1 Introduction

The stability of the hybrid layer is inadvertently compromised during the bonding of resin and dentin. Firstly, etching and rinsing can degenerate the denuded collagen fibrils within the hybrid layer (Carrilho et al., 2007). Secondly, demineralized collagen fibrils may collapse via inter-peptide hydrogen bonding when they are air-dried during bonding, restricting the infiltration of adhesive resin monomers (Pashley et al., 2007; Mai et al., 2009). Thirdly, due to fluid movement within the dentinal tubules during resin infiltration, often aggravated by the presence of a positive pulpal pressure, monomers only inadequately infiltrate the dentin (Wang and Spencer, 2005; Hashimoto, 2010; Lin et al., 2010). These procedural obstacles accelerate in vivo degradation of hybrid layers (Kim J. et al., 2010). Degradation occurs as early as six months after initial intraoral function (Kim J. et al., 2010). It is triggered by hydrolysis of suboptimally polymerized hydrophilic resin components and degradation of water-rich, resin-sparse collagen matrices by matrix metalloproteinases and cysteine cathepsins (Tjaderhane et al., 1998; Pashley et al., 2004; Liu et al., 2011). Subsequently, the compromised hybrid layer causes leakage in the nano-space of the collagen fibers, leading to restoration failure (Li et al., 2000).

Several strategies have been developed to stabilize the bond between composite resin and dentin and thus prevent restoration failure (Liu et al., 2011). Biomimetic remineralization is a recent approach using analogs of matrix proteins to progressively replace water with intrafibrillar and extrafibrillar apatites in order to exclude exogenous collagenolytic enzymes and fossilize endogenous collagenolytic enzymes (Zhang, 2003; Tay et al., 2007; Tay and Pashley, 2008).

Most biomimetic remineralization studies have used transmission electron microscopy to evaluate the interface between adhesive and resin and the degree of remineralization (Kinney et al., 2003; Carrilho et al., 2005; Trebacz and Wojtowicz, 2005; Tay and Pashley, 2009; Gu et al., 2010; Kim Y.K. et al., 2010b). Confocal laser scanning microscopy (CLSM) using Rhodamine B uptake by demineralized or incompletely resin-infiltrated collagen fibers seems to be a promising new method to evaluate remineralization effects (Kim J. et al., 2010). Hence, the current study aimed to use CLSM to evaluate three commercial representative bonding adhesive resins based on three different working mechanisms.

The null-hypotheses with regard to Rhodamine B uptake of the hybrid layers to be tested were: (1) there is no difference between a two-step etch-and-rinse adhesive resin, a self-etching one-step two-bottle adhesive resin, and a self-etching one-bottle adhesive resin; (2) the placement of the experimental specimens in a remineralization medium has no effect in comparison to a control group; and, (3) the duration of immersion has no effect.

2 Materials and methods

2.1 Tooth preparation

Ninety-six intact human premolars were extracted and stored in 0.5% (5 g/L) thymol-saturated isotonic saline solution at 4 °C. The teeth were collected after the patients’ informed consents were obtained under a protocol approved by the Ethics Committee of the Faculty of Dentistry at the Zhejiang University, China. All teeth were used within three months after extraction. A flow chart of the experimental process can be found in Fig. 1.

Fig. 1
figure 1

Experimental set-up used in this study

EL: FluoroBond II; BRM: biomineralization medium; SBF: simulated body fluid; CLSM: confocal laser scanning microscopy

Full size image

The occlusal enamel was removed perpendicular to the longitudinal axis of each tooth using a slow-speed diamond saw (Isomet, Buehler, Lake Bluff, IL, USA) under water cooling (Lin et al., 2014). The resulting flat dentin surface was checked with a stereomicroscope (Wild Makroskop M420, Heerbrugg, Switzerland) at 10× magnification for the absence of enamel (Lin et al., 2014). The dentin bonding surfaces were polished with a 600-grit silicon carbide paper (Buehler) to create a standardized smear layer in mid-coronal dentin (MP-260E, Weiyi Instrument Test Equipment Manufacturing, Laizhou City, China) (Yang et al., 2006). For all specimens, the dentin was kept wet during all preparations by storage in deionized water or placement on a wet dish.

