Introduction

Top Abstract Introduction Materials and Methods Results Discussion Conclusions References

We have recently observed an increase in bone-marrow (BM) interleukin (IL)-5-responsive eosinophil progenitor numbers in association with allergen-induced asthmatic responses, airway inflammation, and airway hyperresponsiveness (1). These results, as well as similar findings in a dog model of allergen-induced airway hyperresponsiveness (2), suggest that communication between the lung and BM may play an important role in the pathogenesis of allergen-induced asthmatic responses. In those studies, because marrow samples were obtained at a single time point after allergen administration, we were not able to determine the time course of the BM response, nor whether it was related to the time course of airway inflammation, airway hyperresponsiveness, or both. Our interpretation of these studies has been that mechanisms responsible for allergen-induced airway inflammation involve increased production of eosinophils by the BM, and that this response is mediated partly through an expansion of the relevant progenitor population.

Several recent publications have shown that allergen-sensitized and -challenged mice develop a transient eosinophilic airway inflammation associated with airway hyperresponsiveness and increased BM eosinophilia (5). Although these murine models lack many of the features of asthma, including chronic inflammation and sustained airway hyperresponsiveness, they are helpful in specifically studying the mechanisms of the acute inflammatory events that take place in allergen-induced asthmatic exacerbations, as well as how these events may contribute to transient physiologic changes in the airway. We have recently observed that sensitized IL-5-deficient mice do not develop airway eosinophilia after allergen exposure to the airway, even when local airway IL-5 production was restored using an IL-5-producing adenoviral construct (9). However, airway eosinophilia was restored in mice exposed to an intramuscular IL-5-producing adenoviral construct, which also elevated circulating IL-5 levels and peripheral blood eosinophils. These observations support our hypothesis that the airway eosinophilic response is dependent on availability of eosinophils in circulation and therefore indirectly dependent on the increased production of eosinophils by the BM.

The purpose of the current study was to determine whether exposure to allergen of sensitized mice results in an expansion of eosinophil progenitor cells, as has been observed in human asthma. A secondary objective was to compare the time course of these changes with measurements of airway eosinophilia and dysfunction.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion Conclusions References

Challenge Protocol

BALB/c mice were purchased (Harlan Sprague Dawley Inc., Indianapolis, IN) at age 10 to 12 wk and housed in specific pathogen-free conditions for 1 wk. All procedures were reviewed and approved by the Animal Research Ethics Board at McMaster University.

The ovalbumin sensitization and challenge protocol (OVA/OVA) is similar to that described by Zhang and colleagues (10) and is illustrated in Figure 1. Days 1 and 11 were considered to be sensitization and Days 19 and 20 to be challenge, although the distinction between the two is likely not dichotomous. Intraperitoneal OVA injections involved precipitating 10% aluminum potassium sulfate with 0.05% OVA, adjusting to pH 6.5, and centrifuging and then resuspending the pellet, followed by a 200-µl intraperitoneal injection. Intranasal OVA involved dissolving 4 mg OVA in 1 ml sterile normal saline, followed by a 25-µl (100 µg OVA) intranasal injection into anesthetized mice. Control mice were sensitized and challenged with diluent (Sham/Sham). End-point measurements, including bronchoalveolar lavage (BAL), airway responsiveness, and BM eosinophil progenitor numbers, were made 2, 12, 24, 48, and 72 h after the final intraperitoneal challenge.

View larger version (6K): [in this window] [in a new window]   Figure 1.   OVA sensitization and challenge protocol.

BAL

Under anesthesia (240 mg/kg intraperitoneally; Avertin, Aldrich Chemical Co., Milwaukee, WI), the chest cavity was opened and the trachea was exposed and cannulated using a blunted 18-gauge needle (n = 5 per group). Two injections of 250 µl phosphate-buffered saline (PBS) were injected and withdrawn through the needle. Mice were then killed via cervical dislocation. The BAL fluid (BALF) was centrifuged for 10 min at 150 × g and 21°C. The cell pellet was resuspended in PBS and a total cell count performed using a hemocytometer. The cells were then diluted to an approximate concentration of 5 × 10⁵/ ml with PBS, cytocentrifuge slides were prepared (Cytospin 3; Shandon Scientific, Sewickley, PA) and stained with Diff-Quik, and cell differential counts were performed based on morphologic and histologic criteria (400 cells counted). All cell counts were performed by one investigator, blind to the experimental condition. Cells were classified as macrophages, neutrophils, lymphocytes, or eosinophils on the basis of morphologic criteria.

