Oxidative Stress Reshapes Porphyromonas gingivalis Outer Membrane Vesicles and Impairs OMV-Mediated Invasion and Persistence in Trophoblast Cells

TL;DR

Oxidative stress changes the size and protein content of Porphyromonas gingivalis outer membrane vesicles, reducing their ability to help the bacteria invade and persist in trophoblast cells.

Contribution

This study reveals how oxidative stress alters OMV properties and impairs their role in bacterial invasion of trophoblast cells.

Findings

Oxidative stress reduces OMV size and surface charge while increasing membrane fluorescence.

OMVs from stressed bacteria fail to enhance P. gingivalis invasion and persistence in trophoblast cells.

Proteomic analysis shows selective enrichment of specific proteins in OMVs under oxidative stress.

Abstract

Background: Porphyromonas gingivalis outer membrane vesicles (OMVs) are key mediators of host–pathogen interactions and have been implicated in both periodontal disease and systemic conditions, including pregnancy complications. Although OMV production and cargo are known to be influenced by environmental stress, how oxidative stress reshapes P. gingivalis OMVs and their functional impact on trophoblast cells remains poorly understood. Here, we investigated how exposure to hydrogen peroxide (H2O2) affects OMV biogenesis, composition, and their ability to modulate bacterial invasion in trophoblast cells. Methods: P. gingivalis was cultured anaerobically and exposed to 30 mM H2O2 during the final 24 h of growth. OMVs were isolated by differential ultracentrifugation and characterized by nanoparticle tracking analysis and transmission electron microscopy and OMV protein cargo was analyzed by proteomics. Functional effects were assessed using invasion and persistence assays in HTR-8/SVneo trophoblast cells pretreated with OMVs. Results: Oxidative stress did not significantly alter total OMV yield but resulted in smaller vesicles (control OMV 168.2 ± 8.7 nm vs. OMV from H2O2-treated cultures 130.0 ± 13.8 nm) with reduced negative surface charge and increased membrane-associated FM4-64 fluorescence. Proteomic analysis revealed a remodeling of the OMV protein cargo under oxidative stress, including the selective enrichment of a von Willebrand factor type A domain-containing protein. Functionally, OMVs from control cultures led to a 2.5-fold increase in P. gingivalis invasion and a 4-fold increase in intracellular persistence in trophoblast cells, whereas OMVs produced under oxidative stress failed to promote these processes. Conclusions: Together, these findings highlight oxidative stress as a key determinant of OMV-mediated host–pathogen interactions at the maternal–fetal interface.

Figures

Full text

1. Introduction

Outer membrane vesicles (OMVs) are nanoscale, spherical blebs released from the outer membrane of Gram-negative bacteria that serve as concentrated packets of proteins, lipids, nucleic acids and small molecules. OMVs can contribute to pathogenesis by delivering toxins and immunomodulatory cargo across biological barriers, modulating host signaling, and reshaping local and systemic immune responses; therefore, they are recognized as mediators of microbe–host crosstalk and disease progression [1,2]. Although OMVs are constitutively produced by Gram-negative bacteria, both their abundance and cargo are modulated by environmental stress. Mutations in stress-responsive pathways and exposure to envelope-perturbing agents increase vesiculation and promote the selective packaging of transporters, efflux-associated proteins, and misfolded periplasmic components [3,4,5,6]. In Pseudomonas aeruginosa, oxidative stress induces OMV production independently of AlgU, MucD, and PQS, but requires the presence of B-band LPS, highlighting the role of specific outer membrane chemical features in stress-induced vesiculation [7]. During infection, bacteria should cope with the oxidative stress produced by the innate immune system cells including hydrogen peroxide (H_2_O_2_) and nitric oxide (NO) species or molecules.

