Research Paper Mediators of Inflammation, 10, 143–154 (2001)

Although airway inflammation and airway hyperreactivity are observed after allergen inhalation both in allergic humans and animals, little is known about the mechanisms by which inflammatory cells can contribute to allergen-induced airway hyperreactivity. To understand how inflammatory cell infiltration can contribute to airway hyperreactivity, the location of these cells within the airways may be crucial Using a guinea pig model of acute allergic asthma, we investigated the inflammatory cell infiltration in different airway compartments at 6 and 24 h (i.e. after the early and the late asthmatic reaction, respectively) after allergen or saline challenge in relation to changes in airway reactivity (AR) to histamine. At 6 h after allergen challenge, a threefold (p < 0.01) increase in the AR to histamine was observed. At 24 h after challenge, the AR to histamine was lower, but still significantly enhanced (1.6-fold, p < 0.05). Adventitial eosinophil and neutrophil numbers in both bronchi and bronchioli were significantly increased at 6 h post-allergen provocation as compared with saline (p < 0.01 for all), while there was a strong tendency to enhanced eosinophils in the bronchial submucosa at this time point (p = 0.08). At 24h after allergen challenge, the eosinophilic and neutrophilic cell infiltration was reduced. CD3+ T lymphocytes were increased in the adventitial compartment of the large airways (p < 0.05) and in the parenchyma (p < 0.05) at 24h post-allergen, while numbers of CD8+ cells did not differ from saline treatment at any time point post-provocation. The results indicate that, after allergen provocation, inflammatory cell numbers in the airways are mainly elevated in the adventitial compartment. The adventitial inflammation could be important for the development of allergen-induced airway hyperreactivity.


Introduction
Allergic asthma in man is clinically characterized by bronchoconstriction in response to inhaled aeroallergens, resulting in an early asthmatic reaction (EAR) that is often followed by a late asthmatic reaction (LAR). The EAR is thought to be primarily caused by an immunoglobulin (Ig)E-mediated release of mast cell mediators, while the LAR is associated with the recruitment and activation of inflammatory cells in the lung, including eosinophils, neutrophils and lymphocytes. [1][2][3][4] Another important feature of allergic asthma is the development of increased reactivity of the airways to non-specific stimuli following allergen exposure, referred to as airway hyperreactivity (AHR). Allergen-induced AHR has been observed in patients with a dual asthmatic reaction, and may already be present before the onset of the LAR. 5 -8 It is generally thought that inflammatory cell influx in the airways is involved in the development of AHR. However, a close relationship between the number of inflammatory cells in the airway lumen or mucosa and chronic AHR has not invariably been found, which may possibly be due to airway remodeling after prolonged inflammation. 9 -11 Although information from biopsy studies is available about inflammation in the airway wall of asthmatics, little is known about the distribution of inflammatory cells throughout the entire airway circumference and lung tissue of these patients. Detailed knowledge about this distribution may be pivotal to understanding the role of inflammatory cells in the development of AHR. Techniques to assess airway inflammatory cell distribution in relation to AHR in asthmatic patients are limited: biopsy samples can only be collected from superficial layers of larger proximal airways, which provide no information about deeper peribronchial areas nor about more distal smaller airways. In addition, cells obtained from bronchoalveolar lavage (BAL) and sputum induction represent primarily luminal inflammation. Detailed histopathologic studies of the airways from patients with fatal and non-fatal asthma with respect to inflammatory cell distribution within and along the airways have been performed, 12 -17 but cannot be quantitatively related with AHR.
