Report

CD95/CD95 Ligand Interactions on Epithelial Cells in Host Defense to Pseudomonas aeruginosa

See allHide authors and affiliations

Science  20 Oct 2000:
Vol. 290, Issue 5491, pp. 527-530
DOI: 10.1126/science.290.5491.527

Abstract

Pseudomonas aeruginosa causes severe infections, particularly of the lung, that are life threatening. Here, we show thatP. aeruginosa infection induces apoptosis of lung epithelial cells by activation of the endogenous CD95/CD95 ligand system. Deficiency of CD95 or CD95 ligand on epithelial cells prevented apoptosis of lung epithelial cells in vivo as well as in vitro. The importance of CD95/CD95 ligand–mediated lung epithelial cell apoptosis was demonstrated by the rapid development of sepsis in CD95- or CD95 ligand–deficient mice, but not in normal mice, afterP. aeruginosa infection.

Several pathogens, including some bacteria, viruses, and parasites, are able to trigger apoptosis of mammalian host cells (1, 2). A paradigm for bacteria-induced apoptosis is Shigella flexneri, which, upon invasion, induces apoptosis by the activation of caspases (in particular, caspase 1) (3). Some other bacteria also induce apoptosis, includingSalmonella typhimurium, some enteropathogenicEscherichia coli, Listeria monocytogenes, andStaphylococcus aureus (4–7). However, the molecular mechanisms involved and their relevance to host pathogen interactions are still largely unknown.

P. aeruginosa is clinically one of the most important classes of bacteria, because it induces pneumonia and sepsis in cystic fibrosis or immunocompromised patients. To gain insight into the molecular mechanisms of P. aeruginosa interaction with mammalian cells, we investigated whether the induction of apoptosis contributes to the previously described cytotoxicity of P. aeruginosa (8). The results (Fig. 1) reveal marked apoptosis of human Chang conjunctiva cells or murine ex vivo lung fibroblasts within 30 min of infection with P. aeruginosa strain 762 (9).

Figure 1

P. aeruginosa induces apoptosis of epithelial cells by the CD95/CD95 ligand system. Human Chang conjunctiva cells or normal murine ex vivo lung fibroblasts undergo apoptosis after infection with P. aeruginosa strain 762 for 30 min (9). Genetic deficiency of CD95 or CD95 ligand in lpr or gld fibroblasts, as well as neutralization of CD95 ligand on Chang epithelial cells by treatment with soluble CD95-Fc protein (5 μg/ml), prevented P. aeruginosa–triggered apoptosis. Treatment with control immunoglobulin G (IgG)–Fc did not affect P. aeruginosa–induced apoptosis. Shown are the mean ± SD of three independent experiments (*P < 0.05,t test).

To look for possible mechanisms by which apoptosis might be initiated during P. aeruginosa infection, we examined the involvement of the CD95/CD95 ligand system, one of the most important systems by which apoptosis is triggered in a variety of mammalian cells (10). Neutralization of the CD95/CD95 ligand system by the addition of CD95-Fc protein (11) to human epithelial cells or by the genetic deficiency of either CD95 or CD95 ligand in ex vivo fibroblasts from lpr (lymphoproliferation) or gld(generalized lymphoproliferative disease) C3H mice (12) almost completely blocked P. aeruginosa–mediated apoptosis (Fig. 1). In contrast, fibroblasts from (isogenic) normal C3H mice were highly sensitive to the induction of apoptosis, indicating a crucial role for the CD95/CD95 ligand system in the initiation of cellular apoptosis by P. aeruginosa (13).

To analyze the importance of P. aeruginosa–triggered, CD95-mediated apoptosis in vivo, isogenic lpr,gld, or wild-type mice were intranasally infected (14) with P. aeruginosa strain 762, and the induction of apoptosis in lung epithelial cells was determined by TUNEL (terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick-end labeling) 6 hours after infection (15). Infection of normal mice resulted in the rapid (within 3 to 4 hours) induction of apoptosis in epithelial cells of small bronchi (Fig. 2A). In contrast, lungs from infected lpr or gld mice did not display apoptotic cells.

Figure 2

Bacteria-mediated apoptosis of epithelial cells protects animals from P. aeruginosa–induced sepsis and death. (A) Infection (14) of normal C3H mice with P. aeruginosa strain 762 results in the apoptosis of lung epithelial cells, particular in small bronchi. In contrast, lung epithelial cells of isogenic gld or lpr mice deficient for CD95 or CD95 ligand are resistant to P. aeruginosa infection. Shown are typical TUNEL assays (15) of the lungs; each assay is from 12 mice. Mice were infected for 6 hours. (B) P. aeruginosa strain 762 induces sepsis (16) and death of lpr andgld mice but not of normal control mice. Twelve mice of each set were infected (14) and continuously observed for 7 days after infection. To determine bacterial sepsis, we killed mice after 36 hours, or if they died earlier, they were examined immediately after death. Shown is the ratio of P. aeruginosa CFUs grown from the spleen to the number of CFUs grown from the lung. This minimizes variations due to the infection procedure. Error bars indicate SD.

To address the question of whether CD95-mediated apoptosis of lung epithelial cells induced by P. aeruginosa functions as part of the immune defense or as part of the bacterial attack against the host, we infected normal, lpr, or gld mice with P. aeruginosa (14) and monitored the development of P. aeruginosa colonization of the spleen (16), indicative of sepsis and, finally, death. None of the wild-type mice showed substantial numbers of P. aeruginosa in the spleen upon infection, and only 10% of these animals died in a 60-day observation period after infection (Fig. 2B). In contrast, all lpr and gld mice developed sepsis, and 100% of the animals died within 96 hours of infection.

