INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Asthma is an inflammatory disease process that affects the whole airway tree from large to small (< 2 mm) airways (1). Up to now, however, the focus has been on the large airways and the small airways have been, not the least because of technical problems, mostly ignored. The current knowledge of the role of small airways in asthma has been reviewed (1, 2). The authors agreed that asthma occurs also in the small airways, but it became clear that the evidence therefore is incomplete and that the contribution of the small airways to asthma remains to be clarified. The present evidence of a role of small airways in asthma is based mainly on three types of observations. First, histological examination of airways from patients with asthma showed that both the inflammation as well as the airway remodeling in these lungs are not confined to airways of any particular size, but occur throughout the entire bronchial tree (3). Second, it is supposedly this inflammation with the structural changes that it produces in the small airways that leads to a 7-fold increase in peripheral lung resistance in subjects with asthma compared with control subjects (4). A comparable increase in peripheral airway resistance in patients with asthma was also confirmed by others (5). Third, there is evidence of a role of increased responsiveness of small airways in asthma. Peripheral airways of patients with asthma but not of control subjects constricted in response to dry air (6) and bradykinin (7). It was also shown that peripheral airways of asthmatics responded more strongly to histamine (8). And finally, in passively sensitized human bronchi of different sizes it was shown that smaller airways responded more strongly to antigen. (9). These authors examined airways down to a diameter of 0.5 mm; their method did not allow them to study smaller airways. Thus, there is as yet only limited direct evidence that small airways can contract in response to allergen and, if so, down to which airway size or airway generation this occurs.

We have developed a method that allows examination of this question, namely the video microscopy of precision-cut lung slices (10). With this method airways are cut into thin (220-µm) viable lung slices that are taken into culture. Constriction of airways is observed under a microscope, which in combination with video microscopy and digital imaging techniques makes it possible to visualize and quantify bronchoconstriction. The smallest airways that we have been able to analyze by this technique have a diameter of only 50 µm, corresponding roughly to the terminal bronchioles (11). In humans, airways with a diameter smaller than 2 mm have been termed small airways (12), that is, airway generation 11 in humans. The corresponding airway generation in rats has a diameter of 780 µm (13), and terminal bronchioles, which start at generation 16, have a corresponding diameter of 200 µm (13). However, such comparisons should be made cautiously, because it is not completely clear how to compare airway size between humans and rats.

To study the response of precision-cut lung slices to allergen, the slices were passively sensitized. Naive cells or tissue can be made sensitive to allergen by incubating them with serum from an allergic subject (animal or human), a process that is referred to as passive sensitization. In the in vitro passive sensitization model, IgE accounts for specific, but not for nonspecific, responses (14).

It was the aim of this study to investigate the immediate allergic response in small airways down to the terminal bronchioles. Therefore, we have examined the effect of passive sensitization and subsequent exposure to the specific antigen ovalbumin on airways < 0.9 mm in diameter. Because our results showed that an immediate allergic response takes place in small airways, we used various inhibitors to investigate the underlying mechanism of antigen-induced bronchoconstriction in small airways.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animals

Lungs were taken from 8-wk-old female Wistar rats (220 ± 20 g) obtained from Harlan Winkelman (Borchen, Germany) and kept under controlled conditions (22° C, 55% humidity, 12-h day/night rhythm) on a standard laboratory chow.

Chemicals

Ovalbumin (Grade V), methacholine, indomethacin, cromolyn, ketanserin, and tripolidine were purchased from Sigma (Deisenhofen, Germany). AA861 was purchased from Biomol (Hamburg, Germany). WEB 2086 was a personal gift from H. O. Heuer (Boehringer Mannheim, Mannheim, Germany). Bosentan was a kind gift from H. Clozel (Actelion, Allschwil, Switzerland).

