Streptococcus pneumoniae - Infection and Immunity - American

INFECTION AND IMMUNITY, Feb. 1995, p. 442–447 0019-9567/95/$04.0010 Copyright q 1995, American Society for Microbiology

Vol. 63, No. 2

Interaction of Pneumolysin-Sufficient and -Deficient Isogenic Variants of Streptococcus pneumoniae with Human Respiratory Mucosa C. F. J. RAYNER,1 A. D. JACKSON,1 A. RUTMAN,1 A. DEWAR,2 T. J. MITCHELL,3 P. W. ANDREW,3 P. J. COLE,1 AND R. WILSON1* Host Defence Unit1 and Electron Microscopy Unit,2 Department of Thoracic Medicine, Royal Brompton National Heart and Lung Institute, London SW3 6LR, and Department of Microbiology and Immunology, University of Leicester, Leicester LE1 9HN,3 United Kingdom Received 22 July 1994/Returned for modification 6 October 1994/Accepted 16 November 1994

Streptococcus pneumoniae is the most common cause of community-acquired pneumonia, and pneumolysin, a hemolytic toxin, is thought to be an important virulence factor. We have studied the interaction of a pneumolysin-sufficient type II S. pneumoniae strain (PL1) and an otherwise identical pneumolysin-deficient derivative (PL2) with human respiratory mucosa in an organ culture with an air interface for up to 48 h. Ciliary beat frequency (CBF) was measured by a photometric technique, and adherence to and invasion of the epithelium were assessed by scanning and transmission electron microscopy. PL1 and PL2 caused a progressive fall in CBF compared with the control which became significant (P < 0.01) at 24 h for PL1 and at 48 h for PL2. At 24 h, there was a significant increase in the percentage of the mucosa of the organ culture that was damaged for PL1 compared with the control (P < 0.01) and PL2 (P < 0.02). At 48 h, there was a significant increase in mucosal damage for both PL1 (P < 0.005) and PL2 (P < 0.05) compared with the control. At 24 and 48 h, PL1 and PL2 adhered predominantly to mucus and damaged cells. PL1 infection alone caused separation of tight junctions between epithelial cells, and at 48 h PL1 cells were adherent to the separated edges of otherwise healthy unciliated cells. PL1 and PL2 both caused damage to the epithelial cell ultrastructure. S. pneumoniae infection caused patchy damage to the respiratory mucosa and a lowered CBF. These changes were more severe and occurred earlier with the pneumolysin-sufficient variant. Colonization of the human nasopharynx by Streptococcus pneumoniae is important as a source of transmission to other susceptible individuals, as a source of infection of other parts of the respiratory tract, and as a site of mucosal invasion leading to systemic disease. S. pneumoniae is the most common cause of community-acquired pneumonia in adults, leading to severe disease (9, 25, 46), and colonizes the upper respiratory tract of children and up to 70% of healthy adults (7, 20). S. pneumoniae is also commonly isolated as a cause of otitis media, acute sinusitis, and exacerbations of chronic bronchitis and, in the absence of meningococcal disease and the advent of the Haemophilus influenzae type b vaccine, is the most common cause of bacterial meningitis in adults and young children (7). Pneumolysin is a thiol-activated toxin of S. pneumoniae (40) and is thought to be an important virulence factor involved in the pathogenesis of pneumococcal disease. We have therefore studied the interaction between S. pneumoniae and human respiratory mucosa in an organ culture with an air interface to determine whether pneumolysin is important for airway colonization and invasion of the mucosa. To investigate the role of pneumolysin in the interaction of S. pneumoniae with respiratory mucosa, we have used isogenic type 2 pneumococci which are sufficient or deficient in the production of pneumolysin.

MATERIALS AND METHODS The pneumolysin-sufficient (PL1) and -deficient (PL2) variants of a type II S. pneumoniae strain were kindly supplied by J. C. Paton (Adelaide Women’s and Children’s Hospital, North Adelaide, Australia) (13). A standard inoculum was prepared as follows. A 10-ml volume of brain heart infusion (Oxoid, Basingstoke, United Kingdom) broth was inoculated with four or five colonies of S. pneumoniae from an overnight culture on blood agar no. 2 (Oxoid). This mixture was incubated for 12 h at 378C and then centrifuged at 2,000 3 g for 15 min. The bacterial pellet was resuspended in 1 ml of fresh broth containing heat-inactivated fetal calf serum at a ratio of 1:5 brain heart infusion. This mixture was diluted with fresh serum-broth to give an optical density at a wavelength of 500 lambda of 0.7 nm. This solution was then incubated for 4 h, a viable cell count was performed in triplicate, and the bulk of the culture was frozen at 2708C. When the viable cell counts were known, the culture was thawed and diluted in fresh serum-broth to give an inoculating dose of 107 CFU/ml. The standard inoculum was stored in 1-ml volumes at 2708C. Prior to inoculation, the 1-ml inocula were thawed and centrifuged at 2,000 3 g for 3 min, washed, and centrifuged three times through 1 ml of phosphate-buffered saline (PBS). At the time of inoculation, viable counts were performed on the inoculum. Organ culture preparation. Human adenoids, resected at operation, were placed in minimal essential medium (MEM) (Gibco, Paisley, United Kingdom) containing antibiotics (penicillin [50 U/ml], streptomycin [50 mg/ml], and gentamicin [50 mg/ml]) to eradicate commensal flora. The tissue was transported to the laboratory, where it was checked by light microscopy, and tissue with a smooth surface and actively beating cilia was transferred to fresh MEM with antibiotics. The tissue was then dissected into 4-mm2 squares. Squares with at least one completely ciliated edge were chosen for use in the organ culture. After 4 h in MEM with antibiotics, the tissue was placed in antibiotic-free MEM for 1 h. To prepare the organ culture system, a 3-cm-diameter petri dish was placed inside a 5-cm-diameter petri dish, and 4 ml of MEM was placed in the outer petri dish and the inner dish was kept dry. A 70-mm-long and 5-mm-wide strip of sterile filter paper (no. 1; Whatman, Maidstone, United Kingdom) was immersed in MEM and then laid across the inner petri dish, with its ends immersed in the medium contained in the outer petri dish. The filter paper acted as a wick to supply nutrients to the tissue. The square of adenoid tissue was placed adventitial cut surface down onto the center of the filter paper. Semimolten 1% agar (no. 1; Oxoid) at 408C was pipetted around the tissue, and as it set it sealed all the cut edges. The organ culture was then incubated at 378C for 1 h in a humidified atmosphere containing 5% CO2 at 378C.

