Intercellular Spreading of Porphyromonas gingivalis Infection in Primary Gingival Epithelial Cells

Porphyromonas gingivalis, an important periodontal pathogen,is an effective colonizer of oral tissues. The organism successfullyinvades, multiplies in, and survives for extended periods inprimary gingival epithelial cells (GECs). It is unknown whetherP. gingivalis resides in the cytoplasm of infected cells throughoutthe infection or can spread to adjacent cells over time. Wedeveloped a technique based on flow cytofluorometry and fluorescencemicroscopy to study propagation of the organism at differentstages of infection of GECs. Results showed that P. gingivalisspreads cell to cell and that the amount of spreading increasesgradually over time. There was a very low level of propagationof bacteria to uninfected cells early in the infection (3 hpostinfection), but there were 20-fold and 45-fold increasesin the propagation rate after 24 h and 48 h, respectively, ofinfection. Immunofluorescence microscopy of infected cells suggestedthat intercellular translocation of P. gingivalis may be mediatedthrough actin-based membrane protrusions, bypassing the needfor release of bacteria into extracellular medium. Consistentwith these observations, cytochalasin D treatment of infectedcells resulted in significant inhibition of bacterial spreading.This study shows for the first time that P. gingivalis disseminatesfrom cell to cell without passing through the extracellularspace. This mechanism of spreading may allow P. gingivalis tocolonize oral tissues without exposure to the humoral immuneresponse.

INTRODUCTION

Top

Introduction

The gram-negative bacterium Porphyromonas gingivalis is an importantcomponent of oral microbiota and is widely acknowledged to bea primary etiological agent of chronic periodontal disease inhumans (34, 43, 44). This successful colonizer of oral tissuesinvades and survives in primary gingival epithelial cells (GECs)(26, 27, 33, 45). Entry of P. gingivalis into GECs, which arenonprofessional phagocytic cells, is rapid and requires thebacteria and host cells to be metabolically active. It has beendemonstrated that adherence of P. gingivalis to GECs is multimodal,and yet the trigger event for subsequent invasion is mediatedprimarily by the binding of major fimbriae to ß1 integrinreceptors (46). Invasion of P. gingivalis is accompanied bythe phosphorylation and activation of putative integrin-signalingand structural proteins, i.e., FAK (focal adhesion kinase) andpaxillin, and by significant remodeling of the actin cytoskeleton,suggesting that rearrangements of the host cell signaling/cytoskeletonproteins permit bacterial entry into GECs (46, 47).

P. gingivalis is nonmotile while in the extracellular space,invades GECs in high numbers, and is then localized in the cytoplasmwithout being constrained by membranous vacuoles (5). The organismis capable of intracellular replication and remains viable forextended periods in GECs cultured in vitro. Despite the burdenof large numbers of intracellular bacteria, infected GECs donot die and are resistant to staurosporine-induced apoptosis.The phosphatidylinositol 3-kinase (PI3K)/Akt (protein kinaseB) pathway appears to be one of several signaling pathways thatcould promote the survival of GECs during P. gingivalis infection(45). Evidently, transcriptomic analysis of gingival cells (humanimmortalized gingival keratinocytes) shows that a variety ofantiapoptotic pathways are also activated by the organism (16).Moreover, recent proteomics and genomics studies have revealedthat P. gingivalis infection induces regulation of a large numberof distinctive P. gingivalis proteins and genes that could beimportant for the adaptation and survival of the microorganismin epithelial cells (19, 35, 48). Hence, the interaction betweenbacteria and GECs leads to the modulation of a wide array ofbiological functions in both host and pathogen, most of whichremain to be characterized.

Several invasive bacterial species have been shown to rely onthe host cell actin cytoskeleton in order to complete theirinfection cycles (2, 6, 10, 29, 42). For example, Shigella flexneriand Listeria monocytogenes multiply in the host cell cytoplasm,move intracellularly, and spread to adjacent epithelial cells.Shigella flexneri and Listeria monocytogenes cells move throughthe cytosol of the infected cells and spread to neighboringcells by the assembly of a propulsive actin tail at one endof the bacteria (38, 40). Rickettsia spp., which are gram negativeand obligate intracellular bacteria, induce the formation ofactin-based, long, twisted protrusions bundled around the microorganisms'outer membrane that allow the organisms to exit the infectedcell and enter adjoining uninfected cells (14). Exploitationof the host cell actin cytoskeleton thus appears to be a commonmechanism for the spread of invasive bacteria (6, 15, 39, 42).Direct cell-to-cell movement of bacteria has been postulatedto facilitate local and systemic spread and avoidance of thehost humoral immune response (18).