Serum

Serum was a generous gift from D. Marx and W. Poppe (Arzneimittelwerke Dresden, Dresden, Germany). Serum was obtained from male Brown Norway rats (200 g) sensitized to ovalbumin or from control animals. Rats were actively sensitized by subcutaneous injections of a suspension of 10 µg of ovalbumin Grade VI (OVA) plus 20 mg of Al(OH)₃ gel and intraperitoneal Bordetella pertussis vaccine (about 400 × 10⁶ heat-inactivated bacteria) on Days 1, 14, and 21. On Day 28 the animals were exposed to a nebulized aerosol of ovalbumin solution (10 mg of ovalbumin/ml) in a nose-only inhalation system for 60 min. Vehicle-treated control animals were exposed to a saline aerosol. The aerosol was generated by a nebulizer driven by compressed air at 0.2 MPa. Forty-eight hours later, animals were killed by urethane overdose (1.5 g/kg body weight, intraperitoneal) and blood was sampled from their hearts.

Serum samples were pooled (n = 5) and the IgE concentration was determined by enzyme-linked immunosorbent assay (ELISA). Total IgE concentration in serum from sensitized rats varied between 800 and 1000 IU/ml. Total IgE levels in control serum were below the detection level.

Precision-cut Lung Slices

Precision-cut lung slices (PCLS) were prepared from isolated perfused lungs (15, 16) and treated and analyzed as previously described for rat (10) and mouse (17) tissue. Briefly, rat lungs were dissected from the animal and perfused blood free with Krebs-Henseleit buffer supplemented with 0.1% glucose and 0.3% HEPES. Subsequently, the lungs were filled with agarose solution (0.75%) via the trachea and put on ice to allow the agarose to cool and solidify. The lung lobes were separated and cut into 5- to 10 mm-thick slabs from which cores along the airways were made with a coring tool. These cores were cut into 200 ± 20 µm-thick slices with the help of a Krumdieck tissue slicer (Alabama Research and Development, Munford, AL). Slices were incubated in minimal essential medium (MEM) and medium was changed every half hour for the first 2 h and then every hour for the next 2 h to remove the agarose from the airways. For analysis, lung slices were moved into a prewarmed 24-well plate. The airways were imaged and digitized with a digital video camera (Visitron 1300; Visitron Systems, Munich, Germany) controlled by the image analysis software Metamorph (Universal Imaging, West Chester, PA). One image of 2.28 mm² was represented by 1,280 × 1,024 pixels. The images were analyzed by the image analysis program Optimas 6.2 (Optimas, Bothell, WA). The lumenal area was taken as the area enclosed by the epithelial border and was quantified after setting the appropriate threshold (a setting in digital imaging software that allows definition of the foreground of the image by selecting a range of gray values). Airway area before addition of the antigen was defined as 100%. Bronchoconstriction was expressed as the percentage decrease in airway area in comparison with the control airway area.

Passive Sensitization

Incubation medium containing penicillin and streptomycin for overnight incubation was supplemented with 1% serum from either sensitized or control rats. Alternatively, serum from sensitized animals was incubated at 60° C for 60 min to inactivate IgE. PCLS were then incubated, each with 1 ml of serum-containing MEM. PCLS were incubated at 37° C and 5% CO₂ in an incubator overnight (approximately 15 h). The next day, the PCLS received fresh medium free of antibiotics. They were kept in an incubator until needed.

Inhibition Experiments

To investigate the mechanism that underlies the immediate allergic response we preincubated PCLS with various inhibitors before challenge with the antigen ovalbumin. In all cases the preincubation time was 10 min.

Statistics

Data are expressed as means ± standard deviation (SD). Data were examined by analyzing the maximum values (Figures 1-4 and 8) or the area under the curve (AUC; Figures 6, 7, and 9) by two-sided t tests. Homoscedasticity was confirmed by F test. In case of heteroscedasticity data were log transformed. The  error resulting from multiple comparisons was adjusted by the method of Hommel (18). p < 0.05 was considered significant.