2.2 Tooth conditioning and bonding procedures

The composition of the main materials used in this study can be found in Table 1. The three adhesive resins were chosen as representatives of commercially available bonding adhesive resins with different bonding methodologies. A two-step etch-and-rinse adhesive resin (FluoroBond II, Shofu, Japan), and two one-step adhesive resins (Xeno-III-Bond, Dentsply, York, PA, USA and iBond, Heraeus, Hanau, Germany) were used according to the manufacturers’ instructions. Following the adhesive procedures, a composite resin filling material was built up incrementally to a height of 5 mm (Valux Plus Restorative, 3M Espe, St. Paul, MN, USA). Each layer was light-cured separately for 20 s (output intensity 1500 mW/cm2; Satelec, Bordeaux, France). Each tooth was sectioned occluso-gingivally. Only the two central 0.9 mm thick slabs were used. Overall, 24 experimental groups with 8 specimens were tested (3 adhesive resins, storage for 1, 2, 3, or 4 months and control group/experimental group; Fig. 1). Before experimentation, all specimens were stored in deionized water at room temperature and used within 24 h.

Table 1 Composition of the three adhesive resins used in this study
Full size table

2.3 Biomimetic remineralization procedures and control group preparation

For the control group, simulated body fluid (SBF) solution was prepared by dissolving 136.8 mmol/L NaCl, 4.2 mmol/L NaHCO3, 3.0 mmol/L KCl, 1.0 mmol/L K2HPO4·3H2O, 1.5 mmol/L MgCl2·6H2O, 2.5 mmol/L CaCl2, and 0.5 mmol/L Na2SO4 in deionized water (Kokubo et al., 1990), and adding 3.08 mmol/L sodium azide to stop bacteria from growing. Tris base at 0.1 mol/L and HCl at 0.1 mol/L were used to buffer the SBF to a pH value of 7.4 and the resulting solution was filtered (Kim J. et al., 2010).

For the experimental groups, initially a Portland cement block was prepared. For that purpose, deionized water was mixed with Type I white Portland cement (Lehigh Cement Company, Allentown, PA, USA) in a water-to-powder ratio of 0.35:1 (w/w). Before it was used for this study, the resulting solution was filtered and stored at 100% relative humidity for one week in silicone molds (Tay et al., 2007; Mai et al., 2009; Sauro et al., 2015). The experimental slabs were then placed on the established blocks of Portland cement (ca. 1 g) inside a glass cylinder (Mai et al., 2009; Kim J. et al., 2010). The biomineralization medium (BRM) was produced by adding 500 µg/ml of polyacrylic acid (PAA; molecular weight (MW) 1800; Sigma-Aldrich, St. Louis, MO, USA) and 200 µg/ml of polyvinylphosphonic acid (PVPA; MW 10177; Sigma-Aldrich) to the SBF. The solution was buffered again to a pH of 7.4 (Mai et al., 2009; Kim J. et al., 2010).

The glass cylinders containing the slabs were either filled with 15 ml of the SBF (control group) or BRM (experimental group), and stored at 37 °C (Mai et al., 2009; Kim J. et al., 2010).

The remineralization medium was changed every month (Kim J. et al., 2010). To ensure that apatite was formed instead of octacalcium phosphate (Meyer and Weatherall, 1982; Eanes, 2001), the pH (after placing on the Portland cement blocks) was monitored weekly and adjusted if necessary to maintain it above 9.25 (Mai et al., 2009; Kim J. et al., 2010).