* Corresponding author. Mailing address: Host Defence Unit, Department of Thoracic Medicine, Royal Brompton National Heart and Lung Institute, Emmanuel Kaye Building, Manresa Rd., London SW3 6LR, United Kingdom. Phone: 071 351 8337. Fax: 071 351 8338. 442

VOL. 63, 1995

FIG. 1. Cell damage and cell debris in an organ culture infected with S. pneumoniae PL1 for 48 h (magnification, 32,500; bar 5 4.0 mm).

S. PNEUMONIAE PNEUMOLYSIN VARIANTS

443

then embedded in araldite. Samples were coded so that the observer was unaware of the experimental conditions. For transmission electron microscopy assessment, an ultrathin section (70 to 90 nm) through the central portion of each specimen was examined. Sections typically contained 150 to 250 cells. Each epithelial cell observed was scored according to the following parameters: extrusion of a cell from the epithelial surface (0, no extrusion; 11, cell fully extruded but in contact with epithelium; 1, intermediate), loss of cilia (for cells bearing cilia, as it was not possible to determine if a totally unciliated cell was originally unciliated or had lost all cilia during the culture; 0, fully ciliated; 11, sparsely ciliated; 1, intermediate), numbers of unciliated cells, presence of cytoplasmic blebbing from the luminal surface of ciliated and unciliated cells (0, nil; 11, severe; 1, intermediate), and mitochondrial damage in ciliated and unciliated cells (present or absent). Statistical analysis (2). Viable cell counts and bacterial adherence were compared by the Mann-Whitney U test. The Kruskal-Wallis analysis of variance was used to determine if there was any difference between the mean CBF or percentage of the surface area of each of the mucosal parameters measured for the control and the two variants at each time point. To investigate the nature of any differences found, the mean CBF or percentage of the surface area of infected tissue was compared specifically with the control by using the Mann-Whitney U test. A P value of ,0.05 was taken as significant. Data are given as means 6 standard deviations.

RESULTS Inoculation and incubation. For each experiment, three organ cultures were prepared. The control tissue was inoculated with 20 ml of PBS, and the second and third organ cultures were inoculated with 20 ml of the prepared standard inoculum of the type II pneumolysin-sufficient and pneumolysin-deficient variants in PBS. The organ cultures were incubated for 4, 24, or 48 h. At the end of the given period, the four edges of each organ culture were touched with a sterile loop and plated onto blood agar no. 2 (Oxoid) to confirm pure growth of S. pneumoniae and sterility of the control. If the cultures were contaminated or the control was not sterile, the whole experiment was discarded. CBF. The tissue was removed from the filter paper and placed in a 3-cmdiameter petri dish containing 1 ml of MEM. Ciliary beating along the edge of the tissue could be visualized by light microscopy (3320 magnification). At 24 and 48 h, ciliary beating could not be visualized in areas where mucus or cell debris had accumulated, and for this reason ciliary beating frequency (CBF) measurements were taken only when cilia were obviously beating. The CBF was measured by a photometric technique (36) and was calculated as the mean of 10 separate areas of beating cilia. Processing and fixation for scanning electron microscopy. The tissue was fixed in 2.5% cacodylate-buffered glutaraldehyde (pH 7.2) for 24 h before being processed through buffer washes and then postfixed in 1% osmium tetroxide in cacodylate buffer for 1 h. The tissue then underwent a graded series of dehydrations through methanol to acetone followed by critical point drying in CO2. The tissue was mounted on aluminum stubs and given a conductive coating of gold. Scanning electron microscopy. Samples were coded so that the observer was unaware of the experimental conditions. The samples were examined with a Hitachi 4000 S field emission scanning electron microscope. A careful morphometric assessment of the mucosal surface of the organ culture was made. A grid (10 by 10 cm) of 100 squares was placed over the image on the screen at a magnification of 350. The grid was used to select a representative sample of the total surface area by a standard preselected method. Two diagonal lines were followed from grid corner to grid corner, each containing 10 squares. A further five squares were selected from a fixed point in the center of each of the four quadrants, for a total of 40 squares. Each of the 40 squares was then examined at a magnification of 32,000 and divided into 100 squares by using the 10-by-10 grid. The numbers of these squares occupied by mucus, extruded cells and cell debris, and ciliated and unciliated cells were determined and used to calculate the percent surface area occupied by that mucosal feature. Therefore, for each sample, a mean percent surface area occupied by each mucosal feature was estimated from an analysis of all 40 squares. Each of the 40 squares had a surface area of 800 mm2. The total area of each sample examined was therefore 0.032 mm2. Areas of cell damage were defined as extruded cells and/or cytoplasmic surface blebbing. Cell debris was defined as cellular material, which was often spherical in appearance (Fig. 1). These two observations were combined in the morphometric assessment as mucosal damage. Unciliated areas were defined as areas not covered by cilia, with or without microvilli. Unciliated cells which had separated tight junctions but were otherwise normal were also scored in this category rather than as damaged cells. When bacteria were seen, they were counted and their position was carefully noted. For each organ culture, the total number of bacteria associated with each of the four mucosal features was calculated. In order to estimate the density of bacteria associating with each mucosal feature, this figure was divided by the percentage of the mucosal surface occupied by that feature. Transmission electron microscopy. For transmission electron microscopy, tissue was fixed in 2.5% cacodylate-buffered glutaraldehyde, postfixed in 1% osmium tetroxide, subjected to a graded series of dehydrations in alcohols, and