View larger version (51K): [in this window] [in a new window]   Figure 1.   Images of a lung slice passively sensitized for 4 h, showing a single airway before, 2 min, and 10 min after administration of 0.1% ovalbumin.

View larger version (25K): [in this window] [in a new window]   Figure 2.   Kinetics of ovalbumin-induced bronchoconstriction in lung slices. Effect of 0.1% ovalbumin on bronchoconstriction 1 h (circles), 2 h (triangles), 3 h (squares), 4 h (diamonds), and 16 h (inverted triangles) after passive sensitization to the specific antigen. Reaction to ovalbumin was monitored for 10 min after administration of the antigen. Data represent means ± SD from three independent experiments. The mean airway diameter of all airways investigated was 506,127 ± 146,465 µm.

View larger version (20K): [in this window] [in a new window]   Figure 3.   Correlation diagram of the extent of airway contraction induced by ovalbumin in sensitized slices plotted against airway size. Slices with airways of different sizes were passively sensitized overnight and challenged with 0.1% ovalbumin the next day. Each dot represents the maximum contraction of a single airway with the given airway size. The top x axis indicates the corresponding airway generation in rat airways according to Yeh and coworkers (1979). A total of 110 single lung slices was investigated. Data were analyzed by calculating the Spearman correlation coefficient (r = 0.74; p < 0.0001).

View larger version (18K): [in this window] [in a new window]   Figure 4.   (A) Correlation between airway size and the time required for maximum airway closure. All slices were challenged with 0.1% ovalbumin. Each dot represents the time it took to reach maximum contraction of a single airway with a given size. A total of 110 slices was investigated. Data were analyzed by Spearman correlation analysis (r = 0.49; p < 0.0001). (B) Correlation between the time required to reach the maximum contraction and the extent of airway closure. Data were analyzed by calculating the Spearman correlation coefficient (r = 0.62; p < 0.0001).

View larger version (13K): [in this window] [in a new window]   Figure 8.   Correlation diagram of the extent of airway contraction induced by serotonin plotted against airway size. Native slices were treated with 105 M 5-hydroxytryptamine and maximum airway closure was determined. Each dot represents the maximum closure of a single airway with the given airway size. Data were analyzed by calculating the Spearman correlation coefficient (r = 0.91; p < 0.0001).

View larger version (11K): [in this window] [in a new window]   Figure 6.   Reactivity of very small airways (< 10,000 µm2) to antigen after overnight passive sensitization.

View larger version (12K): [in this window] [in a new window]   Figure 7.   Concentration-dependent inhibition of ovalbumin-induced bronchoconstriction by the 5-HT2 receptor antagonist ketanserin. Slices were sensitized overnight, and the next day they were pretreated with ketanserin for 10 min before they were challenged with ovalbumin. The log median inhibitory concentration (IC50) of ketanserin was 8.235 ± 0.266. Data represent means ± SD of five independent experiments at each ketanserin concentration. The mean airway areas were not different from each other, i.e., 1 nM ketanserin, 338,302 ± 129,473 µm2; 10 nM, 434,160 ± 144,681 µm2; 100 nM, 427,794 ± 106,983; 1 µM 418,766 ± 62,382 µm2.

View larger version (18K): [in this window] [in a new window]   Figure 9.   Influence of hydrocortisone on ovalbumin-induced bronchoconstriction in rat lung slices. Slices were passively sensitized overnight in the absence (squares; mean airway area, 218,017 ± 61,005 µm2; maximum contraction, 5.1 ± 2.7) or presence (circles; mean airway area, 226,251 ± 96,913 µm2; maximum contraction, 28.6 ± 18.2) of hydrocortisone (0.1 µg/ml) in the incubation medium. Slices were challenged with 0.1% ovalbumin and airway contraction was monitored for 10 min. Data represent means ± SD of three independent experiments. The steroid-pretreated group was significantly different from the group that was challenged with ovalbumin alone.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Passive Sensitization of Precision-cut Lung Slices

To see whether an immediate allergic response can principally be induced in precision-cut rat lung slices, slices were incubated (passively sensitized) with 1% serum from Brown Norway rats sensitized against ovalbumin for 4 h and subsequently exposed to 0.1% ovalbumin (Figure 1). Figure 1 shows that after adding the antigen, airways showed a rapid contraction within 2 min that was partially reversible within the next 8 min. The response to allergen was not repeatable, that is, no response occurred if passively sensitized and subsequently challenged lung slices, whose airway calibers had returned to baseline, were exposed to ovalbumin a second time (data not shown).

