Development of a Co-culture Model for Assessing Competing Host Mammalian Cell and Bacterial Attachment on Zirconia versus Titanium

Abstract

Objectives:

Co-culture models are limited by bacteria rapidly outcompeting host mammalian cells for nutrients in vitro, resulting in mammalian cell death. The goal of this study was to develop a co-culture model enabling survival of mammalian cells and oral bacterial species to assess their competition for growth on dental implant materials.

Methods:

Two early-colonizing oral bacterial species, Streptococcus mutans or Actinomyces naeslundii, were grown in co-culture with primary human macrophages or human gingival fibroblasts for up to 7 days on tissue-culture treated polystyrene or polished titanium and zirconia disks. Chloramphenicol was supplemented in cell culture medium at bacteriostatic concentrations to maintain stable bacterial inoculum size. Planktonic and adherent bacterial growth was assessed via spot plating while mammalian cell growth and attachment were evaluated using colorimetric metabolic assay and confocal fluorescence microscopy, respectively.

Results:

Macrophages and fibroblasts proliferated in the presence of S. mutans and maintained viability above 70% during co-culture for up to 7 days on tissue-culture treated polystyrene and polished titanium and zirconia. In contrast, both mammalian cell types exhibited decreasing proliferation and surface coverage on titanium and zirconia over time in co-culture with A. naeslundii versus control. S. mutans and A. naeslundii were maintained within an order of magnitude of seeding inoculum sizes throughout co-culture.

Significance:

Cell culture medium supplemented with antibiotics at bacteriostatic concentrations can suppress bacterial overgrowth and facilitate mammalian cell viability in co-culture model systems. Within the study’s limitations, oral bacteria and mammalian cell growth in co-culture are comparable on polished titanium and zirconia surfaces.

Graphical Abstract

1. Introduction

Dental implant systems are medical devices that are highly successful in restoring missing and damaged teeth.¹ The placement of dental implant systems is becoming more common, with approximately 800,000 dental implants being placed annually in the United States.² Despite recent improvements in dental implant design focusing on structural function and mechanical strength, approximately 3.6% of dental implant systems fail after 10 years post-implantation due to a variety of complications, such as bacterial infection, excessive loading, improper placement, and impaired healing.^1,3 Bacterial infection and plaque are key etiological factors contributing to implant failure and complications at both early and late stages.^3–6 Bacterial biofilm can form on dental implant surfaces and become established within the peri-implant pocket, triggering an inflammatory cascade that results in soft tissue and bone loss (peri-implantitis) and subsequent implant failure.^6–8 In the case of early stage complications, oral bacteria such as Streptococcus spp. and some Actinomyces spp., typically present in the peri-implant sulcus, can colonize the implant surface and impede soft tissue seal formation.^4,9–13 The lack of this barrier allows for bacteria to penetrate deeper into underlying alveolar bone, which can adversely impact implant integration with surrounding tissue.^14–16 Similarly, in late stage complications, oral bacteria can penetrate the soft tissue seal and cause gradual tissue necrosis at the implant site, causing eventual loss of the entire implant system.^9,14,17

During normal wound healing, initial blood coagulation around a dental implant results in a provisional extracellular matrix (ECM) containing proteins, primarily fibrin, and inflammatory cells like neutrophils.^18–21 The acute inflammatory phase then begins after neutrophils and monocytes are recruited to neutralize invading oral bacteria.¹⁸ Pro-inflammatory (M1 phase) migrating or monocyte-derived macrophages clear the wound of bacteria and cellular debris.¹⁹ Macrophages transition from their M1 phase to an anti-inflammatory (M2) phenotype to resolve inflammation, secreting growth factors that promote the formation of granulation tissue.^19,22 Simultaneously, fibroblasts migrate to and multiply at the wound site during the proliferation phase, degrade the temporary fibrin matrix, and deposit new ECM components, primarily type III collagen.^18–20,22 In the final phase of wound healing, both soft tissue and bone remodeling occurs. Before undergoing apoptosis, activated macrophages and fibroblasts release matrix metalloproteinases, which change type III collagen to type I collagen, resulting in a mature peri-implant mucosal seal.^19,20,22 Meanwhile, mature osteoblasts (osteocytes) and osteoclasts replace and degrade woven bone, respectively, resulting in structured, load-bearing lamellar bone.²³

Normal wound healing around a dental implant may not be resolved under circumstances of bacterial challenge. Streptococcus spp. comprise approximately 60–80% of the bacterial species in dental biofilms.²⁴ Streptococcus mutans is a gram-positive, coccus bacteria that exists as a part of normal oral flora and biofilms associated with periodontal disease.^25,26 As an opportunistic pathogen, S. mutans tends to have a higher rate of sucrose metabolism and production of acids and glycans.²⁵ Moreover, growth of S. mutans is favored when soft tissue seal formation around the implant surface fails due to microaerophilic species thriving in deep pockets with low oxygen tension.²⁶ Actinomyces spp. comprise approximately one-third of bacterial species during early dental biofilm and plaque formation.²⁴ Actinomyces naeslundii is a gram-positive, fimbriated bacteria that coaggregrates with Streptococcus spp. to develop mutualistic growth and stabilize early biofilm communities.²⁴ Thus, A. naeslundii has often been associated with ecological balance and healthy oral biofilm^24,27,28. This species is involved with buffering of local pH by neutralizing acid byproducts created by other bacterial species in the biofilm.^27,29,30 Constant acidic conditions can develop into pathogenic conditions, such as periodontitis or peri-implantitis, causing A. naeslundii to express a more pathogenic phenotype.²⁷. The role of A. naeslundii in the development of pathogenic conditions is still being investigated.²⁷ However, like S. mutans, A. naeslundii can be characterized as an opportunistic pathogen due to expression of this pathogenic phenotype.²⁵ Commensal bacteria expressing pathogenic phenotypes can induce and sustain the M1 phenotype in macrophages, triggering a chronic inflammatory response and, therefore, host tissue damage.³¹ Furthermore, oral pathogens can suppress gingival fibroblast immunomodulatory function and macrophage response to bacterial challenge.^32,33 In the oral cavity, both host mammalian cells and bacteria must compete in a “race for the surface,” and it is critical that host tissue cells win this race against oral bacteria for long-term success and integration with the implant surface.³⁴

