Microbiol. Immunol. Vol. 33 (2), 111-121, 1989 Electron Microscopic Observations on Ciliated Epithelium of Tracheal Organ Cultures Infected with Bordetella bronchiseptica Kachiko SEKIYA,*,1 Yutaka FUTAESAKU,2 and Yasukiyo NAKASE1 1 Department of Microbiology, School of Pharmaceutical Sciences, Kitasato University, Minato-ku, Tokyo 108, and 2Department of Ultrastructure Research, School of Hygienic Sciences, Kitasato University, Sagamihara, Kanagawa 228 (Accepted for publication, December 12, 1988) ôgh ô Abstract ôgs ô Using mouse tracheal organ cultures, the pathogenic effect of ônh ôbordetella bronchiseptica ôns ô to epithelial cells was studied by electron microscopy. The ultra structureof epithelial cells in uninfected tracheal rings was preserved well for longer than 3 days. In mouse tracheal rings infected with graded doses (3 ~10 ôuh ô5 ôus ô to 10 ôuh ô7 ôus ô CFU/ml) of phase I ônh ôb. bronchiseptica ôns ô, the colonization in the interciliary spaces of ciliated epithelial cells was observed after a 20-hr infection period. The infected tra chealrings showed swelling of nonciliated cells as well as ciliated cells, rupture of cell membrane of cilia, swelling and disappearance of cilia, and atrophic cytomorphosis of epithelial cells. The severity of these changes occurred depending on the infec tiondoses. These changes were essentially similar to those observed previously in the tracheal epithelia of the ônh ôb. bronchiseptica ôns ô-infected mice. The usefulness of this ônh ôin vitro ôns ô model was suggested for studying the pathogenesis of ônh ôbordetella ôns ô infection. The specific attachment of Bordetella organisms to cilia of respiratory epithelial cells is shown, in a variety of animal models, to be the initial step in the Bordetella infections (1-3, 10-13, 16-18). Subsequent colonization and destruction of the ciliated cells, in the establishment of infections, have been demonstrated in the susceptible animal hosts. The tracheal cytotoxin (TCT) from B. pertussis has been shown to induce a dose-dependent inhibition of DNA synthesis, disappearance of cilia, and ciliostasis in cultured hamster tracheal cells (6-8). However, the mecha nismof the resulting ciliostasis and destruction of epithelial cells by the multiplying bacteria is poorly understood. It has been reported that in vitro studies with cultured mouse tracheas are inappropriate for studying the respiratory epithelial damage by the Bordetella infection, by reason of the fact that B. pertussis binds specifically to the ciliated cells, but does not reproduce subsequent destruction of these cells (6, 13). Our previous observation revealed that the multiplying bacteria in the tracheal epithelia of B. bronchiseptica-infected mice affected selectively the epithelial cell membranes, resulting in destruction of tracheal epithelial cells (16). These findings were similar to the histological lesion of the trachea from patients of whooping cough (9). In the present study, using cultured tracheas from mice, an attempt 111
112 K. SEKIYA ET AL was made to analyze ultrastructurally, by an electron microscope, whether B. bronchiseptica could induce in vitro to epithelial cells such a specific pathological change, relevant to our understanding of the initial step in the pathogenesis of Bordetella in vivo infection. MATERIALS AND METHODS Mice. Specific pathogen-free female ICR (CLEA Japan Inc., Tokyo) mice were used at 7 weeks of age. Bacteria. A phase I B. bronchiseptica strain H-16 (15) was used. Tracheal ring culture. Tracheal rings were cultured by applying a modified method described by Opremcak and Rheins (13). Tracheas were removed from mice, cut into rings of 0.5 to 1 mm width, and then incubated in 3.5 X 10 mm petri dishes with Hanks medium at 37 C in 5 % C02. Infection. Cultures of B. bronchiseptica grown on Bordet-Gengou medium for 20 hr at 37 C, were suspended in Hanks medium (ph 7.4) to a density of 1010 cells per ml, and 10-fold dilutions were made in the same medium to prepare the bacterial suspensions of three graded doses. The ciliary beating on incubated tracheal rings was confirmed by the phase-contrast microscope, and the rings were removed one by one into each well of a 96-holed plane-bottom microplate which contained 0.15 ml of each. bacterial suspension. The incubation at 37 C in 5 % CO2 was continued Fig. 1. Section of uninfected mouse trachea after 3 days in org an culture. Note flat luminal faces of ciliated cells (C) and nonciliated cells (N).