We next investigated how long slices had to be passively sensitized before addition of ovalbumin elicited bronchoconstriction. Cell culture medium containing 1% serum from rats sensitized to ovalbumin was transferred onto lung slices. One, 2, 3, 4, or 16 h after serum transfer the lung slices were challenged with 0.1% ovalbumin and the reaction was monitored for 10 min. Figure 2 demonstrates that slices had to be passively sensitized for at least 2 h, before they responded to the antigen; the maximum response to ovalbumin, however, did not occur before 4 h. These observations led to the decision to incubate the lung slices overnight with serum before carrying out the antigen exposure in the following experiments. We also investigated the influence of various concentrations of the specific antigen ovalbumin on airway contraction in passively sensitized rat lung slices. At concentrations lower than 0.1% ovalbumin airways showed a weaker and more inconsistent contraction (data not shown). We therefore chose an ovalbumin concentration of 0.1% for the following experiments.

In lung slices incubated with either control serum from nonsensitized rats or heat-inactivated serum (60° C for 1 h) from sensitized rats, ovalbumin did not cause bronchoconstriction. Also, overnight sensitization with 1% serum from sensitized rats and subsequent challenge with a nonspecific antigen (Phleum pratense pollen extract) did not cause bronchoconstriction (Table 1).

Airway Size Dependence of the Immediate Allergic Response

Whether small airways contribute to the pathophysiology of asthma is an important unsolved question. Therefore we examined the dependence of antigen-induced bronchoconstriction in passively sensitized lung slices on airway size. We investigated airways ranging from 10,000 to 650,000 µm² (corresponding to a diameter of 110-910 µm). These airways were sensitized overnight with 1% serum and subsequently challenged with 0.1% ovalbumin. Bronchoconstriction was then assessed for 10 min. Figure 3 shows a scatter diagram of airway size plotted against the maximum degree of airway contraction within this time period. We found a highly significant correlation between airway size and contraction (r = 0.74; p < 0.0001). Forty-two out of 52 airways (80.8%) smaller than 100,000 µm² exhibited complete bronchoconstriction, whereas only 11 airways of 58 (19%) with a size above 100,000 µm² closed completely (p < 0.001, Fisher exact test). We found that smaller airways not only reacted stronger, but also faster. Figure 4A shows a correlation diagram of airway size plotted against the time that was required to reach maximum contraction in that slice (not necessarily complete closure of airway). Forty-three out of 52 (82.7%) airways < 100,000 µm² reached maximum contraction 2 min after adding 0.1% ovalbumin compared with only 32 of 58 (55.2%) in airways with an area larger than 100,000 µm² (p = 0.0022, Fisher exact test). Thus, antigen-induced bronchoconstriction depends on airway size such that the smaller the airways the stronger and quicker the response to specific antigen. Finally, we also found a significant correlation between the time required to reach maximum contraction and the degree of contraction of the individual airways. Airways that reacted rapidly after challenge with ovalbumin also reacted stronger than airways with a slower reaction time (Figure 4B).

The fact that smaller airways contracted to a greater extent and also faster than larger airways, might be explained if large airways were relaxing quicker than smaller airways. However, Figure 5 shows that the opposite was the case, that is, smaller airways not only contracted faster, but they also relaxed faster than larger airways.

