Main

Infectious laryngotracheitis, the so-called croup, is a common childhood disease that may lead to fulminant respiratory distress in children (1). The consequence of several attacks of the disease is huge psychological stress for the patient as well as for his or her parents. Furthermore, airway hyperresponsiveness as a result of croup during childhood is described (2). The most common cause for the disease is a viral infection of the upper airways, but bacterial infections are described also (3). In addition, a psychosomatic contribution is discussed (4). The morphologic basis of the respiratory distress is an edema of the subglottic space. It is still unclear why the subglottic region is affected selectively and other parts of the organ, e.g. the glottic area, are unaffected. One reason might be that the immunologic defense after antigen inhalation differs between the different areas of the larynx, and immunocompetent cells that regulate the local immune response might be composed differently within the local mucosal immune system of these areas. Indeed, for the larynx of healthy rats, it was recently shown that the number of DCs, natural killer cells, and B and T lymphocytes is significantly higher in the subglottic than in the glottic mucosa (5, 6).

To study whether the different microenvironment leads to a differential entry of leukocytes into the subglottic and glottic mucosa during infection, in this study the phenotype of the cells accumulating in the mucosa of different regions of the larynx and the adjacent trachea was studied. Experimental laryngitis was induced using well-established models of experimental upper airway infections in the rat by two different pathogens (7). First, heat-killed Moraxella catarrhalis was used. As shown earlier this represents a strong antigenic stimulus inducing a rapid influx of immunocompetent cells into inflamed mucosa (7). Furthermore, it may cause infectious laryngotracheitis in children (3). To support the biologic relevance of our approach, a second animal model was used in which viable Bordetella pertussis, which clinically causes an infection characterized by a typical whooping cough (6), were applied intralaryngeally.

At various times after either inhalation of heat-killed M. catarrhalis or intralaryngeal application of viable B. pertussis, the larynx was removed. Using quantitative immunohistology (5, 6), the number of immunocompetent cells was determined in the supraglottic, glottic, and subglottic areas of the larynx. Both in the epithelium and subepithelially the appearance of leukocytes, i.e. neutrophils, macrophages, DCs, and T and B lymphocytes, were studied.

Our results demonstrate that after application of either pathogen, the mucosa of the subglottic area of the rat larynx shows a significantly stronger inflammation than that of the glottic area, as demonstrated by the three to five times higher leukocyte accumulation in the former. This strongly resembles the preferential swelling of this region that is seen clinically. Therefore, we provide a useful animal model to further study the pathophysiology of acute laryngeal infections.

METHODS

Animals.

Different rat strains were previously tested as to whether they would react sufficiently after the inhalation of heat-killed M. catarrhalis or inoculation of viable B. pertussis (8). Thus, PVG rats (5 wk old) were chosen to be aerosolized with heat-killed M. catarrhalis, and BN rats (5 wk old) were inoculated with viable B. pertussis. All animals were kept under special pathogen-free conditions until the experiments were started, and they had free access to food and water. The experimental protocols were approved by the Animal Ethics and Experimentation Committee of the Institute for Child Health Research, which complies with the conditions of the Australian National Health and Medical Research Council.

Preparation of M.

catarrhalis and B.pertussis.

M. catarrhalis was obtained from clinical isolates of Princess Margaret Hospital, Perth, Western Australia, and was prepared as described before (7). In brief, the bacteria were cultured on agar plates and in Mueller-Hinton broth (37°C, 24 h). The bacteria were centrifuged and washed three times with sterile water for irrigation (4°C). Then the bacterial mass was suspended in 50 mL of water, and stored at −70°C in 10-mL aliquots (109 CFU/mL) until use. Immediately before aerosolization the bacterial suspension was heated at 60°C for 1 h and passed vigorously through a 26-gauge needle five times. Afterwards, 10 mL of the bacterial suspension (109 CFU/mL) was aerosolized.

