (3). In support of this, ultrastructural studies have demonstrated extensively degranulated eosinophils in the airway tissue during active disease (1, 4). Accordingly, high levels of extracellularly deposited granule proteins have been demonstrated by histochemical analysis of the diseased tissue (7). Due to their potential importance, the degranulating eosinophils should also be an important facet of animal models of allergic airway diseases.

Mouse models are widely and increasingly used for the investigation of immunologic regulation of asthma and allergic rhinitis. In this research field, numerous details regarding the interplay between inflammatory cells and mediators have been described (8). Yet, the basic airway pathologic changes, including the degranulation status of the airway mucosal eosinophils, remain incompletely examined in allergic mice (9). Clearly, significant nasal, peribronchial, and perivascular airway eosinophilia occur in mouse models of asthma (12, 13). It may have been taken for granted that eosinophils in these heavily allergen-exposed mouse airways exhibit the features of human degranulating eosinophils. However, this aspect has not been examined in detail.

Ultrastructural analysis using transmission electron microscopy (TEM) is currently the only technique available that can detect and distinguish between the various eosinophil degranulation pathways (14, 15). Two major modes of degranulation have been described in vivo in asthma and rhinitis: piecemeal degranulation (PMD) (14) and eosinophil cytolysis (ECL) (6, 15, 16). The ultrastructural signs of PMD include partial to complete loss of matrix and/or core from the specific secondary granules (14). ECL is characterized by chromatolysis, rupture of the cell membrane, and release of the complete set of intact intracellular granules into the surrounding tissue (16). Human eosinophils in vitro also exhibit exocytosis (17) whereby secondary granules fuse with the cell membrane to release their contents extracellularly.

We have recently devised novel methods for the quantification of these modes of eosinophil degranulation in human airway tissues (5, 6, 15). Using the same TEM analysis approach, this study determines the extent of ECL and PMD, as well as exocytosis, in airway tissues of ovalbumin (OVA)-sensitized and -challenged mice. In addition, histochemistry is used to identify any extracellular deposition of the granule protein eosinophil peroxidase (EPO). The BALB/c and C57BL/6 strains of mouse were selected for the present study because these strains are commonly used in murine models of asthma. Both TEM analysis and EPO histochemistry are stimulus- and species-independent techniques, thus allowing proper comparisons between findings in the animal model and previous observations in the actual human disorder in vivo. To expand the comparison between mouse and human eosinophils the present in vivo approach was complemented with studies of mouse and human eosinophils in vitro. Interestingly, our data indicate major differences in degranulation of mouse and human eosinophils both in vivo and in vitro.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Reagents

The following compounds were purchased from Sigma Chemical Co. (St. Louis, MO): agarose (type VII), aluminum hydroxide, ammonium chloride, chicken serum albumin (OVA, grade III), cytochalasin B, 3,3-diaminobenzidine tetrahydrochloride (DAB), ethylenediaminetetraacetic acid (EDTA), formaldehyde, formyl-Met-Leu-Phe (fMLP), glutaraldehyde, hydrogen peroxide (H₂O₂), Kaiser's medium, May-Grünwald and Giemsa stain, osmium tetroxide, phorbol myristate acetate (PMA), sodium cyanide, sodium hydrogen carbonate, streptomycin sulfate, and tris(hydroxymethyl)aminomethane (Tris). Türks solution was from Merck (Darmstadt, Germany). Phosphate-buffered saline (PBS) (without calcium and magnesium), fetal calf serum (FCS), and Dulbecco's medium (DMEM) (without phenol red) were from Life Technologies, Ltd. (Paisley, UK). OCT embedding medium was from Miles, Inc. (Elkhart, IN). Benzylpenicillin was from AstraZeneca (Södertälje, Sweden). Poly/Bed 812 was from Polysciences, Inc. (Eppleheim, Germany).

Animals

Male BALB/c or C57BL/6 mice (15 to 20 g) were obtained from Bomholtgaard (Ry, Denmark). The animals were maintained under conventional conditions in constant temperature, humidity, and ventilation, on a 12-h light/dark cycle. Animals had free access to water and food. All animal procedures were reviewed and approved by the Local Animal Research Ethical Committee of Lund/Malmö, Sweden.