View larger version (17K): [in this window] [in a new window]   Figure 5.   Different reactivity of central versus peripheral airways in rat lung slices. Slices were passively sensitized overnight and challenged with ovalbumin (0.1%) the next day. Peripheral airways (squares; mean airway area, 79,009 ± 48,263 µm2) reacted in a more pronounced manner than central airways (circles; mean airway area, 337,789 ± 144,667 µm2). In addition, they also relaxed faster after ovalbumin challenge (p = 0.0003, repeated measurement analysis with polynomial contrasts from 3 to 10 min).

We also sensitized slices containing very small, very distal airways with an area less than 10,000 µm². These airways lacked a distinct muscle layer and were identified as airways by looking at the ciliary lining on the epithelium. Even these very small airways elicited a reaction when challenged with specific antigen. However, as Figure 6 illustrates these airways responded heterogeneously. Because airways of this size are difficult to obtain, we did not further investigate these extremely small airways.

Mechanism of the Immediate Allergic Response in Lung Slices

To investigate the mechanism of ovalbumin-induced bronchoconstriction in lung slices, we preincubated rat PCLS with a variety of inhibitors before exposing them to specific antigen. Table 2 (19) summarizes the data obtained from these experiments. Because it is well known that stimulation with antigen leads to mast cell degranulation (which is also supported by the fact that the effect of ovalbumin was not repeatable as noted above), we tested whether prestored mediators such as histamine or serotonin play a role in antigen-induced bronchoconstriction. We used the histamine H₁ receptor antagonist tripolidine and the serotonin (5-hydroxy-tryptamine, 5-HT) receptor antagonist ketanserin. Whereas preincubation with tripolidine did not prevent bronchoconstriction, ketanserin completely inhibited the ovalbumin-induced bronchoconstriction. A dose-response curve was established for ketanserin to determine the median inhibitory concentration (IC₅₀) in our system. The IC₅₀ value was 5.8 nM (Figure 7). To test whether the released serotonin directly activated airway smooth muscle cells or whether it bound to postganglionic synapses leading to an increased release of acetylcholine, we used atropine to block muscarinic receptors. At a concentration of 100 nM, atropine had only a small effect on ovalbumin-induced bronchoconstriction, indicating that serotonin binds to receptors directly on the airway smooth muscle cells, rather than exhibiting a secondary effect by inducing release of bronchoconstrictory neurotransmitters from nerve endings (Table 2).

Inhibition of leukotriene synthesis with the 5-lipoxygenase inhibitor AA861 or of prostaglandin synthesis with the cyclooxygenase inhibitor indomethacin had no effect on ovalbumin-induced bronchoconstriction (Table 2). Also, blockade of platelet-activated factor (PAF) receptors with WEB 2086 or of endothelin receptors with bosentan failed to prevent the immediate allergic response. Finally, preincubation with cromolyn, an agent that is also used in asthma therapy, failed to block the ovalbumin-induced bronchoconstriction.

These findings indicated that the major mediator responsible for the antigen-induced bronchoconstriction in our model was serotonin. We therefore investigated whether serotonin, like allergen, also contracted smaller airways to a greater degree than larger airways. Serotonin contracted control airways in the same manner (i.e., with a similar time course) as ovalbumin contracted sensitized airways (data not shown). As shown in Figure 8, we observed a similar, significant correlation between airway size and maximum contraction in response to serotonin (10⁵ M).

Effect of Hydrocortisone

In additional experiments, the incubation medium was supplemented with hydrocortisone (HC, 0.1 µg/ml) and subsequently several precision-cut lung slices were incubated overnight with 1% serum from sensitized rats along with HC. The next day, slices were challenged with 0.1% ovalbumin and contraction was compared with that of lung slices sensitized and incubated in MEM free of HC. We found that maximum contractility was diminished in precision-cut lung slices that were preincubated in HC-containing medium (Figure 9). Maximum contractility was 94.9% without HC versus 71.4% for slices treated with HC.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The present study provides direct evidence that immediate allergic reactions can occur in small airways down to the terminal bronchioles. There appears to be no principal difference in the allergic responses of differently sized airways. Of note, the sensitivity to allergen increased along the bronchial tree such that terminal bronchioles were more sensitive to allergen than larger airways. Our pharmacological studies further suggest that serotonin is the major mediator responsible for the allergen-induced bronchoconstriction in rats. In line with this, smaller airways were more sensitive to both allergen and serotonin. These findings suggest that the increased sensitivity of smaller airways to allergen is at least partly explained by an increased sensitivity toward serotonin.

