Enhancing epithelial tissue sealing of titanium alloy implants through CaCl₂-based hydrothermal treatment

Dental implant treatment is widely used as an effective prosthodontic option. However, although implants offer long-term stability and favorable functionality, they also come with the risk of complications.¹ When patients fail to clean their dental implant prosthesis, plaque accumulation would occur.² It has been reported that tissue destruction occurs when plaque accumulates around an implant.³ One review indicated that peri-implant mucositis affected approximately 80% of the studied subjects, while peri-implantitis was observed in 28%–56%.⁴ It has also been reported that peri-implantitis and mechanical overload can contribute to the loss of an implant that has been functioning for many years.⁵ Therefore, dental implants that can stand against bacterial infection would benefit a lot.

Pure titanium has been widely used in dental implants for over half a century due to its excellent biocompatibility, corrosion resistance, and mechanical strength.⁶ However, reported cases of abutment and implant fracture⁷ have necessitated the exploration of higher-strength metals. In this context, materials like Ti–6Al–4V (Ti64), a titanium alloy containing 6% aluminum and 4% vanadium, have garnered considerable attention. Ti64 exhibits superior tensile strength and hardness compared to pure titanium. While Ti64 has been used as a material for artificial hip replacements in orthopedics for many years, there has been insufficient biological evaluation of Ti64, especially for the peri-implant soft tissue.

A peri-implant soft tissue is characterized by supracrestal tissue attachment, closely resembling the characteristics of epithelial and connective tissues found around natural teeth. Laminin-332 (Ln) is one of the main components of hemidesmosomes in the basement membrane between an implant and the epithelial tissue.⁸ The peri-implant tissue is more susceptible to external stimuli due to its weaker epithelial seal, resulting from a longer, scar-like supracrestal tissue attachment compared to natural teeth.^9,10 Rough surfaces have been shown to promote osseointegration, thereby facilitating osteoblast adhesion and differentiation.¹¹ However, rough surfaces exhibit poor adhesion of soft tissue¹² and tend to easily accumulate biofilms,¹³ potentially leading to more aggressive progression of peri-implantitis compared with the case for a smooth surface.¹⁴ As a result, it is common to use smooth surfaces for the transmucosal part of implants, which directly contacts the soft tissue. Although establishing stable soft tissue attachment is important, few products have been developed for improving epithelial tissue sealing.

We reported that Ti64 treated with distilled water (DW) under hydrothermal conditions can improve epithelial cell attachment.¹⁵ Hydrothermal treatment is known to induce various reactions that do not occur at atmospheric pressure. Surface contamination of Ti64 plates can be removed by hydrothermal treatment with minimal changes to surface topography. After the hydrothermal treatment, Ti64 plates showed increased hydrophilicity and Ln adsorption. Additionally, Ti can be chemically modified¹⁶ using calcium (Ca), which is known to enhance cell adhesion.¹⁷ Presence of Ca on the surface of pure titanium after hydrothermal treatment with calcium chloride (CaCl₂) has been reported.¹⁸ However, it remains unclear how Ca bonds with titanium surface. Therefore, the purpose of this study is to clarify the oxidation state and chemical coordination of Ca on Ti64 after the hydrothermal treatment. We also evaluate the effects of hydrothermal treatment on the peri-implant epithelial tissue.

A diameter of 2 mm and a length of 4.5 mm (2.5 mm transmucosal and 2 mm intrabony experimental screw-shaped implants) were used for the in vivo experiments. Machined Ti64 plates, 15 mm in diameter and 1 mm thick, with flat surfaces were used for in vitro experiments. The plates were polished with silicon carbide abrasive (Grit 2000, Struers ApS, Ballerup, Denmark). Subsequently, the plates were cleaned using 100% acetone, 99.5% ethanol, and then DW, in an ultrasonic bath. The plates were sterilized with γ-rays after cleaning. Hydrothermal treatment was then performed using either DW (denoted as HT-DW) or CaCl₂ solution (denoted as HT-Ca). For the HT-DW group, a Ti64 plate immersed in DW (15 ml) was heated at 200 °C for 24 h in a hydrothermal autoclave (HU-50, SAN-AI Kagaku, Nagoya, Japan). For HT-Ca, a Ti64 plate immersed in an aqueous solution of CaCl₂ (15 ml, 10 mmol/l) was heated at 200 °C for 24 h in a hydrothermal autoclave.^18,19 The control group (denoted as Cont) was stored in a vacuum desiccator after washing and sterilization. After hydrothermal processing, each specimen was washed with DW for 10 s and then stored in a desiccator under vacuum conditions to prevent surface contamination.