Viable counts. The mean viable count for PL1 was 1.12 3 108 (95% confidence interval, 0.97 to 1.3) and that for PL2 was 1.07 3 108 (95% confidence interval, 0.92 to 1.2). The growth rates of the two variants in broth did not differ, and there was no statistical difference between the viable counts of the inocula of the two variants in any of the experiments, so a comparison between the effects of PL1 and PL2 could be made. CBF. At 4 h (n 5 6), there was no significant difference in the CBFs of the control and the two pneumolysin variants, with values of 12.1 6 0.76 Hz for the control, 10.8 6 1.1 Hz for PL1, and 11.9 6 0.7 Hz for PL2. At 24 h (n 5 6), the CBF of PL1 (8.8 6 0.9 Hz; P , 0.01) but not that of PL2 (9.7 6 0.8 Hz; P , 0.06) was significantly lower than that of the control (10.7 6 0.7 Hz). Although the CBF of PL1-infected organ cultures was less than that for PL2, this difference was not significant (P , 0.07). At 48 h (n 5 6), there was a further fall in the CBFs of PL1 (5.6 6 1.9 Hz; P , 0.005) and PL2 (7.4 6 1.3 Hz; P , 0.01) compared with that of the control (10.5 6 0.4 Hz).

TABLE 1. Scanning electron microscopy of the interaction between S. pneumoniae and adenoid organ culture % Surface area (mean 6 SD) occupied by:

Time and expta

Mucus

Damaged mucosa

Ciliated epithelium

Unciliated epithelium

4h Control PL1 PL2

48.6 6 7.6 32.3 6 16.7 33.0 6 17.6

4.3 6 0.9 8.4 6 2.9 6.1 6 2.8

22.2 6 16.0 21.2 6 10.5 25.4 6 18.8

24.7 6 11.7 38.1 6 16.9 35.5 6 7.4

24 h Control PL1 PL2

48.4 6 20.6 31.4 6 15.5 41.8 6 21.2

4.3 6 1.4 21.7 6 6.8b,c 8.8 6 2.8

25.1 6 9.1 8.0 6 6.5b 16.2 6 13.2

22.2 6 12.5 38.9 6 11.8 33.2 6 19.6

48 h Control PL1 PL2

39.5 6 22.3 45.4 6 7.8 36.0 6 19.2

5.0 6 2.6 30.0 6 4.7d 23.4 6 10.5e

19.5 6 6.2 4.9 6 2.7d 7.5 6 5.7e

36.0 6 12.3 19.7 6 2.1e 33.1 6 17.0

a At each time, an experiment (n 5 6) consisted of three organ cultures constructed from the same adenoid tissue. b P , 0.01 for PL1 versus control value. c P , 0.02 for PL1 versus PL2 value. d P , 0.005 for PL1 versus control value. e P , 0.05 for PL2 versus control value and for PL1 versus control value.

444

RAYNER ET AL.

INFECT. IMMUN. TABLE 3. Bacterial density on each mucosal feature of adenoid organ culture Bacterial densityb on: Time and expta

FIG. 2. Organ culture infected with S. pneumoniae PL1 for 24 h. The cilia of infected organ cultures appear disorganized and bent in different directions (magnification, 35,000; bar 5 2.0 mm).