Model of Precision-cut Lung Slices

PCLS offer a novel way to study the functional responses of small airways under cell culture conditions. Using various criteria (lactate dehydrogenase [LDH] release, mitochondrial activity, thymidine incorporation, responsiveness to methacholine), we previously have shown that PCLS are viable for at least 1 d under standard cell culture conditions, and for at least 3 d if they are maintained in a dynamic organ culture (10). Investigations of intrinsic tone showed that it is absent in rat PCLS (10).

In a subsequent study using murine tissue, it was demonstrated that the responsiveness of PCLS to a variety of different agonists is qualitatively and quantitatively similar to responses in the whole organ (17). Thus, PCLS have been established as a reliable tool with which to study small airway responses.

It should be noted, however, that it appears mandatory to use thin slices of reproducible thickness. This is illustrated if experiments with PCLS are compared with those with razor-cut lung slices. Whereas PCLS show highly reproducible responses to methacholine, comparable to those seen in other models such as tracheal rings or isolated lungs, the responses observed in razor-cut lung slices vary to a degree that largely precludes their use in functional studies. For example, the EC₅₀ values for methacholine-induced bronchoconstriction vary 5,000-fold in PCLS (only 200-fold if stratified for airway size) (10), but 550,000-fold in razor-cut lung slices (27). The reasons for these differences have been discussed previously (10).

Dandurand and coworkers (28) were the first to study allergic bronchoconstriction in vitro in lung explants prepared from ovalbumin-sensitized Brown Norway rats, although they did not investigate different airway sizes. They used lung slices with a thickness between 0.5 and 1.0 mm, challenged with 10% ovalbumin, and imaged the responses for 10 min. They observed an immediate allergic response in almost all lung explants derived from sensitized animals. The maximal response for sensitized airways was 42 ± 24% versus 4 ± 3% for lung slices from control animals and peaked within 1-8 min. Compared with these findings, passively sensitized precision-cut lung slices reacted more strongly (78.5%), despite a 100-fold reduced ovalbumin concentration. Besides technical issues, the comparatively small reactivity observed by Dandurand and coworkers might also be explained by the presence of hydrocortisone (0.1 µg/ml) in their incubation medium. We found that exposure to hydrocortisone decreased the maximum contractility of airways after challenge with the antigen ovalbumin (Figure 9). Another notable difference between their study and ours is that they used explants from sensitized rats, whereas we used passive sensitization.

It might be assumed that the enhanced responsiveness of smaller airways to allergen is an inherent property of the slice model, because smaller airways are also more sensitive toward both methacholine (10) and thromboxane (29). Therefore it is important to point out the fact that small and large airways react almost equally strong to endothelin 1 (29). That the relative sensitivity of small and large airways in PCLS differs from agonist to agonist indicates that biological mechanisms rather than model properties are responsible for this behavior. Thus the increased sensitivity of smaller airways to allergen is most likely explained by either enhanced production of serotonin and other mediators or by a greater reactivity of the smaller airways to these mediators (see below).

Serotonin Mediates Allergen-induced Bronchoconstriction in PCLS

It is well known that the major mediator in allergic bronchoconstriction in rats is serotonin (30). The present study confirms these findings and extends them to the small airways. In keeping with this, Purcell and coworkers (31) have shown that pulmonary mast cells (which are the most likely source of serotonin in our model) contain biogenic amine at approximately 25 pg/cell, of which more than 80% is serotonin.

