Butyrate-treatment induces gingival epithelial cell death in a three-dimensional gingival-connective tissue hybrid co-culture system

Abstract

Three-dimensional (3D) cell culture systems are reported to be more physiologically similar to the in vivo state than 2-dimensional (2D) models, which are extensively employed in periodontal research. Herein, we developed a 3D gingival tissue model with both epithelial and lamina propria layers using human gingival epithelial Ca9-22 cells and primary gingival fibroblasts. The epithelial layer of the developed 3D gingival tissue culture was treated with butyrate, a metabolite of oral bacteria, and the treatment induced the release of damage-associated molecular patterns, such as DNA and Sin3A associated protein 130 kDa (SAP130). Taken together, butyrate exposure to the epithelium of 3D gingival epithelial-connective tissue hybrid systems could induce epithelial cell death and the subsequent release of damage-associated molecular patterns.

Introduction

Two-dimensional (2D) culture systems have long been employed to examine the pathogenic mechanisms associated with periodontal diseases. Despite numerous reports using conventional 2D culture systems, the molecular mechanisms underlying periodontal disease remain elusive. Recent advances in cell culture techniques have enabled the development of three-dimensional (3D) culture systems, potentially allowing cells to mimic the in vivo environment more accurately.¹ For developing 3D gingival culture systems, it is desirable to use primary gingival epithelial cells and fibroblasts. However, obtaining primary gingival epithelial cells that lack fibroblast contamination can be challenging. Therefore, we developed a gingival 3D culture model consisting human gingival epithelial Ca9-22 cells and primary gingival fibroblasts.

Following the development of mature dental plaque, bacteria produce high concentrations of short-chain fatty acids (SCFAs), such as butyrate.² In a previous study using a 2D culture system, treatment of gingival epithelial cells with butyrate was shown to induce cellular death, accompanied by the release of intracellular molecules called damage-associated molecular patterns (DAMPs), which can induce the production of proinflammatory cytokines.^3,4 Herein, we developed a gingival 3D culture system and examined the effects of butyrate on gingival layers.

Materials and methods

Construction of the 3D gingival tissue culture system

This study was approved by the Ethics Committee of the Nihon University School of Dentistry (#EP19D0011). After obtaining written informed consent, fragments of gingival tissue were obtained from a periodontally healthy patient (a 59-year-old male, non-smoker) during tooth extraction surgery (the left lower third molar) at Nihon University Dental Hospital. Primary gingival fibroblasts were obtained as described previously.^5,6 Briefly, the epithelial layers were separated from the lamina propria and minced lamina propria layers was placed in each well of 6-well plates. Sprouting cells from the lamina propria layers were collected as human primary gingival fibroblasts (hGFBs); we used hGFBs up to the 15th generation.⁶ hGFBs and human gingival epithelial Ca9-22 cells, which are frequently used as a counterpart of gingival epithelial cells,³ were maintained in 10% fetal bovine serum (FBS; Biowest, Nuaillé, France) Minimum Essential Medium α (MEMα; Wako, Osaka, Japan) supplemented with 1% penicillin/streptomycin (Wako). hGFBs suspended in FBS (500 μL, 2.5 × 10⁵ cells/mL) were mixed with 2.3 mL of Cellmatrix type I-A collagen sol (Nitta Gelatin, Osaka, Japan), 760 μL of 5 × Dulbecco’s modified Eagle medium (Nitta Gelatin), and 330 μL of reconstitution buffer (Nitta Gelatin, Fig. 1A, panel a). The mixture was placed in a 0.4 μm-pore-size transwell insert (for 12 well plates, Corning, Corning, NY, USA) and incubated for 30 min at 37 °C (Fig. 1A, panel b). A Ca9-22 cell suspension (1 × 10⁵ cells/mL, 300 μL) in Epilife medium (Thermo Fisher Scientific, Waltham, MA, USA) containing supplement S7 (Thermo Fisher Scientific) was placed on the gel and incubated overnight at 37 °C. MEMα containing 10% FBS was then added to the lower chambers and incubated for a 24 h to form epithelial monolayers on the collagen gels (Fig. 1A, panel c). The epithelial layers were exposed to air for 9 days by aspiration to induce epithelial stratification (Fig. 1A, panel d). During this step, the medium in the lower chambers was replaced every other day.