After 1, 2, 3, and 4 months, respectively, eight slabs for each adhesive resin were retrieved and immersed in a 0.1 mmol/L Rhodamine B solution (CAS81-88-9, MW 479, Shanghai Chemicals, China). After 24 h, the dye-infiltrated slabs were rinsed with deionized water for 30 s and examined using a CLSM that was coupled with a helium neon gas laser (Laser Scanning Confocal Microscope TCS-SP2, Leica, Wetzlar, Germany; 50% of 543 nm excitation, 383 V, and magnification of 40×4). Images were captured starting from 10 µm beneath the surface to avoid superficial specimen preparation artifacts. The total fluorescent area, average fluorescence (AF), and total fluorescence (TF) were measured using Image-ProPlus 6.0 software in the region of interest (ROI). The ROI was defined as the region between the composite and the dentin, including the hybrid layer and adhesive resin layer. The fluorescent area was the area of mineral deficiency in the hybrid layers (Sidhu and Watson, 1998; Watson et al., 2008); the AF represented the fluorescence intensity of the defect area in the hybrid layers; and the TF represented the sum of all fluorescent pixels in the mineral defect area.

2.4 Statistical analysis

The data were analyzed using a statistical software package (SPSS 18.0, Chicago, IL, USA). Since data were distributed normally (Shapiro-Wilks and Kolmogorov-Smirnov tests), parametric tests were employed (three-way analysis of variance (ANOVA) (least significant difference (LSD)) and Tukey’s honest significant difference (HSD) tests). All testing was performed at a confidence level of 95%.

3 Results

Three-way ANOVA results are presented in Table 2. For each adhesive resin, the fluorescent area, AF, and TF are presented in Tables 3–5. The choice of adhesive resin, the use of a remineralizing medium (SBF vs. BRM), and the duration of immersion had a significant impact on the fluorescent area and TF (P≤0.05). For the AF, only the duration of immersion was significantly influential (P≤0.05).

Table 2 Three-way ANOVA of the data to evaluate the influences of the adhesive resins, the remineralizing medium, and the immersion time on fluorescent area, average fluorescence (AF), and total fluorescence (TF)
Full size table
Table 3 Confocal laser scanning microscopy (CLSM) analysis using Image-Pro-Plus 6.0 software to measure the fluorescent area, average fluorescence (AF), and total fluorescence (TF) for FluoroBond II
Full size table
Table 4 Confocal laser scanning microscopy (CLSM) analysis using Image-Pro-Plus 6.0 software to measure the fluorescent area, average fluorescence (AF), and total fluorescence (TF) for Xeno-III-Bond
Full size table
Table 5 Confocal laser scanning microscopy (CLSM) analysis using Image-Pro-Plus 6.0 software measuring the fluorescent area, average fluorescence (AF), and total fluorescence (TF) for iBond
Full size table

iBond specimens and FluoroBond II specimens showed the largest fluorescent area (P>0.05), followed by Xeno-III-Bond with a significantly lower fluorescent area (P≤0.05). With the exception of Xeno-III-Bond specimens after one month and FluoroBond II specimens after two months, all BRM groups showed a lower Rhodamine B uptake than the SBF groups (fluorescent area, Fig. 2). For FluoroBond II and iBond, no statistical differences could be found for SBF groups among the different storage times (P>0.05). In contrast, using Xeno-III-Bond, the fluorescent area was significantly larger after 2, 3, and 4 months than after 1 month (P=0.002). Most importantly, a statistical significant difference could be found for all adhesive resins when comparing the SBF and BRM groups at the 4th month (P≤0.05). This finding was supported by CLSM images (Figs. 3–5), in which Rhodamine B uptake was visually equal for all SBF specimens, while BRM specimens at the 4th month showed less Rhodamine B infiltration and stabilization of the dentinal collagen fiber network.

Fig. 2
figure 2

Relationship between fluorescent area in the hybrid layer and infiltration time for FluoroBond II (a), Xeno-III-Bond (b), and iBond (c) in SBF and BRM

All data are expressed as mean±SD, with n=8

Full size image
Fig. 3
figure 3

Representative confocal laser scanning microscopy (CLSM) images (magnification 160×) obtained from one to four months old FluoroBond II specimens