Ti64 plates with flat surfaces (length: 30 mm, width: 5 mm, and height: 1 mm) were used for x-ray absorption fine structure (XAFS) measurements. These plates were subjected to the same treatment as those used for in vitro experiments, as described above. Except for titanium foil, all reference samples were prepared as pellets using powder samples of calcium chloride dihydrate (CaCl₂⋅2H₂O), anhydrous calcium chloride (CaCl₂), calcium titanate (CaTiO₃), calcium oxide (CaO), titanium (III) oxide (Ti₂O₃), rutile titanium (IV) oxide (rutile TiO₂), and anatase titanium (IV) oxide (anatase TiO₂). Each reference powder sample was mixed with boron nitride (BN) powder, which is commonly used as a binder in XAFS experiments and then pelletized using a 10 mm KBr pellet die (PT-10, JASCO, Tokyo, Japan). The appropriate powder sample weight for making the pellets was calculated using the SAMPLEM4M tool²⁰ after fixing the thickness of BN powder at 0.2 and 0.4 mm for Ca K-edge and Ti K-edge measurements, respectively, to ensure that the pellets had thicknesses suitable for XAFS measurements.

The surface form and structure of the transmucosal area of the implants was probed using a scanning electron microscope operated at 15 kV accelerating voltage and magnification of 50× and 3000× (SEM; S-3400N, Hitachi High-Technologies, Tokyo, Japan). Surface roughness (Ra) together with maximum roughness height (Rt) were measured using a stylus profilometer (SV-3100, Mitutoyo, Kanagawa, Japan). Five random spots were chosen per sample, and two samples were measured per group (i.e., for each group n = 10). The concentration of calcium on the HT-Ca implant was determined using x-ray photoelectron spectroscopy (XPS; K-alpha, Thermo Fisher Scientific, East Grinstead, UK). To investigate the oxidation state, symmetry, and coordination chemistry of Ca and Ti in HT-Ca, XAFS measurements were conducted at the Ca K-edge and the Ti K-edge using both surface-sensitive conversion electron yield XAFS (CEY-XAFS) and bulk-sensitive transmission XAFS techniques. XAFS measurements were recorded at Kyushu University Beamline (SAGA Light Source/BL06, Saga, Japan). XAFS spectra can be divided into two regions: (i) x-ray absorption near-edge structure (XANES), which provides information about the electronic and the geometric structure of the absorbing element, and (ii) extended x-ray absorption fine structure (EXAFS), which provides information about the bond lengths and the coordination number of the neighboring atoms. The regions of focus were the vicinity of the absorption edges ranging from 4034 to 4068 eV for the Ca K-edge and 4960 to 5020 eV for the Ti K-edge, which are the XANES regions for Ca and Ti, respectively. In the case of the Ti K-edge, the EXAFS region was also studied. All reference samples were measured using transmission XAFS at the Ca K-edge for pellet samples of CaCl₂ dihydrate, anhydrous CaCl₂, CaTiO₃, and CaO and at the Ti K-edge for Ti foil, Ti₂O₃ pellets, rutile TiO₂ pellets, anatase TiO₂ pellets, and CaTiO₃ pellets. In contrast, the HT-Ca, Cont, and HT-DW flat plate samples were measured in the CEY-XAFS mode. ATHENA software was used to analyze the XAFS data.²¹ 