P. gingivalis is a Gram-negative bacterium and a major etiologic agent of periodontal disease. Although it is an anaerobic bacterium, P. gingivalis can survive in the periodontal pocket characterized by inflammation [8]. While catalase is absent, P. gingivalis relies on a broad battery of antioxidative defenses including superoxide dismutase, redox-active and DNA-protective proteins, and oxidative stress–induced factors such as PG_0686, which contribute to survival under H_2_O_2_ exposure [9]. However, there is a gap in knowledge regarding the OMV cargo when P. gingivalis is exposed to oxidative stress and the consequences for bacterial virulence.

P. gingivalis has also been linked to systemic clinical disorders, including pregnancy complications such as preeclampsia and intrauterine growth restriction [10,11]. Bacterial DNA and viable P. gingivalis have been detected in placental tissues and amniotic fluid from complicated pregnancies [12,13,14]. In addition, animal models demonstrate that oral infection with P. gingivalis can induce fetal growth restriction, placental inflammation, and pregnancy loss [15,16]. At the maternal–fetal interface, previous studies have shown that P. gingivalis whole bacteria and purified virulence factors can impair trophoblast cell function, disrupt barrier function, and alter cytokine production, suggesting multiple pathways by which periodontal pathogens may compromise placental health. Particularly, OMVs can directly target trophoblast cells, modulate their inflammatory and metabolic responses, and potentially contribute to placental dysfunction [8,17,18]. Notably, the role of P. gingivalis OMVs in bacterial invasion and persistence within trophoblast cells remains poorly understood and may represent a key mechanism facilitating placental colonization.

Here, we addressed these gaps by exposing P. gingivalis to H_2_O_2_ and analyzing the resulting OMVs in terms of rate of production, physicochemical properties, cargo composition, and its impact on trophoblast cell function. While OMVs derived from control cultures promoted P. gingivalis invasion and intracellular persistence in trophoblasts, OMVs produced under oxidative stress conditions lost this ability, indicating that oxidative stress profoundly alters OMV-mediated host–pathogen interactions.

2. Results

2.1. Characterization of P. gingivalis OMVs Under Oxidative Stress

To evaluate the impact of oxidative stress on OMV production and properties, P. gingivalis was cultured anaerobically for 4 days and exposed to H_2_O_2_ during the last 24 h. OMVs were isolated from control and H_2_O_2_-treated cultures by differential ultracentrifugation and subsequently characterized. Nanoparticle tracking analysis revealed no significant differences in total OMV particle concentration between control and H_2_O_2_-treated conditions (Figure 1a). In contrast, OMVs derived from H_2_O_2_-treated cultures displayed a significantly reduced median size (Figure 1b). Specifically, OMVs derived from control bacteria exhibited a mean diameter of 168.2 ± 8.7 nm, whereas OMVs produced under H_2_O_2_ exposure showed a smaller mean diameter of 130.0 ± 13.8 nm. Consistent with these observations, transmission electron microscopy (Figure 1c) confirmed differences in OMV morphology and apparent size between control and H_2_O_2_-treated samples. These findings suggest that oxidative stress alters vesiculation dynamics or selectively affects larger OMV populations, which may be more susceptible to oxidative damage.

Further characterization revealed that oxidative stress markedly altered the surface properties of P. gingivalis OMVs. Zeta potential analysis showed that OMVs derived from H_2_O_2_-treated cultures exhibited a significantly less negative surface charge compared to control OMVs (Figure 2a), indicating changes in membrane composition or charge distribution. Consistent with this finding, FM4-64 fluorescence, used here as an indicator of membrane-associated signal, normalized to total protein, was significantly increased in OMVs produced under oxidative stress conditions (Figure 2b). The increased FM4-64 fluorescence may result from differences in OMV abundance and/or size distribution when equal amounts of total protein are analyzed, as well as potential changes in membrane organization. Laser-scanning confocal microscopy revealed stronger FM4-64–associated fluorescence in OMVs derived from H_2_O_2_-treated cultures at the population level, providing qualitative support for an increased membrane-associated signal without excluding size-dependent surface-to-volume effects (Supplementary Figure S1). Together, these results indicate that H_2_O_2_ exposure affects OMV size and surface-associated properties, which may influence OMV stability and interactions with host cells.