C: composite; D: dentin. (a) The CLSM image of a specimen immersed in SBF for one month shows that most of the hybrid layer (between short arrows, the same below) and the superficial dentinal tubules were permeable to Rhodamine B, but not the mineralized peritubular dentin; (b) The CLSM image of a specimen immersed in BRM for one month shows that the hybrid layer was permeable to Rhodamine B, with no significant difference from that of (a); (c) The CLSM image of a specimen immersed in SBF for two months shows uniform Rhodamine B uptake throughout the hybrid layer and the dentinal tubules; (d) The CLSM image of a specimen immersed in BRM for two months shows the same fluorescence stripe as in (c), just slightly wider; (e) Three-month CLSM image of a specimen immersed in SBF shows a 25-µm thick hybrid layer infiltrated with Rhodamine B; (f) Three-month CLSM image of a specimen immersed in BRM exhibits a 25-µm thick red fluorescent stripe, with no significant difference from that of (e); (g) Four-month CLSM image of a specimen immersed in SBF shows no difference from that of (a), (c), and (e); (h) Four-month CLSM image of a specimen immersed in BRM displays areas of remineralization in the hybrid layer (long thin arrows), which could indicate mineral re-deposition

Full size image
Fig. 4
figure 4

Representative confocal laser scanning microscopy (CLSM) images (magnification 160×) obtained from one and three months old Xeno-III-Bond specimens

C: composite; D: dentin. (a) One-month CLSM image of specimens immersed in SBF shows that the entire hybrid layer (between short arrows, the same below) was permeated by Rhodamine B; (b) One-month CLSM image of specimens immersed in BRM exhibits no difference from that of (a); (c) Three-month CLSM image of specimens immersed in BRM exhibits a solidly permeated hybrid layer and tubular dentin (long thin arrows); (d) Three-month CLSM image of specimens immersed in SBF shows the entire hybrid layer infiltrated with Rhodamine B, but the tubular dentin mainly left unpermeated

Full size image
Fig. 5
figure 5

Representative confocal laser scanning microscopy (CLSM) images (magnification 160×) obtained from one and three months old iBond specimens

C: composite; D: dentin. (a) One-month CLSM image of a specimen immersed in SBF shows that the entire hybrid layer (between short arrows, the same below) of about 30 µm thickness was completely permeable to Rhodamine B; (b) One-month CLSM image of a specimen immersed in BRM exhibits no difference from that of (a); (c) Three-month CLSM image of a specimen immersed in SBF shows that the entire hybrid layer of about 20 µm thickness was infiltrated by Rhodamine B; (d) Three-month CLSM image of a specimen immersed in BRM shows zones of reduced fluorescence in irregularly shaped and varied sizes (long thin arrows)

Full size image

4 Discussion

The null hypotheses could not be accepted, as the factors adhesive resin, remineralization medium, and duration of immersion were each found to have significant effects on Rhodamine B uptake of the hybrid layer.

Because collagen fibers are reinforced with minerals, healthy dentin can provide the tooth with very good strength and toughness (Kinney et al., 2003; Mai et al., 2009). To be incorporated into the gap zones of the collagen fibers, these minerals need to be sufficiently small and be arranged in a specific way (Xu et al., 2007; Mai et al., 2009).

Collagen fibers are first formed during the mineralization of biological hard tissues, in which water from the intrafibrillar compartments is replaced by apatites (Glimcher, 2006). This process is dependent on the modulation of type I collagen fibers for growth of crystals, in which apatite crystals, assembled by type I collagen fibers, and hydroxyapatite are arranged along the long axis within the mineralized collagen fibers (Glimcher, 2006). The passage of molecules is dramatically reduced when demineralized dentin collagen is replaced by apatite minerals (Glimcher, 2006). Water-soluble molecules below a molecular mass of 6 kDa can diffuse into the water compartments of a collagen fiber, while those greater than 40 kDa cannot (Toroian et al., 2007). As the channels become narrower during mineralization, the speed and volume of water-soluble molecules diffusing through mineralized collagen matrix decrease (Toroian et al., 2007). With a molecular mass of 0.479 kDa, Rhodamine B molecules easily diffuse into the demineralized or incompletely resin-infiltrated collagen fibers. After three months, the amount of Rhodamine B molecules penetrating the resin-sparse regions of the hybrid layer was reduced for all adhesive resins immersed in remineralizing solution, indicating mineral deposition in the collagen matrices. In contrast, Rhodamine B uptake from specimens immersed in SBF remained almost constant.