Animals were kept in accordance with the ethical guidelines for animal care as stipulated by the Animal Experiment Committee, Kyushu University (Approval No. A19-385-0). Oral implants were placed as described previously.²² Thirty-six four-week-old male Wistar rats (80–90 g) from Kyudo (Saga, Japan) were used [18 for immunohistochemistry, 18 for horseradish peroxidase (HRP) infusion]. The maxillary right first molar of each rat was extracted under systemic and local anesthesia using an intraperitoneal injection of a triple-combination anesthetic (10 ml/kg) consisting of medetomidine hydrochloride (0.15 mg/kg), midazolam (2 mg/kg), and butorphanol tartrate (2.5 mg/kg). Two weeks later, an implant cavity of 1.2 mm diameter was formed using a dental reamer (#80-#120, Torpan, Maillefer, Ballaigues, Switzerland). The experimental screw-shaped implant was screwed into the implant cavity such that the platform of the implant matched the level of the gingival edge. Following the surgical procedures, each rat was administered buprenorphine (0.05 mg/kg intramuscularly) for postoperative pain management. The experimental rats were housed in a temperature-controlled environment with ad libitum food and water until euthanasia, which was performed four weeks post-implantation [Fig. 1(a)].

To assess the strength of peri-implant epithelial sealing, four weeks after implantation, we topically applied 50 mg/ml HRP (Type 11, molecular weight ∼40 000 Da; Sigma-Aldrich, St. Louis, MO) around the implant following our reported procedure.²² Under systemic anesthesia, 50 mg/ml HRP solution was periodically dripped onto cotton balls placed on the gingival margin around the implant body every 10 min for 60 min [Fig. 1(b)].

Peri-implant tissues were prepared according to our previously published methods.²² Four weeks post-implantation, the rats were deeply anesthetized and fixed via intracardiac perfusion with heparinized phosphate-buffered saline (PBS, pH 7.4) followed by 4% paraformaldehyde (pH 7.4). The maxillae were demineralized in a 5% solution of ethylenediaminetetraacetic acid at 4 °C for four days, followed by removal of the peri-implant soft tissue or oral mucosa from the bone, implant, or tooth. The samples were immersed in a 20% sucrose solution at 4 °C overnight, followed by embedding in an optimal cutting temperature compound (O.C.T. Compound, Sakura Finetek, Tokyo, Japan). 10 μm thick coronal sections were then obtained using a cryostat (CM-1860, Leica Biosystems, Nußloch, Germany) at −20 °C.

Immunohistochemical staining was performed, on sections, using the avidin-biotin complex (ABC; Nichirei Bioscience, Tokyo, Japan) procedure, as previously described.²³ Sections treated with 10% normal goat serum for 30 min were incubated overnight at 4 °C with a rabbit polyclonal anti-rat Ln antibody (1:100 dilution, Clone2778, Scripps Research Institute, La Jolla, CA), followed by incubation with biotinylated goat anti-rabbit IgG (Nichirei Bioscience) for 10 min, and then by the ABC reagent for 60 min. Ln localization was visualized using a peroxidase staining 3,3′-diaminobenzidine (DAB) kit (Nacalai Tesque, Kyoto, Japan). After lightly counterstaining with hematoxylin (Muto Pure Chemicals, Tokyo, Japan), the sections were visualized using a fluorescence microscope (BZ-9000, Keyence, Osaka, Japan).

The localization of HRP using the DAB method has been described previously.²² In this study, 0.01M PBS was used to wash the sections, which were then incubated in the peroxidase-staining DAB kit (Nacalai Tesque) at room temperature (RT) for 30 s, followed by light counterstaining with hematoxylin.

The frozen section located at the center of the mesiodistal implant was studied using images taken by the light microscope (BZ-9000). The depth of HRP penetration was quantified by measuring the distance between the top of the peri-implant sulcular epithelium and the bottom of the HRP-stained area.

For in vitro studies, the GE1 mouse-derived gingival epithelial cell line,²⁴ sourced from the RIKEN Cell Bank in Tsukuba, Japan, was utilized. A mixture of 10 ng/ml mouse epidermal growth factor (Corning, NY) and 1% fetal bovine serum (Biowest, Nuaillé, France) in a basal serum-free medium (SFM-101, Nissui, Tokyo, Japan) was used for culturing the GE1 cells, in a humidified atmosphere with 5% carbon dioxide at 33 °C.