Molecular and cellular changes associated with P. gingivalisinvasion have been subject to extensive studies during the lastdecade (17, 29). Surprisingly, it is unknown whether the pathogenresides in the cytoplasm of the same infected cell throughoutthe infection, whether it needs to be released into the extracellularspace in order to initiate infection of new host cells, or whetherit can spread directly to adjacent cells over time. Furthermore,the consequences of prolonged infection by P. gingivalis inGECs, in the absence of host cell death, for the disseminationof the organism within deeper layers of the oral epitheliumand colonization of adjacent tissues remain to be investigated.

The present study represents the first report that P. gingivaliscan spread from cell to cell without passing through the extracellularspace, and that spreading is dependent on the host cell actincytoskeleton. Little cell-to-cell spreading can be detectedat early times of infection, but the spreading process gainsmomentum after 24 h of infection. These findings suggest a novelmechanism for P. gingivalis colonization of host tissues andevasion of the humoral immune response.

MATERIALS AND METHODS

Top

Materials and Methods

Bacteria and culture conditions. P. gingivalis ATCC 33277 was cultured anaerobically for 24 hat 37°C in Trypticase soy broth supplemented with yeastextract (1 mg/ml), hemin (5 µg/ml), and menadione (1 µg/ml).Bacteria were harvested by centrifugation at 6,000 x g and 4°Cfor 10 min, washed, and resuspended in Dulbecco's phosphate-bufferedsaline (PBS; Sigma, St. Louis, MO), pH 7.3. The number of bacteriawas determined using a Klett-Summerson photometer (26).

Culture of GECs. Primary cultures of GECs were generated as described previously(28). Briefly, healthy gingival tissue was obtained after oralsurgery, and surface epithelium was separated by overnight incubationwith 0.4% dispase. Cells were cultured as monolayers in serum-freekeratinocyte growth medium (KGM) (Cambrex, New Jersey) at 37°Cin 5% CO₂. GECs were used for experimentation at $~$ 75% confluenceand reacted with bacterial cells or other test reagents in KGM.

Fluorescent labeling and coculturing of GECs for examination of bacterial spreading. GECs, grown in six-well plates, were labeled with CellTrackerBlue CMAC (7-amino-4-chloromethylcoumarin) (Molecular Probes,Inc., Eugene, OR). CellTracker dyes are fluorescent chloromethylderivatives that freely diffuse through the membrane of livecells. Once inside the cells, these mildly thiol-reactive probescan react with thiols on proteins and peptides and thus becomemembrane impermeant. The labeling allows subsequent fixationand probing with fluorescence-labeled antibodies (12). Briefly,the cells are incubated in a warm, fresh KGM medium containing2.5 µM CellTracker Blue for 30 min and then rinsed threetimes with Dulbecco's PBS (Sigma). The cells were then incubatedwith fresh KGM medium, and the monolayers were infected withP. gingivalis ATCC 33277 at a multiplicity of infection of 100for the indicated time periods at 37°C in 5% CO₂. At theend of each experimental condition, the cells were dissociatedby mild trypsinization using 0.05% trypsin-0.53 mM EDTA (GibcoBRL, California) and collected by centrifugation. New sets ofuninfected GECs were similarly labeled with CellTracker GreenCMFDA (5-chloromethylfluorescein diacetate) (Molecular Probes),mixed with approximately equal numbers of previously collectedinfected blue-labeled cells, and coplated onto the new plates.The growth medium was supplemented with metronidazole (200 µg/ml)and gentamicin (300 µg/ml) for each condition. Uninfectedblue-labeled cells, uninfected green-labeled cells, uninfectedunlabeled cells, and the infected blue-labeled cells were usedseparately as controls. Cytochalasin D (1 µg/ml) (Sigma)was added to control cocultured samples and maintained throughoutthe coculturing period.

