INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Spirochetes are commonly detected in very low numbers in healthy gingival crevices, but their population increases sharply in periodontal diseases. Not only are spirochetes related to the progression of periodontal diseases (4, 5, 7, 35), but their presence in health-associated plaque may be associated with increased susceptibility to the development of periodontitis (32). They may also serve as prognostic markers for disease recurrence following treatment (36). Spirochetes are often found in direct contact with the sulcular epithelium lining the periodontal pocket and the junctional epithelium maintaining the epithelial attachment to the tooth (23). Therefore, they have an opportunity to contribute directly to the development of gingival lesions. In acute necrotizing ulcerative gingivitis and in advanced lesions of periodontitis, spirochetes may invade into the connective tissue (24, 33). Hence, they and their metabolic products or cellular fragments which penetrate the gingiva may also have an opportunity for direct contact with immunocompetent cells and cells which maintain connective tissue homeostasis, including fibroblasts (10).

Treponema denticola is the spirochete most frequently isolated from human periodontal pockets. Several cytopathic responses of fibroblasts have been documented following their contact with T. denticola (2, 3, 11, 41, 44). In vitro, late-stage cellular responses include formation of plasma membrane folds or blebs, rounding and shrinkage from the normally stellate shape, a reduction in proliferative capacity in some studies, and, ultimately, detachment of a subset of fibroblasts from the substratum. We have shown that T. denticola induces disruption of human gingival fibroblast (HGF) F-actin, its rearrangement into a perinuclear array, and the reduction in total F-actin content, especially in the ventral third of the cells (2, 46). We have also shown that T. denticola suppresses the inositol phosphate pathway and inhibits stretch-induced calcium transients as well as oscillations of calcium ions in resting cells (20, 46). These profound effects on the actin cytoskeleton and related signaling pathways would be expected to alter critical fibroblast functions like locomotion, which is important for effective wound healing, or phagocytosis, which is crucial for physiologic remodeling and wound repair in the periodontium. As the cytoskeletal changes were shown to be induced by outer membrane (OM) extracts as well as whole T. denticola cells, our specific aim was to determine the effects of T. denticola OM extract on collagen phagocytosis by HGFs. We were interested in collagen uptake because (i) collagen fibers are the primary protein component in the connective tissue matrix of the periodontium and are particularly important in providing structural stability due to their tensile strength (37) and (ii) a net loss of gingival connective tissue collagen is characteristic of periodontal diseases (27).

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Culture conditions and preparation of OM extracts. T. denticola type strain ATCC 35405 stock cultures were maintained in a complex spirochete broth medium containing brain heart infusion, tryptic peptone, yeast extract, and volatile fatty acids, supplemented with 2.0% rabbit serum (8). This strain has been shown to stimulate F-actin rearrangement, plasma membrane blebbing, and degradation of endogenous fibronectin in HGFs (2, 10, 11, 46). OM extract was prepared by a modification of the detergent extraction method of Penn et al. (31), as we have previously described (46). Briefly, the bacteria were harvested at late stationary phase, washed twice in 0.01 M phosphate-buffered saline, pH 7.2 (PBS), resuspended, and dispersed uniformly in PBS containing 10 mM MgCl₂ at a ratio of 10 ml of PBS per g (wet weight) of bacteria. After Triton X-100 (Surfact-Amps X-100; Pierce, Rockford, Ill.) extraction and repeated centrifugation, the supernatant was dialyzed against distilled water (molecular mass cutoff, 50 kDa; Spectra/Por; Spectrum, Houston, Tex.) for several days until the OM precipitated. The contents of the dialysis tubing were centrifuged at 25,000 × g for 45 min at 4°C. The pellet was then resuspended in distilled water to the predialysis volume and stored at 70°C until used. The dry weight concentration of the OM extract was 1.6 mg/ml after freeze-drying. A single batch of OM extract was used for all experiments reported herein.