The adhesion strength of GE1 cells was determined as reported previously.²⁵ Cell culture (1 ml) was seeded on each Ti64 plate in a 24-well culture plate (Multiwell 24 well, Corning), at a density of 5 × 10⁴ cells per well. After 1 day of culturing, a cell counting kit (Cell Count Reagent SF, Nacalai Tesque) was used to determine the number of attached cells, representing the initial cell count. A spectrophotometer (NJ-2300, Biotech, Tokyo, Japan) was used to measure the absorbance at 450 nm. Subsequently, the plates were shaken for 10 min at 160 rpm six times on a rotary shaker (NX-20, Nissin, Tokyo, Japan) to remove any weakly attached cells. The remaining adhered cells were then quantified using the cell counting kit. The cell adhesion ratio was defined as the percentage remaining after shaking, relative to the initial cell count (n = 6 for each group).

4% paraformaldehyde (Merck, Darmstadt, Germany) was used to fix the one day cultured GE1 cells on the plates, blocked with 1% bovine serum albumin (Bovine Serum Albumin Fraction V, Roche Diagnostics, Basel, Switzerland) for 30 min at RT. The specimen was then incubated overnight at 4 °C with 1:100 dilution of mouse anti-rat Ln monoclonal antibody (P3H9-2, Santa Cruz Biotechnology, Dallas, TX). After washing with PBS for 5 min three times, the cells were labeled for 2 h at RT using a 1:100 dilution of fluorescein isothiocyanate-conjugated goat anti-mouse IgG secondary antibody (Invitrogen, Portland, OR). 1:100 dilution of tetramethylrhodamine isothiocyanate-conjugated phalloidin (Sigma-Aldrich) was used for staining the actin filaments for 1 h at RT. An anti-fade reagent containing 4′,6-diamidino-2-phenylindole (Vectashield, Vector Laboratories, Burlingame, CA) was then used to mount the cells for nuclear staining. Then, observation of the stained cells was performed using a fluorescence microscope (BZ-9000).

The mean ± standard deviation (SD) of the data was determined. One-way analysis of variance (ANOVA) and Tukey’s methods were utilized to compare multiple groups and identify pairwise differences. Values of p < 0.05 were considered statistically significant.

Photographs revealed that the HT-DW implants were gold and the HT-Ca implants were purple [Fig. 2(a)]. Statistical analysis revealed no significant differences in the Ra and Rt values among the various groups (Table I). SEM images of the transmucosal region of the implants at 3000× were similar for all groups, and the grooves of mechanical polishing were visible [Fig. 2(b)]. Ca was detected on the surface of the HT-Ca implant in both XPS and CEY-XANES Ca K-edge measurements [Figs. 3(a) and 3(b), respectively]. The Ca concentration calculated from the XPS results was 0.22 at. %. The CEY-XANES spectrum of HT-Ca showed a pre-edge peak at 4040.7 eV, similar to those of anhydrous CaCl₂ and CaTiO₃. The features on the rising edge at 4044.7 keV for HT-Ca resembled those detected for anhydrous CaCl₂ and CaTiO₃. Furthermore, the main peak at 4050.5 eV for the HT-Ca matched that of anhydrous CaCl₂. The Ti K-edge XANES spectrum for HT-Ca [Fig. 3(c)] shows a pre-edge at 4967.0 eV, similar to those detected for Cont, HT-DW, and Ti foil. There was a positive shift of the rising edge in the Ti K-edge spectra for both HT-Ca and HT-DW, which was attributed to an increase in the oxidation state of Ti. Meanwhile, the main peak at 4987.1 eV observed for HT-Ca coincided with those for Ti₂O₃, rutile TiO₂, and anatase TiO₂, suggesting that HT-Ca was oxidized. Figure 3(d) shows the EXAFS results for the implants, which revealed the radial distance of neighboring atoms from the Ti atoms. The peak at 1.5 Å was attributed to Ti–O coordination, indicating that HT-Ca was oxidized, which agrees with the XANES data presented above. The diffuse peak at 3.4 Å (also attributed to Ti–O) observed for HT-Ca was consistent with the presence of CaTiO₃, which was likely on the surface of HT-Ca. These results revealed that the surface of the hydrothermally treated implants was oxidized. Moreover, Ca existed on the surface of HT-Ca, probably mostly in the form of anhydrous CaCl₂ and CaTiO₃ structures.