2.2. Oxidative Stress Alters the Protein Cargo of P. gingivalis OMVs

To determine whether oxidative stress modifies the protein composition of P. gingivalis OMVs, SDS–PAGE followed by Coomassie Blue staining was performed (Figure 3a). OMVs from both conditions displayed a characteristic protein profile, with qualitative differences in band intensities, suggesting stress-dependent changes in OMV cargo. To further characterize these differences, proteomic analysis was performed, and the identified proteins were compared between conditions. A core set of 21 proteins was shared between control and H_2_O_2_-derived OMVs, whereas distinct protein subsets were detected in each condition (Figure 3b; Table 1, Table 2 and Table 3). OMVs produced under oxidative stress exhibited a markedly expanded repertoire, with 36 proteins uniquely identified in the H_2_O_2_ condition (Table 1). Notably, among the proteins detected in H_2_O_2_-OMVs, Von Willebrand factor type A (vWA) domain-containing protein, was consistently identified in all three biological replicates and was completely absent in control OMVs. Given that vWA domain-containing proteins are often involved in protein–protein interactions and adhesion-related processes, its selective enrichment in stress-induced OMVs may have functional implications for OMV-mediated interactions with host cells.

2.3. OMVs Modulate P. gingivalis Invasion and Persistence in Trophoblast Cells

Considering that trophoblast cells play a central role at the maternal–fetal interface and that P. gingivalis-derived OMVs impact their metabolic and inflammatory programs, we investigated whether OMVs modulate P. gingivalis invasion and persistence within trophoblast cells. To address this, HTR-8/SVneo trophoblast cells were pretreated for 24 h with control OMVs or OMVs derived from H_2_O_2_-treated cultures (H_2_O_2_ OMVs). Pretreatment with control OMVs significantly increased P. gingivalis invasion (Figure 4a) and persistence (Figure 4b) compared to non-treated cells. In contrast, OMVs obtained from H_2_O_2_-treated bacteria failed to promote bacterial invasion or persistence and showed a significantly reduced effect compared to control OMVs.

3. Discussion

In this study, we demonstrate that oxidative stress profoundly reshapes the physical properties, protein cargo, and functional activity of P. gingivalis-derived outer membrane vesicles, with direct consequences for host–pathogen interactions at the maternal–fetal interface. This conclusion is supported by several key observations. First, exposure of P. gingivalis to H_2_O_2_ altered OMV biophysical characteristics, resulting in smaller vesicles with modified surface charge and enhanced membrane-associated FM4-64 fluorescence, indicating stress-dependent remodeling of OMV. Second, oxidative stress markedly expanded and diversified the OMV protein cargo. Finally, these stress-induced changes were functionally relevant, as OMVs derived from control cultures enhanced P. gingivalis invasion and intracellular persistence in trophoblast cells, whereas OMVs produced under oxidative stress conditions lost this ability. Together, these findings demonstrate that oxidative stress not only modifies OMV composition but also impairs OMV-mediated mechanisms that facilitate bacterial invasion and persistence in trophoblast cells.