Compared to transmission electron microscopy, CLSM requires no preparation of thin specimens and is therefore simple and fast (Fontana et al., 1996; Pfarrer et al., 1996; Zhong et al., 2015). When using CLSM, the intensity of a fluorescent dye penetrating an artificial carious dentin lesion is proportional to the porosity or loss of mineral density within the lesion (van der Veen et al., 1996; Jardine et al., 2016). For this reason, re- or demineralization can be quantitatively measured by the fluorescence area, TF, and AF (Fontana et al., 1996; Jardine et al., 2016).

The Portland cement-based remineralization protocol introduced in 2008 uses biomimetic analogs to guide apatite remineralization (Tay et al., 2007; Tay and Pashley, 2008). According to this approach, a PAA with a low MW is used to imitate the stabilizing function of dentin matrix protein 1 (DMP1) on amorphous calcium phosphate precursors (He et al., 2005; Mai et al., 2009; Zhong et al., 2015). The interaction of the Portland cement with SBF reduces these precursors to a nanoscale (Olszta et al., 2003; Mai et al., 2009) and thus prevents their aggregation and precipitation (Cai and Tang, 2008; Mai et al., 2009; Zhong et al., 2015). PVPA, the other biomimetic analog, is a polyanion that imitates the negative charges of phosphoproteins (e.g. DMP1), phosphophoryn, or bone sialoprotein and their roles in organizing parts of the mineralization process (Gajjeraman et al., 2007; Baht et al., 2008; Mai et al., 2009). In our study, Figs. 3h, 4d, and 5d and the quantitative analysis from specimens immersed in BRM show reduced fluorescence zones, suggesting mineral re-deposition, in agreement with the literature (Kim J. et al., 2010; Kim Y.K. et al., 2010a; Zhong et al., 2015).

Some fluoride-releasing adhesive resins may induce crystal growth within fluid-filled gaps of resin-dentin interfaces without biomimetic remineralization (Hashimoto et al., 2008). However, note that in the current study apatite deposition occurred using non-fluoride-releasing adhesive resins in the absence of interfacial gaps. The remineralization effect of iBond was most obvious and had the fastest emergence time. Most of the remineralization with iBond occurred between one and three months and did not improve further after four months. This finding is similar to an apparent self-limitation of remineralization in primary dentin observed by Kim J. et al. (2010) using Adper Prompt L-Pop, a one-step self-etch adhesive resin. The remineralization effect on specimens bonded with Xeno-III-Bond was not as obvious as for iBond. In FluoroBond II, one-month-old specimens, Rhodamine B-free areas could be found in the hybrid layer for specimens immersed in SBF and BRM alike (Figs. 3a and 3b). These areas were significantly reduced in two- and three-month-old specimens (Figs. 3c–3f), while for the four-month-old specimens immersed in SBF the fluorescence of the hybrid layer remained unchanged. BRM specimens showed less Rhodamine B uptake, indicating a possible remineralization.

Rhodamine B uptake was reduced for all adhesive resins tested in the current study when placed in the biomineralizing medium, which is a promising result. However, no solution-based systems can yet be used clinically and basic science may here only pave the way for the development of future usable techniques (Zhong et al., 2015; Luo et al., 2016).

5 Conclusions

Hybrid layers of the adhesive resins used in this study could be remineralized by placement in a BRM. However, this method cannot yet be used clinically. Rhodamine B uptake in conjunction with CLSM is a useful tool to quantify remineralization effects on hybrid layers. iBond hybrid layers showed the highest Rhodamine B uptake, followed by FluoroBond II and then Xeno-III-Bond, indicating variation in the quality of the hybrid layers following the use of different adhesives.

Compliance with ethics guidelines

Hui-ping LIN, Jun LIN, Juan LI, Jing-hong XU, and Christian MEHL declare that they have no conflict of interest.

All procedures followed were in accordance with the ethical standards of the responsible committee on human experimentation (institutional and national) and with the Helsinki Declaration of 1975, as revised in 2008 (5). Informed consent was obtained from all patients for being included in the study.