Viable B. pertussis (strain W28 provided by Dr. P. Nowotny, Kent, England) were stored at −70°C until use. They were incubated for 72 h on Charcol agar (Oxoid Australia, Melbourne, Victoria, Australia) containing cephalexin (40 μL/mL) and 10% defibrinated horse blood (State Health Lab, Perth, Western Australia, Australia). Then they were harvested, washed in PBS, and centrifuged (Beckman high-speed GPR centrifuge, Beckman, Munich, Germany) at 3500 ×g for 7 min. The last step was repeated twice before the bacteria were resuspended in 3 mL PBS to the final concentration of 6 × 109 bacteria/mL.

Application of heat-killed M.

catarrhalis and viable B. pertussis.

Because M. catarrhalis was heat-killed, PVG rats could be exposed for 60 min to an aerosol of the organisms using an inhalation exposure apparatus (Tri-R Instruments, New York, NY). Control animals were aerosolized with 0.9% NaCl. This method could not be used when viable B. pertussis were applied because the viable bacteria would contaminate the whole apparatus. Therefore, the larynx of the BN rats was adjusted with a conventional otoscope under ether anesthesia, and B. pertussis (60 μL; 4 × 108 bacteria) was injected directly into the lumen of the larynx of the BN rats using a 26-gauge cannula. Control animals were injected with 60 μL of PBS.

Removal of the larynx and detection of immunocompetent cells.

At various times after application of either M. catarrhalis or B. pertussis (1 to 96 h), the animals were exsanguinated under ether anesthesia, and the larynx including the first four tracheal rings was removed. The organs were filled with OCT compound (Miles, Naperville, IN), snap frozen in liquid nitrogen, and stored at −70°C. Cryosections, 6 μm thick, were made in a frontal direction, air dried, wrapped in aluminum foil, and stored at −20°C until further processing (2, 3). Staining of surface antigens was performed as described elsewhere (10, 11). Briefly, the slides were fixed in 100% ethanol at 4°C for 10 min, washed in Tris-buffered saline (TBS) containing 0.05% Tween 20 (Serva, Heidelberg, Germany), and incubated for 60 min at room temperature in a moist chamber with the primary antibodies [DCs, Ox62 (12); neutrophilic granulocytes, RP3 (13); macrophages, ED2 (14); α/β T lymphocytes, R73 (15); and B lymphocytes, HIS14 (16)]. Because the antibody Ox62 also identifies γ/δ T lymphocytes, these T cells were separately stained with V65 (17). Negative controls were performed by omitting the first antibody. After washing in TBS, the slides were incubated for 60 min with the biotinylated second antibody (sheep anti-mouse IgG; Amersham, UK). Then the slides were washed in TBS and incubated for 60 min with a streptavidin–horseradish peroxidase conjugate (Amersham) to detect secondary biotinylated antibody. After washing, enzyme-linked antibody was revealed by reacting with 3′3-diaminobenzidine (Sigma Chemical Co., Munich, Germany) and 0.015% hydrogen peroxide for 10 min. Finally, slides were counterstained with hematoxylin (Sigma) and embedded using glycergel (DAKO).

Tissue evaluation and statistical analysis.

The number of neutrophils, macrophages, DCs, and T and B lymphocytes were determined in different regions of the larynx by conventional light microscopy at 500× magnification as described before (5, 6). In the supraglottic, the glottic (i.e. the vocal folds), the subglottic area, and the trachea both the epithelium and the subepithelial space were analyzed. Using a graticule in the eyepiece, cells were counted within the epithelium and within the subepithelial area that was located under the basal lamina. The subepithelial space analyzed was approximately two times that of the covering epithelium. Cells were determined per area (0.1 mm2;Fig. 1). At least 100 cells of each subpopulation were counted in each region of the larynx per animal. To obtain this, one or more stained sections of the respective larynx was evaluated. Means and standard deviations were determined (SPSSPC+, Vers.4.0.1; SPSS Inc., Chicago, IL). Differences between the time points were taken as significant at p≤ 0.05 using the Man-Whitney U test. If the number of the subpopulations was too low (less than or equal to 5 cells/mm2) the data of these animals were pooled, which means the total number of cells within the four larynxes was determined and calculated per area counted.