Lung tissue was prepared from separate mice (n = 5 per group) by inflating the lungs with periodiate-lysin- paraformaldehyde to a pressure of 30 cmH₂O and then embedded in paraffin. Sections 3 µm thick were cut and stained with Congo red for identification of eosinophils upon viewing with light microscopy.

Airway Responsiveness

Airway responsiveness was measured on the basis of the response of total respiratory system resistance (R_RS) to increasing intravenous doses of methacholine (MCh) (n = 8 per group). R_RS was measured using the flow-interrupter technique, as modified for use with mice (11).

Mice were anesthetized (Avertin, 240 mg/kg intraperitoneally). When anesthesia was established, the trachea was exposed and cannulated using a blunted 18-gauge needle. The needle was then attached to a ventilator (RV5; Voltek Enterprises, Inc., Toronto, ON, Canada) designed to deliver constant inspiratory flow despite the disturbances in the respiratory impedance that occur during the MCh challenge. The initial pattern of ventilation was a tidal volume of 0.1 ml/kg delivered over 45 ms, with a 530-ms end-inspiratory pause and a 95-ms period of passive expiration (breathing frequency of 90 breaths/min). Heart rate and oxygen saturation were monitored via infrared pulse oxymetry (Biox 3700; Ohmeda, Boulder, CO) using a standard ear probe placed over the proximal portion of the mouse's hind limb. After the mouse was stabilized on the ventilator, the internal jugular was cannulated using a 25-gauge needle. Paralysis was achieved using pancuronium (0.03 mg/kg intravenously) to prevent respiratory effort during measurement.

The response of R_RS was measured after intravenous injections of saline, then 10, 33, 100, and 330 µg/kg of MCh (ACIC [Can], Brantford, ON, Canada), each delivered as a 0.2-ml bolus. To establish a constant volume history, mice were subjected to three inspirations to total lung capacity (TLC) (end-inspiratory pressure of 30 cm H₂O) followed by 60 s of 90 breath/min ventilation before each dose. Upon injection, the ventilatory pattern was changed so that the time allowed for passive expiration was extended to 1,425 ms, thus reducing the breathing frequency to 30 breaths/min as suggested by Volgyesi and associates (12). This change was to prevent the dynamic hyperinflation (also termed breath-stacking) that was observed during MCh challenge when mice were ventilated with shorter expiratory times. After the peak in R_RS (20 to 30 s) the breathing pattern was returned to 90 breaths/min. When R_RS returned to baseline, the mouse was again inflated three times to TLC and ventilated for 90 s before beginning the next dose.

During each MCh dosing the mouth-pressure signal from the ventilator was converted to a digital signal (Dash 16; Metrabyte, Staughton, MA) and recorded at 400 Hz on a PC computer. R_RS and respiratory system elastance were calculated as described previously (11). Evaluation of airway responsiveness was based on the peak R_RS measured in the 30 s after the saline and MCh challenges. An index of airway reactivity was calculated as the slope of the straight-line regression between peak R_RS and the log₁₀ of the MCh dose, using the data from only the 10, 33, and 100 µg/kg doses. The data at the 330-µg/kg dose was not included in this regression because peak R_RS had frequently reached a plateau at this dose. An index of airway sensitivity was calculated as the MCh dose at which the above regression intersected with the baseline R_RS (peak R_RS after the saline challenge).