TRACHEAL EPITHELIUM AND B. BRONCHISEPTICA INFECTION 113 Fig. 2. Sections of ciliated and nonciliated epithelial cells at 20 hr after infection with bacterial suspension of 3 x 105 CFU/ml (2a) or 3 x 107 CFU/ml (2b). Note bacteria (arrows) infiltrated into interciliary spaces of ciliated cells (C), and swelling of nonciliated cells (N) and ciliated cells (2a and 2b). Lipid droplets (arrowheads) appeared in swelled nonciliated cell (2b).
114 K. SEKIYA ET AtL Fig. 3. Section of epithelial cells at 20 hr after infection with 3 x 106 CFU/ml. Note cilia swelled like a balloon. for varying times. The uninfected control rings were incubated at the same condition in equal volumes of Hanks medium alone. Electron microscopy. After the incubation, the tracheal rings were fixed doubly with 2.5% glutaraldehyde in 0.05 m phosphate buffer for 1.5 hr at 4 C, and then I % OSO4 in the same buffer for 1 hr at 4 C. After being dehydrated with a series
TRACHEAL EPITHELIUM AND B. BRONCHISEPTICA INFECTION 115 Fig. 4. Cross and longitudinal sections of ciliated cells at 20 hr after infection with 3 x 105 CFU/ml. Note bacteria in the interciliary spaces, and cilia shared one cell membrane (arrows).
116 K. SEKIYA ET AL Fig. 5. Sections of cilia at 20 hr after infection with 3 x 105 CFU/ml (5a), 3 x 106 CFU/ml (5b and 5c). Note rupture (bars in Fig. 5b) and bare ciliary structures without cell membrane (arrows in Fig. 5c).
TRACHEAL EPITHELIUM AND B. BRONCHISEPTICA INFECTION 117 Fig. 6. Sections of cilia at 20 hr after infection with 3 X 107 CFU/ml. Note single cuboidal (6a) and squamous epithelia (6b) with lipid droplets (arrows) and without cilia.
118 K. SEKIYA ET AL of graded alcohols, they were embedded in Epon 812. The ultrathin sections were stained with uranyl acetate and lead citrate, and then examined by the trans missionelectron microscope, Hitachi H-500, at 75kV. Viable count. The viable count of the bacterial suspension was done by the method previously described (15). RESULTS The morphological changes of the ciliated epithelium of the tracheal rings infected in vitro with graded doses of B. bronchiseptica were compared with those of uninfected controls, after a 20-hr infection period. In the uninfected tracheal rings the ultrastructure of the ciliated epithelium was well preserved for 3 days of the observation period and no remarkable change was observed (Fig. 1). The ciliary beating remained normal for longer than 2 weeks. In tracheal rings infected with the lower (3 ~105 CFU/ml) (Fig. 2a) and middle (3 ~106CFU/ml) (data not shown) doses of B. bronchiseptica, many growing bacteria were seen, in close proximity to cilia at the interciliary spaces, suggesting colonization. In those infected with the higher dose (3 ~107CFU/ml) the dis appearanceof cilia from ciliated cells was remarkable (Fig. 6), and there were many bacteria at the interciliary spaces of a few intact ciliated cells which remained (Fig. 2b). There was no evidence of intracellular invasion. Both ciliated and nonciliated cells extruded out into the luminal spaces (Fig. 2a), occasionally forming a balloon-like structure (Fig. 2b). In those with the higher dose, many lipid droplets were observed in the cytoplasm of the extruded cells (Fig. 2b). The cilia also showed noticeable swelling like a balloon (Fig. 3). In those infected with the lower dose, a small batch of cilia fusing to each other with their swollen cell membranes was remarkably observed (Fig. 4, a and b). The cell membrane of cilia became lean, and ruptured in certain places (Fig. 5a). In those with the middle dose, the rupture was more remarkable (Fig. 5b) and the membrane in certain Table 1. Histopathological findings on organ-cultured trachea infected with B. bronchiseptica a) Histopathological changes: grade and positive rate o n magnification ~5,000. ++, severe, >80%; +, marked, 50%-80%; }, infrequent, <50%; -, none. b) Questionable because cilia had disappeared.