The aim of the present study was to investigate the distribution of the inflammatory infiltrate both within the airway wall (i.e. the adventitia and submucosa) and between the central and peripheral airways and the parenchyma in relation to AHR after the allergeninduced EAR and LAR. To this purpose, we used a guinea pig model of allergic asthma with great qualitative and quantitative similarities in allergeninduced EAR, LAR and AHR after both reactions, and airway inflammation as assessed by BAL compared with asthmatic patients. 18

Animals
Specified pathogen-free tri-colored bastard male guinea pigs (Charles River; SAVO, Kiszlegg, Germany) weighing 400-500 g were actively sensitized to ovalbumin as described previously. 19 To obtain a shift to IgE class antibodies, an allergen solution containing 100 m g of ovalbumin and 100 mg of Al(OH) 3 per ml saline was used. The allergen solution was gently rotated for 60 min to obtain an alu-gel, and 0.5 ml was injected intraperitoneally, while another 0.5 ml was divided over seven intracutaneous injection sites in the proximity of lymph nodes in the paws, lumbar regions and neck. Animals were operated on 3 weeks after sensitization and used experimentally 4-8 weeks after sensitization. The animals were housed in individual cages in climate-controlled animal quarters, and were given water and food ad libitum. All protocols described were approved by the University of Groningen Animal Ethical Committee.

Measurement of airway function
Airway function was assessed by measuring pleural pressure (P pl ) as described previously. 20 In short, a small saline-filled latex balloon connected to a salinefilled canula was surgically implanted in the thoracic cavity. The free end of the canula was subcutaneously driven to the neck of the guinea pigs, where it was exposed and attached permanently. The canula was connected to a pressure transducer (Ohmeda DTX™; SpectraMed, Bilthoven, The Netherlands) using an external saline-filled canula. P pl (cmH 2 O) was measured continuously using an on-line computer system. We have previously shown that changes in P pl are linearly correlated to changes in airway resistance and hence can be used as a sensitive index for allergic and non-allergic bronchoconstriction. 20 In this way, airway function can be monitored repeatedly and continuously for prolonged periods of time, while the animals are unaware of the measurements being taken.

Provocation procedures
Ovalbumin and histamine provocations were performed by inhalation of aerosolized solutions, produced by a DeVilbiss nebulizer (type 646; DeVilbiss, Somerset, PA, USA), with an airflow of 8 l/min resulting in an output of 0.33 ml/min. Provocations were carried out in a specially designed perspex cage (internal volume, 9 l) in which the guinea pigs could move freely.
Before the start of the experiment, the animals were habituated to the experimental conditions and the provocation procedures on two sequential days at least 1 week after surgery, when preoperative weight was restored. On the first day, the animals were placed in the provocation cage unconnected to the pressure transducer. After an adaptation period of at least 30 min, three consecutive provocations with saline were performed, each provocation lasting 3 min and separated by a 7 min interval. The next day, this procedure was repeated with the animals connected to the measurement system During the experimental protocol following the habituation procedure, all provocations were preceded by a 30 min adaptation period and followed by two consecutive control provocations with saline (3 min) as already described. A baseline P pl value was calculated by averaging the P pl values from the last 20 min of the adaptation period.
To assess the airway reactivity for histamine, histamine provocations were started with a solution containing 25 m g/ml of histamine in saline, followed by increasing dosage steps of 25 m g/ml. Histamine provocations lasted 3 min, separated by 7 min intervals. Animals were challenged until P pl was increased by more than 100% above baseline, for at least three consecutive minutes. P pl returned to baseline value within 15 min after the last histamine provocation. The airway reactivity (AR) to histamine was expressed as the concentration causing a 100% increase in P pl (PC 100 ), which was calculated by linear intrapolation of the concentration-P pl response curve. Changes in AR towards histamine were expressed as the ratio PC 100 pre/post allergen challenge.