Scanning electron microscopy. The percentages of surface area occupied by mucus, mucosal damage, unciliated cells, and ciliated cells at the three time points are shown in Table 1. At 4 h, there was no difference between the control, PL1, and PL2 in the measured parameters. At 24 h, in PL1 the surface area covered by mucosal damage had increased, and this was associated with a decrease in the surface area covered by ciliated epithelium. PL1 also demonstrated a significant (P , 0.02) increase in mucosal damage compared with PL2. In the PL1-infected organ cultures alone, cilia appeared to be disorganized (Fig. 2) and had an increased granularity on their shafts, and extruded ciliated cells were seen. At 48 h, there was a significant increase in the surface area covered by mucosal damage for both PL1 (P , 0.005) and PL2 (P , 0.05) compared with the control. There was also a significant (P , 0.01) decrease in ciliated surface area for both PL1 and PL2 compared with the control. Bacterial adherence. At 4 h, bacteria were not seen in association with the mucosal surface despite the inoculation of large numbers of bacteria at the start of the experiment. The total numbers of bacteria adhering to the mucosal surface at 24 and 48 h are shown in Table 2. The density of bacteria adhering to each mucosal feature has been calculated by dividing the total number of bacteria by the percentage of the mucosal surface occupied by each feature, and these results are shown in Table 3. At 24 h, PL1 and PL2 were most commonly found in association with mucus, and there was no significant differ-

Mucus

Damaged mucosa

Ciliated epithelium


24 h PL1 PL2

3.7 6 1.0c 2.3 6 2.2

2.0 6 1.8c 1.5 6 1.4

0.8 6 1.7 0

0.43 6 0.27 0.16 6 0.19

48 h PL1 PL2

7.7 6 2.0d 8.6 6 6.0e

10.3 6 1.5d 8.9 6 0.7e

1.3 6 1.0 0.4 6 0.8

3.0 6 1.2 1.7 6 1.4

a At each time, an experiment consisted of three organ cultures (including a control) constructed from the same adenoid tissue. b Results are the mean (n 5 6) total number of bacteria adhering to each mucosal feature divided by the percentage of the mucosal surface occupied by that feature 6 standard deviation. c For adherence to mucus and damaged cells compared with ciliated and unciliated tissues, P , 0.05. d For adherence to mucus and damaged cells compared with ciliated and unciliated tissues, P , 0.02. e For adherence to mucus and damaged cells compared with ciliated and unciliated tissues, P , 0.05.

ence in adherence to mucus for the two variants (P , 0.7). PL1 and PL2 adherence to mucus was significantly greater than their adherence to ciliated and unciliated tissues (P , 0.05). In infected cultures, the mucus became fibrogranular, a result which was not seen in control organ cultures, and appeared to contain both cellular material and bacteria (Fig. 3). PL2 bacteria were not seen in association with cilia, and association with cilia was very uncommon for PL1. Both PL1 and PL2 were seen to adhere to damaged cells. Adherence to unciliated epithelium was uncommon for both variants, although the total number of PL1 cells adhering to unciliated epithelium was greater than that for PL2 cells (P , 0.05). The reason for this was PL1 adherence to normal unciliated cells at sites where separation of tight junctions had occurred. This separation of cells that otherwise appeared normal was seen exclusively in PL1-infected cultures. At 48 h, the two variants again demonstrated tropism for mucus and damaged cells compared with ciliated and unciliated tissues (Table 2). PL1 cells were more frequently seen adhering to unciliated cells where a separation of tight junction integrity could be seen (Fig. 4 and 5). The total numbers of bacteria adherent to the mucosa were not significantly different for the two variants (Table 2). Table 3 shows the density of bacteria adhering to each mucosal feature, and the results

TABLE 2. Bacterial adherence to the mucosal surface of adenoid organ culture assessed by scanning electron microscopy No. of bacteria (mean 6 SD) adhering tob: Time and expta Mucus

a

24 h PL1 PL2

19.2 6 8.9 16.7 6 9.9

48 h PL1 PL2

142.0 6 14.7 128.7 6 23.6

Damaged mucosa

Ciliated epithelium


Total no. of bacteriac

9.8 6 4.5d 4.8 6 3.0

1.1 6 1.6 0

6.2 6 2.7d 2.3 6 2.3

35.8 6 12.9 23.7 6 11.5

121.7 6 8.4d 100.1 6 26.6

2.7 6 2.4 1.0 6 2.7

23.8 6 8.7d 14.2 6 8.5

290.7 6 21.1 241.3 6 47.5

At each time, an experiment consisted of three organ cultures (including a control) constructed from the same adenoid tissue. n 5 6. c Mean total number of bacteria adhering to the area of mucosa examined, which was 0.032 mm2. d P , 0.05 in comparison with the value for PL2. b


VOL. 63, 1995

FIG. 3. Organ culture infected by S. pneumoniae PL1 for 24 h. Diplococci are seen, and the mucus appears fibrogranular (magnification, 36,250; bar 5 1.6 mm).

emphasize the tropism of PL1 and PL2 for mucus and damaged cells. Transmission electron microscopy results. The effects of S. pneumoniae infection on epithelial ultrastructure are shown in Table 4. Both PL1 and PL2 caused an increase in the number of cells extruding from the epithelial surface, compared with that of the control, and an increase in the number of poorly ciliated cells. Infected organ cultures also had an increase in the number of unciliated cells with mitochondrial damage compared with that of the control. Cell extrusion and toxic changes in cells were greater for tissues infected with PL1 than with PL2, but this was not significant except for a significant increase in cytoplasmic blebbing on unciliated cells (P , 0.05). DISCUSSION Bacterial adherence to mucosal surfaces is thought to be an important determinant of colonization and the pathogenesis of most infections (11, 28). Previous studies have shown that S. pneumoniae adheres to suspended buccal and nasopharyngeal cells (4–6, 37). These studies have identified an adhesin that

FIG. 4. Organ culture infected by S. pneumoniae PL1 for 48 h. PL1 bacteria are seen adhering to both areas of cell damage and areas of unciliated epithelium where there is separation of the tight junctions between cells (magnification, 33,750; bar 5 2.6 mm).