3

Fibroblast culture. Primary explant cultures of HGFs were derived from biopsies of normal gingiva as described elsewhere (1). Flow cytometry experiments were conducted on cells grown to confluence in six-well plates (Corning Glass Works, Corning, N.Y.) obtained as follows. Confluent cells in T-75 flasks were harvested with 0.01% trypsin. The cell concentration was determined by electronic particle counting (Coulter Counter ZM; Coulter Electronics of Canada Ltd.), and the HGFs were plated at a concentration of 3 × 10⁵ cells per well. The cells were grown until confluent at 37°C in a humidified 5% CO₂ incubator in alpha minimal essential medium (-MEM) supplemented with 10% (vol/vol) antibiotic solution (penicillin G [145.0 U/ml; Sigma], gentamicin sulfate [50 µg/ml; Gibco BRL, Grand Island, N.Y.], and amphotericin B [0.25 µg/ml; Gibco BRL] in -MEM) and 15% (vol/vol) heat-inactivated fetal bovine serum (FBS; ICN Biomedicals Inc., Costa Mesa, Calif.). Cells between passages 3 and 12 were used for all experiments.

Co-I and BSA coating of fluorescent beads. To test the effect of OM extract on specific as well as nonspecific bead phagocytosis, collagen type I (Co-I)- and BSA-coated beads were used. Coated beads were prepared by the method of Lee et al. (22). Briefly, 80 µl of yellow-green fluorescence polystyrene microbeads (2.0-µm diameter; excitation maximum = 490 nm; emission maximum = 515 nm; catalog no. L5301; Molecular Probes, Eugene, Oreg.) were incubated for 10 min at 37°C in 1 ml of an acidic bovine collagen solution (Vitrogen 100; 3.0 mg/ml, adjusted by 100 µl of 1 N NaOH to pH 7.4; Collagen Corp., Palo Alto, Calif.). At pH 7.4, the collagen solution undergoes fibril assembly on the beads. For BSA-coated beads, 80 µl of the beads was incubated for 30 min at room temperature in 1 ml of BSA (1.54 mg/ml; Sigma). In both cases, the beads were centrifuged and resuspended in 1 ml of PBS containing no Ca²⁺ or Mg²⁺ (PBS,Ca,Mg). To estimate the resulting concentrations of beads per ml, the beads were first sonicated to produce single-bead suspensions, and an aliquot was then counted with the use of a hemacytometer. Bead sonication was also performed prior to each flow cytometry experiment.

Treatment of HGF with OM extract. To compensate for the variable growth rates of different HGF cell populations, and to keep the OM-to-cell ratio constant at 0.192 ng/cell, the number of cells in one representative well was determined by electronic particle counting in each experiment. The samples from the remaining wells were then incubated with the appropriate amount of OM in CO₂-independent medium (CIM; Gibco BRL) at 37°C for 1 h (except for time course studies), and the control wells received CIM without OM extracts. This OM-to-cell ratio caused minimal changes in HGF morphology (as determined by inverted light microscopy) and minimal cell loss due to detachment (as determined by electronic particle counting). The cells were then washed three times with PBS,Ca,Mg to remove unbound OM, and fresh growth medium plus 15% FBS was added.

PI. The number of cells containing internalized beads in a given HGF population was determined and expressed as a percentage of total cells, the phagocytosis index (PI) (19, 22, 28, 29). Briefly, the number of HGF cells in representative samples of the control as well as the OM-treated groups was determined by electronic particle counting. Beads, previously coated with either collagen or BSA, were added at the saturation level of 4:1 (beads/cell [22]) and incubated for 3 h at 37°C in a humidified air and CO₂ incubator. The cells were then washed twice with PBS,Ca,Mg and detached by incubation in 0.01% trypsin for 5 to 10 min, and the trypsin was neutralized with growth medium plus 15% FBS. This procedure removes all loosely bound but not internalized beads (19, 29). The cells were centrifuged and resuspended in PBS,Ca,Mg. Verification that single cell suspensions were produced was done by light microscopy. Unstained cells (n = 10,000/sample) were analyzed at a rate of 200 cells/s in a FACStar Plus flow cytometer (Becton Dickinson FACS Systems, Mountain View, Calif.). Cells containing ingested beads, cells without ingested beads, and unbound beads that had not been internalized by cells were discriminated on the basis of forward light scatter and side scatter. By setting specific electronic gates previously established for whole cells, only cells with forward and side scatter characteristics similar to those of whole, intact cells were included in the analysis. Histograms of the number of gated cells versus fluorescence intensity (corresponding to bead number) were plotted, and markers were set to evaluate the PI and the average number of beads per phagocytosing cell, the phagocytic capacity (PC).