Sensitization and Allergen Challenge Protocols

The OVA sensitization and challenge protocols I to III (BALB/c mice) were modified from previous studies (12, 13, 18, 19). Protocol IV (C57BL/6 mice) has been described in detail previously (20, 21). Briefly, BALB/c mice were sensitized with an intraperitoneal (i.p.) injection of a 0.3-ml suspension of OVA (25 µg/ml) and aluminum hydroxide (5 mg/ml) in 0.9% saline. Control animals were given saline only. An identical i.p. booster injection was administered 7 d later. The animals were challenged by an OVA aerosol (1% in PBS) for 1 h at Days 14 and 21 (Protocol I); Days 14, 21, and 22 (Protocol II); or Days 14, 21, and 28 (Protocol III). For Protocol IV, C57BL7/6 mice were sensitized with an i.p. injection of a 0.4-ml suspension of OVA (25 µg/ml) and aluminum hydroxide (2.5 mg/ml) in 0.9% saline. From Day 14 to Day 20, the mice were challenged daily for 30 min by a 1% OVA aerosol. Control animals were given saline only. For all protocols, six animals from each group (control or OVA-challenged) were used. An ultrasonic nebulizer driven at 4 bar (Bird 500 ml Inline Micronebulizer; Bird Co., Palm Springs, CA), produced the aerosol in all experiments. At 24 h after the last allergen challenge, the mice were killed by an overdose of sodium thiopental (Abbott, North Chicago, IL) and the lungs, nasal septum, bone marrow, and blood were collected for histopathologic examination.

Histochemistry

The airway tissue samples (lungs or nasal septum) were placed in 4% formaldehyde/PBS. After fixation for 4 h, the specimens were rinsed in buffer (Tyrode buffer supplemented with 10% sucrose), embedded in OCT mounting medium, and snap-frozen at 70°C. Tissue eosinophils were identified by histochemical staining of cyanide-resistant EPO activity (20). Briefly, cryosections (7 µm) were placed in Tris-HCl buffer (pH 7.6) for 2 min and then incubated for 8 min at room temperature in the same buffer, supplemented with DAB (75 mg/100 ml), H₂O₂ (100 µl/100 ml), and NaCN (50 mg/100 ml). Slides were washed in PBS, counterstained with hematoxylin, mounted in Kaiser's medium, and examined with a Zeiss light microscope. Tissue eosinophils were counted at a depth of 0 to 250 µm from the epithelial basement membrane and expressed as total numbers per 1 mm² tissue area. The presence of extracellular EPO activity in eosinophil-rich areas (with no signs of mechanical crush artifacts) was used as a criterion for eosinophil degranulation, as previously described for human tissue samples (6).

Intracellular Distribution of EPO

The intracellular distribution of EPO activity was evaluated by TEM analysis (22). Briefly, in these experiments the tissue tissues were fixed in 3% formaldehyde/PBS, washed three times for 2 h in PBS, and incubated for 12 min in the same DAB-containing PBS buffer described earlier. After rinsing in PBS, the specimens were prepared for TEM analysis (see the following section). The DAB-reaction product appeared as a dark, electron-dense precipitation (22).

Ultrastructural Evaluation of Eosinophil Degranulation Activity

TEM analysis. Airway tissue samples (lungs or nasal septum) were fixed in TEM fixative (1% glutaraldehyde and 3% formaldehyde/PBS) overnight. For each specimen (tissues or agarose-embedded cells), two blocks (each ~ 3 × 3 × 5 mm, representing two separate regions) were rinsed in buffer, postfixed in 1% osmium tetroxide for 1 h, dehydrated in graded acetone solutions, and embedded in Poly/Bed 812 (6). Plastic sections (1 µm thick) were cut on an Ultracut E ultratome (Leica, Solms, Germany), stained with toluidine blue, and examined in an Axiscope light microscope (Zeiss, Göttingen, Germany). Areas with intact surface epithelium were selected for further electron microscopic analysis to avoid studying artifacts induced by the sampling procedure. Ultrathin sections (~ 90 nm) were cut and placed on 200-mesh, thin-bar copper grids and stained with uranyl acetate and lead citrate (6). The specimens were examined using a Philips EM10 transmission electroscope (Philips, Eindhoven, The Netherlands).