Periodontal inflammation is characterized by sustained oxidative stress, driven primarily by neutrophil-derived reactive oxygen species such as hydrogen peroxide [19]. Previous studies have shown that exposure of P. gingivalis W83 strain to H_2_O_2_ elicits a rapid and coordinated transcriptional response, including the early induction of DNA repair pathways followed by protein folding and stress adaptation mechanisms [20]. In this study, we employed a higher H_2_O_2_ concentration (30 mM H_2_O_2_) based on the work of Leke et al. (1999), who reported that this oxidative stress had minimal effects on bacterial growth in rich medium while significantly altering virulence-associated activities, such as gingipain and hemagglutinin function [21]. Rather than examining global bacterial stress responses, we specifically focused on how oxidative stress remodels the OMV compartment, which is increasingly recognized as a critical mediator of P. gingivalis pathogenicity, especially for the systemic effects of periodontal infection. In accordance with MISEV2023 guidelines [22], OMV preparations were characterized as vesicle-enriched fractions based on size distribution, morphology, zeta potential, and protein composition. Proteomic analysis confirmed the presence of canonical OMV markers including gingipains, outer membrane proteins, TonB-dependent receptors, and Type IX secretion system components, validating the vesicle-enriched nature of the preparations. However, we cannot entirely exclude co-purification of non-vesicular components such as released fimbriae or trace bacterial lysis products. Importantly, both control and H_2_O_2_-treated samples were processed identically, ensuring that observed differences reflect genuine stress-dependent OMV remodeling rather than technical artifacts.

OMV cargo can be different depending on the used culture medium, culture stage, treatment, and more importantly the strain used [23,24]. For example, Mantri et al., 2015 compared the vesicle content of P. gingivalis ATCC33277 and W83, an afimbrial strain, reporting differences in the fimbrial content of OMVs, with FimA, Mfa1, and the minor fimbrial subunits FimC–E as a differential component between these OMVs that were largely absent from W83 vesicles [25]. These fimbria components are key mediators of adhesion, and their enrichment in 33,277 vesicles is consistent with the greater efficiency of these vesicles in entering host cells compared with W83, despite similar levels of other major outer membrane and gingipain-associated proteins [26,27,28]. P. gingivalis expresses long (FimA) and short (Mfa1) fimbriae, with Mfa fimbriae forming a tip-associated complex in which accessory proteins Mfa3–5 assemble with the Mfa1 shaft [29]. Among these, Mfa5 is structurally unique, containing a von Willebrand factor (vWF) domain and being exported to the cell surface via the Type IX Secretion System (T9SS), a specialized protein export pathway in P. gingivalis that translocate virulence factors across the outer membrane [30,31]. Notably, prior work has shown that vesicles produced by strain ATCC 33277 are internalized by host cells at markedly higher levels than those from W83 or mfa-locus mutants, indicating that components of the minor fimbrial system are key determinants of vesicle–host interactions [32].

In the present study, we detected vWF exclusively in OMVs from H_2_O_2_-treated cultures, along with Mfa4 and other minor fimbrial components. Importantly, this detection occurred alongside the presence of multiple additional proteins, including FimB, hemagglutinins, and internalin-related proteins involved in adhesion and receptor engagement, as well as TonB-dependent receptors, RagA, and gingipain-associated proteins linked to heme and iron acquisition [33]. Together, this protein signature suggests that oxidative stress induces a coordinated OMV remodeling program that integrates host interaction, nutrient scavenging, and secretion functions, rather than reflecting the action of a single dominant virulence factor. The simultaneous enrichment of outer membrane β-barrel proteins, lipoproteins, and T9SS C-terminal domain-containing proteins further supports the notion that stressed OMVs undergo active surface and cargo reprogramming, favoring extracellular signaling and host modulation. In this context, Mfa5 and associated minor fimbrial proteins may cooperate with hemagglutinins and TonB-dependent receptors to engage β1-integrins and innate immune receptors on trophoblast cells, thereby priming inflammatory and autophagy-related pathways or occupying adhesion sites and limiting subsequent bacterial internalization [34,35].

Although this model is supported by the coordinated enrichment of multiple adhesion- and receptor-binding proteins in stressed OMVs and by the observed functional effects on host cells, direct mechanistic evidence for specific ligand–receptor interactions has not yet been reported. Future studies employing antibody-blocking approaches, ELISA-based binding assays, or receptor-specific signaling readouts will be required to directly validate the contribution of individual fimbrial, vWA-domain-containing, or hemagglutinin proteins to OMV–host cell interactions.