Thus, the high susceptibility of lpr or gldmice to P. aeruginosa infections was critically associated with the failure of lung epithelial cells to be activated by CD95 and to undergo apoptosis upon infection. However, the high susceptibility of these mice might also have been caused by a malfunction of the immune system, thereby permitting the bacteria to induce generalized infection. To differentiate between these possibilities, we ablated the immune system of normal, gld, or lpr mice by irradiation and reconstituted these animals with bone marrow cells (BMCs) from syngenic lpr or gld mice, in the case of irradiated wild-type mice, or with BMCs from syngenic normal mice, in the case of irradiated lpr and gld mice (17); the success of the transplant was confirmed by the induction or absence, respectively, of activation-induced cell death (AICD) (18) in ex vivo splenic lymphoblasts (19) (Fig. 3A). Mice were intranasally infected (14) with P. aeruginosa and monitored for the development of sepsis (Fig. 3B). Normal mice transplanted with BMCs from lpr orgld mice were as resistant to infection as untransplanted normal mice were (Fig. 3B). Likewise, lpr or gldmice transplanted with normal BMCs were as susceptible to infection with P. aeruginosa as untransplanted lpr orgld mice were and died of sepsis. Consistent with this, apoptosis of lung epithelial cells was only observed in normal mice, regardless of whether they had been transplanted with BMCs fromlpr or gld mice or had been left untransplanted (Fig. 3C) (14). lpr or gldmice remained resistant to the induction of apoptosis in lung epithelial cells, despite transplantation with normal BMCs. This excludes a predominant role for the immune system and suggests a pivotal function of lung epithelial cells in the primary response toP. aeruginosa in the lung.

Figure 3

CD95-mediated apoptosis of lung epithelial cells, but not of immune cells, is required for the host response against P. aeruginosa. (A) Lymphoblasts (19) from animals transplanted with BMCs from normal mice show CD95-dependent AICD upon activation with antibodies to CD3, 2C11 (0.5 μg/ml). Transplantation of BMCs from lpr orgld mice abolishes AICD, indicating the success of the transplant. Data for irradiated animals are labeled with an asterisk; the transplanted BMCs are indicated after the slash. Solid bars represent untreated lymphoblasts; open bars represent cells stimulated for 12 hours with antibody to CD3. Shown is the mean ± SD for each set of 24 mice (*P < 0.05, ttest). Each individual animal was successfully transplanted. (B) Mice were infected with P. aeruginosa strain 762, and the survival of animals was determined (12 animals in each set). The mice responded independently of their reconstituted immune system: lpr or gld mice transplanted with normal BMCs died at an equivalent rate to the untransplanted controls. In contrast, normal mice transplanted with BMCs from either lpror gld mice were still resistant to P. aeruginosainfection. Transplanted BMCs are indicated as in (A). (C) Representative TUNEL assays (15) of lungs from 12 animals each killed 6 hours after infection show that lung epithelial cells in normal mice transplanted with lpr or gld BMCs still undergo apoptosis. lpr or gld mice transplanted with normal BMCs behave as the untransplanted controls and do not show apoptotic lung epithelial cells after infection.

Several studies have demonstrated the cytotoxicity of P. aeruginosa toxins (20); however, the mechanisms of cell death triggered by these factors are unknown. Our work suggests that the up-regulation of the CD95/CD95 ligand system constitutes one of the major mechanisms mediating this cytotoxicity. It could be speculated that bacterial polysaccharides (21) induce a translocation of secretory vesicles, which have been previously shown to store CD95 and CD95 ligand (22, 23), resulting in increased CD95/CD95 ligand surface expression and the induction of apoptosis.

Activation of CD95 in lung epithelial cells may protect the host from P. aeruginosa infection by at least two mechanisms: First, apoptosis of infected cells results in the targeting of P. aeruginosa into apoptotic bodies, which are rapidly internalized by other cells. Fusion of those phagosomes with lysosomes then results in digestion of the bacteria. In contrast, internalization of P. aeruginosa without apoptosis of the host cell (as inlpr or gld mice) might permit the bacterium to block maturation of the phagosome, promote intracellular survival and even growth of the bacterium before transcytosis, and protect bacteria from the host immune system. Second, P. aeruginosa–mediated activation of CD95 might stimulate nuclear factor κB (24), c-Jun NH2-terminal kinase (25), GADD153 (26), and phospholipase A2 (27) or may interfere with the functions of growth factor receptors (28), resulting in the secretion of defensins and/or cytokines from epithelial cells. CD95-triggered secretion of defensins and/or cytokines into bronchi may then kill extracellular bacteria. Those factors may also prevent the penetration of bacteria through the epithelial cell barrier, even in situations where excessive apoptosis of epithelial cells occurs.

The concerted action of CD95-dependent lysosomal degradation ofP. aeruginosa in apoptotic bodies and the secreted bacteriotoxic substances may impede intra- and extracellular bacteria. This would explain the pivotal role observed for the CD95/CD95 ligand system in host defense against infection with P. aeruginosa.

  • * Present address: Department of Immunology, St. Jude Children's Research Hospital, 332 North Lauderdale Street, Memphis, TN 38105, USA.

  • To whom correspondence should be addressed. E-mail: erich.gulbins{at}stjude.org

REFERENCES AND NOTES

View Abstract

Navigate This Article