Figure 1.

Construction of a three-dimensional gingival tissue system. (A) Human primary gingival fibroblasts (hGFBs) suspension in collagen sol (a) was poured into transwell inserts (b). After collagen gelation, the Ca9-22 cell suspension was placed on the gel and incubated for 2 days (c). Gingival epithelial monolayers were exposed to the air for 9 days to trigger epithelial stratification (d). The multilayered epithelium was treated with butyrate-containing media (e). (B) The three-dimensional culture was fixed, embedded in paraffin, sectioned, and stained with hematoxylin and eosin. Scale bar: 100 μm. (C) Immunohistochemistry analysis of a representative section. Green: Cytokeratin. Red: Vimentin. Blue: 4′,6-diamidino-2-phenylindole counterstain.

Hematoxylin and eosin (HE) staining and immunohistochemistry

Blocks of 3D cultures were fixed in 10% formalin and embedded in paraffin. The blocks were sectioned vertically, and sections were stained with HE. Immunohistochemical triple staining was performed using an anti-human cytokeratin mouse monoclonal antibody (clone AE1/AE3, DAKO Cytomation, Glostrup, Denmark), anti-human vimentin mouse monoclonal antibody (clone V9, Leica-Novocastra, Newcastle upon Tyne, UK), and 4′,6-diamidino-2-phenylindole as a nuclear counterstain.

Stimulation with butyrate in 3D and 2D culture systems

Sodium butyrate (Wako) was diluted in 1%FBS Roswell Park Memorial Institute 1640 (RPMI1640, Wako). Sodium-butyrate-containing media (0, 2.5, 5, 10, 15, and 20 mM butyrate; 600 μL) was added to the upper chamber of the co-cultures (Fig. 1A, panel e). After incubation for 0, 24, 48, and 72 h at 37 °C, conditioned media in the upper chambers were collected and centrifuged at 3000 rpm to remove cellular debris. Supernatants (550 μL) were collected and used for DNA analysis and western blotting. Ca9-22 cell 2D cultures were also treated with or without 20 mM butyrate. After incubation, collected supernatants were used for DNA analysis.

Measurement of released DNA

Aliquots of cell culture supernatant (100 μL/well) were placed in a 96-well black plate (Greiner Bio-One, Kremsmünster, Austria), and 100 μL/well of 400 nM SYTOX-green dye (Thermo Fisher Scientific) in RPMI1640 was added. Fluorescence intensity (Ex/Em = 485/535 nm) was measured using a Wallac ArvoSX1420 spectrofluorometer (PerkinElmer, Waltham, MA, USA).

Western blot

Aliquots of cell culture supernatant (20 μL/lane) were subjected to western blotting using an anti-Sin3A associated protein 130 kDa (SAP130) rabbit polyclonal antibody (GeneTex, Irvine, CA, USA) as the primary antibody and WestVision secondary antibody labeled with horseradish peroxidase (Vector Laboratories, Burlingame, CA, USA). Clarity Western ECL substrate (Bio-Rad, Hercules, CA, USA) and ChemiDoc XRS (Bio-Rad) were used for visualization.

Statistical analysis

Each experiment was repeated more than twice, and similar results were obtained from each independent experiment. Non-normal distribution of data was determined using the Shapiro–Wilk test. The Kruskal–Wallis test, followed by the Steel test, was used to examine differences. P-values < 0.01 were considered statistically significant.

Results

Based on our HE staining we detected the construction of a stratified epithelium on collagen gel containing fibroblasts (Fig. 1B). Immunohistochemical staining showed that cytokeratin, an epithelial marker, was present only in the stratified epithelial layer (Fig. 1C, upper left panel). Vimentin, a fibroblast marker, was detected only in the lamina propria layer (Fig. 1C, upper right panel). These findings demonstrated that gingival epithelial cells and fibroblasts did not invade each other (Fig. 1C, lower right panel).