Consistent with this interpretation, vesicles derived from stressed cultures did not stimulate P. gingivalis invasion of trophoblast cells, in contrast to the pro-invasive effect of OMVs from control bacteria. Invasion is a well-established strategy for immune evasion and is promoted by fimbrial and other surface proteins that interact with β1-integrins and activate downstream signaling pathways involving calcium fluxes, cytoskeletal rearrangements, and MAPK activation [36]. Different strains of P. gingivalis have been classified as high, moderate, or low invaders, and several studies propose that intracellular localization offers P. gingivalis a protected niche. The intracellular niche would provide access to host-derived proteins and iron, thereby supporting long-term persistence while remaining largely shielded from humoral and cellular immune mechanisms and may also confer a degree of tolerance to antimicrobial therapies [37]. We propose that the oxidative pulse experienced by P. gingivalis led to smaller OMVs with a lower zeta-potential enriched with proteins such as Mfa5, Mfa4, and RagA, along with heme-acquisition and Rgp-associated factors, which may interact with β1-integrins and innate immune receptors at the host cell surface. This interaction could prime inflammatory and autophagy pathways or saturate key adhesion sites, thereby limiting subsequent bacterial internalization and shifting the pathogenic strategy from intracellular invasion toward extracellular virulence mechanisms such as immune modulation and barrier disruption. Although direct evidence for enhanced alternative pathogenic mechanisms remains to be demonstrated in future studies, this interpretation is particularly relevant at the maternal–fetal interface, where excessive inflammation and oxidative stress are hallmarks of pathological pregnancies [38,39].

Previous studies have shown that P. gingivalis OMVs from control cultures induce metabolic reprogramming of trophoblast cells toward a resting state with reduced migration and invasion, together with disruption of vascular and neutrophil regulation [17,18]. Our current findings raise the possibility that OMVs generated under oxidative stress, derived from the action of immune system and/or antibiotic treatment, instead promote an alternative metabolic and inflammatory profile, shifting trophoblast cells toward a more glycolytic and pro-inflammatory state, which may contribute to limiting bacterial persistence while potentially promoting placental dysfunction, a possibility that warrants further investigation. The fact that OMVs from anaerobic bacteria can efficiently reach the placenta even during healthy pregnancies underscores the potential clinical relevance of these OMV-mediated effects in both physiological and pathological gestations [40].

Taken together, our findings highlight oxidative stress as a critical determinant of OMV functional specialization in P. gingivalis. Future studies should therefore extend beyond invasion assays to assess how stress-conditioned OMVs influence trophoblast immune signaling, barrier integrity, metabolism, and crosstalk with decidual immune cells, providing a more comprehensive understanding of OMV-mediated pathogenic mechanisms at the maternal–fetal interface.

4. Materials and Methods

4.1. Porphyromonas gingivalis Outer Membrane Vesicle Isolation

Porphyromonas gingivalis ATCC 33277 was grown anaerobically at 37 °C in ATCC Medium 2722 (Tryptic Soy Broth 30.0 g/L and Yeast Extract 5.0 g/L) supplemented with 5% L-cysteine, 5 mg/L hemin, and 1 mg/L vitamin K1 (Sigma-Aldrich, St. Louis, MO, USA) [18]. Bacteria were cultured anaerobic conditions for a total of 5 days until it reached the early stationary phase (OD_600nm_). For the final 24 h, cultures were divided into two experimental conditions: control cultures, which continued under basal conditions, and oxidative stress-treated cultures, which were exposed to 30 mM hydrogen peroxide (H_2_O_2_, Sigma-Aldrich).