Martin and colleagues (32, 33) have shown that in addition to serotonin, leukotriene D₄ may also contribute to the early response in rats. However, in our model lipoxygenase inhibition was not effective and thus in rat PCLS serotonin appears to be the sole mediator of antigen-induced bronchoconstriction. Nagase and coworkers (34) investigated the role of serotonin and leukotriene D₄ (LTD₄) in central airways and in parenchymal tissue in sensitized rats. Whereas in their system the bronchial response was completely abolished by serotonin receptor antagonism (as in our system), the parenchymal strip response was only partially affected and the authors showed that in parenchymal tissue LTD₄ is involved in the anaphylactic response. However, the parenchymal strip is made up of a number of different anatomic structures, including small airways, small vessels, and alveolar walls. Therefore, contractile responses of parenchymal strips are likely to reflect a complex interplay among these various elements. Hence, parenchymal strips have some limitations in studying the role of small airways in anaphylactic bronchoconstriction.

The above-described findings suggest that serotonin is the major mediator of allergen-induced bronchoconstriction in rats. In addition, we observed that serotonin, like allergen, preferentially contracted small airways. Unfortunately, not much is known about the distribution of serotonin receptors along the bronchial tree. Das and Steinberg (35) identified several serotonin-binding receptors in the rat lung. However, their analysis of extracted lung tissue revealed only a general localization of serotonin receptors on microsomal and mitochondrial membrane fractions. Serotonin receptors have been attributed to epithelial cells, neurons, and smooth muscle cells. Although the distribution of serotonin receptors in the airways is unknown, examples of such heterogeneity are given by vasoactive intestinal peptide (VIP) receptors that prevail in proximal airways, in contrast to tachykinin receptors, which are more prominent in distal airways (36). Bradley and Russell (40) investigated the distribution of histamine receptors in isolated canine airways of different size. They observed that responses to histamine were more pronounced in intrapulmonary airways than in tracheal strips and concluded that histamine sensitivity in canine airways varies within the bronchial tree. In line with this Wagner and colleagues observed that even in healthy humans small airways were "surprisingly responsive" to histamine (8). All these examples lend support to the hypothesis that serotonin receptors might also be distributed heterogeneously along the airways.

Airway Size and Immediate Allergic Responses

Previously the small airways have been dubbed the silent zone. Only a few previous studies have examined early allergic airway responses as a function of airway size (9, 34, 41). Among these studies, this is the first in which airways smaller than 500 µm in diameter including terminal bronchioles have been investigated. The relationship between airway size and airway generation in rat lungs is indicated in Figure 3, where it can be seen that airway responsiveness to allergen increased continuously from central to distal airways. It should be noted, that the diameter of these small airways was in the range of 100 µm, which corresponds to airway generations 16 and 17, that is, the terminal bronchioles. Thus, the terminal bronchioles are those lung structures that showed the strongest response to allergen. On occasion, we also detected airways that were smaller than 100 µm, and although most of them were principally responsive to allergen these extremely small airways did not react in a predictable fashion (Figure 6). Interestingly, the terminal bronchioles have also been identified in other models and by other methods as the site in the bronchial tree where the bronchoconstriction takes place, that is, in lungs treated with endotoxin (42) or thromboxane (43). Thus, this region of the lung appears to be particularly reactive.

Because the responses to both serotonin and allergen in smaller airways were quicker and stronger than those in larger, more central airways, it could be hypothesized that either the distribution of serotonin receptors varies within the bronchial tree or that mast cell numbers are greater in small airways. It seems also possible that mast cells alongside small airways contain more mediators and thus elicit a faster and stronger response after degranulation. As discussed above, there are at present, however, no data on the serotonin receptor distribution in the airways. With respect to mast cells, Bachelet and coworkers (44) as well as Du and colleagues (41) found fewer mast cells in peripheral than in large membranous airways. Thus the greater sensitivity to allergen does not appear to be related to the numbers of mast cells present.