BM Eosinophil Progenitor Responses

BM colony assays were performed as previously described for dog (2) and human studies (1), with modifications for measurements in mice (n = 10 per group). After mice were killed by terminal exsanguination via cardiac puncture, one femur was removed from each mouse and freed of soft tissue. BM cells were flushed using 3 ml McCoys 3+ buffer injected through a 25-gauge needle. To break up cell clumps, the suspension was passed through needles of 18, 20, and 22 gauge. Suspended cells were separated by density gradient centrifugation over 65% Percoll for 30 min at 1,500 rpm, and the cells at the interface were removed. The mononuclear cells were washed once with McCoys 3+, then incubated overnight in plastic flasks at 37°C and 5% CO₂ to remove adherent cells. The nonadherent cells were cultured in microassay 24-well plates (Becton Dickinson, Lincoln Park, NJ), 7.5 × 10⁴ cells per well. The culture medium was made up of 0.9 % methylcellulose (Caledon Lab, Georgetown, ON, Canada), 30% fetal calf serum (GIBCO BRL, Grand Island, NY), and recombinant mouse IL-5 (R&D Systems Inc., Minneapolis, MN). A range of concentrations of IL-5 was used to establish a dose-response curve, after which all cultures were performed using an optimal IL-5 concentration (5 ng/ml). After 7 d, colonies greater than 40 cells were counted using inverse microscopy and identified using morphologic and histologic criteria. To confirm the identification of colonies as eosinophil colony-forming units (Eo-CFU), selected samples were removed from the culture wells for viewing under light microscopy after staining with Diff-Quik, or for identification using electron microscopy (EM). For EM, selected colonies were pooled in 2% glutaraldehyde containing 0.1 M sodium cacodylate (pH 7.4). After fixation for 2 h at 4°C, the sample was washed in 0.2 M sodium cacodylate (4°C) for 1 h, dehydrated in graded ethanol followed by propylene oxide, and embedded in Spurr's resin. The block was sectioned using a Reichert Ultracut E ultramicrotome and sections were stained for 5 min with uranyl acetate and then for 2 min with lead citrate. Stained sections were examined using a JEOL 1200EX Biosystem electron microscope.

Analysis

Comparisons between Sham/Sham and OVA/OVA mice with respect to airway reactivity (slope of the R_RS  MCh [logged] dose-response curve), airway sensitivity (initial MCh dose [logged] to induce bronchoconstriction), BALF cell numbers, and BM Eo-CFU numbers were made using analysis of variance. All post hoc comparisons were carried out using the Neuman-Keuls test for significant effects (13). All comparisons were two-tailed, with critical  set at 0.05.

	Results

Top Abstract Introduction Materials and Methods Results Discussion Conclusions References

BM Eo-CFU

At IL-5 concentrations of 0.1 ng/ml or less, no colony growth was observed from any mice. Rather than observing a dose-dependent increase in colony numbers at higher concentrations, we observed optimal growth at a concentration of 1 ng/ml, which was maintained until a concentration of 80 ng/ml (Figure 2). The cells in all colonies appeared morphologically homogenous, and all selected colonies stained positive with Congo red. Further, the cultured cells showed the characteristic morphology of eosinophils when examined under electron microscopy (Figure 3). On the basis of these observations, we concluded that colonies were uniformly eosinophilic and thus we identified each counted colony as an Eo-CFU.

View larger version (10K): [in this window] [in a new window]   Figure 2.   Eo-CFU growth from four allergen-naive BALB/c BM aspirates, incubated for 7 d in a range of IL-5 concentrations (M ± SEM). For reference, all other results were obtained using an IL-5 concentration of 5 ng/ml.

View larger version (123K): [in this window] [in a new window]   Figure 3.   Electron micrograph of a cell plucked from an Eo-CFU colony (original magnification ×5,500). The ringed nucleus and multiple cytoplasmic granules containing an electron-dense core and a translucent matrix that identify this cell as an eosinophil are characteristic of murine eosinophils.

Eo-CFU numbers from Sham/Sham mice ranged between 15.3 and 23.6 colonies (per 250,000 nonadherent mononuclear cells) at all time points after challenge (Figure 4). No increase in Eo-CFU numbers was observed in OVA/OVA mice, either 2 or 12 h after the second OVA challenge, where there were 16.2 (standard error of the mean [SEM] 5.13) and 21.7 (6.87) colonies, respectively (Figure 4). However, at 24 and 48 h after this challenge there were 35.1 (11.11) and 28.4 (18.99) colonies, respectively, significantly more than the 23.6 (7.47) and 18.2 (5.76) colonies grown from Sham/Sham mice at the same time points (P < 0.05) (Figure 4). The number of Eo-CFU 72 h after the second allergen challenge was 18.1 (5.73), which was not significantly different than the 17.3 (5.60) colonies grown from Sham/Sham mice.