445

FIG. 5. Organ culture infected by S. pneumoniae PL1 for 48 h. Bacteria are seen adhering to separated tight junctions between unciliated cells that are otherwise normal (magnification, 33,750; bar 5 2.6 mm).

forms a link between pneumococci and the carbohydrate receptors on the host cell (3). S. pneumoniae adheres to GalNac b1-4 Gal sequences of glycosphingolipids found in the respiratory tract (4, 23). However, little is known about the interaction of S. pneumoniae with intact respiratory mucosa. By using the frog palate, it has been demonstrated that S. pneumoniae adheres rapidly to mucus but not to normal ciliated epithelium (31). S. pneumoniae has also been shown to occupy a gelatinous layer formed above the epithelial surface of immersed human nasal turbinate tissue in organ culture; however, adherence of bacteria to normal epithelium was rare (17). The nature of the gelatinous layer was uncertain, although it was thought that it might contain both mucus and bacterial products. S. pneumoniae adherence may change if the epithelium is damaged. Influenza A virus infection of murine respiratory epithelium caused damage which led to a significant increase in S. pneumoniae adherence compared with adherence to non-virus-infected tissue (32). Respiratory pathogens are known to release products which interfere with mucosal defences by slowing the ciliary beat, causing ciliary dyskinesia, and causing epithelial disruption and cell death (15, 22, 33, 42, 44, 45). Pneumolysin, an oxygenlabile intracellular cytolysin, is thought to be directly involved in the pathogenesis of pneumococcal disease (29). Pneumolysin causes slowing of the ciliary beat and ciliostasis (26, 42) and has been shown to be toxic for the guinea pig cochlea, causing disorganization and loss of cilia, damage to the hair bundles, and change in the surface of the cilia (14). In mice, pneumolysin-deficient mutants have reduced virulence, and immunization with pneumolysin increases survival time after the intranasal challenge with S. pneumoniae (12, 13, 30). Instillation of pneumolysin into the rat lung induces the salient histological features of a pneumococcal pneumonia (16). In the present study, we have shown that pneumolysin-sufficient S. pneumoniae caused a progressive lowering of the CBF over 48 h, and there was a loss of ciliated epithelium. The cilia appeared disorganized, with an increased granularity on the surfaces of the cilium shafts, similar to that produced by pneumolysin in the guinea pig cochlea. PL1-infected organ cultures showed a progressive increase in the surface area covered by damaged cells, which included extruded ciliated and unciliated cells and cells with damage to their surface, such as blebbing, pitting, and craters. Transmission electron microscopy showed toxic changes to cell ultrastructure. Mucosal damage occurred

446

RAYNER ET AL.

INFECT. IMMUN.

TABLE 4. Transmission electron microscopy of adenoid organ culture infected by S. pneumoniae for 24 h % of cells showing indicated degree of given histological featureb

Culture (n)a

Cells extruding from cell surface 0

Control (297.2) PL1 (254.2) PL2 (298.7)

89.2 80.0 84.5

1

11

9.3 12.2 11.0

1.5 7.8c 4.5e

Presence of cilia Ciliated cells 0

1

11

69.4 66.7 74.2

6.4 11.1c 10.5c

0.6 1.8c 0.8

Blebbing on: Ciliated cells

Mitochondrial damage in:

Unciliated cells

Ciliated cells

Unciliated cells

Unciliated cells

0

1

11

0

1

11

Absent

Present

Absent

Present

23.6 20.4 14.5

98.7 96.3 97.5

1.3 2.2 2.1

0 1.5 0.4

97.2 92.5 98.2

2.8 6.8d 1.8

0 0.7 0

97.5 93.2 94.6

2.5 6.8 5.4

94.3 80.7 86.8

5.7 19.3c 13.2e

a The cells were from six separate experiments, each consisting of three organ cultures constructed from the same adenoid tissue. n, mean number of cells examined per experiment. b Each cell in a tissue section was examined for extrusion from the epithelial surface (0, cell normally positioned in the epithelium; 11, cell fully extruded but in contact with epithelium; 1, intermediate), the presence of cilia (0, full complement of cilia on the cell surface; 11, sparsely ciliated; 1, intermediate), cell blebbing (cytoplasmic projections from the luminal cell surface) (0, none; 11, severe; 1, intermediate), and mitochondrial damage. c P , 0.01 in comparison with the control value. d P , 0.05 in comparison with the control and PL2 values. e P , 0.05 in comparison with the control value.