The Ln expression, an index of epithelium adhesion, was observed as a band at the interface between the implant and the peri-implant epithelium (PIE) for all groups (Fig. 4). For the HT-Ca group, Ln-positive staining was apparent throughout the implant–PIE interface, similar to the junctional epithelium around a natural tooth (Nt), with strong Ln expression in the lower to the upper third of the pocket and into deep layers. However, in the Cont and HT-DW groups, Ln expression was confined to the lower and surface layers of the epithelial interface.

Four weeks after implantation, the implants were exposed to HRP as a foreign substance. After HRP exposure, HRP was found on the surface of the PIE [Fig. 5(a)]. The penetration depth of HRP into the epithelial structure around the implants was decreased by hydrothermal treatment, with HT-Ca showing the smallest HRP penetration depth and Cont showing the largest of the three implants [Fig. 5(b)].

The cell adhesion ratio for GE1 cells on control and hydrothermally treated Ti64 plates was greater for the HT-Ca and HT-DW groups than for the Cont group. Statistical analysis revealed no significant differences in cell adhesion ratios between the HT-Ca and HT-DW groups (Fig. 6).

GE1 cells were cultured on Ti64 plates for one day and observed with a fluorescence microscope. The immunoreactive expression is shown in Fig. 7. Actin filaments at the intracellular margin of the cells were observed for all groups. Signals for Ln in the Cont group were weaker than those for the hydrothermally treated groups. Additionally, a stronger signal for Ln was observed at the edge of the cells for the HT-Ca group.

Several studies have proven that the biologic width of epithelial and connective tissue seals around implants shares common features with the epithelial and connective tissues of natural teeth.^8,10 Epithelial cells that constitute the PIE adhere to the adjacent cells or the extracellular matrix (ECM) through proteins on the cell membrane, which are called adhesion molecules.²⁶ The basement membrane of the ECM of epithelial cells exists at the interface between tissues and is composed of collagen IV, proteoglycan, and Ln.²⁷ Ln in the basement membrane is involved in cell–cell adhesion and is known to have Ca²⁺-binding properties.²⁸ 

Despite the presence of a basement membrane-like structure between an implant and a PIE, it is reported to be a scar-like tissue.⁹ The accumulation of plaque can easily cause the destruction of tissue structures.³ It is believed that the peri-implant soft tissue exhibits inferior resistance to foreign factors compared with that of the natural gingivae.^29,30 Therefore, improving the soft tissue sealing of implants should help to lower the risk of infection.³¹ Hydrothermal treatment with CaCl₂ was reported to improve the epithelial sealing of a pure titanium implant.³² 

Pure titanium has been widely used as an implant material because it is biocompatible and supports the proliferation and adhesion of a variety of cell types.³³ Furthermore, pure titanium implants exhibit high corrosion resistance.³⁴ Titanium is classified as commercially pure titanium and titanium alloys. Ti64, a titanium alloy, has higher tensile strength and hardness than pure titanium. In situations where high strength, hardness, and biocompatibility are required, titanium alloys are often used.³⁵ Its high hardness prevents abutments in transmucosal areas from becoming worn or rough surfaces from long-term brushing. Increasing risk of fracture in abutments and implant bodies over time has been reported;⁷ so Ti64 has been used in applications where high strength is necessary.

In the present study, our aim was to investigate whether Ti64 implants subjected to hydrothermal treatment with DW or CaCl₂ solution would improve PIE tissue sealing. We previously reported that hydrothermal treatment of a Ti64 plate with DW increased the adsorption of Ln and improved the adhesion of epithelial cells.¹⁵ It is also known that hydrothermal treatment of anatase TiO₂ induces oxidative decomposition of surface hydrocarbons to form hydroxyl groups with high water compatibility; i.e., hydrothermal treatment increases the hydrophilicity of titanium surfaces.³⁶ 