HGF viability. To test for cell viability, propidium iodide (Calbiochem, La Jolla, Calif.) at 0.5 mg/ml (final concentration) in PBS was added to the single-cell suspensions following the phagocytosis assay and 5 min prior to flow cytometry analysis. A positive control of dead HGFs was obtained by treating a sample at 4°C with cold 70% ethanol for 10 min.

LY uptake. Lucifer yellow, a fluid-phase pinocytosis marker, was chosen as a control to test for the effect of T. denticola OM extracts on endocytic activity other than the uptake of large particles (39, 40). Confluent HGFs in six-well plates were treated with T. denticola OM in CIM, or CIM alone for the control samples, as in the phagocytosis assay. After three washes with PBS,Ca,Mg to remove the unbound OM, 0.5 ml of lucifer yellow CH-potassium salt (LY; Molecular Probes) was added at 0.5 mg/ml (final concentration) in -MEM containing 10% FBS, as described by Swanson et al. (39, 40). The samples were incubated for 1, 10, 30, and 60 min. The cells were washed three times with PBS,Ca,Mg and detached as described above. The cells were centrifuged and resuspended in PBS,Ca,Mg. For each gated sample run by the standard flow cytometry protocol, the mean fluorescence intensity was evaluated from histograms of gated cells versus LY fluorescence intensity.

Statistical analysis. Each experiment included triplicate samples. Means and SE were calculated for group comparisons within and among experiments. The experiments were repeated at least two to three times except for the pinocytosis assay with the consecutive OM-LY treatment and the dose-response experiments for V. atypica OM extracts, which were done once. For comparisons between control and OM-treated HGFs, the percent difference between corresponding samples within each experiment was calculated. For comparison of key variables, probability values cited in the text were based on mean percent differences from several independent experiments. Student's t test for unpaired samples was used to test significance in most cases. Analysis of variance was used for analysis of multiple samples for the heated OM extract experiment. A P value of <0.05 was set as a standard for statistical significance. Results shown represent examples of the repeated experiments rather than an average.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

T. denticola increases HGF collagen phagocytosis. OM treatment of HGFs induced a consistent, statistically significant, concentration-related increase in collagen-coated bead uptake (Fig. 1 shows results of a representative experiment; n = 3). The average of 11 independent experiments showed that the overall mean untreated HGF uptake of collagen-coated beads was increased from 20.4% ± 1.9% to 61.1% ± 3.6% (mean PI ± SE) following T. denticola OM challenge, a difference of 202% (P < 0.00001). In addition to the increase in the percentage of cells ingesting the beads, T. denticola OM challenge induced a statistically significant increase in the number of beads taken up per cell (Fig. 2, inset). Uptake of collagen-coated beads increased from 155 ± 3 to 230 ± 6 fluorescence units (mean PC of three independent experiments ± SE) following T. denticola OM challenge, an increase of 48% (P < 0.001). This finding indicates that OM challenge induces an increase in the phagocytosis phenotype as well as the phagocytic capacity of HGFs. The increase observed in PI (Fig. 3 and 4) and PC (data not shown) of collagen-coated beads varied directly with the duration of incubation with OM. An exposure of HGF to T. denticola OM as brief as 15 min induced an obvious increase in the PI compared with controls.

View larger version (20K): [in this window] [in a new window]   FIG. 1.   Concentration-related effect of T. denticola OM challenge on HGF-specific (Co-I) and nonspecific (BSA) bead uptake. HGFs were treated with OM at final concentrations of 0.096, 0.192, and 0.288 ng (dry weight) of OM/HGF cell for 1 h. PI is expressed as mean ± SE (three independent samples of 10,000 cells each; note that error bars are very small). *, P < 0.01; **, P < 0.001 versus OM-free control.