Definitions of ultrastructural features. Several investigators have defined the ultrastructural features of tissue eosinophils (14, 15, 17, 23). Accordingly, the eosinophil features in this study were defined as follows: resting eosinophils, intact eosinophils with no ultrastructural signs of degranulation (i.e., all secondary granules had an electron-dense core surrounded by a translucent matrix compartment); exocytosis, intact eosinophils releasing whole membrane-free granules; ECL, eosinophils displaying chromatolysis and loss of plasma membrane integrity (late-stage ECL was characterized by the occurrence of clusters of three or more extracellular membrane-bound secondary granules); PMD, eosinophils with intact cell membranes undergoing partial to complete loss of secondary granule contents. On the basis of previous observations in human eosinophils, the granule changes during PMD include a marked loss of matrix and/or core electron density (5).

Quantification of eosinophil degranulation. For each specimen, individual eosinophils were analyzed regarding the ultrastructural morphology of their secondary granules (the granules in a cross section of an analyzed eosinophil were considered representative of the whole cell). At least 200 secondary granules were examined in each sample and classified into either resting (as defined earlier) or altered (i.e. various morphologic changes in the granule core and/or matrix). The percentage of altered secondary granules in each eosinophil was calculated as follows: Granule Alteration (%) = 100 × (numbers of altered secondary granules/total numbers of secondary granules in a given cell).

Bone-Marrow Sampling and Blood Leukocyte Preparation

For bone-marrow sampling the right femur was removed and freed of soft tissue. Bone marrow (BM) cells were flushed using 3 ml PBS injected through a 25-gauge needle (12). The cells were suspended and centrifuged at 4°C for 10 min at 1,200 rpm (300 × g) and resuspended in cell medium (DMEM/10% FCS; supplemented with 0.1 g/liter benzylpenicillin, 0.1 g/liter streptomycin sulfate), and the viability and total cell number were determined by trypan blue exclusion. A differential count was carried out on cytospun May-Grünwald- and Giemsa-stained cells.

For leukocyte preparation, heparinized whole blood from at least three animals was collected via cardiac puncture, pooled, differentiated, and counted for total cells using Türks staining. The pooled blood was incubated with cold ammonium chloride buffer (0.15 M NH₄Cl, 0.01 M NaHCO₃, and 1 mM EDTA) (2 ml/50 ml) for 10 min at room temperature. The erythrocyte-depleted leukocyte fraction was centrifuged at 4°C for 10 min at 1,100 rpm (250 × g) and washed twice with PBS/2% FCS. The cell pellets were pooled and resuspended in cell medium and the viability and total cell number were measured as described earlier. Human heparin-anticoagulated peripheral blood was obtained from healthy, nonallergic subjects. The leukocytes were prepared and stimulated according to the same protocol as for the mouse blood cells.

For TEM analysis, BM cells or blood leukocytes (3 to 5 × 10⁶ cells) were centrifuged, resuspended in 300 µl TEM fixative, and kept for 1 h at room temperature. The cell suspensions were added to microcentrifuge tubes and centrifuged at 4°C for 10 min at 2,800 rpm (1,630 × g). The obtained pellets were gently embedded in warm (40 to 50°C) 3% agarose/PBS. The cell-rich agarose gels were placed in TEM fixative overnight and processed for TEM analysis.