Cultures were harvested at early stationary phase (OD_600nm_), and bacterial cells were removed by two centrifugations at 6000× g for 15 min at 4 °C. The resulting supernatants, containing extracellular vesicles, were passed through a 0.22 µm membrane filter (GE Healthcare Life Sciences, Boston, MA, USA) to eliminate residual cells and debris. OMVs were pelleted by ultracentrifugation at 100,000× g for 70 min at 4 °C. The resulting vesicle-containing pellet was gently resuspended in PBS and purified by a second ultracentrifugation step at 100,000× g for 70 min at 4 °C. The final pellet was resuspended in PBS and stored at −80 °C until use. Particle-free PBS processed through the same isolation protocol was used as a negative control to exclude buffer- or system-derived particulate contamination, in line with MISEV recommendations. For SDS-PAGE analysis, equal amounts of protein were loaded (10 μg). Bands were visualized using Coomassie Blue G-250 staining. Molecular weights were estimated using the Precision Plus Protein™ molecular weight marker 10–250 kDa (Bio-Rad, Hercules, CA, USA).

4.2. Transmission Electron Microscopy

For structural characterization, OMVs were examined by transmission electron microscopy. Samples were applied onto carbon-coated copper grids, washed with PBS, and negatively stained with 1% (w/v) uranyl acetate.

4.3. NTA and Zeta Potential

The size distribution, particle concentration, and zeta potential of OMVs were analyzed using a ZetaView PMX-230 Twin Laser system (Particle Metrix, Inning am Ammersee, Germany) equipped with a 488 nm laser. Measurements were based on particle Brownian motion and light scattering detection. For nanoparticle tracking analysis, OMV samples were diluted in particle-free Milli-Q water to reach the optimal concentration range and analyzed according to the UC_EV_Scatter standard operating procedure (SOP). Data acquisition was performed at medium resolution, with 11 positions, one cycle per position, a frame rate of 30 frames per second, sensitivity set to 80, and a shutter value of 100. Zeta potential measurements were performed using the same instrument following the EVS_ZP SOP. Measurements were acquired at high resolution, using two positions, two cycles, and a frame rate of 30 frames per second, with sensitivity set to 80, shutter value of 100, and continuous pulse mode (<2 mS/cm). Zeta potential values were calculated using a class width of 2.0 mV, within a maximum range of ±200 mV. All data were processed using ZetaView software (version 8.05.16 SP7) (Particle Metrix), and reported values correspond to the mean of all recorded positions. The ZetaView is in the Instituto de Investigaciones en Microbiología y Parasitología Médica (IMPAM), Facultad de Medicina, Universidad de Buenos Aires, Argentina.

4.4. Membrane-Associated Fluorescence and Protein Concentration

The membrane-associated signal of Pg-OMVs was assessed using the lipophilic fluorescent dye FM4-64 (Invitrogen, Carlsbad, CA, USA). OMV samples were incubated in the working staining solution containing 2 µg/mL of FM4-64 and incubated 10 min before fluorescence measurement. For each condition, 200 µL of the stained samples were transferred to individual wells of a black flat-bottom 96-well plate (Greiner Bio-One, Kremsmünster, Austria). Fluorescence was measured immediately using a plate reader. Protein concentration was determined using the Pierce μBCA protein assay kit (Thermo Fisher Scientific, Waltham, MA, USA).

4.5. LC-MS Analysis

Peptide separations were carried out on a nanoHPLC Ultimate3000 (Thermo Fisher Scientific) using a C18 nano column (15 cm, Thermo Fisher Scientific). The mobile phase flow rate was 300 nL/min using 0.1% formic acid in water (solvent A) and 0.1% formic acid and 100% acetonitrile (solvent B). The gradient profile was set as follows: 4–30% solvent B for 64 min, 30–80% solvent B for 7 min and 90% solvent B for 1 min. 3 microliters of the sample were injected. MS analysis was performed using a QExactive HF mass spectrometer (Thermo Fisher Scientific). For ionization, 1.9 kV of liquid junction voltage and 300 °C capillary temperature was used. The full scan method employed a m/z 200–2500 mass selection, an Orbitrap resolution of 120,000 (at m/z 200), a target automatic gain control (AGC) value of 1 × 10^6^, and maximum injection times of 100 ms. After the survey scan, the 15 most intense precursor ions were selected for MS/MS fragmentation. Fragmentation was performed with a normalized collision energy of 27 eV and MS/MS scans were acquired with a dynamic first mass, AGC target was 5 × 10^5^, resolution of 30,000 (at m/z 200), isolation window of 1.4 m/z units and maximum IT was 55 ms. Charge state screening was enabled to reject unassigned, singly charged, and equal or more than six protonated ions. A dynamic exclusion time of 25 s was used to discriminate against previously selected ions.