Another factor which can influence host tissue growth and oral bacterial colonization is implant surface condition.^9,14,35 Due to its mechanical strength and osseointegrative properties, titanium (Ti) has traditionally been used to manufacture dental implant materials.^35,36 Ti is conducive to various surface treatments that stimulate soft tissue seal formation and osseointegration.^9,14,35,37 At the same time, these surface modifications may encourage oral bacterial adhesion, potentially promoting biofilm formation.^35,38,39 Smooth Ti surfaces favor soft tissue cell attachment over oral bacterial adhesion, which is ideal on the abutment and implant collar where the “race for the surface” first begins.^9,37,40 Although Ti-based dental implants have many advantages, oral bacteria and their metabolites can corrode the protective passive oxide layer (TiO₂) providing Ti biocompatibility.^35,36,41,42 The buildup of Ti corrosion byproducts within host tissue can exacerbate the inflammatory response and lead to implant failure.⁴² The bioceramic zirconia (ZrO₂) has been proposed as an alternative material to overcome limitations associated with Ti.^41,43 ZrO₂ surfaces have shown less biofilm formation than Ti ones due to their lower surface free energy.^41,44 Also, host cells like osteoblasts had greater viability on ZrO₂ than Ti after bacterial challenge in vitro.¹⁴

Although oral bacteria and mammalian cells compete against each other in vivo, many in vitro studies have evaluated bacterial and mammalian cell response to surfaces separately.^44–51 Thus, in vitro co-culture models that can simulate in vivo competition between bacteria and host cells are needed to assess the biological response to dental implant materials. Previous in vitro co-culture models have advanced the understanding of interactions between bacteria and mammalian cells.^{14,34,35,46,52–58} However, the inability of mammalian cells to survive beyond short time periods (4–72 h) was a major limitation of these previous models. Due to the relatively fast growth rate of bacteria, the microenvironment rapidly changes within 24 h, resulting in limited availability of nutrients and the production of organic acidic wastes.⁵⁷ Host mammalian cells quickly die under these conditions, which prevents long-term evaluation of mammalian cell proliferation under bacterial challenge. Even when fluid flow conditions or bioreactors are used to replenish nutrients and remove metabolic waste, bacteria can quickly outcompete mammalian cells for attachment due to excess growth nutrients, which are generally not readily available in the oral cavity (saliva).^34,52,57 To suppress bacterial growth, a previous study utilized spectinomycin as a bacteriostatic agent in cell culture medium, allowing for mammalian cell survival up to 48 h.⁵³ However, no other studies have used this approach to evaluate mammalian and bacterial co-culture growth to date. Thus, the goal of the present study was to develop an in vitro co-culture model using a bacteriostatic agent to assess survival of early-colonizing oral bacteria and host mammalian cells involved in the early tissue healing response and inflammation post implantation.

2. Materials and Methods

2.1. Specimen Preparation

Disk-shaped specimens of commercially pure Ti and partially yttria-stabilized ZrO₂ (Ø 9.5 mm × 6 mm) were surface-treated by sequential polishing until mirror-like finish with 50 nm alumina (Pace Technologies) to resemble smooth implant surfaces (e.g., abutment and collar) and cleaned via ultrasonication and dried in an oven at 60 °C prior to testing as previously described.⁵⁹

2.2. Development of Mammalian and Bacterial Co-culture Model

Streptococcus mutans UA159 (ATCC 700610) and Actinomyces naeslundii (ATCC 12014) were routinely grown in Brain Heart Infusion (BHI) medium supplemented with 5 mg/mL hemin and 0.1 mg/mL vitamin K1. Primary human macrophages (Celprogen) and human gingival fibroblasts (Lifeline Cell Technology) were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM) supplemented with 10% fetal bovine serum (FBS) as previously described.^4,59 Incubation for all mono- and co-culture experiments was carried out in 5% CO₂ at 37 °C. To prepare co-culture, overnight cultures of S. mutans or A. naeslundii were grown to late log/early stationary phase. For S. mutans, optical density readings at 600 nm wavelength (OD₆₀₀) readings were recorded and used to dilute to low (2 × 10⁴ colony-forming units (CFU)/mL) or high (2 × 10⁷ CFU/mL) bacterial loads per previous methods.⁴ S. mutans culture was diluted in DMEM supplemented with chloramphenicol (CAM) at 5 or 5.5 μg/mL CAM for low or high bacterial loads, respectively. CAM was selected amongst other bacteriostatic agents for modified co-culture medium based on preliminary testing demonstrating stable bacterial counts over 7 days for several early-colonizing species without severely reducing host mammalian cell proliferation (≥ 70% viability; data not shown). Ultraviolet-visible spectroscopy was used to periodically monitor and account for minor differences between CAM stock concentrations (nominal: 100 μg/mL); co-culture medium with CAM was freshly made just prior to starting an in vitro test. Mammalian cells were also diluted in co-culture medium based on cell type and growth period being tested as specified in Table 1. Different seeding amounts were chosen to ensure (i) that mammalian cell counts were not too low for metabolic assay absorbance measurements at the end of each growth period and (ii) not too high resulting in saturation of metabolic activity assay readings. Macrophages were diluted to 5, 1, or 0.1 × 10⁴ cells/mL while fibroblasts were diluted to 2, 1, or 0.2 × 10⁴ cells/mL for 1-, 3- or 7-day proliferation tests, respectively. Afterward, 625 μL of diluted mammalian and bacterial cell culture were each placed into individual wells of 24-well plates, yielding a working volume of 1.25 mL (1.25 × 10⁴ and 1.25 × 10⁷ CFU for low and high bacterial loads, respectively). For 3- and 7-day experiments, 50% (625 μL) of cell culture medium was replaced every 24 h by gently mixing (4×) to remove ~50% of planktonic bacteria and replaced with 650 μL of fresh medium (to account for evaporation and handling). Wells containing only mammalian or bacterial cell monocultures served as positive controls. Three independent biological trials were performed with four technical replicates per test group. Separate co-culture wells were used for assessing bacterial adhesion and mammalian cell viability. Culture wells with and without bacteria were grown in separate 24-well plates to prevent cross-contamination during medium changes.