Ovalbumin provocations were performed by inhalation of increasing concentrations of 1.0, 3.0 and 5.0 mg/ml of ovalbumin in saline for 3 min each, separated by 7 min intervals. Allergen inhalations were discontinued when an increase in P pl of 100% or more was observed. Using these conditions, none of the animals developed anaphylactic shock after allergen provocation. For the quantitative assessment of the EAR and LAR, airway function was continuously measured up to 24 h after allergen challenge. The magnitudes of the allergen-induced EAR and LAR were expressed as the area under the P pl timeresponse curve (AUC) between 0 and 6 h after allergen provocation for the EAR, and between 8 and 24 h after provocation for the LAR. P pl was expressed as the percentage change from baseline, and AUC was calculated by trapezoid integration over discrete (5 min) time periods. Based on saline control provocations, threshold values of AUC (mean + 2 SD; 99% confidence interval) were defined as 1185% ´5 min for a positive early response and 2790% ´5 min for a positive late response, respectively. 18 Histamine PC 100 values were assessed 24 h before, and 6 h (after the EAR) and 24 h (after the LAR) after ovalbumin provocation. Inhalations with saline instead of ovalbumin served as controls. For tissue sampling, animals were terminated at 6.5 h (group 1) or at 24.5 h (group 2) after ovalbumin or saline provocation.

Tissue processing
Guinea pigs were killed by cervical dislocation, after which the lungs were immediately resected and kept on ice in a petri dish during further processing. Additionally, the spleen was removed for qualitative assessment of antibody specificity. The superior and median lobes of the right lung and the spleen were snap frozen at -80°C in isopentane, and tissue was stored at -80°C. Frozen sections were cut at a thickness of 4 m m and stored, wrapped in aluminium foil, at -20°C until use (within 2 weeks).

Inflammatory cell characterization
Inflammatory cells were identified using histochemical and immunohistological techniques. Eosinophils were identified using hematoxylin-eosin staining. Neutrophils were identified by a combination of morphologic criteria and tissue non-specific alkaline phosphatase enzyme histochemistry, as described previously. 21 Infiltration of T cells in the lung was analyzed by immunohistological techniques using a rabbit antihuman CD3 polyclonal antibody (Dakopatts Ltd, High Wycombe, UK) as a pan T-cell antibody, and the murine monoclonal CT6 as a CD8 + T-cell antibody 22 (generously provided by Prof. R.J. Scheper, Free University of Amsterdam). In brief, frozen sections were fixed at room temperature (RT) for 10 min in 100% acetone and washed for 5 min in phosphatebuffered saline (PBS) (pH 7.4). To prevent nonspecific staining, the slides were incubated for 30 min with 10% normal goat or normal rabbit serum, respectively, diluted in PBS. Then, the sections were incubated for 1 h at RT with the primary antibody diluted (1:100 and 1:5000, respectively) in 1% bovine serum albumin in PBS. After rinsing three times with PBS for 5 min, endogenous peroxidase activity was blocked by incubating the slides in 0.075% hydrogen peroxide in PBS for 30 min. Following three washes with PBS, the slides were incubated for 30 min with peroxidase-conjugated secondary antibodies (Dakopatts, Glosstrup, Denmark), diluted in 2% normal guinea pig serum in PBS. A peroxidase-conjugated goat-anti-rabbit IgG antibody (dilution 1:100) was used for the detection of the CD3 polyclonal antibodies, and a peroxidase-conjugated rabbit-antimouse IgG (dilution 1:50) was used for the detection of the CT6 monoclonal antibodies. Next, the sections were washed in PBS and the peroxidase activity was developed in a freshly prepared solution of 3-amino-9-ethylcarbazol, containing 0.03% hydrogen peroxide for 10 min. Counterstaining was performed using Mayer's hematoxylin. Antibody specificity was checked by incubating adjacent lung sections with non-relevant antibodies. Guinea pig spleen tissue served as a positive control for specificity.

Cell quantification analysis
Using a light microscope (Olympus BH-2, Paes, Zoeterwoude, The Netherlands), positively stained cells were enumerated in three anatomical areas as indicated in Fig. 1. The numbers of eosinophils, neutrophils, total T cells and CD8 + T cells were counted employing an eyepiece graticule comprising 100 squares of known area. The number of positive cells present in the submucosa was expressed as cells per millimeter of basement membrane. The length of the basement membrane was estimated using the formula derived from Buffons needle problem: B a = [(p /2)/I L ] * d, 23 where B a is the length of the boundary (mm), I L is the number of intersections formed by the boundary with the parallel lines of the grid, and d is the distance between two parallel grid lines (mm).