In Vitro Assay for Eosinophil Degranulation

Purified blood leukocytes, from nonallergic humans or control (nonsensitized, nonchallenged) or sensitized, OVA-challenged mice, were plated in 48-well culture plates at 0.5 × 10⁶ cells/well. The cells were incubated with fMLP (10⁶ M) or PMA (10 ng/ml) at 37°C for 30 min in 5% CO₂. In a separate experiment we demonstrated that preincubation of leukocyte preparations for 10 min with cytochalasin B (5 µg/ml) did not have any additional effect on the granule morphology of the mouse eosinophils (results not shown). Control cells were incubated in cell medium alone. For each stimulus, cells from six wells (comprising ~ 3 × 10⁶ cells) were pooled, centrifuged, and resuspended in 300 µl TEM fixative and examined by TEM analysis as described earlier.

Statistical Analysis

All calculated values are expressed as means with a 95% confidence interval (CI). Throughout, n refers to the number of animals studied. The statistical analyses, Wilcoxon-Mann-Whitney U test or Kruskal-Wallis one-way analysis of variance, were performed using Astute 1.5, a statistics Add-in for Microsoft Excel (DDU Software, Leeds, UK). Probability values (P) < 0.05 were considered statistically significant.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

OVA-Induced Airway Eosinophilia

Eosinophils were scarce in nasal or lung tissues of saline-challenged control mice. In contrast, all sensitization and challenge protocols induced a marked airway tissue eosinophilia (P < 0.01 compared with controls). The number of tissue eosinophils was  115/mm² lung tissue (< 4/mm² for controls) or  90/mm² nasal tissue (< 2/mm² for controls) for all protocols and there was no difference in tissue eosinophilia between the strains. Most of the lung-tissue eosinophils had a peribronchial and perivascular distribution (Figure 1A). In the nose, the tissue eosinophils were mainly restricted to the subepithelial tissue (Figure 1B). Both in the nose and lung, very few eosinophils were present in the airway epithelium.

View larger version (96K): [in this window] [in a new window]   View larger version (117K): [in this window] [in a new window]   Figure 1.   Representative airway sections showing lung (A) and nasal septum (B) from OVA-challenged mouse. Eosinophils are stained for cyanide-resistant EPO (dark stain). The eosinophils are mainly distributed around bronchi, bronchioles (b), and larger vessels (v) in the lung (A), and in the subepithelial (Ep) tissue in the nose (B). Note the absence of extracellular EPO activity. Bars: A, 180 µm; B, 70 µm.

Infiltrated Airway Eosinophils Retain Their Granule Matrix Protein EPO

Regardless of protocol, all EPO reactivity was localized within eosinophils (Figure 1). Further, detailed analysis of the extracellular tissue could not detect any released EPO. Increasing the incubation time and temperature during the DAB-staining procedure (up to 25 min and 37°C) did still not reveal any presence of extracellular EPO. In addition, TEM analysis confirmed the lack of extracellular EPO (Figure 2A) and showed that the EPO activity was restricted to the matrix compartment of the eosinophil secondary granules (Figure 2B).

View larger version (169K): [in this window] [in a new window]   View larger version (154K): [in this window] [in a new window]   View larger version (163K): [in this window] [in a new window]   View larger version (139K): [in this window] [in a new window]   View larger version (151K): [in this window] [in a new window]   View larger version (148K): [in this window] [in a new window]   Figure 2.   Electron micrographs showing eosinophils in OVA-challenged mouse lungs (A-C, E, and F ), and bone marrow (D). EPO-stained eosinophils with a normal morphology, including intact nuclear, organellar, and plasma membranes and condensed peripheral chromatin, are observed in the cell-rich area around the bronchi (A and B). The EPO activity is restricted to the secondary granules, shown as dark reaction product (polymerized DAB) in the granule matrix (arrow in B). Scattered secondary granules of airway mucosal (C) and bone-marrow (D) eosinophils exhibit a coarsening of matrix electron density (see arrows in C and D). High-power analysis of granule cores revealed that the MBP-containing crystal was present in both normal electron-dense cores (E) and in the few light cores ("reversed cores") that were observed (F ). The image of the regularly patterned core crystal is shown in insets (E and F ). Bars: A, 6 µm; B and C, 1.7 µm; D, 1 µm; E and F, 0.25 µm (insets, 20 nm).