2.3. Bacterial and Mammalian Co-culture Preparation on Ti and ZrO₂ Specimens

After establishing a co-culture model within 24-well plates, co-culture experiments were repeated to assess bacterial adhesion and mammalian cell proliferation on Ti and ZrO₂ disk specimens. For co-culture experiments on disks, a single intermediate load (1 × 10⁶ CFU/mL) for S. mutans and A. naeslundii was chosen after preliminary testing (data not shown). S. mutans was grown in DMEM with 5 μg/mL CAM while A. naeslundii was grown in DMEM with 2 μg/mL CAM. Mammalian cells were seeded at different amounts based on cell type and growth period. Mammalian and bacterial monocultures were prepared per Section 2.2. Macrophages were seeded at 12500 and 2500 cells/disk for 1- and 3-day tests while fibroblasts were seeded at 5000, 2500, and 500 cells/disk for 1-, 3-, and 7-day tests, respectively. Bacterial culture was diluted to 2 × 10⁷ CFU/mL in co-culture medium. Macrophages were diluted to 20, 4 or 0.4 × 10⁴ cells/mL while fibroblasts were diluted to 8, 4, or 0.8 × 10⁴ cells/mL for 1-, 3- or 7-day tests, respectively. Polished Ti and ZrO₂ disks were then placed in a 24-well plate with 4 disks (technical replicates) per test group. 62.5 μL of mammalian and bacterial culture dilutions were directly seeded on disks, yielding a 125 μL droplet of cell culture suspension covering the entire disk surface (containing 1 × 10⁷ CFU/mL bacteria). Specimens were carefully transferred and incubated at 37 °C and 5% CO₂ for 1 or 2 h to allow for macrophage or fibroblast attachment, respectively. After incubation, 1.125 mL of co-culture medium was added to each of the wells (diluting initial bacterial inoculum to 1 × 10⁶ CFU/mL), collapsing the cell culture medium droplets on disks. Specimens immersed in plain medium served as negative controls while disks seeded with only bacteria or only mammalian cells served as positive controls. Half of the well volume (625 μL) was replaced with fresh medium (650 μL) every 24 h, and both the well and disk surfaces were each gently mixed (4×) by pipetting to suspend and remove ~50% of planktonic/loosely attached bacteria.

2.4. Evaluation of Bacterial and Mammalian Cell Growth

Both planktonic and adherent bacterial counts were quantified for co-culture experiments on disks (planktonic bacteria only for tests in tissue culture-treated polystyrene wells) after 1, 3, or 7 days of growth. Planktonic bacteria were aliquoted from each culture well after gently mixing the well and disks surfaces (4× each) with pipetting. Adherent bacteria (disks only) were defined as those which remained after gentle mixing and washing (1×) in 1.5 mL of 1× phosphate-buffered saline (PBS). To detach adherent bacteria, disks were aseptically transferred to glass test tubes and incubated for 30 min at 37 °C and 5% CO₂ in 1.5 mL of 0.08% saponin and 0.05% trypsin-EDTA solution (to detach and lyse mammalian cells) followed by 5 min of ultrasonication. Preliminary tests ensured that ultrasonication in saponin + trypsin-EDTA solution did not significantly affect bacterial viability and counts (data not shown). After ultrasonication, planktonic and adherent bacteria were diluted 10-fold in PBS (50/450 μL) to mitigate bacteriostatic/bactericidal effects of CAM and saponin + trypsin-EDTA solution on bacterial growth when plated on BHI agar. For low S. mutans loads (1 × 10⁴ CFU/mL), 100 μL of diluted bacterial suspension was plated directly on BHI agar. For high S. mutans loads (1 × 10⁷ CFU/mL), bacterial suspensions were further serially diluted (up to 1000-fold) in 96-well plates (20/180 μL) prior to plating 10 μL on BHI agar. For the intermediate bacterial load (1 × 10⁶ CFU/mL) test on disk specimens, 10 μL of ten-fold serial dilutions of S. mutans or A. naeslundii were plated on BHI or supplemented Brucella blood (5% defibrinated sheep blood, 10 mg/mL hemin, 10 mg/mL vitamin K1) agar, respectively. After incubating agar plates for 48–72 h at 5% CO₂ and 37 °C, bacterial counts were enumerated and used to calculate planktonic and adherent bacteria CFU counts for mono- and co-culture samples.

To assess potential intracellular invasion of macrophages and fibroblasts which could interfere with mammalian cell viability measurements, co-culture tests were prepared using low (1 × 10⁴ CFU/mL) and high (1 × 10⁷ CFU/mL) bacterial loads for S. mutans or intermediate load (1 × 10⁶ CFU/mL) for A. naeslundii in 24-well plates. Planktonic/loosely adherent bacteria were quantified as described previously. To quantify intracellular bacteria, co-culture samples were incubated with DMEM supplemented with 200 μg/mL of gentamicin or 300 μg/mL of metronidazole for 4 h to neutralize extracellular S. mutans or A. naeslundii, respectively.⁶⁰ Co-culture medium post-antibiotic treatment was collected in addition to any intracellular bacteria obtained after washing (3×) with PBS and lysing mammalian cells with 250 μL of 0.08% saponin and 0.05% trypsin-EDTA solution. After 30 min incubation at 37 °C and 5% CO₂, all bacteria samples (pre-antibiotic, post-antibiotic, and intracellular) were diluted 10-fold in PBS (50/450 μL) and plated directly (low bacterial loads) or were serially diluted (up to 1000-fold) in 96-well plates (20/180 μL) and plated.