Figure 1
figure 1

Schematic presentation of the rat larynx. Areas that were investigated in the current study are indicated on the left side. TC, thyroid cartilage;CC, cricoid cartilage;VF, vocal fold.

Full size image

RESULTS

With the techniques used, all investigated cell types could be clearly identified in all regions of the larynx (Fig. 2). In addition, it was possible to differentiate between cells within the epithelium and subepithelially located cells (Fig. 2). In the laryngeal mucosa of the control animals, all investigated cell types were present with the exception of neutrophils.

Figure 2
figure 2

Neutrophilic granulocytes, DCs, T lymphocytes, and macrophages in the subglottic mucosa of control animals, and 2 h after inhalation of heat-killed M. catarrhalis. Cryosections were incubated with the respective antibodies RP3 (neutrophilic granulocytes), Ox62 (DCs), R73 (α/β T lymphocytes), and ED2 (macrophages). EP, epithelium;SE, subepithelium. Bar = 50 μm.

Full size image

Neutrophils, DCs, and T and B lymphocytes were found inconsiderable numbers in the laryngeal mucosa 1 h after applicationof M.

catarrhalis.

Already 1 h after application of M. catarrhalis many neutrophils were found in the mucosa, reaching their maximum number 2 h after application (Figs. 2 and 3). Even at this early time, DCs and T and B lymphocytes were also present in large numbers (Figs. 2 and 3). The increase of the DCs was not because of immigration of γ/δ T lymphocytes (also detected by the applied antibody Ox62) because staining with the antibody V65 (γ/δ T lymphocyte-specific, not shown) showed no increase of this cell type. The pattern of cellular infiltration observed in the mucosa of the subglottic area (Fig. 3) was comparable to that found in the mucosa of the supraglottic area, glottic area, and trachea (Table 1). However, the number of neutrophils, DCs, and T and B lymphocytes detected in the subglottic area after instillation was always higher than that in the glottic area (Figs. 3 and 4), with the exception of the number of macrophages (see below).

Figure 3
figure 3

Number of immunocompetent cells within the subglottic and the glottic mucosa of control animals (time = 0) and at various times after inhalation of M. catarrhalis (n= 4 at each time), mean ± SD. When the number of cells was less than 5 cells/0.1 mm2, data of the animals were pooled; *indicates values that are significantly different from control values of the same region (p≤ 0.05).

Full size image
Table 1 Distribution of immunocompetent cells in the laryngeal and tracheal mucosa (epithelium and subepithelial tissue) of control animals and 2 h after the inhalation of M. catarrhalis (cells/0.1 mm2) * Mean ± SD;n= 4. † Because the number of cells was too low (≤5 cells/0.1 mm2) the data were pooled from four animals (the total number of all counted cells in four animals/area counted).
Full size table
Figure 4
figure 4

Two hours after the inhalation of heat-killed M. catarrhalis, the number of DCs (arrows) detectable subglottically (B) exceeds the number of cells within the glottic mucosa (A). EP, epithelium;SE, subepithelium. Bar = 100 μm.

Full size image

At the peak of inflammation, about three to five times as many neutrophils, DCs, and T lymphocytes were present in the subglottic area compared with the glottic area (Fig. 3). A comparable relationship was found for B lymphocytes, although on a lower level (Table 1). The number of neutrophils, DCs, and T lymphocytes first increased subepithelially and then in the epithelium (Fig. 5). In addition, in all areas of the larynx the number of cells in the epithelium was lower than that subepithelially.

Figure 5
figure 5

Number of immunocompetent cells within the epithelium and the subepithelial layer of the subglottic mucosa of control animals (time = 0) and at various time after inhalation of M. catarrhalis (n= 4 at each time), mean ± SD. When the number of cells was less than 5 cells/0.1 mm2, data of the animals were pooled; *indicates values that differ significantly from control values of the same region (p≤ 0.05).