Adventitial cell counts were carried out in the same airways. Positive cells were expressed as cells per millimeter of external smooth muscle perimeter. This perimeter was also estimated using Buffons formula. In the large airways, the adventitia was defined as the area between the smooth muscle and the cartilage; in the small airways, the adventitia was defined as a 50 m m border outside and directly adjacent to the smooth muscle. In the parenchyma, positively stained cells were expressed as percent of total cell counts. Per specimen, at least three large and three small airways were enumerated and, in the parenchyma, four random fields of 6.25 ´10 4 m m 2 (the total area of the graticule) were counted, and average values were calculated per animal.

Statistical analysis
Statistical analysis on the magnitudes of the early and the late asthmatic responses, and on the cell counts between treatment groups, was performed using the unpaired Mann-Whitney rank sum test. PC 100 values before and after allergen challenge were compared using a one-way analysis of variance. When significance was observed (p < 0.05), a complementary Student-Newman-Keuls test was performed. For comparison of AHR at different time points within treatment groups, a Wilcoxon signed-rank test was used.
For the correlation between AHR and inflammatory cells, correlation coefficients and p values were calculated using Spearman rank order correlation analysis.

Results
Ovalbumin-induced early and late asthmatic reactions Figure 2 shows a typical example of a dual asthmatic reaction, induced by ovalbumin provocation. The EAR starts with a sharp increase in P pl , which is maximal within 20 min and lasts 2-4 h. The LAR starts at approximately 8 h after allergen provocation, being reversible within 24 h after the challenge. The airway responses to allergen and saline control provocations in animals terminated after the EAR and the LAR are presented in Table 1. Table 2 shows that an approximately threefold (p < 0.01) increase in AR to histamine was observed at 6 h after allergen provocation in both groups of animals, while a smaller but still significant AHR (1.57-fold increase, p < 0.05) was observed at 24 h after allergen challenge in the animals terminated after the LAR. No change in airway reactivity was observed in salinechallenged control animals.

Airway reactivity
Distribution of inflammatory cells in ovalbumin and saline challenged animals d Eosinophils. At 6 h after allergen provocation, a significant fourfold increase in eosinophils was observed in the adventitial layer of the bronchi (p < 0.01; Fig. 3A), whereas in the adventitial layer of the bronchioli an eightfold increase in eosinophil numbers was observed at the same time point (p < 0.01; Fig. 3C). The adventitial eosinophil numbers were not significantly different between large and small airways. Twenty-four hours after allergen challenge, adventitial eosinophil numbers still tended to be elevated but no longer reached statistical significance. In the submucosal layer of the bronchi, eosinophils tended to be enhanced at 6 h after allergen exposure compared with saline (p = 0.08; Fig. 3B), while at 24 h post-allergen submucosal eosinophil numbers did not differ from controls. In addition, no significant change was observed in the eosinophil number of the bronchiolar submucosa (Fig. 3D). Furthermore, in the parenchyma, no changes in eosinophil numbers were observed either at 6 or 24 h after allergen (Fig. 3E). Cell numbers in saline treated groups did not differ from one another at any time point. well as in the bronchioli (p < 0.01; Fig. 4C). Twentyfour hours after allergen exposure, some neutrophils were still present in the adventitia of the large airways, but these numbers were not significantly different from the saline-treated group (Fig. 4A). Few neutrophils were observed in the submucosal compartment of both large and small airways, after allergen provocation (Fig. 4B,D, respectively); however, no significant differences were observed compared with saline. No neutrophils were present in the submucosal compartment at any time point post-saline. Figure 4E shows that neutrophil numbers were not significantly elevated in the parenchyma at any time point post- Data are presented as area under the percent change in P p l time-response curve (AUC) between 0 and 6 h after the allergen provocation for the early asthmatic reaction (EAR) and between 8 h and 24 h after allergen provocation for the late asthmatic reaction (LAR). Results represent mean values ± SEM of (n) animals. Group 1, guinea pigs terminated after the EAR; group 2, guinea pigs terminated after the LAR. * p < 0.05, † p < 0.01 compared with saline inhalation, Mann-Whitney rank sum test.