MS data analysis: MS data were analyzed with Proteome Discoverer (V: 2.4.1.15) using standardized workflows. Mass spectra *.raw files were searched against a database from Porphyromonas gingivalis (UP000000588). Precursor and fragment mass tolerance were set to 10 ppm and 0.02 Da, respectively, allowing 2 missed cleavages. The following modifications were set:

Dynamic Modifications:

• -Max. Equal Modifications Per Peptide: 3.
• -Max. Dynamic Modifications Per Peptide: 4.
• -1. Dynamic Modification: Oxidation/+15.995 Da (M).

Dynamic Modifications (protein terminus):

• -1. N-Terminal Modification: Acetyl/+42.011 Da (N-Terminus).
• -2. N-Terminal Modification: Met-loss/−131.040 Da (M).
• -3. N-Terminal Modification: Met-loss + Acetyl/−89.030 Da (M).

Static Modifications:

• -1. Static Modification: Carbamidomethyl/+57.021 Da (C).

4.6. Trophoblast Cell Culture

The human first trimester trophoblast HTR-8/SVneo cell line (HTR-8) provided by Dr. Gil Mor (Wayne University, Detroit, MI, USA) was cultivated as is described in [17,18]. Briefly, cells were cultured in flasks at 37 °C, 5% CO_2_ in Dulbeco’s modified Eagle’s medium and Nutrient Mixture F-12 (DMEM-F12) medium (Gibco, Grand Island, NY, USA) supplemented with 10% heat-inactivated fetal bovine serum (FBS) (Internegocios, Buenos Aires, Argentina) and 100 μg/mL penicillin solution (Life Technologies, Carlsbad, CA, USA). Cells were subcultured upon reaching 80–90% confluence to be used in different assays.

4.7. Bacterial Invasion and Persistence Assay

The invasion assay was performed as described previously [41,42]. Briefly 5 × 10^4^ control or pretreated cells during 24 h with P. gingivalis OMV (1 μg/mL) were washed twice with phosphate-buffered saline (PBS) without any supplements prior to incubation with Porphyromonas gingivalis. Bacterial cells were cultured anaerobically for 3 days as described above. Bacterial suspension was adjusted to achieve a multiplicity of infection (MOI) of 30 and afterwards was incubated with host cells for 1.5 h for invasion or 24 h for persistence assays. Afterwards, to eliminate non-invasive Pg cells were treated with 300 μg/mL gentamicin (Invitrogen) for an additional 1.5 h. The host cells were washed twice with PBS and lysed with 0.1% Triton X-100, 0.5% trypsin, and 0.3 mg/mL DNase in PBS. Aliquots (5 μL) of several dilutions were dropped on blood agar plates (ATCC Medium 2722; Tryptic Soy Broth 30.0 g/L, Yeast Extract 5.0 g/L, and agar 15.0 g/L) supplemented with 5% L-cysteine, 5 mg/L hemin, 1 mg/L vitamin K1 and 5% ovine blood and incubated seven days under anaerobic conditions at 37 °C. The number of internalized bacteria was determined based on the calculation of bacterial colonies grown on the agar plates. Bacterial persistence was calculated as the ratio of the number of bacteria recovered after 24 h to the number of internalized bacteria.

4.8. Statistical Analysis

The data were analyzed using GraphPad Prism 9 software (GraphPad, San Diego, CA, USA). The unpaired t-test or one-way ANOVA with multiple comparisons was used for parametric analysis. Results are expressed as mean ± S.E.M. Differences between treatments were considered significant at p < 0.05.