TRACHEAL EPITHELIUM AND B. BRONCHISEPTICA INFECTION 119 places had completely disappeared (Fig. 5c). The disappearance of cilia, and atrophic cytomorphosis were locally observed only in the rings infected with the higher dose (Fig. 6). Many lipid droplets were also observed in these cells. The cell layers seemed like a simple cuboidal (Fig. 6a) or squamous (Fig. 6b) epithelium. These alterations became exceedingly severe thereafter up to 48hr (data not shown). Table 1 summarizes the results on 20-hr postinfection. DISCUSSION In the experimental bordetellosis including B. pertussis in various animal models, it has been shown that bacteria superficially attach to the cilia of tracheal epithelium (2, 3, 7, 9-12, 15), and subsequently induce the selective destruction of ciliated cells. In a previous in vivo study (16), we demonstrated in mice that a rupture of the cell membrane of cilia, resulting in destruction of tracheal epithelia, is an initially important step of the pathogenesis of B. bronchiseptica subsequently after the colo nization. In the present in vitro study, we have observed the ciliated epithelium of mouse tracheal cultures infected with B. bronchiseptica. The observation was initiated at 20hr after infection; from the results of preliminary experiments, we have found this to be appropriate because it facilitates various histopathological findings depend ingon the infection doses and also avoids artifacts as much as possible. We could thereby find the characteristic changes of epithelia, similar to those observed in the infected mice (16), such as colonization in the interciliary spaces of ciliated cells, alteration and rupture of the cell membrane of cilia, extrusion of epithelial cells, swelling and disappearance of cilia. It is noteworthy that the characteristic lesions similar to those observed in the trachea of whooping cough in the earlier stages, such as the existence of many bacteria between cilia, the lateral spreading or mush roomingof the cilia, and the cilia reduced short stubs or entirely wanting (9), were also reproduced in B. bronchiseptica-infected mouse tracheas in vitro. In the present study, these changes occurred reproducibly within the first 20hr of incubation. Goldman et al (6-8) have demonstrated that TCT induces the noticeable lesions such as extrusion of ciliated cells, in hamster tracheal cultures, but not in mouse tracheas, after a longer incubation period than 60hr. Whether these discrepancies are due to different materials, to different animal species, or to different bacterial strains, remains to be determined. The present results indicate that mouse tracheal cultures have adequacies for the in vitro studies on Bordetella infection as far as using B. bronchiseptica. In this observation, we found two different patterns of cell-membrane damages in epithelial cells-spreading and rupturing. The former seems to result in an aggregation of cilia as observed previously by SEM (16), and in the extrusion of both ciliated and nonciliated cells. The purified TCT, a soluble peptidoglycan fragment common to all Bordetella species (4, 5, 14), induces selectively ciliated-cell damages but it does not act on nonciliated cells (6-8). However, in the previous in vivo study (16), we noticed that the destruction occurred not only in ciliated cells but