allergen provocation as compared with salinetreated animals.
d T-lymphocytes. CD3 + T-cell numbers showed no quantitative changes at 6 h after allergen provocation in any of the compartments investigated in the allergen-treated group as compared with controls ( Fig. 5A-E). At 24 h, however, CD3 + cells were increased in both the adventitial layer of the large airways (p < 0.05) and in the parenchyma (p < 0.05), as is shown in Fig. 5A,E, respectively. CD8 + T cells showed no significant change at any time point, after either ovalbumin or saline treatment (not shown).
For all cell numbers, no significant differences between large and small airways were observed. No significant correlations between allergeninduced cell infiltration in the different compartments and AHR were observed, except for CD3 + cells in the parenchyma at 24 h after ovalbumin (r = 0.89, p < 0.05).

Discussion
In a guinea pig model of allergic asthma, characterized by allergen-induced EAR, LAR and AHR after both reactions, we have explored the infiltration of eosinophils, neutrophils, and T lymphocytes in the submucosal and adventitial compartments of large (cartilaginous) and small (non-cartilaginous) airways at 6 and 24 h after allergen challenge. At 6 h after challenge (i.e. after the EAR, but before the initiation of the LAR; Fig. 2), a significant increase in eosinophil and neutrophil numbers was observed in the adventitia of both large and small airways, while there was a tendency to an increased eosinophil number in the bronchial submucosa at this time point. The infiltration of eosinophils and neutrophils was reduced at 24 h after challenge, whereas a significant T-cell infiltration was observed at this time point both in the bronchial adventitia and in the parenchyma. The allergen-induced increase in inflammatory cells was associated with the development of AHR to histamine, which, as shown previously in this model, 18 was maximal at 6 h after challenge and decreased, but still significantly present, at 24 h after the provocation. The temporal association between the degree of, mainly adventitial, infiltration of eosinophils and neutrophils and the degree of AHR to histamine at 6 and at 24 h after allergen challenge may indicate a causal relationship between these events. However, no significant correlations between infiltration of these cells in any of the airway compartments and AHR were observed. In addition, the enhanced T-cell infiltration at 24 h after allergen challenge was not significantly correlated with AHR. Apart from the relatively low number of experiments, this may indicate that other factors may be involved in the development of allergen-induced AHR.
Likewise, in human asthmatics, no or only weak correlations were found between AHR and inflammatory cell numbers in BAL, 9,24 -27 induced sputum, 9,28 -31 and bronchial biopsies. 9,32 -37 However, inflammatory cells found in BAL or sputum represent luminal instead of tissue inflammation, while endobronchial biopsies only provide information about the inflammatory response close to the surface of the proximal airways (mucosa and submucosa), so a possible relationship between AHR and inflammatory cells in the airways could be easily overlooked using these techniques.
Indeed, in previous studies using our guinea pig model of allergic asthma, no significant correlation was found between eosinophil and neutrophil numbers in BAL and AHR at 6 and 24 h after allergen provocation. 18 In this model of acute allergic asthma, it was found that BAL eosinophils and neutrophils were increased at 6 h after allergen provocation, and even further increased at 24 h after allergen. In the current study, it was observed that tissue eosinophils and neutrophils peak at 6 h post-allergen and are reduced at 24 h, supporting the idea that these cells migrate from the tissue to the lumen. A similar relationship between the time course of tissue and Mediators of Inflammation · Vol 10 · 2001 Airway reactivity is expressed as the ratio of histamine provocation concentration causing 100% increase in pleural pressure (PC 100 ) values obtained at 24 h before, and 6 and 24 h after inhalation of saline or ovalbumin. Data represent mean values ± SEM of (n) animals. Group 1, guinea pigs terminated after the early asthmatic reaction; group 2, guinea pigs terminated after the late asthmatic reaction. * p < 0.05, † p < 0.01 compared with before treatment, one-way analysis of variance followed by the Student-Newman-Keuls test; ‡ p < 0.05 compared with 6 h post-ovalbumin, Wilcoxon signed rank test.