Quantitative evaluation of mammalian cell proliferation in co-culture with bacteria was performed using a colorimetric assay (R&D Systems) for metabolic activity. Wells and disk specimens were washed gently mixed (4× each) prior to washing with 1.5 mL of PBS. 1.5 mL of DMEM containing 200 μg/mL gentamicin, 300 μg/mL metronidazole, and 10 μg/mL penicillin G was added to each well and incubated in 37 °C and 5% CO₂ for 24 h. The antibiotics and incubation time were chosen based on previous antibiotic protection assays and preliminary testing to ensure neutralization of planktonic and adherent bacteria which could interfere with the metabolic activity assay for mammalian cell growth.^61,62 After incubation, all specimens were washed with 1.5 mL of PBS, and disks were aseptically transferred to individual wells in a 48-well plate. Next, 200 μL of DMEM containing 1% penicillin-streptomycin was added to each well, and 20 μL of 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide (MTT) reagent was added. The plates were then incubated for 2 h in the dark at 37 °C and 5% CO₂. Afterward, 200 μL of detergent was added to all wells and incubated overnight at 37 °C and 5% CO₂ to solubilize the reduced product (formazan). Lastly, 200 μL of solubilized suspension was transferred carefully to 96-well plates, and absorbance measurements were taken at 570 nm using an automated plate reader (Synergy Mx, BioTek). Cell viability was calculated by subtracting the average blank values (negative control) from absorbance reads and normalizing values against the positive control (mammalian cells only) set to 100% to allow for comparison across different co-culture growth periods.

2.5. Confocal Imaging of Mammalian Cell Attachment in Co-culture

Macrophage and fibroblast growth and morphology were qualitatively assessed under mono- and co-culture conditions (Sections 2.2 and 2.3) in polystyrene wells or on disks. After 1, 3 or 7 days, samples in 24-wells or Ti and ZrO₂ disks were gently mixed (4× each) and washed twice with 1.5 mL of PBS. Mammalian cells were fixed with 4% paraformaldehyde for 30 min followed by two more PBS washes. The nuclei and F-actin of mammalian cells were stained using 0.3 μM of 4′,6-diamidino-2-phenylindole (DAPI, Invitrogen) and Alexa Fluor 488 dye (Invitrogen) for 30 min, respectively. All samples were washed once and maintained in 1.5 mL of PBS prior to fluorescent imaging using a confocal laser scanning microscope (Olympus FV3000RS) at 20× magnification.

2.6. Statistical Analysis

Two-way analysis of variance (ANOVA) followed by post hoc Tukey tests were performed at a significance level of 95% (α = 0.05). For co-culture experiments in 24-wells (no disks), bacterial counts for low and high bacterial loads were each analyzed across two factors: growth time (1, 3 or 7 days) and culture condition (mono- vs. co-culture). Mammalian cell viability was analyzed across two factors: growth time and bacterial load (low vs. high). For co-culture experiments with disks, planktonic and adherent bacterial counts were analyzed across two factors: growth time and substrate + culture condition (Ti vs. ZrO₂ + mono- vs. co-culture). Mammalian cell viability was analyzed across two factors: growth time and substrate (Ti vs. ZrO₂). For bacterial invasion tests, statistical analysis was performed separately for growth time periods across two factors: bacterial load and bacterial samples (pre-antibiotic planktonic, post-antibiotic planktonic, and intracellular).

3. Results

3.1. Mammalian and Bacterial Co-culture Growth

Bacterial counts of S. mutans in mono- and co-culture with macrophages are shown in Fig. 1 A and B. For S. mutans seeded at a low density (10⁴ CFU/mL), S. mutans counts in mono- and co-culture with macrophages remained statistically the same across 1, 3 and 7 days of growth (Fig. 1 A). At higher seeding density (10⁷ CFU/mL), S. mutans counts decreased over time (Fig. 1 B): S. mutans in co-culture at 3 days was significantly lower than corresponding 1-day time group, and both 7-day mono- and co-culture groups significantly decreased by 10-fold vs. 1- and 3 days groups (p < 0.05). Macrophage viability in co-culture with low S. mutans load was comparable to macrophage monoculture viability (100%) after 1 day of growth, significantly decreased after 3 days, and recovered by 7 days (Fig. 1 C). In contrast, macrophage viability in co-culture with high S. mutans load significantly increased (~2-fold) after 7 days vs. all other test groups (p < 0.05). Preliminary invasion tests confirmed that S. mutans did not significantly invade macrophages as intracellular counts were near the limit of detection and/or at least 2 orders of magnitude lower than extracellular bacterial counts (Fig. S1).

Figure 1.

Bacterial count of S. mutans in mono- and co-culture with human macrophages under (A) low (10⁴ CFU/ml) or (B) high (10⁷ CFU/mL) bacterial loads and (C) cell viability (%) of macrophages in co-culture with low and high loads after 1, 3, or 7 days of growth in 24-well plates. The black dotted horizontal line denotes (A-B) initial S. mutans seeding amounts and (C) normalized viability (100%) of macrophage monoculture control groups at each time point (n = 3). Significant difference is denoted as follows: * versus test groups at all other time points. ^# versus corresponding 1-day group. ^& versus all other test groups. ^$ versus 1-day high S. mutans load group (α = 0.05).

Bacterial counts of S. mutans in mono- and co-culture with fibroblasts under low and high bacterial loads are shown in Fig. 2 A and B. At a low load (10⁴ CFU/mL), mono- and co-culture S. mutans counts at 3 and 7 days were significantly lower than corresponding 1-day groups (Fig. 2 A), but by no more than ~10-fold (p < 0.05). At a high load (10⁷ CFU/mL), no significant difference between S. mutans count in mono- vs. co-culture was observed within each time point (Fig. 2 B). However, S. mutans counts significantly decreased between 1 and 3 days and between 3 and 7 days (p < 0.05). Fibroblast viability (Fig. 2 C) at all time points and S. mutans loads were lower on average than monoculture (100%), excluding 1-day co-culture group under high S. mutans load. However, fibroblast viability remained above 70% for all test groups. Preliminary tests confirmed minimal S. mutans invasion of fibroblasts with intracellular counts near the limit of detection and/or at least 2 orders of magnitude lower than extracellular bacterial counts (Fig. S2).

Figure 2.

Bacterial count of S. mutans in mono- and co-culture with human gingival fibroblasts under (A) low (10⁴ CFU/ml) or (B) high (10⁷ CFU/mL) bacterial loads and (C) cell viability (%) of fibroblasts in co-culture with low and high loads after 1, 3, or 7 days of growth in 24-well plates. The black dotted horizontal line denotes (A-B) initial S. mutans seeding amounts and (C) normalized viability (100%) of fibroblast monoculture control groups at each time point (n = 3). Significant difference is denoted as follows: * versus test groups at all other time points. ^& versus all other test groups. ^{^} versus corresponding 3- and 7-day groups (α = 0.05).