View larger version (23K): [in this window] [in a new window]   Figure 4.   The time course of airway responsiveness, BAL eosinophils, and BM Eo-CFU after airway OVA challenge (M ± SEM). Mouse numbers per group were 8 (airway responsiveness), 5 (BAL), and 10 (BM). *P < 0.05 compared with Sham/Sham at the same time point.

Airway Responsiveness

There was a significant leftward shift in the MCh dose- response curve 24 h after OVA challenge in sensitized mice (Figure 5). Airway reactivity, expressed as the slope of the dose-response curve, was significantly altered at 24 and 48 h after the allergen challenge, but not at earlier or later time points (Table 1 and Figure 4). Airway sensitivity, as measured by the intercept of the dose-response curve with baseline resistance, was not affected at any time points after allergen administration.

View larger version (15K): [in this window] [in a new window]   Figure 5.   The dose-response curves of respiratory system resistance (RRS) against the intravenous MCh dose in Sham/Sham and OVA/OVA mice, 24 h after the final OVA challenge (M ± SEM; 8 mice per group).

Airway Inflammation

There was a significant increase in total cell count from OVA/OVA mice at all time points except 7 d, compared with Sham/Sham (Table 2). BALF eosinophils were significantly raised at 2 h, peaked at 24 h, remained markedly increased until at least 72 h, and were still slightly elevated at 7 d (Table 2 and Figure 4). BALF neutrophils were significantly increased at 2 h and appeared to peak at 12 h, after which they returned toward baseline levels. Lymphocytes were also elevated at 2 h and remained increased at all subsequent time points up to 72 h, but had returned to normal at 7 d.

The pattern of eosinophilic inflammation in bronchial sections from Sham/Sham and OVA/OVA mice at all time points is illustrated in Figure 6. Eosinophils were not evident in Sham/Sham mice. At 2 h, eosinophils were evident in the perivascular region but not peribronchially. At 12 h, eosinophils were evident in parenchymal tissue adjacent to the airway but only occasionally seen within the epithelial or subepithelial regions. At subsequent time points, eosinophils were evident in the perivascular and peribronchial regions, as well as, to some extent, in the parenchyma.

View larger version (184K): [in this window] [in a new window]   Figure 6.   Mouse airway sections showing bronchi (AW) and in some cases blood vessels (BV) from (A) Sham/Sham, (B) OVA/OVA 2 h, (C ) OVA/OVA 12 h, (D) OVA/OVA 24 h, (E ) OVA/OVA 48 h, and (F ) OVA/OVA 72 h. Staining is with Congo red, and eosinophils can be identified by the red-staining cytoplasm.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion Conclusions References

In this study we have observed, for the first time, that OVA sensitization and challenge of BALB/c mice are associated with increased numbers of IL-5-responsive Eo-CFU in BM aspirate. Further, the time courses of the increased BM progenitors appear to be similar to that of allergen-induced airway hyperresponsiveness, both being present in the 24-to-48-h after-challenge period. Increases in airway eosinophils, however, are evident earlier and persist longer than the increase in BM progenitors.

Our findings demonstrate that allergen-induced eosinopoiesis is mediated in part by an expansion of the BM Eo-CFU progenitor population, rather than simply an increased rate or frequency of division/maturation of an already-established number of Eo-CFU progenitors. These findings are in agreement with observations we have made in humans with asthma, where allergen-induced hyperresponsiveness was associated with increased BM Eo/basophil-CFU numbers (1). These findings support the hypothesis that mechanisms of eosinopoiesis in mouse models of allergen challenge are similar to those operating in the BM of asthmatics. Previous authors have demonstrated that allergen challenge of sensitized mice results in increased BM mature eosinophil numbers (14), and increases in eosinophil peroxidase-positive cells (i.e., mature eosinophils) when BM was placed in liquid cultures with IL-5 (15). Further, Gaspar Elsas and colleagues have reported increased numbers of IL-3-responsive mixed myeloid colonies, containing eosinophils, grown from mouse BM aspirates after allergen challenge (15). Increased numbers of IL-5-responsive Eo-CFU have also been cultured from mouse BM in response to parasitic infection (16).