Surface analysis of the transmucosal part of the implants revealed that hydrothermal treatment with CaCl₂ solution deposited Ca on the implant surface with minimal change to surface topography (Table I). Although XPS measurements confirmed the presence of Ca, the amount of Ca detected was so small that the chemical state of Ca atoms could not be determined. Therefore, XAFS measurements were used to analyze the valence and interatomic distances of elements from the x-ray absorption coefficients of materials. We focused on the spectral structures in which the absorption coefficient increased in the range of 4035–4065 eV for the Ca K-edge and 4960–5020 eV for the Ti K-edge [Figs. 3(b) and 3(c), respectively] and analyzed the interatomic distance between Ti and O [Fig. 3(d)]. The Ca K-edge spectrum of HT-Ca showed similar edge features to those of anhydrous CaCl₂ and CaTiO₃, while the corresponding EXAFS (from the Ti K-edge spectrum) displayed a scattering peak at 3.4 Å similar to that of CaTiO₃. The Ti K-edge spectrum obtained for HT-Ca showed similar features to those of Ti₂O₃, rutile TiO₂, and anatase TiO₂, proving that the surface was oxidized. Based on the explanation above, we deduced that the surface of HT-Ca was oxidized and anhydrous CaCl₂ and CaTiO₃ structures formed on the HT-Ca surface during the hydrothermal treatment process.

A rough implant surface inhibits the proliferation and adhesion of gingival epithelial cells and fibroblasts,³⁷ and biofilm and bacteria are more likely to adhere to a rough implant surface than a smooth one.³⁸ The present study revealed that hydrothermal processing of Ti64 implants did not alter the implant surface topography, so hydrothermal treatment may not interfere with the biocompatibility of the Ti64 implants for epithelial cells.

Immunohistochemical staining results revealed that the distribution of Ln expression for the HT-Ca group was comparable to that of natural teeth throughout the whole epithelial interface (Fig. 4). This suggests that the presence of Ca on the implant surface may have influenced the formation of the basement membrane. Furthermore, evaluation of the penetration of HRP as a foreign factor revealed that the HT-Ca group showed a much shorter penetration depth than those of the Cont and HT-DW groups. The molecular weight of HRP is similar to that of lipopolysaccharides, so it was used here as an alternative to endotoxins.

The immunofluorescent images showed stronger expression of Ln, which was investigated as an adhesion molecule on HT than on the other groups (Fig. 7). It was speculated that the epithelial cells on the HT plates had stronger adhesion properties than the cells on the Cont plates. This speculation is also supported by our adhesion assay and HRP penetration data. It can be inferred that the strong adhesion between the epithelial tissue and the HT-Ca implant body might have inhibited HRP penetration. This implies that hydrothermal treatment with CaCl₂ and its subsequent formation of CaTiO₃ on the HT-Ca implant surface resulted in the formation of a PIE with a laminin-rich basement membrane, indicating improved epithelial sealing of Ti64.

Hydrothermal treatment of Ti64 implants with calcium chloride resulted in the deposition of Ca on the implant surface with minimal change in surface morphology. The presence of Ca on the implant surface was shown to increase the expression of Ln in the basement membrane between the implant and the PIE, subsequently restricting the invasion of foreign factors. These results suggest that hydrothermal treatment of Ti64 implants with CaCl₂ solution may improve the epithelial sealing of the peri-implant soft tissue.

This work was supported by JSPS KAKENHI (Grant No. JP22K17094).

The authors have no conflicts to disclose.

This study was approved by the Animal Experiment Committee, Kyushu University (Approval No. A19-385-0).

Yasushige Sakamoto: Data curation (lead); Formal analysis (lead); Investigation (lead); Methodology (lead); Software (lead); Validation (lead); Visualization (lead); Writing – original draft (lead); Writing – review & editing (lead). Akihiro Furuhashi: Writing – review & editing (equal). Albert Mufundirwa: Formal analysis (equal); Investigation (equal); Writing - review & editing (equal). Takeharu Sugiyama: Formal analysis (equal). Ikiru Atsuta: Conceptualization (equal). Yasunori Ayukawa: Project administration (lead); Supervision (lead).

The data that support the findings of this study are available from the corresponding author upon reasonable request.