in the absence of bacterial adherence, which suggests that a diffusible bacterial factor released on the mucosal surface may mediate these changes. Pneumolysin has previously been shown by electron microscopy to exert a toxic effect on epithelial cells (14, 42) and to cause endothelial and alveolar epithelial cell injury (34, 35). Although at 24 h PL2 variant-infected organ cultures showed less mucosal damage than PL1-infected organ cultures, at 48 h they also showed a lowering of the CBF and an increase in the surface area associated with cell damage, suggesting that other bacterial factors besides pneumolysin damage the respiratory mucosa. However, mucosal damage was less severe in PL2 variant-infected organ cultures, and it is possible that host inflammatory mediators released by the organ culture in response to the infection, such as cytokines or prostaglandins, could be involved in lowering the CBF and causing epithelial cell damage (10, 39, 43). Pneumolysin has been shown to stimulate the production of tumor necrosis factor and interleukin 1 from human mononuclear phagocytes (21). This suggests that pneumolysin may mediate inflammation, not only indirectly by damaging tissue and activating complement (16, 29, 42), but also by stimulating the production of these two cytokines. Both variants of S. pneumoniae were seen to adhere to damaged cells and extruded cells and rarely to normal epithelium. This lack of S. pneumoniae adherence to normal epithelium has been noted previously (17, 31, 32). Separation of tight junctions between otherwise apparently normal unciliated cells and bacterial adherence at this site were seen exclusively in PL1-infected organ cultures. This might represent an important mechanism of invasion which appears to be dependent on pneumolysin. Pneumococcal adherence to human endothelial cells has been shown to cause damage and cell separation, which was also caused by purified pneumococcal cell wall components. These changes were inhibited by antibodies to interleukin 1 and tumor necrosis factor. In these studies, pneumococci were similarly seen to accumulate along the borders of endothelial cells where separation had occurred (18). Bacteria were not seen associated with the mucosa at 4 h, but there were more PL1 than PL2 bacteria at 24 and 48 h, although the difference did not reach statistical significance. The two variants did not differ in their growth rates in broth culture. The higher numbers of PL1 may be accounted for by bacterial tropism for damaged cells and sites of separated tight junctions, which were more frequent in the PL1-infected or-

gan cultures. Damaged cells may also release nutrients which might influence the growth of PL1 bacteria. A pneumolysin-sufficient type III S. pneumoniae infection of an immersed organ culture caused a reduction of the CBF but only minor damage to the epithelial surface at 24 h (17). Secretions formed a thick gelatinous layer, in which chains of S. pneumoniae were found, overlying the epithelium. Immersion of tissue in culture medium removes the physiological air-mucosa interface, and the medium may dilute the concentration of bacterial toxins produced. This may explain the increased damage seen in the present study using an organ culture with an air interface. Pneumolysin is easily oxidized (8), and this inactivates the toxin, but despite the air interface the damage caused by the PL1 variant was greater, suggesting that pneumolysin is active in the microenvironment on the mucosal surface. Respiratory mucus contains a heterogeneous mixture of mucus glycoprotein molecules which may differ in the amount and extent of sialyation and sulfation of the constituent oligosaccharides (24, 38). The multiple carbohydrate chains of the mucin molecule may represent sites for adhesion of microorganisms (27). Bacterial infection of guinea pig organ cultures caused increased mucus production (1). This may be mediated by bacterial products (42) or host inflammatory mediators (27). The character of the mucus may also be changed by the infection (19). Another pneumococcal toxin, neuraminidase, might be expected to decrease the viscosity of mucus through cleavage of sialic acid residues. In the present study, there was no change in the surface area covered by mucus, but the appearance of the mucus changed in the infected organ cultures. The mucus appeared fibrogranular, and bacteria were seen in association with the mucus. These appearances were similar to those seen previously in an organ culture immersed in medium (17), but the numbers of bacteria were lower in the present study, and bacteria were usually single cocci or diplococci rather than the chains seen in the presence of medium. This study, using an organ culture with an air interface, shows that ciliary beat slowing, ciliary disorganization, epithelial damage, and abnormal mucus occur during infection of the respiratory mucosa by S. pneumoniae in vitro. The ciliary beat slowing and epithelial damage caused by a variant deficient in pneumolysin had a delayed onset and reduced severity. S. pneumoniae can therefore perturb the mucociliary defense of the epithelium and cause damage. Pneumolysin makes a major contribution to these changes. Bacteria adhere to mucus, dam-