3.2. Mammalian and Bacterial Co-culture Growth on Ti and ZrO₂

Planktonic and adherent S. mutans counts in mono- vs. co-culture with macrophages after 1 and 3 days are shown in Fig. 3 A and B. In general, planktonic bacterial counts remained within 10-fold of initial seeding amount (1.25 × 10⁶ CFU/disk) while adherent bacterial counts were about 10- to 50-fold lower than corresponding planktonic groups after 1 and 3 days. However, after 3 days, planktonic S. mutans count (Fig. 3 A) in mono-and co-culture on ZrO₂ were significantly lower (~5-fold) than 3-day Ti and 1-day groups (p < 0.05). Correspondingly, adherent S. mutans counts (Fig. 3 B) in co-culture with macrophages on ZrO₂ were on average higher than other 3-day groups (p > 0.05). Macrophage viability (Fig. 3 C) in mono- and co-culture on Ti and ZrO₂ were statistically similar after both 1 and 3 days (p > 0.05) and were only marginally higher than macrophage monoculture control (100%).

Figure 3.

(A) Planktonic and (B) adherent bacterial counts of S. mutans in mono- and co-culture with human macrophages and (C) cell viability (%) of macrophages in co-culture with S. mutans after 1 or 3 days of growth on Ti versus ZrO₂ disks. The black dotted horizontal line denotes (A) initial S. mutans seeding amount (10⁶ CFU/mL) and (C) normalized viability (100%) of macrophage monoculture control groups at each time point (n = 3). * denotes significant difference versus test groups at all other time points (α = 0.05).

Planktonic and adherent S. mutans counts in mono- vs. co-culture with fibroblasts after 1, 3, and 7 days are depicted in Fig. 4 A and B. In general, adherent S. mutans counts were ~10 to 100-fold lower than corresponding planktonic counts. Planktonic S. mutans counts (Fig. 4 A) were significantly higher (~5-fold) after 3 days vs. 1- and 7-day groups (p < 0.05) but decreased after 7 days for all groups, which was significant for monoculture Ti and ZrO₂ groups (p < 0.05). For adherent S. mutans counts (Fig. 4 B), Also, 7-day adherent S. mutans counts in mono- and co-culture on ZrO₂ were higher (~5-fold) than 7-day Ti groups (Fig. 4B). In contrast, adherent S. mutans counts were the lowest in monoculture on polished Ti after 3 and 7 days. In Fig. 4 C, fibroblast viability in mono- and co-culture was not significantly different on Ti and ZrO₂ across all timepoints (p > 0.05) and remained above 70% across all groups.

Figure 4.

(A) Planktonic and (B) adherent bacterial counts of S. mutans in mono- and co-culture with human gingival fibroblasts and (C) cell viability (%) of fibroblasts in co-culture with S. mutans after 1, 3 or 7 days of growth. The black dotted horizontal line denotes (A) initial S. mutans seeding amount (10⁶ CFU/mL) and (C) normalized viability (100%) of fibroblast monoculture control groups at each time point (n = 3). Significant difference is denoted as follows: * versus test groups at all other time points. ^# versus 3-day S. mutans + Ti group. ^& versus 3- and 7-day S. mutans + Ti groups. (α = 0.05).

Planktonic A. naeslundii counts were statistically similar after 1 day in mono- or co-culture with macrophages (Fig. 5) or fibroblasts (Fig. S4) but generally decreased (≤ 10-fold) after 3-days. Adherent A. naeslundii counts were lower than corresponding planktonic counts and also decreased by 3 days of growth (Figs. 5 and S4 B). In general, adherent A. naeslundii counts were statistically similar in mono- and co-culture with macrophages or fibroblasts within each timepoint. Macrophage and fibroblast exhibited decreasing viability (< 70%) after 1- and 3-day co-culture with A. naeslundii (Figs. 5 and S4 C) vs. control (100%) and to a significant extent for 3-day fibroblast co-culture (Fig. S4 C).

Figure 5.

(A) Planktonic and (B) adherent bacterial counts of A. naeslundii in mono- and co-culture with human macrophages and (C) cell viability (%) of macrophages in co-culture with S. mutans after 1 or 3 days of growth on Ti versus ZrO₂ disks. The black dotted horizontal line denotes (A) initial A. naeslundii seeding amount (10⁶ CFU/mL) and (C) normalized viability (100%) of macrophage monoculture control groups at each time point (n = 3). Significant difference is denoted as follows: * versus test groups at all other time points. ^$ versus corresponding 1-day timepoint. ^% versus A. naeslundii + Ti and A. naeslundii + Macrophages + ZrO₂ group (α = 0.05).

3.3. Mammalian Cell Attachment in Mono- and Co-culture

Confocal images of macrophages in mono- and co-culture with S. mutans on tissue-culture treated polystyrene (24-wells) and on polished Ti and ZrO₂ disks are shown in Fig. 6. At all timepoints, macrophages exhibited a round morphology across all groups with no apparent visual differences between mono- and co-culture groups. After 1 day of growth, macrophages in both mono- and co-culture doubled in number across all substrates. By 3 days, macrophages had experienced 3–4 population doublings across all groups and surfaces. Additionally, clusters of adherent S. mutans colonies (tiny scattered blue dots) were observed on ZrO₂ in co-culture with macrophages. By 7 days, macrophages reached nearly complete confluency with no qualitative difference in cellular morphology or number between mono vs. co-culture groups on all substrates.

Figure 6.

Confocal fluorescent images of human macrophages in mono- or co-culture with S. mutans after 1, 3 or 7 days of growth (20× magnification). Stained nuclei (mammalian and any bacteria) appear as blue while stained F-actin appear as green. Scale (red bar): 100 μm.