luminal infiltration of eosinophils was also found in another guinea pig study 38 and in asthmatic patients. 39 In addition to cellular infiltration in the different airway compartments, the activation state of the infiltrated cells may be an important determinant of allergen-induced AHR. Thus, in biopsy studies in patients, a positive correlation was found between (activated) EG2 + eosinophils and AHR. 34 In addition, in mild asthmatics, a significant correlation was found between the AHR to histamine and the concentration of major basic protein (MBP) in the BAL. 25 Accordingly, in our guinea pig model, a possible relationship between eosinophil activation and AHR was indicated by the observation that levels of eosinophil peroxidase (EPO) in BAL fluid were already enhanced at 6 h after allergen challenge. 40 Moreover, in the same model, we have recently observed, in contrast to the number of BAL eosinophils, that the activation state of the cells (measured as EPO release) is higher at 6 h after provocation than at 24 h after the challenge, while a significant correlation was found between the increased EPO release from the BAL cells and histamine PC 100 values at 6 h, but not at 24 h, after allergen provocation (De Boer et al., unpublished results).
A third cause of possible dissociation between inflammatory cell numbers and AHR in asthmatics could be progressive airway remodeling, charac-terized by thickening of the basement membrane, subepithelial fibrosis, and hypertrophy and hyperplasia of airway smooth muscle, as occurs in chronic asthma with persistent airway inflammation. 41 In the presence of these morphological changes, the presence of particular inflammatory cells in the airways is not a prerequisite for (baseline) AHR.
Similar to the findings in the present study, the occurrence of preferential adventitial rather than submucosal eosinophilia after allergen challenge has been reported previously in guinea pigs. 42,43 However, in the study of Frew et al., 43 tissue eosinophilia was observed between 17 and 48 h after the challenge instead of after 6 h as observed in the present study. Lapa e Silva et al. 42 found no increase in submucosal eosinophil numbers at 2 and 24 h after challenge, while adventitial eosinophils were elevated at 24 h but not at 2 h after allergen challenge. Different sensitization and provocation procedures may account for the different kinetics of infiltration of inflammatory cells in these models compared with our model. Lapa e Silva et al. 42 used an additional booster for sensitization and challenged by the intraperitoneal route instead of the inhalational route. Frew et al. 43 sensitized guinea pigs by inhalation, challenged under protection of the H 1 receptor antagonist mepyramine and measured airway resistance in a whole-body plethysmograph, while in our model measurements were performed under freely moving conditions, without antihistamine protection. Very importantly, we have previously demonstrated that mepyramine treatment may affect allergeninduced asthmatic reactions, AHR and infiltration of inflammatory cells. 44 In lung tissue from patients with fatal asthma, as well as lung specimens from asthmatics obtained at surgery for lung cancer, adventitial eosinophilia has also been described. In lung specimens obtained from patients with asthma, Hamid et al., 45 found greater numbers of eosinophils in the adventitia as compared with the submucosa of airways with an internal perimeter greater than 2 mm; however, activated eosinophils seemed more numerous in the submucosa. In a recent study, Haley et al. 46 reported a difference in eosinophil distribution between large and small airways in lung tissue from patients who died from status asthmaticus. As in our study, the eosinophilic infiltrate in the small airways was predominantly found in the adventitia. On the contrary, large airways predominantly showed submucosal eosinophilia, while both adventitial and submucosal accumulation was found in our study. However, both for the adventitial and the submucosal compartments, the eosinophil infiltration was not significantly different between large and small airways in our study. In line with these results, Carroll et al. 15 also found no difference in eosinophil densities between large and small airways in the inner wall area between the basement membrane and the outer border of the smooth muscle in patients dying from asthma. Since studies performed on autopsy material of patients represent cellular infiltration in a state of chronic asthma, the relevance of the actual inflammatory cell counts in this tissue in the contribution to AHR is unknown.