A high butyrate concentration can affect the gingival epithelium at the mature dental plaque in close contact with gingival epithelial cells. Thus, we examined the effect of butyrate on the epithelial layers of 3D gingival cultures. Treatment of epithelial layers with butyrate induced DNA release in a time-dependent manner, and the amount of released DNA was significantly higher than that induced by control treatment (Fig. 2A). In addition, butyrate induced DNA release in a dose-dependent manner (Fig. 2B). Furthermore, butyrate induced SAP130 release in a dose-dependent manner (Fig. 2C). To compare the effects of butyrate treatment between 2D and 3D systems, the time-course of DNA release from the 2D culture during butyrate treatment was examined. Treating epithelial layers of the gingival 3D model with butyrate time-dependently increased DNA release until 72 h (Fig. 2A). Although the amount of DNA released from the 2D system during butyrate treatment increased in a time dependent manner until 48 h (Fig. 2D), it tended to decrease after 48 h (Fig. 2D).

Figure 2.

Butyrate treatment induces the release of cellular components from gingival epithelial layers of the 3D gingival culture system. (A) Time-course of the amount of DNA released during butyrate treatment. The amount of DNA in the conditioned media was measured using SYTOX-green dye (n = 4, ∗, P < 0.01 compared to the control treatment). (B) Butyrate treatment for 72 h induces dose-dependent DNA release from epithelial cells of the 3D gingival culture system (n = 4, ∗, P < 0.01 compared to control treatment). (C) Western blot analysis using conditioned media showed that butyrate treatment for 72 h induced SAP130 release from the epithelial layer of the 3D gingival culture system. (D) Time-course of the amount of DNA released during butyrate treatment of the 2D Ca9-22 cell culture (n = 4, ∗, P < 0.01 compared to control treatment).

Discussion

In the present study, we developed a 3D gingival culture system consisting of both stratified gingival epithelium and lamina propria layers. As shown in Fig. 1B and C, Ca9-22 cells were not observed in the lamina propria layer of the 3D culture system, and fibroblasts were not detected in epithelial layers. In addition, this system allows gingival epithelial cells to obtain nutrients via the lamina propria. Therefore, this double-layered tissue system is anatomically similar to the actual gingival tissue. However, in our system, we used a gingival epithelial Ca9-22 cell line as a source of gingival epithelial cells. Although the cell line is easy to handle, cell characteristics, such as cell shape and growth speed, occasionally differ from those of primary gingival epithelial cells. We attempted to obtain primary gingival epithelial cells from clinical samples as described previously,^5,6 gingival epithelial cells failed to sprout from pieces of clinically obtained gingival epithelium and sometimes dead during maintenance. The condition of the clinically obtained tissues, coverslip placement, passage timing, or other unidentified reasons could underlie these failures. Xiao et al. have developed a similar 3D model. The authors used the immortalized human skin epidermal keratinocyte cell line HaCaT, which reportedly differentiates into stratification layers.^7,8 However, this cell line is aneuploid and not derived from gingival tissue.⁷

Although the development of a 3D culture system using primary gingival epithelial cells is anticipated, a realistic approach for continuous use of a human gingival 3D culture model could be the model developed in the present study or that of Xiao et al.

Similar to previous reports that used 2D gingival culture systems,⁴ treating the epithelial layer of the 3D culture systems with butyrate could induce release of intracellular molecules such as DNA and SAP130 (Fig. 2). Although butyrate treatment of 3D epithelial layers resulted in an increased release of DNA over time until 72 h (Fig. 2A), the amount of released DNA decreased after 48 h in the 2D model (Fig. 2D). These observations could be attributed to the following reasons: (i) enzymes that the 2D model may release cleave DNA after 48 h, (ii) the 3D model may not release DNA digesting enzymes, and (iii) the 3D culture may release molecules that inhibit DNA digestion.

DAMPs released following butyrate treatment can bind to receptors on surrounding cells and induce the production of proinflammatory cytokines.⁹ This indicates that SCFA production from mature dental plaque bacteria may be one of the most important inducers of gingivitis. If molecular mechanisms controlling DAMP release after or during SCFA action can be established, effective methods to suppress periodontal disease can be developed. Therefore, a more complex 3D gingival culture system, which has blood vessels in the lamina propria and so on, is required to evaluate the precise mechanism underlying the development of gingivitis.

Declaration of competing interest

The authors declare no conflict of interest associated with this manuscript.