120 K. SEKIYA ET AL also in nonciliated cells, suggesting the possibility of the participation of different toxins from TCT, capable of acting on both ciliated and nonciliated cells. We are currently attempting to isolate the factors responsible for the epithelial cell-damage, to clarify the mechanism of the pathogenesis in Bordetella infection using the present in vitro models. Compared with the previous finding of in vivo infection (16), the extrusion of epithelial cells was much more striking in the infected cultured tracheal rings, though the rupture of the cell membrane of cilia was almost similar. This difference in response may be due to either the participation of certain host-defence factors or to the host-parasite relationships. REFERENCES 1) Bakaletz, L.O., and Rheins, M.S. 1985. A whole-organ perfusion model of Bordetella pertussis adherence to mouse tracheal epithelium. In Vitro Cell. Dev. Biol. 21: 314-320. 2) Bemis, D.A., and Kennedy, J.R. 1981. An improved system for studying the effect of Bordetella bronchiseptica the ciliary activity of canine tracheal epithelial cellṣ J. Infect. Dis. 144: 349-357. 3) Collier, A.M., Peterson, L.P., and Baseman, J.B. 1977. Pathogenesis of infection with Bordetella pertussis in hamster tracheal organ culture. J. Infect. Dis. 136 (Suppl.): 196-203. 4) Cookson, B.T., and Goldman, W.E. 1987. Tracheal cytotoxin: a conserved virulence determinant of all Bordetella species. J. Cell Biochem. 11B (Suppl.): 124. 5) Gentry-Weeks, C.R., Cookson, B.T., Goldman, W.E., Rimler, R.B., Porter, S.B., and Curtiss III, R. 1988. Dermonecrotic toxin and tracheal cytotoxin, putative virulence factors of Bordetella avium. Infect. Immun. 56: 1698-1707. 6) Goldman, W.E. 1988. Tracheal cytotoxin of Bordetella pertussis. Pathogenesis and immunity in pertussis, p. 231-246. In Wardlaw, A.C., and Parton, R. (eds), John Wiley and Sons Ltd., Chi chester. 7) Goldman, W.E., Klapper, D.G., and Basman, J.B. 1982. Detection, isolation, and analysis of a released Bordetella pertussis product toxin to cultured tracheal cellṣ Infect. Immun. 36: 782-794. 8) Goldman, W.E., and Herwaldt, L.A. 1985. Bordetella pertussis tracheal cytotoxin. Dev. Biol. Stand. 61: 103-111. 9) J Mallory, F.B., and Hornor, A.A. 1912. Pertussis: the histological lesion in the respiratory. Med. Res. 27: 115-123 tract.. 10) Matsuyama, T., and Takino, T. 1980. Scanning electronmicroscopic studies of Bordetella bronchi septicaon the rabbit tracheal mucosạ J. Med. Microbiol. 13: 159-161. 11) M hamster use, K.E., Collier, A.M., and Baseman, J.B. 1977. Scanning electron microscopic study of tracheal organ cultures infected with Bordetella pertussis. J. Infect. Dis. 136: 768-777. 12) Muse, K.E., Findley, D., Allen, L., and Collier, A.M. 1978. In vitro model of Bordetella pertussis infection: (eds), pathogenic and microbicidal interactionș p. 41-50. In Manclark, C.R., and Hill, J.C. International symposium on pertussis, U.S. Department of Health, Education and Welfare, Washington, D.C. 13) Opremcak, L.B., and Rheinș M.S. 1983. Scanning electron microscopy of mouse ciliated oviduct and tracheal epithelium infected in vitro with Bordetella pertussis. Can. J. Microbiol. 29: 415-420. R 14) osenthal, R.S., Nogami fragment, W., Cookson, B.T., Goldman, W.E., and Folkening, of soluble peptidoglycan released from growing Bordetella pertussis W.J. 1987. Major Infecṭ is tracheal cytotoxiṇ Immun. 55: 2117-2120. 15) Sekiya, K., Kawahira, M., and Nakase, Y. 1983. Protection against experimental Bordetella bronchiseptica 598-603 infection in mice by active immunization with killed v. accine. Infect. Immun. 41: 16) Sekiya, K., Futaesaku, U., and Nakase, Y. 1988. Electron microscopic observations on tracheal
TRACHEAL EPITHELIUM AND B. BRONCHISEPTICA INFECTION 121 epithelia of mice infected with Bordetella bronchiseptica. Microbiol. Immunol. 32: 461-472. 17) Tumanen, E.I., and Hendley, J.O. 1983. Adherence of Bordetella pertussis to human respiratory epithelial cells. J. Infect. Dis. 148: 125-130. 18) Yokomizo, Y., and Shimizu, T. 1979. Adherence of Bordetella bronchiseptica to swine nasal epithelial cells and its possible role in virulence. Res. Vet. Sci. 27: 15-21. (Received for publication, September 7, 1988)