Quantitation of bacterial spreading by flow cytofluorimetry. Following coculturing of the initially infected and uninfectedcells for the indicated time periods, samples were collectedand centrifuged. Pellets were fixed in 10% neutral bufferedformalin and permeabilized using 0.1% Triton X-100. Sampleswere then incubated with anti-P. gingivalis ATCC 33277 rabbitpolyclonal antibody (1:2,000) followed by red fluorescence (Alexa-Fluor633)-conjugated goat anti-rabbit immunoglobulin G (IgG) (heavyplus light chains [H+L]) highly cross-absorbed antibody (1:500)(Molecular Probes). Samples were washed in Dulbecco's PBS twice.Measurements were taken and analyzed by flow cytometry withexcitation at 350 nm and 450 nm using a band-pass filter forBlue CMAC detection, excitation at 488 nm and 525 nm using aband-pass filter for Green CMFDA detection, and excitation at638 nm using a 645-nm long-pass filter for Alexa-Fluor 633 (redfluorescence) detection. For each dye, appropriate electroniccompensation of the instrument was performed to avoid overlappingof the three emission spectra. Samples were calibrated for lightscattering and fluorescence properties using controls (uninfectedblue- or green-labeled cells or uninfected unlabeled cells treatedwith or without anti-P. gingivalis plus red fluorescence antibodies).Samples were then analyzed by setting gates for forward scatter,side scatter, and red fluorescence staining. The total percentageof events that produced red staining (cells that are infectedby bacteria) was further analyzed, and the cells were separatedaccording to their green and blue fluorescence intensities.The percentage of the blue and green cells that contain P. gingivaliswas then quantitated. Uninfected cells, incubated with anti-P.gingivalis ATCC 33277 rabbit polyclonal antibody and Alexa-Fluor633-conjugated goat anti-rabbit IgG (H+L) antibody, were usedas controls to determine the red staining threshold level. Positiveevents for the infected cells were then gated based on the determinedthreshold.

Imaging of bacterial spreading by fluorescence microscopy. GECs labeled with CellTracker Blue CMAC and preinfected withP. gingivalis were mixed with the uninfected CellTracker GreenCMFDA-labeled cells and grown on four-well chambered slides(Nalge-Nunc International, Rochester, NY) under conditions similarto those described above. The growth medium was supplementedwith metronidazole (200 µg/ml) and gentamicin (300 µg/ml)for each condition. After incubation, the slides were washedwith ice-cold PBS and fixed in 10% neutral buffered formalinfor 20 min. The cells were permeabilized for 10 min with 0.1%Triton X-100, and the slides were incubated with anti-P. gingivalisATCC 33277 rabbit polyclonal antibody (1:2,000) followed byred-orange fluorescence (Alexa-Fluor 546)-conjugated goat anti-rabbitIgG (H+L) highly cross-absorbed antibody (1:500) (MolecularProbes) for 1 h at room temperature. After consecutive washes,the slides were mounted in Vectashield mounting medium (VectorLaboratories, Burlingame, CA) and examined with an epifluorescencemicroscope (Zeiss Axioskop) equipped with band-pass opticalfilter sets appropriate for imaging of the dyes. The imageswere captured by multiple exposures with a cooled charge-coupled-devicecamera controlled by QCAPTURE software, version 1394.

Staining of bacteria and the actin cytoskeleton. Labeling of intracellular bacteria and organization of actinfilaments was performed employing an immunofluorescence techniquedescribed previously (47). Briefly, the GECs grown on coverslideswere incubated with anti-P. gingivalis ATCC 33277 antibody andthen reacted simultaneously with Oregon Green 488 secondaryantibody, phalloidin-tetramethylrhodamine B isothiocyanate,and 4,6-diamidino-2-phenylindole (DAPI) (1 µg/ml) (Sigma).This was followed by visualization with the epifluorescencemicroscope described above. The images are representative ofthree independent experiments.