Confocal images of fibroblasts in mono- and co-culture with S. mutans on tissue-culture treated polystyrene (24-wells), Ti, and ZrO₂ are depicted in Fig. 7. Overall, fibroblasts presented characteristic spindle-shaped morphology across all test groups and timepoints. After 1 day of growth, fibroblast morphology appeared disorganized with relatively low cell count. Furthermore, larger clusters and streaking patterns of S. mutans colonies were observed in co-culture on polished Ti while a more scattered, smaller S. mutans colonies were visible on polished ZrO₂. By 3 days, fibroblasts increased several folds in number but still appeared disorganized. Qualitatively, fibroblast organization appeared more aligned and structured for fibroblast monoculture groups as compared to co-culture groups. Among co-culture groups, less adherent S. mutans colonies were visible on Ti after 3 days, but these colonies appeared to be larger. On the other hand, smaller and more scattered colonies were observed on ZrO₂ surfaces after 3 days, similar to their corresponding 1-day group. By 7 days, fibroblasts reached nearly complete confluency and exhibited organized structures with most fibroblasts aligned along a single axis on all substrates. In co-culture, larger clusters of S. mutans colonies were visible on both Ti and ZrO₂ after 7 days.

Figure 7.

Confocal fluorescent images of human gingival fibroblasts in mono- or co-culture with S. mutans after 1, 3 or 7 days of growth (20× magnification). Stained nuclei (mammalian and any bacteria) appear as blue while stained F-actin appear as green. Scale (red bar): 100 μm.

Confocal images of macrophages or fibroblasts in mono- and co-culture with A. naeslundii on tissue-culture treated polystyrene (24-wells) and on polished Ti and ZrO₂ disks are shown in Figs. 8 and S5. As observed for mono- and co-culture with S. mutans, macrophages appeared round in morphology (Fig. 8) while fibroblasts exhibited characteristic spindle-like shape (Fig. S5). Macrophage and fibroblast monoculture appeared to proliferate to a greater extent by 3 days of growth vs. macrophages and fibroblasts in co-culture with A. naeslundii which appeared to be less dense and fewer in number on 24-wells and on Ti and ZrO₂. In contrast to S. mutans, stained A. naeslundii colonies were only occasionally observed and appeared as a faint biofilm in regions devoid of macrophages or fibroblasts (red arrow in Fig. 8).

Figure 8.

Confocal fluorescent images of human macrophages in mono- or co-culture with A. naeslundii after 1, 3 or 7 days of growth (20× magnification). The red arrow depicts faint A. naeslundii biofilm. Stained nuclei (mammalian and any bacteria) appear as blue while stained F-actin appear as green. Scale (red bar): 100 μm.

4. Discussion

In the present study, a co-culture model in which both mammalian cells and oral bacteria could not only survive but proliferate up to 7 days on Ti and ZrO₂ surfaces in vitro was accomplished. Specifically, human macrophages or gingival fibroblasts were simultaneously grown with early colonizers, S. mutans or A. naeslundii. Chloramphenicol (CAM), an antibiotic with bacteriostatic activity due to its reversible binding to bacterial 50S ribosomal unit,⁶³ was added in cell culture medium (DMEM) to prevent S. mutans and A. naeslundii overgrowth. Among several classes of antibiotics tested (data not shown), CAM (1–6 μg/mL) had a bacteriostatic effect on a wide range of S. mutans and A. naeslundii loads (10⁴-10⁷ CFU/mL). In these preliminary tests, S. mutans and A. naeslundii monocultures persisted at stable planktonic loads (10^±1 CFU/mL) for at least 7 days in DMEM supplemented with CAM while performing 50% daily medium changes every 24 h. Planktonic bacterial counts increased until culture medium was acidified (yellow medium color) when using low CAM concentrations (< 4 μg/mL for S. mutans and < 2 μg/mL for A. naeslundii) or excluding daily 50% medium changes. In contrast, higher CAM concentrations (> 6 μg/mL for S. mutans and > 3 μg/mL for A. naeslundii) were bactericidal and/or inhibited recovery of bacteria to its initial inoculum size along with daily medium changes. Moreover, smaller inoculum sizes (10⁴ CFU/mL) were more sensitive to CAM at longer time points (7 days) than larger ones (10⁷ CFU/mL). Thus, 5 μg/mL of CAM in DMEM was established for S. mutans inoculum sizes ranging from 10⁴-10⁶ CFU/mL while 5.5 μg/mL of CAM in DMEM was used for 10⁷ CFU/mL. For an intermediate load of 10⁶ CFU/mL for A. naeslundii, 2 μg/mL CAM in DMEM was established for use in co-culture.

Co-culture tests with S. mutans or A. naeslundii vs. macrophages or fibroblasts were performed after establishing use of CAM in co-culture medium. The number of bacteria in a sample could be quantified by plating on agar without interference of mammalian cells. However, mammalian cell viability was evaluated using a metabolic assay in which the reagent (MTT) could be metabolized by both mammalian cells and bacteria, which would overestimate mammalian cell viability in co-culture. Therefore, prior to the metabolic assay, all samples were incubated with gentamicin, metronidazole, and penicillin G at bactericidal concentrations. Incubation for 24 h with antibiotic-supplemented DMEM was optimal for neutralizing bacteria in tissue culture-treated polystyrene wells. Another potential limitation in determining mammalian cell viability was the potential for intracellular bacterial invasion. Invasion tests were performed in co-culture with macrophages or fibroblasts for 1, 3, or 7 days to account for potential intracellular invasion by S. mutans or A. naeslundii. After neutralizing extracellular bacteria with gentamicin, mammalian cells were detached and lysed using 0.05% trypsin-EDTA + 0.08% saponin per protocols in the literature.^64,65 Saponin served as a mild detergent capable of lysing mammalian cells while maintaining gram positive (S. mutans) and gram negative (e.g., Aggregatibacter actinomycetemcomitans) bacteria viability.⁶⁴ In the end, S. mutans or A. naeslundii invasion of both mammalian cell lines was minimal, near the limit of detection in most cases, and several orders of magnitude lower than planktonic bacterial counts for all bacterial load sizes (Figs. S1–3). Thus, these preliminary tests confirmed that MTT assays quantified only mammalian cell metabolic activity in co-culture after neutralizing bacteria under co-culture growth test conditions for up to 7 days.