View larger version (20K): [in this window] [in a new window]   FIG. 2.   Flow cytographs of collagen-coated bead uptake. Cytographs (log fluorescence on the x axis and cell number on the y axis) illustrate the relative proportion of cells without beads (left-hand peak in each cytograph) versus the proportion of cells with internalized beads (right-hand peaks [M5] in each cytograph). Within M5, each peak corresponds approximately to a group of cells with a finite number of beads: 1 bead (M1); 2 beads (M2); 3 beads (M3); and 4 beads (M4) phagocytosed. (A) Control untreated HGFs. (B) HGFs challenged with 0.192 ng of OM/HGF for 1 h. The inset graph illustrates the relationship between increasing OM concentrations (x axis; nanograms/HGF) and the weighted mean fluorescence units (y axis) of internalized collagen-coated beads (solid line) or BSA-coated beads (dashed line), expressed as mean ± SE (three independent samples of 10,000 cells each; note that error bars are very small). *, P  0.05; **, P < 0.01; ***, P < 0.001 versus OM-free control.

View larger version (23K): [in this window] [in a new window]   FIG. 3.   Effect of duration of T. denticola OM challenge on HGF-specific (Co-I) and nonspecific (BSA) bead uptake. HGFs were treated with T. denticola OM (0.192 ng [dry weight] of OM/HGF cell) for 15, 30, or 60 min. PI is expressed as mean ± SE (three independent samples of 10,000 cells each; note that error bars are very small). *, P < 0.05; **, P < 0.001 versus OM-free control.

View larger version (30K): [in this window] [in a new window]   FIG. 4.   T. denticola OM-induced phagocytosis enhancement is a long-lasting effect. Group 1, phagocytosis assay tested immediately following washing of the HGFs free of the OM; group 2, treatment of HGFs with OM followed by washing and overnight incubation with serum-free medium. The phagocytosis enhancement persisted. PI is expressed as mean ± SE (three independent samples of 10,000 cells each). *, P < 0.05; **, P < 0.01; **, P < 0.001; ***, P < 0.0001 versus OM-free control.

Increased BSA-coated bead uptake, a test of specificity. To determine whether the T. denticola OM-mediated enhancement was due to a general increase in the phagocytosis of protein-coated beads, cells were incubated with BSA-coated beads, which are generally taken up nonspecifically in very low numbers by HGFs (19). OM challenge to HGF monolayers induced a consistent and significant concentration-related increase in BSA-coated bead uptake, which was also persistent for at least 16 h (Fig. 1 and 4). The untreated control HGF uptake of BSA-coated beads was increased from 6.5% ± 0.9% to 29.7% ± 3.8% (mean PI of eight independent experiments ± SE) following T. denticola OM challenge, an increase of 357% (P < 0.001). Analysis of the mean number of BSA-coated beads ingested per cell revealed a dose-response increase smaller than that for the PC of collagen-coated beads (Fig. 2, inset). Time course experiments revealed a positive, linear relationship between the duration of OM incubation and BSA-coated PI (Fig. 3 and 4) and PC (data not shown). The OM-induced specific and nonspecific phagocytic enhancement was evidently an early-onset, long-lasting response.

Increased LY uptake. LY was used to test the effect of T. denticola OM extract on pinocytosis in contrast to the uptake of relatively large particles (~2-µm diameter). A time course study of LY incubation yielded a positive concentration response for accumulation of dye in control cells. Coincubation of HGFs with LY and T. denticola OM simultaneously yielded a significant, saturable, time-related enhancement of LY accumulation (Fig. 5A). In contrast, a 1-h pretreatment with T. denticola OM followed by washing and LY exposure did not seem to affect the subsequent fluid-phase endocytosis activity of HGFs (Fig. 5B).

View larger version (21K): [in this window] [in a new window]   FIG. 5.   Effect of T. denticola OM challenge on HGF pinocytic activity. (A) HGFs were incubated with LY at a final concentration of 0.5 mg/ml and T. denticola OM (0.192 ng/HGF; dashed line), or without the OM extract (solid line), for 1, 10, or 15 min. P < 0.01 for all data. (B) HGFs were pretreated with buffer (negative control; solid line) or T. denticola OM (0.192 ng/HGF for 1 h; dashed line), washed, and incubated with LY up to 60 min. PI is expressed as mean ± SE (three independent samples of 10,000 cells each; note that error bars are very small).