Mouse Airway Eosinophils Exhibit Marginal Changes in Their Secondary Granule Matrix, but Lack the Major Features of Human Eosinophil Degranulation

For all sensitization and challenge protocols used, no signs of eosinophil degranulation by exocytosis or ECL could be detected in nasal or lung tissues. To examine the occurrence of PMD we analyzed in detail the morphology of the eosinophil secondary granules. Regardless of protocol, about half of the examined eosinophils in the lung tissue lacked altered secondary granules, whereas the remaining eosinophils contained a few granules displaying a marginal coarsening of matrix electron density (Figures 2C and 3A, and Table 1). In the nasal septum a similar proportion of eosinophils, 55% (95% CI: 44 to 65), displayed a marginal coarsening of granule matrix, represented by an average of 1.3 (95% CI: 1.0 to 1.6) altered secondary granules per affected eosinophil. A ragged loss of granule cores (as frequently observed in degranulating human tissue eosinophils) could not be detected in any of the examined eosinophils; i.e., all core compartments remained intact (Figure 2E). However, translucent granule cores (i.e., "reversed-core" phenomenon) were seen in a few scattered eosinophils (Figure 2F). Despite this loss of electron density, the crystalline core structure was clearly visible and intact (Figure 2F, inset).

Marginal Morphologic Changes of the Secondary Granule Matrix Also Occurs in Bone-Marrow and Blood Eosinophils of Naive Mice

To investigate whether the marginal coarsening of granule matrix was restricted to airway eosinophils of OVA-challenged mice, blood and bone-marrow eosinophils from both control and OVA-treated animals were examined. Regardless of treatment, blood and bone-marrow eosinophils had the same low extent of granule changes as airway tissue eosinophils derived from OVA-challenged mice (Figures 2D and 3B).

fMLP and PMA Induce Degranulation of Human Blood Eosinophils but Not of Mouse Blood Eosinophils

To investigate whether mediators known to stimulate human eosinophils also induce degranulation in mouse eosinophils, we stimulated blood leukocyte preparations from both species with fMLP or PMA. Despite marked effects on neutrophils (increase in cytoplasmic protrusions and vacuoles) fMLP or PMA did not induce any granule changes in naive mouse eosinophils (Figures 3C and 4A). A lack of effect was also seen using leukocyte preparations from OVA-challenged mice. In comparison, upon the same stimulation the human eosinophils exhibited pronounced changes in granule morphology (Figures 3C and 4B). The granule changes in human eosinophils included a marked loss of core, matrix, or both core and matrix content (Figure 5).

View larger version (11K): [in this window] [in a new window]   Figure 3.   The occurrence of morphologic changes of secondary granules of lung tissue eosinophils in OVA-challenged mice (A), bone-marrow and peripheral blood eosinophils from control animals (B), and mouse and human blood eosinophils incubated in vitro with PMA or fMLP (C). The granule changes are presented as the percentage of structurally altered secondary granules of total calculated granules. **P < 0.001 compared with mouse eosinophils. All values are means ± standard error of the mean of n  250 granules for each sample. For details regarding sensitization and challenge Protocols I through IV, see MATERIALS AND METHODS.

View larger version (154K): [in this window] [in a new window]   Figure 5.   High-power electron micrographs exemplifying characteristic structural changes of degranulating human eosinophil secondary granules after in vitro stimulation. Granule core changes include a ragged loss of the central core content (A), completely dissolved secondary granule matrix (B), coarsening of the core and matrix within the same granule (C), and extensive loss of both core and matrix (D). Bar: 0.5 µm.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The present investigation represents the first quantitative ultrastructural study examining the occurrence of degranulating eosinophils in mouse models of asthma and rhinitis. Importantly, we demonstrate that eosinophils recruited to the lungs and the nose of heavily allergen-exposed mouse airways do not exhibit any sign of degranulation. Further, we demonstrate in in vitro studies that human eosinophils degranulate extensively whereas mouse eosinophils remain unaffected. The present study thus provides evidence for significant differences between human and mouse eosinophils, and highlights a limitation of current mouse models of airway disease.