RESULTS

Top

Results

Cell-to-cell spreading of P. gingivalis in GECs and potential role of the actin cytoskeleton. In preliminary studies, we examined the effects of early andprolonged incubation of P. gingivalis with GECs by fluorescencemicroscopy, which suggested that P. gingivalis could be transferredbetween the host cells through protrusions containing tightlypacked actin filaments after 24 h of infection. The localizationof intracellular bacteria and the organization of actin filamentswere visualized through immunofluorescence microscopy (Fig.1). The nuclei were also labeled with DAPI in order to confirmthe subcellular cytoplasmic location of P. gingivalis. The resultsdemonstrate that P. gingivalis is often localized within actin-basedmembrane protrusions, which could facilitate the organism'sspreading to neighboring cells after 24 h of infection. Theprotrusions frequently demonstrated a long cylindrical appearance,as shown in Fig. 1A and B. In addition, the infected cells sometimesassembled multiple filaments into larger-scale protrusions appearingas bundles where the filaments (red) align in parallel withcolocalized P. gingivalis cells (green) merging to the adjacentcell (Fig. 1C).

View larger version (34K):
[in this window]
[in a new window]
FIG. 1. Intercellular translocation of P. gingivalis through actin fibers at 24 h postinfection. Image sets (A to C) denote three independent experiments to visualize the localization of P. gingivalis ATCC 33277 and the organization of actin filaments using immunofluorescence microscopy. GECs were fixed and stained with phalloidin-tetramethylrhodamine B isothiocyanate (red) and anti-P. gingivalis antibody plus Oregon Green 488-conjugated secondary antibody to visualize the intracellular P. gingivalis (green) (left columns). DAPI (blue) was used in staining the nucleus to determine the localization of P. gingivalis in the cytoplasm (images in the right columns represent overlay combinations of three different experiments). White arrows indicate actin projections and bacterial translocation between the host cells. Bar, 10 µm.

Time/actin-dependent kinetics of intercellular P. gingivalis spreading analyzed by flow cytofluorometry. In order to quantitatively measure the spreading of P. gingivaliscells to adjacent cells and begin examining the cellular basisfor spreading, we followed the time course of bacterial spreadingand evaluated whether actin participates in this process. Forthe purpose of quantification, we developed a novel techniquebased on cytofluorimetry, which is described in Materials andMethods. GECs were labeled with CellTracker Blue, which exhibitsbright blue fluorescence in the cytoplasm, and the fluorescentcells were infected with P. gingivalis and then cocultured withuninfected CellTracker Green-labeled GECs for the indicatedtime periods. Finally, the samples were probed for P. gingivalisinfection by immunostaining (red fluorescence) and analyzedby flow cytofluorometry (Fig. 2A). Cells were gated based onmorphological profiles (size and light refraction properties)determined by side scatter-versus-forward scatter measurements(Fig. 2A1). Bacterium-containing cells within the whole populationwere quantitated by setting a gate for red fluorescence staining(Fig. 2A2). The total percentage of bacterium-containing cells(Fig. 2A2, gated region R2) was then further analyzed, and thecells were separated according to their green (gated regionR3 in Fig. 2A3) and blue (Fig. 2A4, gated region R4) intensitiesto measure the percentage of newly infected cells.

View larger version (33K):
[in this window]
[in a new window]
FIG.2. Intercellular spreading analyzed and quantified by flow cytofluorometry. The CellTracker Blue fluorescence-labeled cells infected with P. gingivalis ATCC 33277 were cocultured with uninfected CellTracker Green fluorescence-labeled cells, and samples were probed with anti-P. gingivalis antibody plus Alexa-Fluor 633-conjugated secondary antibody (red fluorescence). (A) Samples were analyzed by setting gates for forward scatter, side scatter, and red fluorescence staining (A1 and A2), with cells in the R2 window being defined as infected (A2). The total percentage of cells stained with red (bacterium-containing cells in gate R2) were further analyzed, and the cells were separated according to their green (A3) and blue (A4) intensities to measure the percentage of newly infected cells. (B) Histograms of red fluorescence distributions in blue and green cells over time show increasing infection of green cells. Green cells at 0 h of coincubation with 1-h-infected blue cells (B1), green cells after a 2-h coincubation with 1-h-infected blue cells (B2), green cells after a 24-h coincubation with 1-h-infected blue cells (B3), and green cells after a 24-h coincubation with 24-h-infected blue cells (B4) are shown.