V. atypica OM decreases collagen but does not affect BSA-bead uptake. To test whether the HGF phagocytosis phenotype can be enhanced by an OM extract of a bacterium other than T. denticola, the OM extract of V. atypica ATCC 17744 was used. In contrast to the T. denticola OM extract, V. atypica OM extract induced a significant decrease in the collagen-coated bead PI (Fig. 6A). The mean untreated control HGF uptake of collagen-coated beads decreased by 47% (P < 0.001) following V. atypica OM challenge (three independent experiments). In fact, V. atypica OM concentrations up to four times the standard concentration (0.192 ng/cell) yielded a concentration-dependent decrease in HGF uptake of collagen-coated beads (Fig. 6B). V. atypica OM extract did not induce a significant change in the PI for BSA-coated beads (Fig. 6A). The mean uptake of BSA-coated beads decreased by 19% following V. atypica OM challenge (P > 0.05, mean change over three independent experiments). Propidium iodide staining indicated that the V. atypica OM challenge did not induce HGF cell death (data not shown). These findings indicate that the T. denticola OM-enhanced BSA- and collagen-coated bead uptake by HGFs is not a general phenomenon typical for any bacterial OM extract.

View larger version (32K): [in this window] [in a new window]   FIG. 6.   Effect of V. atypica OM challenge on HGF phagocytosis phenotype. (A) HGFs were exposed to CIM (negative control), T. denticola OM (positive control), or V. atypica OM (0.192 ng [dry weight] of OM/HGF cell for 1 h). In contrast to T. denticola, V. atypica OM pretreatment induced a decrease in collagen (Col)-coated bead phagocytosis and no statistically significant change in BSA-coated bead uptake. For each condition, three independent samples of 10,000 cells each were assayed (mean ± SE). Data shown are expressed as percent difference from negative control HGFs. *, P = 0.001; **, P < 0.01 versus control. (B) Dose response. HGFs were exposed to CIM (negative control) or V. atypica OM at final concentrations of 1× (Va×1), 2× (Va×2), and 4 × (Va×4) 0.192 ng (dry weight) of OM/HGF cell for 1 h. Analysis was identical to that for panel A. Data are expressed as percent difference from negative control HGFs.

Heat sensitivity of T. denticola OM extract activity. At the concentrations used here, a 1-h T. denticola OM extract challenge increased collagen-coated bead uptake 3-fold and BSA-coated bead uptake 4.5-fold. Heating at either 60°C for 30 min or 100°C for 10 min completely abolished the OM enhancement of both collagen- and BSA-coated bead uptake (data not shown). Lipopolysaccharides (LPS) are known to be stable at these conditions. Heating the OM at 60°C for 30 min should denature most proteins, yet the extract retained greater than 65% of its SAAPNA-degrading activity, as in our previous studies (20, 46). The SAAPNA-degrading activity was reduced to <10% of the control level after boiling of the OM extract for 10 min. The results suggest that the PI increase was mediated by a heat-sensitive component(s) of the OM, probably neither LPS-like molecules nor the native chymotrypsin-like protease.

Cell viability. Preliminary studies led us to choose an OM concentration of 0.192 ng/cell, as it seemed to induce minimal HGF detachment. Indeed, several experiments confirmed this finding. The mean cell loss after 60 min T. denticola OM challenge was only 11.4% ± 4.8% (mean loss over 14 independent experiments), which is comparable to the cell loss under control conditions (2).

View larger version (14K): [in this window] [in a new window]   FIG. 7.   Cell cycle analysis following T. denticola OM challenge. Flow cytometry indicated no change in the sub-G1 population. Cells were exposed to OM-free CO2-independent medium (A) or T. denticola OM (0.192 ng/cell) for 15 min (B), 30 min (C), or 60 min (D). Cells were washed free of the extract and assayed for cell cycle phases. Prior to flow cytometry analysis, cells were fixed with 70% ethanol at 4°C for 5 min and then centrifuged and resuspended in sulforadamine with DAPI. The flow cytometer was configured to analyze only cells with the forward scatter and side scatter characteristics of whole cells. M1 indicates the sub-G1 population.