The BALB/c and C57BL/6 strains of mice used in this study are commonly used in development of asthma models (24, 25). Likewise, the present sensitization and challenge procedures are also representative of the widely used protocols of OVA-induced airway inflammation in mice (12, 13, 19). Analyses of eosinophil degranulation in this study were carried out at 24 h after the last allergen challenge when a marked tissue eosinophilia was established (12). This is also the time at which airway hyperresponsiveness has been reported (9, 21). Hence, any eosinophil degranulation in the airway tissues of allergic mice would likely be revealed by the present study design.

Activation of mouse eosinophils in vitro, as reflected by an increased release of O₂ and leukotriene C₄, has been reported (26) but to our knowledge so far degranulation of mouse eosinophils ex vivo has not been shown. The limited number of eosinophils obtained from normal mouse blood has impeded the characterization of mouse eosinophil degranulation responses. Previous in vitro studies have, therefore, used cultured bone-marrow cells (27), eosinophils from tissue granulomas or the peritoneal cavity (accumulated in response to parasite infection [26]), or eosinophil-rich blood from interleukin-5 transgenic mice (28). In this study, however, we chose to examine eosinophils in leukocyte preparations from blood of normal and allergen-challenged mice because this allows a direct comparison with human blood cells. Thus, by combining in vivo and in vitro investigations, the consistency of any observed differences between human and mouse eosinophils was explored.

The present work confirms the general ultrastructural features of mouse eosinophils, including the presence of elongated secondary granules carrying distinct matrix and core compartments (23). We failed, however, to detect any sign of degranulation. This observation is in sharp contrast to studies of human diseased airways, where the eosinophils exhibit clear signs of degranulation through PMD and ECL (1, 4, 15, 16). For example, using the same TEM approach as in the present study, it was recently demonstrated that during active allergic rhinitis 33% of the mucosal eosinophils underwent ECL and that the remaining eosinophils displayed signs of extensive PMD (6). Indeed, the extent of PMD displayed by the airway tissue eosinophils in allergic rhinitis was comparable to that observed in the human eosinophils in vitro in this study (Figure 4B). Our TEM observations on lack of PMD and ECL of mouse airway eosinophils as well as confinement of EPO reactivity to the granule matrix are corroborated by the absence of extracellular EPO, as assessed in this study by airway histochemistry. In accord, previous studies have reported lack of extracellular major basic protein (MBP) (19, 29) and EPO (24) in mouse models of allergic asthma. The differing features of mouse and human eosinophils demonstrated in this study in vitro, in the presence of fMLP or PMA, further suggest that species-dependent differences rather than an insufficient allergic reaction have contributed to the present in vivo findings. Supporting the possibility of fundamental differences in cell regulation, it has already been demonstrated that there are differences between species in gross morphology and content of the eosinophil secondary granules (30) as well as in the expression of cell-surface receptors (27). The present comparison between mouse and human eosinophils, together with the previous demonstrations of extensive eosinophil degranulation in human diseased airway tissues (15), highlight profound differences between mouse and human eosinophils. Hence, eosinophil degranulation-driven pathologic events should not unconditionally be expected to occur in mouse models of allergic airway disease.

View larger version (131K): [in this window] [in a new window]   View larger version (152K): [in this window] [in a new window]   Figure 4.   Electron micrographs showing representative morphology of PMA-treated mouse (A) and human (B) blood eosinophils. The mouse eosinophils exhibit marginal changes in their secondary granules (A) whereas the corresponding human eosinophils display signs of extensive PMD (i.e., abundant loss in secondary granule core, matrix, and both core and matrix contents) (B). Bars: 1 µm.

A few nonquantitative observations of the ultrastructure of eosinophils in the airways of allergic mice have previously been reported. Depending on the nature of the individual observations these have been interpreted as a sign of either poor or significant degranulation. Thus, eosinophils occurring in mouse allergen-exposed airways of "high responder immunoglobulin E" mice (BP2) and C57BL/6 mice have been reported to be nondegranulating (20, 24, 31). On the other hand, several morphologic features of mouse airway eosinophils have been interpreted as reflecting degranulation (10, 13). However, as discussed later, these features have not been quantified nor have they been convincingly associated with actual degranulation of eosinophils.