After validating the use of antibiotic-supplemented medium to neutralize bacteria and quantify mammalian cell viability, co-culture tests between S. mutans and macrophages or fibroblasts were performed for 1, 3 or 7 days under low (10⁴ CFU/mL) or high (10⁷ CFU/mL) bacterial loads. As shown in Figs. 1 and 2, S. mutans counts under mono- or co-culture conditions were maintained near initial inoculum sizes after 1 day but gradually declined by 7 days, which was significantly lower in some cases between corresponding 1-, 3-, and 7-day groups (p < 0.05). However, this decrease was no larger than 10-fold. For macrophages (Fig. 1 C), no significant difference in cell viability was observed between mono- and co-culture conditions after 1 and 3 days (p > 0.05). On the other hand, after 7 days, macrophage viability in co-culture under high bacterial loads was nearly double that of monoculture and co-culture under low S. mutans load. The relatively high multiplicity of infection (MOI) by S. mutans could have activated macrophages and acted as a positive mitogenic stimulus on macrophage growth over longer (7 day) time points.^66,67 Previous studies demonstrated that human macrophages can recognize S. mutans surface lipoproteins, and a greater pro-inflammatory macrophage profile was observed when in contact with S. mutans.^68,69 For co-culture of fibroblasts with S. mutans (Fig. 2 C), average cell viability was marginally lower than monoculture and decreased to a greater extent under higher S. mutans load by 7 days but remained above the 70% threshold for standard cytotoxicity tests per ISO-10993. The greater susceptibility of fibroblasts to S. mutans in co-culture correlates with findings of a previous study in which fibroblasts were susceptible to Streptococcus oralis biofilms in co-culture after 4 h, resulting in proinflammatory cytokine release.⁵² However, further experimentation to assess gene expression and inflammatory response are required to validate changes observed in macrophage and fibroblast viability in the present study.

After using tissue culture-treated polystyrene, co-culture experiments were performed on polished Ti and ZrO₂ disks. A polished surface was chosen to mimic the smooth implant abutment and collar surfaces on which early-colonizing bacteria compete with immune and soft tissue cells for attachment. As illustrated in Figs. 3–5, planktonic and adherent S. mutans and A. naeslundii counts in co-culture with macrophages or fibroblasts were generally equal to monoculture on Ti and ZrO₂ within each timepoint. Over time, some planktonic bacterial counts did significantly fluctuate (decrease with macrophages, increase with fibroblasts), but the change was not more than 10-fold. In contrast, adherent S. mutans count on ZrO₂ in Fig. 4 (mono- and co-culture with fibroblasts) was 5-fold higher than Ti after 7 days. Corroborating this trend, larger clusters of adherent S. mutans colonies were visible after 7-day co-culture with fibroblasts on ZrO₂ and Ti (Fig. 7). Interestingly, this behavior may be attributed to qualitative differences in attachment on Ti vs. ZrO₂ as S. mutans formed larger adherent clusters on polished Ti initially (1 day) but was smaller and more uniformly dispersed on ZrO₂. Bacterial attachment is governed by surface properties like surface free energy (wettability), roughness, and electrostatic charge.⁷⁰ Although Ti and ZrO₂ have oxide surfaces, Ti has a more negatively charged surface due to its metallic bulk composition that attracts Ca²⁺ ions and subsequent Ca-mediated attachment of bacteria, which have negatively charged cell walls.⁷¹ In contrast, ZrO₂ has a more neutral surface charge, which may explain its uniform dispersion of S. mutans attachment after 1 day. A recent study also observed no significant difference in S. mutans attachment after 1 day on polished Ti and ZrO₂ with surface roughness and wettability similar to disks used in present study.⁷¹ However, over longer growth periods (3 and 7 days), ZrO₂ retained more adherent S. mutans (< 10-fold) vs. polished Ti which had the lowest average adherent S. mutans count (Fig. 4). Likewise, Zhao et al. also demonstrated that polished Ti under fluid flow exhibits the least amount of adherent bacteria.⁹ Additional tests including zeta potential measurement will elucidate differences in initial bacterial adhesion on ZrO₂ versus Ti.

In contrast to S. mutans, adherent A. naeslundii counts declined over time (Figs. 5 B and S4 B) in addition to planktonic counts. Though an initial colonizer, A. naeslundii thrives in biofilms through interactions and metabolization of byproducts from other bacterial colonizers.^27,29,30,72 Also, A. naeslundii expresses higher growth rates and pathogenic phenotype under acidic conditions.^27,29,30,72 Therefore, the decrease of adherent A. naeslundii counts by 3 days (Fig. 5B) could be due to poor biofilm growth under the present culture conditions. However, A. naeslundii attachment was still observed during confocal fluorescent imaging (Fig. 8).

For mammalian cellular proliferation on Ti or ZrO₂ in co-culture with S. mutans, no significant difference was observed between macrophage or fibroblast proliferation on polished Ti vs. polished ZrO₂ (Figs. 3 C and 4 C). Specifically, under suppression of bacterial growth, both Ti and ZrO₂ surface yielded comparable macrophage and fibroblast proliferation. In contrast, Zhao et al. demonstrated that human gingival fibroblasts could only win the “race for the surface” on smooth Ti surfaces when co-cultured with Streptococcus and Staphylococcus spp. under parallel plate fluid flow conditions.⁹ Thus, the low surface roughness of polished Ti and ZrO₂ in this study was expected to reduce bacterial adherence while facilitating mammalian cell attachment, a function crucial for soft tissue seal formation on smooth dental implant collars and abutments. As opposed to co-culture with S. mutans, macrophage and fibroblast viability declined vs. monoculture when exposed to A. naeslundii on both Ti and ZrO₂ surfaces and decreased over time (Figs. 5 C and S4 C). This behavior could be attributed to virulence factors specific to A. naeslundii, which persisted despite decreasing adherent A. naeslundii counts after 1 and 3 days on Ti and ZrO₂. For example, peptidoglycan of A. naeslundii was shown previously to stimulate pro-inflammatory cell response and activate osteoclasts⁷³ and will be investigated future studies. On the other hand, Zhao et al. demonstrated greater U2OS osteoblast cell survival and attachment on ZrO₂ surfaces vs. Ti variants under challenge by anaerobic pathogens (static fluid).¹⁴ Thus, differences in mammalian cell growth on Ti and ZrO₂ between this and previous co-culture models can be attributed to use of other mammalian cell lines, bacterial species, and growth conditions.