Alternatively, the greater responsiveness of smaller airways may be explained by structural differences between large bronchi and small bronchioles such as the relatively greater muscular thickness in peripheral airways; airway muscles are most strongly developed in the terminal airways (45). Hence, smooth muscle shortening may result in greater narrowing in small airways than in larger airways (46, 47). Another structural difference between large and small airways may be the number of mucosal folds in the airways. Wiggs and coworkers (48) reasoned that airways with fewer mucosal folds narrow to a greater extent than airways with more folds. Because larger airways may be quite folded, this might account for the decreased contraction after stimulation with the antigen. However, in the only experimental study in this area that is known to us, Mitchell and coworkers (who also observed a greater airway narrowing of small versus large airways) counted mucosal folds in large and small porcine bronchi, but found no correlation with muscle shortening in either group (46). Finally, the increased reactivity of small airways could also be due to other structural features such as a different arrangement and orientation of smooth muscle fibers in large and small airways (49).

The observation that smaller airways responded quicker to allergen is in line with a study of mouse lung explants, where it was not only shown that smaller airways responded quicker to methacholine, but more importantly that the velocity of airway contraction in lung explants from different mouse strains correlated with the degree of bronchial responsiveness in vivo (50). Therefore the smooth muscle shortening velocity may be a major determinant of airway responsiveness. It has even been suggested that the problem in asthma may be that the smooth muscle operates too fast (see [1]).

Potential Relevance to Human Asthma

Thus, compared with larger airways, several different factors may make small airways a preferred target in the early allergic reaction: differences in neural controls, endogenous regulation, structure, and smooth muscle shortening velocity. Regardless of the precise mechanism, if this phenomenon occurs in humans, it has both therapeutic and mechanistic implications. Therapeutically, it would make small airways an important target in the therapy of asthma. Mechanistically, it could explain the increased muscle thickness in small airways from patients with asthma (49, 51), because it is thought that such hypertrophy may be a consequence of overworked airways. Interestingly, we have preliminary evidence obtained in sensitized human lung slices that after antigen challenge greater airway narrowing occurs in smaller airways.

Data based on human tissue that are in line with the present findings were reported by Ellis and coworkers who investigated antigen-induced bronchoconstriction in human airways of different sizes by using isolated central (5- to 12-mm outer diameter) or peripheral airways (0.5 to 2 mm) (9). In that study, tissue was passively sensitized with serum from ragweed-allergic individuals and stimulated after a 16-h incubation period. The authors found that ragweed antigen was 14-fold more potent in peripheral compared with central airways. They also observed a more rapid response to ragweed antigen in peripheral airways. To explain these findings they provided evidence that there is a larger percentage of unoccupied IgE receptors on mast cell in peripheral airways than central airways. However, because it is also known that mast cell numbers decrease along the bronchial tree to the peripheral airways (44, 52), an additional factor for the increased sensitivity of smaller airways probably is the fact that peripheral airways exhibit a greater sensitivity to the released mediators as demonstrated for serotonin in the present study. Further human data that support the notion that smaller airways are more sensitive toward various stimuli are reports showing that in patients with asthma smaller airways are more sensitive to methacholine (53), histamine (8), dry air (6), and bradykinin (7).

A further piece of evidence suggesting that contraction of small airways may be important under physiologic conditions is studies comparing the airway-relaxing effects of bronchodilatory aerosols of different particle sizes. The diameter of aerosol particles determines where in the lung particles are deposited; the smaller the particles the smaller the airways they can reach (54). In such studies it was noted that aerosols made of small particles are more efficient in relaxing airways than those containing larger particles, suggesting that the smaller airways were the major determinant of airway resistance (55).

In summary, we have shown that rat small airways respond in a quicker and stronger fashion to allergen and to serotonin, which appears to be the principal mediator of allergen-induced bronchoconstriction in small and large airways in rats. These findings show that small airways are anything but a silent zone.