The reversed-core phenomenon of secondary granules, seen as an electron-light core surrounded by an unchanged matrix, has been suggested to reflect eosinophil degranulation (32). Tissue eosinophils exhibiting reversed cores also occurred occasionally in our present in vivo models (Figure 2F). Although the electron density of the core crystals (of which the major component is MBP [33]) was diminished, the "reversed" cores retained the regular pattern as well as the size and the sharp edges of the MBP-rich crystal structure (Figure 2F). This image differs entirely from the human eosinophils displaying PMD, which characteristically show ragged losses of core contents (Figure 5, and Ref. 5). Accordingly, we interpret this reversed-core phenomenon to be compatible with a fully retained granule core.

Hypodense eosinophils (i.e. eosinophils distributed to lower density fractions than normal in Percoll or Metrizamide gradients) have been demonstrated to contain significantly smaller, but not fewer, secondary granules than do normodense eosinophils (34). Importantly, there is no evidence to indicate that hypodensity, diminished size, or even diminished numbers of secondary granules reflect degranulation (35). Speculatively, diminished numbers of the secondary granules may merely reflect the release of premature eosinophils from the bone marrow that occurs during an allergic inflammation (12). Also, we recently demonstrated a poor correlation between the numbers of secondary granules and ultrastructural degranulation in eosinophil-rich human airways (5). Without proof of concept, the presence of pseudopods in mouse eosinophils has also been forwarded as a sign of degranulation (13, 36). It may be assumed that pseudopods or cell-surface protrusions serve general purposes, such as locomotion, rather than reflecting eosinophil degranulation activities.

Although we failed to demonstrate signs of "human-like" degranulation of mouse eosinophils, the present detailed TEM analysis revealed that occasional secondary granules of mouse eosinophils displayed a marginal coarsening of their matrix. Notably, this particular alteration occurred in this study to the same extent in lung tissues, bone marrow, and blood eosinophils under control as well as under allergen-challenged conditions. Further, it occurred in mouse eosinophils in vitro regardless of stimuli. Thus, it is unlikely that the present coarsening of the granule matrix represents any allergen-induced degranulation in the mouse airways. A possible explanation for this phenomenon could be that some eosinophil granules are incompletely filled during the eosinopoiesis. Alternatively, the coarsening of matrix may in part represent an artifact due to sampling and processing.

Increased levels of EPO and MBP in bronchoalveolar lavage (BAL) samples from OVA-challenged mice have been reported (25, 37, 38). However, BAL samples from OVA-exposed animals may contain high numbers of eosinophils, which makes it difficult to rule out the possibility of granule protein release due to accidental cell rupture during the sample processing. Also, it may be expected that extended dwell times of eosinophils in the airway lumen lead to degenerative processes involving release of granule proteins. Indeed, cytolytic eosinophils have been depicted in the airway lumen of OVA-challenged BP2 mice (24). In general, eosinophil-derived proteins in the lumen may not always be a reliable indication of tissue eosinophil degranulation (39). Further studies are warranted to validate the possibility that mouse eosinophils degranulate in the airway lumen in vivo, and to determine the extent to which the luminal proteins may penetrate the mucous layer and epithelium to influence the underlying tissue.

In conclusion, we demonstrated in this study that airway mucosal eosinophils in allergen-challenged mice lack signs of degranulation such as cytolysis and PMD. Moreover, this mouse-human dichotomy was corroborated by in vitro experiments indicating a fundamental difference in the regulation of mouse and human eosinophils. Because the present findings cannot support the assumption that there is a significant release of cytotoxic granule proteins from mouse eosinophils, caution is advised when generating concepts about eosinophil-driven airway pathology in mouse models of airway inflammation. Further investigations using different combinations of strains and sensitization/challenge protocols are urgently warranted in the search for mouse models that exhibit significant and human-like eosinophil degranulation.