The mechanisms responsible for the increase in BM Eo-CFU are not yet known. Clearly IL-5 is a likely candidate because it is detected in circulation after allergen challenge in mice (14), it is known to support eosinopoiesis (Figure 2), and we have recently shown that systemic IL-5 is sufficient for restoration of airway eosinophilia after allergen challenge in IL-5-deficient mice (9). An interesting observation reported by Minshall and coworkers (17) is an increase in IL-5 immunostaining and expression of IL-5 messenger RNA within CD3⁺ cells resident in mouse BM after allergen challenge, suggesting that signaling between lung and BM may result in an increased local production of eosinopoietic cytokines within the BM compartment. Although part of this response may be due to cell trafficking, Gaspar-Elsas and colleagues (15) have demonstrated that plasma taken from mice after allergen challenge was capable of increasing BM mature eosinophils when injected into naive recipient mice. Thus it is clear that serum factors, present after allergen challenge, are capable of stimulating eosinopoiesis in the BM, possibly through the increased local production of IL-5.

Although it is likely that IL-5 production, either at the lung or locally, at the BM, is required for allergen-induced increased eosinopoiesis, it is not clear at which stage of eosinophil differentiation IL-5 exerts its effect, or whether other mediators are involved. It is well known that pre- exposure of blood- or BM-derived progenitors to IL-3 or granulocyte macrophage colony-stimulating factor is often required before IL-5-responsive Eo-CFU can be detected (18). It is therefore possible that mediators other than IL-5 are required for the expansion of the IL-5-responsive Eo-CFU progenitor population that we observed after allergen challenge. This possibility is supported by our failure to observe a dose-response effect of IL-5 on Eo-CFU numbers (Figure 2), suggesting that IL-5 alone is not sufficient for expansion of the Eo-CFU pool, although it can support eosinopoiesis from an already existing pool.

Measuring the complete time course of airway hyperresponsiveness, airway inflammation, and BM eosinophil progenitors has allowed us to make comparisons that have not been practical in studies with human subjects, and that have not been made in previous studies using mouse models. These observations suggest that airway hyperresponsiveness and increases in marrow eosinophil progenitors follow a similar time course that begins approximately 24 h after the second of two allergen challenges and persists until a time 24 to 48 h later. Airway eosinophilia, as detected using both BAL and lung histology, was present earlier and persisted longer than airway dysfunction or BM responses, suggesting a dissociation between allergen-induced eosinophilia and airway hyperresponsioveness in BALB/c mice. This dissociation has been reported by others, who have observed that allergen challenge of IL-5-deficient or anti-IL-5-treated BALB/c mice does not prevent airway hyperresponsiveness, despite blocking eosinophilic inflammation (22). Whether our observation of similar time courses for airway hyperresponsiveness and BM Eo-CFU expansion is based on a common underlying cause remains to be seen.

The protocol we adopted for OVA sensitization involves two intraperitoneal injections with adjuvant and one intranasal instillation. Sensitization alone with two intraperitoneal injections of OVA/Alum (Days 1 and 11) and a single intranasal OVA challenge (Day 11) did not result in airway hyperresponsiveness on Day 20 (data not shown). This is evident from Figure 4, where airway responsiveness in the OVA/OVA mice at the 2-h time point was not different from that of the Sham/Sham mice.

	Conclusions

Top Abstract Introduction Materials and Methods Results Discussion Conclusions References

We observed an increase in BM Eo-CFU in a model of airway exposure to OVA in sensitized mice. These findings indicate that, as in asthma, allergen-induced eosinopoiesis by the BM involves expansion of the relevant eosinophil progenitor population. These findings suggest that the hemopoietic mechanisms in place in asthma are reproduced by exposure of sensitized mice to allergen. Uncovering the specific mechanisms of increase in Eo-CFU and of eosinophil proliferation in these mice should increase our understanding of the mechanisms of asthma and potentially uncover new therapeutic interventions.