To rule out the possibility that externalized bacteria may contributeto cell-to-cell spreading, high levels of gentamicin and metronidazolewere included in the growth medium throughout the coculturingperiods. Therefore, any bacteria present in the extracellularmedium will be rendered nonviable and thus unable to invade.

The red fluorescence distribution analysis of blue and greencells over time showed increasing infection of green cells afterincubation with blue cells (Fig. 2B1 to B4). Initially, bacteriallevels for green cells cocultured for 2 h with 1-h-infectedblue cells did not indicate a significant amount of infection(Fig. 2B2) compared with that for green cells that had not beenincubated with infected blue cells (Fig. 2B1). After coculturingof green cells for 24 h with 1-h-infected blue cells, the bacterialevels in the green cells increased by almost an order of magnitude(Fig. 2B3), as indicated by the rightward shift of the red fluorescencedistribution peak. The red fluorescence peak for green cellscocultured for 24 h with 24-h-infected blue cells shifted evenfurther to the right on the logarithmic x axis scale, implyingan even higher level of infection in the green cells (Fig. 2B4).

Quantitative analysis of the percentage of initially infectedcells (blue) with initially uninfected cells (green) by flowcytofluorometry (Fig. 2A1 to A4 and 3) demonstrated that coculturingof cells that had been infected with P. gingivalis for 1 h withthe uninfected cells for 24 h resulted in 22 (± 1.6)%newly infected cells. In contrast, a 2-h coculture with thesame uninfected cells failed to promote new infection (1.1 [±0.4]%), as seen in Fig. 3. Furthermore, the cells that had beeninfected for 24 h and then cocultured with uninfected cellsfor another 24 h produced a 49 (± 4.8)% infection ratefor the initially uninfected cells (Fig. 3). The latter infectionrate is approximately 2.5-fold higher than that for the cellsinfected for 1 h and then cocultured for an additional 24 hand approximately 45-fold higher than that for the cells infectedfor 1 h and then cocultured for an additional 2 h. These resultsreinforce the interpretation that P. gingivalis can spread directlyfrom cell to cell over the time course of infection.

View larger version (18K):
[in this window]
[in a new window]
FIG. 3. Cells exhibiting bright blue fluorescence in the cytoplasm were infected for 1 h or 24 h and then coplated with uninfected green-labeled cells for 2 h or 24 h, as indicated in the histogram. Cytochalasin D was used to determine whether the actin cytoskeleton was required for spreading. (Cytochalasin D was added to cocultured control samples and maintained throughout the coculturing period.) The bars show the percentages of GECs stained for P. gingivalis infection. Error bars represent standard deviations of at least three independent measurements.

Cytochalasin D, a fungal toxin that disrupts actin filamentsand inhibits actin polymerization, was used to evaluate therole played by the actin cytoskeleton in the propagation ofP. gingivalis through GECs. Cytochalasin D was employed onlyduring the period when the two groups of host cells were coculturedand not during the primary infection of the CellTracker Blue-stainedcells (cytochalasin D was maintained in the medium throughoutthe coculturing period).

Treatment of the cocultured cells with cytochalasin D revealedsignificant inhibition of the level of new infection (6.8 [±1.4]%) (Fig. 3), confirming that the stimulation of actin polymerizationis an underlying factor for dissemination of P. gingivalis withinGECs.