Previous co-culture models characterized the competition between bacterial and mammalian cells on biomaterial surfaces in vitro.^{9,14,34,35,46,52–58} However, a key limitation of these models was the time scale of growth (0.5–96 h), which may not be sufficient for studying processes such as wound healing or bacterial biofilm formation, which takes place during the span of days to weeks. Thus, prior co-culture models focused only on initial adhesion of mammalian cells and oral bacteria on surface-treated and coated Ti, ZrO₂, and other biomaterials. Previous strategies to increase the duration of co-culture growth included using smaller inoculum sizes (10³-10⁵ CFU/mL)^34,54–56 and/or fluid flow conditions to reduce exposure time to bacterial metabolites.^9,34 Nevertheless, the potential for bacterial overgrowth in vitro and adverse changes in the microenvironment leading to mammalian cell death prevented study of these co-culture systems under longer time periods.⁵⁶ To suppress bacterial growth, antibiotics within bacteriostatic concentrations, such as CAM in this study or spectinomycin in a previous study, allowed for assessment of bacterial and mammalian cell co-culture proliferation at longer time scales.^53,57 Additionally, some co-culture systems supplemented mammalian culture medium with bacterial culture medium like 2–10% BHI, Todd Hewitt Broth (THB) or Tryptic Soy broth (TSB) for growth fastidious species (e.g., Prevotella intermedia and Porphyromonas gingivalis).^14,34,54,55 However, exposure to these medium supplements, as low as 5%, significantly reduced macrophage and fibroblast viability (< 70%) after 3 days in our preliminary tests (data not shown). Also, the use of anaerobic conditions, while necessary for obligate anaerobe (e.g., P. gingivalis) survival, in several co-culture models eventually led to mammalian cell death (≤ 96 h).^14,46 Although these models are used to study the initial interaction between pathogenic anaerobes and mammalian cells on dental biomaterials, the apoptotic mammalian cell profile is not representative of in vivo outcome in which host cells overcome bacterial challenge and integrate with the implant surface. Finally, prior co-culture models have relied on imaging, fluorescent staining, and counting methods to elucidate mammalian cell proliferation due to presumed interference of bacteria with quantitative metabolic assays.^9,14,54 Although imaging and cell counting methods can facilitate studies on mammalian cellular attachment, these methods may not accurately represent cell viability based on cellular morphology (i.e., dead or dying cells may appear attached on a surface and counted as viable). In the present study, this obstacle was overcome by neutralizing bacteria by incubation with antibiotic-supplemented culture medium prior to assessing mammalian cell activity.

Overall, a co-culture model was developed which simulated the outcome in which mammalian cells proliferate on dental implant surfaces even under constant bacterial challenge. Despite its success, several limitations must be addressed. The present study utilized a single bacterial species in co-culture which is not representative of oral microbial diversity. Thus, future studies should utilize multispecies and/or biofilms as explored previously to better simulate bacterial challenge to host tissue integration with the dental implant surface in vivo.^52,58 Secondly, bacteriostatic suppression of bacterial growth may inadvertently induce stress and alter bacterial gene expression regulating metabolism and adhesion. For S. mutans and A. naeslundii, no apparent differences in colony morphology after agar plating were observed between CAM and CAM-free medium, but potential changes in gene expression need to be further investigated. Furthermore, only mammalian cell viability was investigated in the present work. Future mechanistic studies will assess mammalian cell gene expression and phenotypic changes such as macrophage polarization under bacterial challenge via quantitative PCR and confocal fluorescent imaging. Moreover, DAPI staining in the present study enabled visualization of mammalian cellular and bacterial attachment on each substrate but complicated quantification of bacterial vs. mammalian cell coverage. Future work will utilize fluorescence in situ hybridization (FISH) staining to assess bacterial coverage under co-culture conditions. Finally, the present study was limited to assessing polished surfaces. Future studies with this co-culture model will elucidate inflammatory cell response on rough surfaces using our novel ionic liquid-based coating to promote immunomodulation of Ti implants for constructive host tissue healing under bacterial challenge and diabetic conditions.

5. Conclusion

Overall, an oral bacterial and mammalian co-culture model was successfully developed and used to study cellular growth on clinically relevant Ti and ZrO₂ surfaces. In general, no significant differences in S. mutans planktonic growth and adhesion or mammalian cell viability were observed on Ti vs. ZrO₂. However, average S. mutans counts were marginally higher on ZrO₂ for mono- and co-culture with fibroblasts after 7 days. In contrast, polished Ti had the lowest average adherent S. mutans count in monoculture after 7 days. A. naeslundii counts were generally comparable on Ti vs. ZrO₂ after 1 and 3 days of growth but demonstrated decreased planktonic and adherent counts over time under mono- and co-culture conditions. Macrophage and fibroblast viability remained relatively high (> 70%) when co-cultured with S. mutans at all time points but declined (< 70%) in co-culture with A. naeslundii. Correspondingly, macrophages and fibroblasts were observed via confocal microscopy to proliferate equally well on Ti and ZrO₂, despite the presence of adherent clusters of S. mutans colonies, especially after 7 days. On the other hand, fewer macrophages and fibroblasts were observed attached to Ti or ZrO₂ surfaces in co-culture with A naeslundii as compared to monoculture. Thus, within the limitations of the present study, ZrO₂ is a suitable alternative to Ti in terms of bacterial and mammalian cell co-culture growth.

Supplementary Material

The following figures in the Supporting Information file are available free of charge:

• Figures S1-S3: Planktonic (pre- and post-antibiotic treatment) and intracellular bacterial counts of S. mutans and A. naeslundii in mono- and co-culture with macrophages or fibroblasts.

• Figure S4: Planktonic and adherent A. naeslundii counts of in mono- or co-culture with fibroblasts on Ti and ZrO₂ after 1 or 3 days of growth.

• Figure S5: Confocal fluorescent images of human gingival fibroblasts in mono- or co-culture with A. naeslundii after up to 7 days of growth.