VOL. 63, 1995

aged epithelial cells, and the edge of unciliated cells where separation of tight junctions has occurred. Separation of epithelial cell tight junctions between otherwise apparently normal unciliated cells occurred exclusively with the pneumolysinsufficient strain and may provide a route of invasion by S. pneumoniae which is pneumolysin dependent. ACKNOWLEDGMENTS This work was funded by the Wellcome Trust. T.J.M. is a Royal Society Research Fellow. We thank Jane Burditt for the preparation of the manuscript. REFERENCES 1. Adler, K. B., D. Hendley, and G. S. Davis. 1986. Bacteria associated with obstructive pulmonary disease elaborate extracellular products that stimulate mucin secretion by explants of guinea pig airway. Am. J. Pathol. 125: 501–514. 2. Altman, D. G. 1991. Practial statistics for medical research. Chapman & Hall, Ltd., London. 3. Andersson, B., E. H. Beachey, A. Tomasz, E. Tuomanen, and C. SvanborgEden. 1988. A sandwich adhesin on Streptococcus pneumoniae attaching to human oropharyngeal epithelial cells in vitro. Microb. Pathogen. 4:267–278. 4. Andersson, B., J. Dahmen, T. Frejd, H. Leffler, G. Magnusson, G. Noori, and C. S. Svanborg-Eden. 1983. Identification of active disaccharide unit of a glycoconjugate receptor for pneumococci attaching to human pharyngeal epithelial cells. J. Exp. Med. 158:559–570. 5. Andersson, B., B. Eriksson, E. Falsen, E. Frogh, L. A. Hansen, O. Nylen, H. Peterson, and C. Svanborg-Eden. 1981. Adhesion of Streptococcus pneumoniae to human pharyngeal epithelial cells in vitro: differences in the adhesive capacity among strains isolated from subjects with otitis media, septicemia, or meningitis or from healthy carriers. Infect. Immun. 32:311–317. 6. Andersson, B., D. Porras, L. A. Hanson, and C. Svanborg-Eden. 1985. Non-antibody containing fractions of breast milk inhibit epithelial attachment of Streptococcus pneumoniae and Haemophilus influenzae. Lancet i:643–644. 7. Austrian, R. 1984. Pneumococcal infections, p. 257–288. In R. Germanier (ed.), Bacterial vaccines. Academic Press, London. 8. Avery, O. T., and J. M. Neill. 1924. Studies in oxidation and reduction by pneumococcus. Oxidation of haemotoxins in sterile extracts of pneumococcus. J. Exp. Med. 39:745. 9. Bath, J. C. J. L., C. P. D. Boissard, M. A. Caulder, and M. A. J. Moffett. 1964. Pneumonia in hospital practice in Edinburgh 1960–1962. Br. J. Dis. Chest 58:1–16. 10. Bedard, M., C. D. McClure, N. I. Schiller, C. Francoeur, A. Cantin, and M. Denis. 1993. Release of interleukin-8, interleukin-6, and colony-stimulating factors by upper airway epithelial cells: implications for cystic fibrosis. Am. J. Respir. Cell. Mol. Biol. 9:455–462. 11. Beechey, E. H. 1981. Bacterial adherence: adhesion receptor interactions mediating the attachment of bacteria to mucosal surfaces. J. Infect. Dis. 143:325–345. 12. Berry, A. M., J. C. Paton, and D. Hansman. 1992. Effect of insertional inactivation of the genes encoding pneumolysin and autolysin on the virulence of Streptococcus pneumoniae type 3. Microb. Pathogen. 12:87–93. 13. Berry, A. M., J. Yother, D. E. Briles, D. Hansman, and J. C. Paton. 1989. Reduced virulence of a defined pneumolysin-negative mutant of Streptococcus pneumoniae. Infect. Immun. 57:2037–2042. 14. Commis, S. D., P. M. Osbourne, J. Stephen, M. J. Tarlow, T. L. Hayward, T. J. Mitchell, P. W. Andrews, and G. J. Boulnois. 1993. Cytotoxic effect on hair cells of the guinea pig cochlea produced by pneumolysin, the thiol activated toxin of Streptococcus pneumoniae. Acta Otolaryngol. (Stockholm) 113:152–159. 15. Denny, F. W. 1974. Effect of a toxin produced by Haemophilus influenzae on ciliated respiratory epithelium. J. Infect. Dis. 129:93–100. 16. Feldman, C., N. C. Munro, P. J. Jeffery, T. J. Mitchell, P. W. Andrew, G. J. Boulnois, D. Guerreiro, J. A. L. Rohde, H. C. Todd, P. J. Cole, and R. Wilson. 1991. Pneumolysin induces the salient histological features of pneumococcal infection in the rat lung in vitro. Am. J. Respir. Cell. Mol. Biol. 5:416–423. 17. Feldman, C., R. Read, A. Rutman, P. K. Jeffery, A. Brain, V. Lund, T. J. Mitchell, P. W. Andrew, G. J. Boulnois, H. C. Todd, P. J. Cole, and R. Wilson. 1992. The interaction of Streptococcus pneumoniae with intact human respiratory mucosa in vitro. Eur. Respir. J. 5:576–585. 18. Geelen, S., C. Bhattacharyya, and E. Tuomanen. 1993. The cell wall mediates pneumococcal attachment to and cytopathology in human endothelial cells. Infect. Immun. 61:1538–1543. 19. Griod, S., J. M. Zham, M. C. Plotkowski, G. Beck, and E. Puchelle. 1992. Role of physicochemical properties of mucus in the protection of respiratory epithelium. Eur. J. Respir. Dis. 5:477–487. 20. Hendley, J. O., M. A. Sande, P. M. Stewart, and J. M. Gwaltney, Jr. 1975. Spread of Streptococcus pneumoniae in families—carriage rate and distribu-