Examination of intercellular P. gingivalis infection by fluorescence microscopy. We next employed the labeling strategy used in the flow cytofluorometryanalysis to visualize the distribution of P. gingivalis disseminationin GECs by immunofluorescence microscopy. Figure 4A shows 1-h-infectedblue cells after a 2-h coincubation with uninfected green cells.After 24 h of coincubation, many of the green cells acquiredred-labeled bacteria (Fig. 4B). This technique allowed us notonly to distinguish the newly infected cells but also to quantifytheir numbers (Fig. 4C1 and C2). In agreement with the flowcytofluorometry data, the results of the fluorescence microscopyexhibited similar infection rates for early and later stagesof invasion, as shown in Fig. 4A and B. A count of the newlyinfected green cells revealed a 42 (± 2)% infection rateamong previously uninfected cells (21 [± 2]% infectionrate in the whole population of blue and green cells), as shownin the conceptual diagram (4C1 and C2). Treatment with gentamicinand metronidazole was used to inactivate the bacteria that mayhave escaped outside the cells, and fluorescence immunostainingverified the absence of extracellular bacteria (Fig. 4A and B).It is also noteworthy that the newly infected cells (green-labeledcells infected with red-labeled bacteria) displayed much lowerlevels of intracellular P. gingivalis than the cells that hadbeen previously infected for 48 h (blue cells) (Fig. 4B). Indeed,quantitative analysis of intracellular bacterial staining byNIH ImageJ analysis software indicated that initially infectedcells (blue) harbor at least four times more bacteria than thenewly infected cells (green) (data not shown). This is consistentwith the ability of P. gingivalis to successfully replicatein GECs over time (26). Since the main focus of this initialstudy was to characterize intercellular spreading of P. gingivalis,the comprehensive examination of the replication will be undertakenin future follow-up studies. Overall, the data demonstrate thatP. gingivalis can efficiently spread within GECs. More importantly,the capacity to disseminate intercellularly appears to be acquiredlater in the intracellular infectious process.

View larger version (38K):
[in this window]
[in a new window]
FIG. 4. P. gingivalis spreading visualized by fluorescent microscopy. The images of cocultured cells confirmed the flow cytofluorometry results. (A) Blue cells that were infected with P. gingivalis ATCC 33277 for 1 h and incubated with uninfected green cells for 2 h displayed no visible transmission of bacteria to green cells. (B) Coculturing (24 h) of cells initially infected for 24 h showed transmission of bacteria to newly infected cells (green), which displayed red-labeled P. gingivalis in their cytosol. White arrows indicate newly infected cells. Bar, 10 µm. (C) The conceptual diagram illustrates the quantitative distributions of initially infected, initially uninfected, and newly infected cells as detected by fluorescence microscopy. At the outset, infected and uninfected GECs were present in approximately equal amounts (C1). Coculturing for 24 h of uninfected cells with 24-h-infected cells resulted in approximately 21% newly infected cells (C2). The ratio of newly infected cells translates to an approximately 42% infection rate among initially uninfected cells (green). This infection rate is similar to that detected by flow cytofluorometry. At least 10 separate fields containing an average of 25 GECs were studied in each of two independent experiments performed in duplicate.

DISCUSSION

Top

Discussion

Epithelial cells in mucosal membranes act as an initial barrierto bacterial infection. They produce a variety of innate antimicrobialdefenses, cytokines, and other factors integrating signals frommicroorganisms to the host immune system (1, 13, 24). Therefore,the challenge for bacterial pathogens is to devise strategiesthat enable them to breach that barrier, enter the host cells,and colonize host tissues. Extensive bacterial host invasionstudies have revealed that the internalization of invasive bacteriainto epithelial tissues involves a concerted interplay of manybacterial and epithelial cell proteins (10, 22, 36). While variousbacterial pathogens can invade epithelial cells, their abilityto disseminate and colonize host tissues depends on their abilityto survive and replicate in epithelial cells (2, 30, 40).

P. gingivalis, an important pathogen of severe/chronic casesof periodontal disease, successfully invades, multiplies in,and survives for extended periods in primary GECs, which arethe lining cells of the gingival crevice and natural targetsof this subgingival organism. It has been documented that theorganism can also invade dendritic cells, fibroblasts, and heartor aortic endothelial cells (3, 11, 20, 37). Although the virulencemechanisms of P. gingivalis and invasion by the organism ofepithelial cells and other cell types are under intense scrutiny(17, 25, 29, 41), it is unknown whether the intracellular bacteriareside exclusively in the cytosol of the same cell throughoutthe infection or spread to neighboring cells to propagate.