Both human and guinea pig studies have indicated that various mechanisms may be involved in eosinophil-induced AHR. Although eosinophilic infiltration was mainly adventitial and epithelial damage was not investigated in this study, it is well known that eosinophilic granular cationic proteins may cause epithelial damage, thereby causing enhanced airway permeability and exposing the underlying sensory nerves, which may lead to enhanced cholinergic and excitatory non-adrenergic non-cholinergic reflex activities. [47][48][49] Enhanced cholinergic reflex activity after allergen challenge may also develop from a dysfunction of autoinhibitory prejunctional muscarinic M 2 receptors induced by eosinophil-derived MBP. 50 -52 In addition, it has recently been established that eosinophilic polycations may be involved in a deficiency of bronchodilating NO in the airways, presumably by inhibition of cellular uptake of L-arginine, the substrate for NO synthase. 53 -57 Finally, it is known that eosinophil-derived mediators like LTC 4 and reactive oxygen species may cause swelling of the airway tissue by enhanced vascular permeability. 58 Since adventitial eosinophilia was most pronounced, it is tempting to speculate that uncoupling of the airway smooth muscle from the parenchyma caused by eosinophil-induced outer wall swelling, as proposed by Lambert et al., 59 is an important determinant of the observed AHR.
In line with our results, several studies have demonstrated increased neutrophil numbers in BAL fluid from asthmatics and from guinea pigs after allergen challenge. 2,18,38,60 Moreover, there are increasing indications that neutrophils may play a role in acute severe asthma. Thus, enhanced infiltration and/or activation of neutrophils has been demonstrated in fatal asthma of sudden onset, 61 status asthmaticus, 62 and during exacerbations of asthma. 63 In addition, recent evidence from bronchial biopsies 64 and induced sputum 65 indicates that neutrophils may also be involved in severe persistent asthma.
The role of neutrophils in the development of allergen-induced AHR is presently unclear. Neutrophils have the capacity to generate enhanced levels of superoxide anion (O 2 -) in patients with asthma, which correlates with the airway reactivity to methacholine in these patients. 66 In the presence of high concentrations of inducible nitric oxide synthase (iNOS)-derived nitric oxide in the inflamed airways, O 2 may cause formation of the highly cytotoxic oxidant peroxynitrite (ONOO -), which has been found to induce AHR. 67 Moreover, ONOOwas demonstrated to increase the release of MBP from eosinophils in tracheal preparations, 67 while neutrophils may also enhance eosinophilic LTC 4 production. 68 In addition, neutrophils may enhance eosinophil recruitment and degranulation through secretion of IL-8. 65,69,70 Elevated CD3 + cells were found only at 24 h after allergen provocation, both in the parenchyma and in the bronchial adventitia. In addition, the number of CD3 + cells in the parenchyma was correlated with AHR. The observed rise in CD3 + cells was not associated with a change in CD8 + cells. In the absence of a specific antibody directed against guinea pig CD4 + cells, the likely CD4 + character of these CD3 + CD8cells, as in humans, remains to be established. Elevated CD3 + CD8lymphocytes in guinea pig bronchus, but not parenchyma, were also observed by Frew et al. 43 at 17 and 48 h after allergen. CD4 + Th2-lymphocytes could contribute to AHR by activation of eosinophils by the production of IL-5. 71 In conclusion, our findings indicate that, in addition to (sub-)mucosal inflammation, inflammation in the adventitia of both large and small airways could be importantly involved in the development of AHR in a guinea pig model of acute allergic asthma. Although investigations of the entire human airway circumference in relation to allergen-induced AHR are hampered by technical limitations, this finding may be of importance for asthmatic patients, and merits further investigation.