447

tion of types. J. Infect. Dis. 132:55–61. 21. Houldsworth, S., P. W. Andrew, and T. J. Mitchell. 1994. Pneumolysin stimulates production of tumor necrosis factor alpha and interleukin-1b by human mononuclear phagocytes. Infect. Immun. 62:1501–1503. 22. Johnson, A. P., and T. J. Inzana. 1986. Loss of ciliary activity in organ cultures of rat trachea with lipooligosaccharide isolated from Haemophilus influenzae. J. Med. Microbiol. 22:265–268. 23. Kirvan, H. C., D. D. Roberts, and V. Ginsburg. 1988. Many pulmonary pathogenic bacteria bind specifically to the carbohydrate sequence Gal NAC b 1-4 Gal found in some glycolipids. Proc. Natl. Acad. Sci. USA 85:6157–6161. 24. Lamblin, G., M. Lhermitte, A. Klein, N. Houdret, A. Scharfman, P. Ramphal, and P. Roussel. 1991. The carbohydrate diversity of respiratory mucins: a protection of the underlying mucosa? Am. Rev. Respir. Dis. 144:S19–S24. 25. Macfarlane, J. T., R. L. Finch, M. J. Ward, and A. D. Macrae. 1982. Hospital study of adult community acquired pneumonia. Lancet ii:255–258. 26. Munro, N. C., A. Barker, A. Rutman, G. Taylor, D. Watson, W. J. McDonaldGibson, R. Towart, A. Taylor, R. Wilson, and P. J. Cole. 1989. The effect of pyocyanin and 1-hydroxyphenazine on in vivo tracheal mucus velocity. J. Appl. Physiol. 76:316–323. 27. Nadel, J. A. 1991. Role of mast cell and neutrophil proteases in airway secretion. Am. Rev. Respir. Dis. 144:S48–S51. 28. Neiderman, M. S. 1989. Bacterial adherence as a mechanism of airway colonisation. Eur. J. Clin. Microbiol. Infect. Dis. 8:15–20. 29. Paton, J. C., P. W. Andrew, G. J. Boulnois, and T. Mitchell. 1993. Molecular analysis of the pathogenicity of Streptococcus pneumoniae: the role of pneumococcal proteins. Annu. Rev. Microbiol. 47:89–115. 30. Paton, J. C., R. A. Lock, and D. J. Hansman. 1983. Effect of immunization with pneumolysin on survival time of mice challenged with Streptococcus pneumoniae. Infect. Immun. 40:548–552. 31. Plotkowski, M. C., G. Beck, J. Jacquot, and E. Puchelle. 1989. The frog palate mucosa as a model for studying bacterial adhesion to mucus coated respiratory epithelium. J. Comp. Pathol. 100:37–46. 32. Plotkowski, M. C., E. Puchelle, G. Beck, J. Jacquot, and C. Hannoun. 1986. Adherence of type I Streptococcus pneumoniae to tracheal epithelium of mice infected with influenza A/PR8 virus. Am. Rev. Respir. Dis. 134:1040– 1044. 33. Read, R., P. Roberts, N. Munro, A. Rutman, A. Hastie, T. Shryock, R. Hall, W. McDonald-Gibson, V. Lund, G. Taylor, P. J. Cole, and R. Wilson. 1992. Effect of Pseudomonas aeruginosa rhamnolipids on mucociliary transport and ciliary beating. J. Appl. Physiol. 72:2271–2277. 34. Rubbins, J. B., P. G. Duane, D. Charboneau, and E. N. Janoff. 1992. Toxicity of pneumolysin to pulmonary endothelial cells in vitro. Infect. Immun. 60: 1740–1746. 35. Rubins, J. B., P. G. Duane, D. Clawson, D. Charboneau, J. Young, and D. E. Niewoehner. 1993. Toxicity of pneumolysin to pulmonary alveolar epithelial cells. Infect. Immun. 61:1352–1358. 36. Rutland, J., and P. J. Cole. 1980. Non-invasive sampling of nasal cilia for measurement of beat frequency and study of ultrastructure. Lancet ii:564– 565. 37. Selinger, D. S., and W. P. Reed. 1979. Pneumococcal adherence to human epithelial cells. Infect. Immun. 23:545–548. 38. Sheehan, J. K., D. J. Thorton, M. Sommerville, and I. Carlstedt. 1991. The structure and heterogeneity of respiratory mucus glycoproteins. Am. Rev. Respir. Dis. 144:S4–S9. 39. Simpson, S. Q., R. Singh, and D. E. Bice. 1994. Heat-killed pneumococci and pneumococcal capsular polysaccharides stimulate tumor necrosis factor-a production by murine macrophages. Am. J. Respir. Cell. Mol. Biol. 10:284– 289. 40. Smythe, C. J., and J. L. Duncan. 1978. Thiol activated (oxygen labile) cytolysins, p. 129–183. In J. Jeljaszewicz and T. Waldstrom (ed.), Bacterial toxins and cell membranes. Academic Press, London. 41. Somerville, M., A. Rutman, R. Wilson, P. J. Cole, and P. S. Richardson. 1991. Stimulation of secretion into human and feline airways by Pseudomonas aeruginosa proteases. J. Appl. Physiol. 70:2259–2267. 42. Steinfort, C., R. Wilson, T. Mitchell, C. Feldman, A. Rutman, H. Todd, D. Sykes, J. Walker, K. Saunders, P. W. Andrews, G. J. Boulnois, and P. J. Cole. 1989. Effect of Streptococcus pneumoniae on human respiratory epithelium in vitro. Infect. Immun. 57:2006–2013. 43. Wanner, A., D. Maurer, W. M. Abraham, Z. Szepfalusi, and M. Sielczak. 1983. Effects of chemical mediators of anaphylaxis on ciliary function. J. Allergy Clin. Immunol. 72:663–667. 44. Wilson, R., T. Pitt, G. Taylor, D. Watson, J. MacDermot, D. Sykes, D. Roberts, and P. J. Cole. 1987. Pyocyanin and 1-hydroxyphenazine produced by Pseudomonas aeruginosa inhibit the beating of human respiratory cilia in vitro. J. Clin. Invest. 79:221–229. 45. Wilson, R., D. A. Sykes, D. Watson, A. Rutman, G. W. Taylor, and P. J. Cole. 1988. Measurement of Pseudomonas aeruginosa phenazine pigments in sputum and assessment of their contribution to sputum sol toxicity for respiratory epithelium. Infect. Immun. 56:2515–2517. 46. Woodhead, M. A. 1990. BTS study of severe community acquired pneumonia in the intensive care unit. Thorax 45:302.