The findings of this study demonstrate that intracellular P.gingivalis can spread directly from cell to cell, and that theintercellular translocation of the organism is facilitated byan actin network-dependent process, where the cortical actinfilaments are assembled to form membranous projections to transmitthe bacteria to adjacent host cells. The levels of the spreadingincrease markedly around the first 24 h postinfection ( $~$ 22%)compared to the levels at early stages of infection (a totalof 3 h postinfection), which showed a transfer rate of $~$ 1.5%.Longer incubations (a total of 48 h postinfection) exhibiteda significant increase in the overall new-infection rate ( $~$ 49%).The results suggest that the intercellular spreading of P. gingivalisis potentially associated with the organism's stage of infectioncycle and/or the bacterial load carried by the GECs. The utilizationof cytochalasin D confirmed that polymerization (elongation)of actin filaments is required for the promotion of P. gingivalisinfection cell to cell, since disruption of actin polymerizationby toxin treatment had a significant inhibitory effect on theamount of spreading. The exploitation of the host actin cytoskeletonis customary for other bacterial pathogens. A typical exampleis seen with Listeria, which promotes the appearance of an actintail that projects the bacterium through the cytoplasm and movesthrough host cells rapidly (38). Although P. gingivalis invasionresembles that in the Listeria model, in which the organismdoes not remain in a membrane vacuole, the utilization of anactin cytoskeleton for translocation of the bacteria intercellularlydiffers noticeably. P. gingivalis does not use an actin tailbut instead takes advantage of actin-dependent protrusions toreach new host cells. It is tempting to speculate that the associationof P. gingivalis with the actin cytoskeleton may enable thebacterium to evade the host immune system response and propagatesuccessfully.

A common theme that has been observed for an increasing numberof bacteria is for them to invade epithelial cells, survive,replicate, and eventually spread to adjacent host cells. Forexample, Actinobacillus actinomycetemcomitans, an importantperiodontopathogen, invades epithelial cells, undergoes rapidmultiplication, and subsequently propagates intercellularly(31). Salmonella enterica serovar Typhimurium also survivesand begins to replicate after a 3- to 5-h lag period and continuesto do so up to 24 h, when the organism exits the host cell (8,30). Interestingly, Salmonella infection is known to cause apoptosisin macrophages but not in epithelial cells (23, 32), and directcell-to-cell spread has not been observed for Salmonella inmacrophages. Similarly, invasion of GECs by P. gingivalis doesnot induce apoptosis and instead inhibits apoptosis in thesecells (33, 45). Our previous studies indicate that the organismexerts this effect at the level of mitochondria (e.g., blockingof mitochondrial membrane potential depolarization and cytochromec release), and that P. gingivalis-dependent inhibition of apoptosisis likely mediated by the PI3K/Akt pathway (45). Accordingly,the resistance to apoptosis could represent a coordinated planutilized by intracellular bacteria to acquire adequate timeto adapt and to express necessary effector molecules to disseminateinfection (7).

To date, several mechanisms have been proposed for the organism'sability to invade into deeper layers of gingival epitheliumand gain access to subepithelial tissues, such as through transcellularor paracellular pathways (4). P. gingivalis proteases may contributeto the disintegration of the epithelial transmembrane proteins,E-cadherin, ß₁ integrin, and occludin, which are essentialcomponents of the epithelial function as a mechanical barrieragainst microbes (9, 21). This suggests that, in addition tointercellular spreading, P. gingivalis may employ other complementarydissemination mechanisms.

The observations of intercellular spreading of P. gingivalisin later phases of infection may result from the bacterium'sability to diminish the host apoptotic response. Indeed, thisobservation presents a logical link to the bacterium's capacityto multiply successfully and suppress host cell death. The delayor interruption of apoptosis may enable the bacterium to shiftfrom intracellular multiplication to intercellular dissemination.It will be intriguing to examine whether the same pathways involvedin suppression of host cell apoptosis also play a role in determiningthe timing of cell-to-cell spread.

ACKNOWLEDGMENTS

This work was supported by NIDCR grants DE14168, Universityof Washington, and DE11111, University of Florida and Universityof California, Merced.

FOOTNOTES

* Corresponding author. Mailing address: Department of Pathobiology, University of Washington, Box 357 238, Seattle, WA 98195. Phone: (206) 543-6427. Fax: (206) 543-3873. E-mail: ozlem{at}u.washington.edu.

Editor: D. L. Burns

REFERENCES

Top