Full size image

Local macrophages differ from all other cell types investigated intheir distribution and their kinetics during M.

catarrhalis-mediated inflammation.

In contrast to the other subpopulations, which were found in higher numbers in the subglottic than in the glottic area, local macrophages showed the reverse pattern, their number being lowest in the subglottic area (Table 1, Fig. 3). During infection local macrophages were the only population that did not increase (Fig. 3) and did not penetrate the epithelium (Fig. 5). This was true for all time points and all regions analyzed.

Inoculation of viable B.

pertussis induces cellularchanges in the laryngeal mucosa comparable to that of heat-inactivated M. catarrhalis.

After inoculation of viable B. pertussis, four major observations were made, which are comparable with those observed after inhalation of M. catarrhalis. First, not only neutrophils but also DCs and T and B lymphocytes increased with similar kinetics in the laryngeal mucosa during infection (Fig. 6). Second, this increase was most pronounced in the subglottic area (and the tracheal mucosa), and very much less seen in the glottic area. Third, the cell number present subepithelially was higher than that within the epithelium (data not shown). Fourth, in contrast to neutrophils, DCs, and T and B lymphocytes, local macrophages were less numerous in the subglottic area and were not observed within the epithelium at any time during the infection. The major difference between the application of B. pertussis and M. catarrhalis was the prolonged time course (first peak of infiltration, 24 h for B. pertussis and 2 h for M. catarrhalis) and the preferential accumulation of T lymphocytes 96 h after inoculation of viable B. pertussis.

Figure 6
figure 6

Number of immunocompetent cells within the subglottic and the glottic mucosa of control animals (time = 0) and at various times after inoculation of B. pertussis (n= 4 at each time), mean ± SD. When the number of cells was less than 5 cells/0.1 mm2, data of the animals were pooled; *indicates values that differ significantly from control values of the same region (p≤ 0.05).

Full size image

DISCUSSION

In the current study, acute laryngitis in the rat was induced by the application of heat-killed M. catarrhalis or viable B. pertussis. These models were well established to study the influx of immunocompetent cells into the mucosa of the trachea (7). The pathogenicity of heat-killed M. catarrhalis is mainly because of the lipooligosaccharides contained in the cell wall (18). In a second animal model, viable B. pertussis were used, which exert their pathogenicity via different adhesins and toxins (19). After inhalation of M. catarrhalis, inflammation of the laryngeal mucosa peaks within 2–3 h. In contrast, after B. pertussis inoculation, a first peak of the inflammation is seen only after 24 h. This different time course might be related to the mode of application (inhalation versus inoculation). In addition, in the current study, B. pertussis was inoculated directly after warming up to 37°C, and it is known that B. pertussis needs about 8 h after reaching body temperature to develop its full pathogenicity (19). However, both pathogens caused similar cellular changes in the laryngeal mucosa. In addition, the inflammation preferentially occurred in the subglottic area of the larynx and to a lesser extent in the adjacent glottic area. This pattern closely resembles the situation in the clinic, in which also during acute laryngotracheitis mainly the subglottic area is affected (1). This indicates that both animal models are helpful in elucidating the pathophysiology of acute laryngotracheitis.

In the early phase of inflammation, the well-known increase of neutrophils in the laryngeal mucosa was observed in both experimental models. However, DCs entered the mucosa with similar kinetics and as early as 2 h after application of M. catarrhalis. This is in agreement with experiments showing that early after initiating inflammation the number of DCs in the mucosa of the trachea (7) and middle ear (20) increases considerably. Functionally, this shows that even in the very early phase of local inflammation not only is phagocytosis of invading pathogens by neutrophils of importance but also the initiation of a specific immune response. It is known that DCs take up antigen, transport it to the draining lymph node, and present the processed antigen to T lymphocytes, which then provide help for cytotoxic T cells and the production of antibodies by B cells (21). This scenario is supported by two other observations. First, DC numbers also increase in the epithelium of the laryngeal mucosa, a localization in which invading pathogens are found in high numbers very early in the course of infection. Secondly, the involvement of the specific immune system even at very early times during infection is also indicated by the increase of T and B cells both subepithelially and within the epithelium. Although on a lower level, their kinetics are comparable with those of neutrophils and DCs. These events are very fast: Within an hour after M. catarrhalis inhalation, cells pass the endothelium of the subepithelially located vessels, transmigrate the lamina propria, and cross the basal lamina to enter the epithelium.