Effect of cytochalasin. The involvement of actin in T. denticola OM-stimulated phagocytosis was determined by using cytochalasin D, an actin cytoskeleton-disrupting agent (6). At a concentration of cytochalasin D which disrupted actin stress fiber arrangement (indicated by fluorescein isothiocyanate-conjugated phalloidin staining), cytochalasin D strongly inhibited collagen-coated bead phagocytosis in controls but did not affect the BSA-coated bead uptake (Fig. 8). Cytochalasin D treatment greatly reduced uptake of both the T. denticola-enhanced collagen- and BSA-coated beads (Fig. 8). These results were not due to cytochalasin D-induced toxicity since the cells were not stained by propidium iodide.

View larger version (30K): [in this window] [in a new window]   FIG. 8.   Cytochalasin D treatment of HGFs abolished the T. denticola OM-induced enhancement of bead uptake. HGFs were counted and exposed to CIM control (Cont) or T. denticola OM (0.192 ng of OM/HGF for 1 h). Cells were washed free of the extract, recounted, and then incubated in either -MEM or 1 µM cytochalasin D (Cyt D) for 1 h. PI is expressed as mean ± SE (three independent samples of 10,000 cells each). *, P < 0.01; **, P < 0.0001 versus the corresponding non-cytochalasin D-treated group. Col, collagen.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

We have provided clear evidence that exposure of HGFs to the OM of T. denticola causes rapid enhancement of their phagocytosis of protein-coated beads. At the concentration used, there was minimal cell detachment and cytotoxicity; the control, T. denticola-treated, and V. atypica-treated cells all showed negligible uptake of propidium iodide. Moreover, cell cycle analysis showed no increase in two of the hallmarks of apoptosis, namely, cell shrinkage and the development of cells with sub-G₁ DNA content. These results are consistent with previous findings that both fibroblasts and epithelial cells undergoing cytoskeletal rearrangement but remaining attached to the substratum after challenge with whole T. denticola cells are viable, as measured by lactate dehydrogenase release, propidium iodide staining, colony formation, and limiting dilution assays (2, 9).

Factors affecting the collagen phagocytosis enhancement. The observed enhancement of phagocytosis was OM concentration and challenge duration related. The significant increase in phagocytosis was rapid, preceding any apparent morphological changes or previously documented, significant actin rearrangement or inositol phosphate pathway suppression (2, 9, 46). The swift onset suggests that the stimulation of phagocytosis is due to T. denticola OM components that affect the surface receptors or membrane fluidity of HGFs, rather than through up-regulation of transcriptional events of cytokine production. In this same system, the only change that we have documented to be as swift is OM suppression of calcium transients (20). OM pretreatment for just 15 min, followed by washing, was sufficient to induce a long-lasting (>16-h), irreversible PI increase at a level equivalent to the PI of 60-min-pretreated cells. Therefore, the T. denticola OM-induced enhancement in HGF phagocytosis function is apparently persistent.

Hypothesis for the mechanism of OM enhancement of bead uptake. T. denticola OM induced a prolonged enhancement of both specific and nonspecific bead uptake, as well as a transient increase in pinocytic activity. It is unlikely that the pinocytic and phagocytic enhancement were due to the same mechanism(s). Pinocytosis is a constitutive and spontaneous event occurring in most eukaryotic cells by internalization of plasma membranes as vesicles, probably involving clathrin-coated and non-clathrin-coated pits (34, 38). In contrast, phagocytosis of large particles is more specific, involving activated receptors and transmembrane signaling to initiate particle engulfment (13, 19, 22). The LY accumulation was probably not due to T. denticola OM permeabilization of the HGF plasma membranes. Previous work indicated that HGF cells loaded with the fluorochrome fura-2 showed no significant dye leakage following T. denticola OM treatment at an OM concentration higher than the one used in this study (20). Moreover, examination of the HGFs by fluorescence microscopy found bright yellow fluorescence distributed as small discrete points over the entire cell surface (not shown), which is suggestive of uptake into pinosomes rather than through pores or unrestricted channels. Irrespective of the mechanism of pinocytic enhancement, we hypothesize that the increased pinocytic activity may mediate the uptake of the OM or fragments of the OM, which could lead to the observed persistence of bead phagocytosis.