It is known that T lymphocytes play an essential role in the defense against B. pertussis. This might be supported by the second increase of the T cells (96 h), which probably represents the specific immune response against B. pertussis (22).

Although the pattern of cellular changes in the laryngeal mucosa was similar in all regions of the larynx investigated, the maximum cell numbers differed considerably. The subglottic area (comparable with the trachea) was always significantly more affected than the glottic area, an observation also frequently made in the clinical situation. The present study shows that about three to five times more neutrophils, DCs, and T and B lymphocytes were present in the mucosa of the subglottic region compared with that of the glottic region. Because we determined the number of cells per area, and the actual volume of the glottic and subglottic area is unknown, we cannot provide the absolute number of cells entering the two compartments. However, the data clearly show that during infection the relative increase in the number of neutrophils, DCs, and T lymphocytes is much higher in the subglottic area compared with the glottic area. Such a huge difference in cellular influx into the mucosa of two directly adjacent locations is surprising, and indicates that the inflammatory response is specifically regulated in the respective microenvironment by mechanisms largely unknown so far. Our data suggest that local macrophages might be involved in creating the two different microenvironments. In contrast to DCs and T and B lymphocytes, which are present in higher numbers in the subglottic than in the glottic area, the mucosa of the subglottic area contains significantly lower numbers of macrophages compared with the other regions of the larynx. This observation has been made in three different rat strains [PVG and BN in the current study; Lewis (5)]. Thus, a higher number of local macrophages in the glottic area may provide a more efficient early defense against invading pathogens, and would therefore not depend so much on the recruitment of other leukocytes. This would result in reduced swelling of the glottic area, which is essential for life because the glottic area is the narrowest part of the respiratory system in adults. The induction of the specific immune response probably occurs in the subglottic area. Because of the lower number of macrophages present, leukocytes have to be recruited to fight the invading pathogens. Although a measurement of mucosal swelling was not performed in the present study, it can be assumed that the high number of immigrating leukocyte subsets into the subglottic area is accompanied by a greater amount of swelling in this region. This can be tolerated up to a certain degree because the subglottic area has a larger diameter than the glottic area and therefore swelling of the mucosa in this region is not immediately life-threatening. In contrast, the infant's subglottic area represents the narrowest part of the upper airways, and an edema of 1 mm reduces the normal infant's subglottic cross-sectional area by approximately 35% (1). Therefore, swelling may result in fulminant respiratory distress. Thus, the mucosa of the larynx fulfills two opposite tasks during infection: permitting a continuous oxygen supply by allowing only minimal swelling (glottic area) and induction of specific immune responses leading to a more efficient elimination of the acute infection and providing memory for future challenges (subglottic area).

However, the current data do not exclude other possible causes for a region-specific response. For example, the epithelium of the subglottic area itself might produce larger amounts of cytokines and chemokines (23), or might permit the preferential adherence of the pathogens (19).

In summary, the mucosa of two directly adjacent locations in the larynx reveal a huge difference in their responsiveness during infection. This enables both continuous oxygen supply by preventing swelling of the glottic area, and induction of a specific immune response by allowing massive recruitment of leukocytes, especially DCs, to the subglottic area. With the present animal model, it is possible to investigate the role of local macrophages in creating specific microenvironments within the laryngeal mucosa, and to study viral infections of the larynx. In addition, this animal model allows us to study how the course of acute laryngotracheitis can be modified, e.g. by injecting antibodies directed against adhesion molecules, and therefore might help to develop new therapeutic strategies for the treatment of acute laryngeal infections (24).