Introduction
In recent years, composite resin becomes the most popular treatment for repairing dental defects due to its excellent aesthetics. Resin-dentin bonds mainly depend upon sufficient infiltration of adhesive into collagen matrix to produce the hybrid layers (HLs) [
1,
2]. Actually, adhesive fails to completely infiltrate into the collagen matrices [
3], causing a series of issues. The most striking thing is that the exposed collagen fibrils are degraded by endogenous matrix metalloproteinases (MMPs) that can be activated by the acid-etching phase of dentin bonding [
4‐
7]. Additionally, these unprotected collagen fibrils at the bottom of the HLs are prone to fatigue breakdown after repetitive loads due to the loss of intrafibrillar minerals [
8]. These adverse factors deteriorate the durability of resin-dentin bonding [
9‐
11].
Currently, several strategies have been used to address the aforementioned issues. For example, chlorhexidine, as the most commonly used MMPs inhibitors, has been widely used to inhibit collagenolytic activity within the HLs [
6]. However, chlorhexidine is water-soluble and only binds to demineralized dentin through weak electrostatic interaction, and its long-term anti-proteolytic activity is compromised [
12]. Another approach is to use collagen cross-linking agents such as riboflavin. The use of cross-linkers has been confirmed to improve the collagen stiffness, tensile strength and compressive modulus by strengthening the interfacial layer through additional hydrogen bonds and formation of covalent intra- and intermolecular crosslinks, and by inhibiting dentin MMPs via masking the cleavage sites of collagen [
13‐
21]
. Although cross-linker can maintain the integrity of the mineral-denuded collagen matrices, the water-filled, resin-sparse and flaccid collagen fibrils are prone to cyclic fatigue breakdown after repetitive loads [
8].
The remineralization of demineralized dentin collagen within HLs has been proposed as a strategy superior to the aforementioned methods. By inducing the deposition of mineral crystals in intrafibrillar compartments of collagen fibril, the biomimetic remineralization makes the water in intrafibrillar compartments be replaced by apatite and fossilizes dentin proteases [
22‐
24], which restores the flaccid collagen fibril to the natural mineralization state and improves the resistance of collagen matrices to cyclic fatigue rupture and enzymatic corruption. The process of biomimetic remineralization mainly depends on three factors, including non-collagenous proteins (NCPs), collagen scaffold and extraneous calcium and phosphorus. NCPs are very important for the modulation of the biomineralization process to biomimetically induce the hierarchical intrafibrillar mineralization [
25]. However, it is not easy to obtain native NCPs and enable them to be commercially available. Therefore, many NCPs analogues with well-defined steric structure have been studied to mimic the functions of natural NCPs [
26].
Poly(amidoamine) (PAMAM) dendrimers are kind of hyperbranched polymeric macromolecules consisting of a core, internal cavity, highly branched structure and functional groups on the exterior [
27]. With a great number of reactive groups on their exterior, well-defined size and symmetrical structure, PAMAM dendrimers enable themselves to be a desirable candidate to mimic the natural NCPs [
28]. The application of PAMAM dendrimers in the field of biomineralization has become hot research recently, and PAMAM dendrimers bearing different types of terminal groups on the surface are able to modulate the size, morphology and dimension of mineral crystals in vitro [
25,
29,
30]. Polyhydroxy-terminated PAMAM (PAMAM-OH) dendrimer has been applied to act as NCPs to induce biomimetic remineralization on the dentinal tubule due to its relatively cheap cost and superior biocompatibility [
30]. However, there has been no report whether PAMAM-OH can induce intrafibrillar mineralization of dentin.
Furthermore, dentin collagen matrix is the fundamental scaffold for the growth of dentin minerals, which plays a significant role in biomimetic remineralization [
2]. However, the remineralization process in vivo is time-consuming, during which the denuded collagen fibrils are vulnerable to enzymatic degradation, resulting in unsatisfactory remineralization. Therefore, how to protect the exposed collagen fibrils within HLs from degrading of MMPs during remineralization becomes the urgent problem for the clinical application of biomimetic remineralization [
31,
32]. Obviously, if NCPs analogues (e. g. PAMAM-OH) themselves possess concomitant anti-proteolytic activity during the induction of remineralization, it would be very beneficial to obtain satisfactory remineralization.
Thus, the present study aimed to explore whether PAMAM-OH can inhibit soluble and dentin-bound MMPs using functional enzyme activity assays and in-situ zymography. Besides, the effects of PAMAM-OH on resin-dentin interface before and after thermal cycling were also evaluated. The null hypotheses tested were as follows: (a) PAMAM-OH does not inhibit the activity of endogenous MMPs; (b) PAMAM-OH has adverse impact on the resin-dentin bonds.
Methods
Cytotoxicity
The cytotoxicity of PAMAM-OH on the viability of human dental pulp stem cells (HDPCs) and mouse fibroblast cells L929 was assessed using Cell Counting Kit-8 (CCK-8). HDPCs and L929 were cultured as described below. Following the patients’ informed consent, three teeth were collected and transported to the lab. Pulp tissues were carefully extracted by blunt non-cutting forceps and immersed into collagenase/dispase solution for 1.5 h. The cells were subsequently resuspended in Dulbecco’s Modified Eagle’s Medium (DMEM) containing necessary supplements. The seeded cells were grown at 37 °C in a humidified incubator with 5% CO2 until achieving 85% confluency.
The third passage cells were seeded in a 96-well plate at a density of 104 cells per well and incubated for 24 h. Then the medium was replaced with fresh medium containing different concentrations of PAMAM-OH. After 24 h of incubation, the required CCK-8 solution was added to each well according to the instructions. The micro titration plate was transferred to a constant temperature incubator for 4 h under light-proof conditions. The absorbance was measured at 570 nm.
Characterization of adsorption capacity of PAMAM-OH on demineralized dentin
Adsorption capacity of PAMAM-OH on hydroxyapatite (HA) powders
The PAMAM-OH solutions with concentrations from 1 to 10 mg/ml (1, 2, 4, 6, 8, 10 mg/ml) were firstly prepared. HA powders (100 mg) were immersed in 1 ml of each PAMAM-OH solution, stirred in 10 ml tubes at 37 °C overnight, and subsequently centrifuged at 10000 rpm for 3 min. The supernatant was collected, filtered and determined at a wavelength of 282 nm to assess the amount of PAMAM-OH. The amount of PAMAM-OH adsorbed on HA powders was calculated by the decrease of PAMAM-OH in the supernatant.
Synthesis of FITC-labeled PAMAM-OH
To visually observe whether PAMAM-OH can bind to demineralized dentin, a threefold molar amount of FITC solution (in acetone) and G4-PAMAM-OH aqueous solution was mixed and stirred in the dark for 24 h. Then the mixture was air-sparged to remove excess acetone, dialyzed against deionized water overnight and filtered to separate free FITC. The conjugate was finally lyophilized for subsequent experiment.
Assessment adsorption capacity of PAMAM-OH on demineralized dentin
FITC-labeled PAMAM-OH solution (8 mg/mL, 50 μL) and free FITC solution were separately applied to the demineralized dentin disk. After being dried at room temperature, each sample was washed with deionized water and dried again. The specimens were scanned by CLSM (LSM 780, Carl Zeiss). An imaging software (Image-Pro Plus 6.0, MD, USA) was used to quantify the green fluorescence intensity that represents the adsorption capacity of PAMAM-OH to demineralized dentin.
Inhibition of gelatinolytic activity
Soluble rhMMP-9
Various concentrations of PAMAM-OH solutions as test agents (0.25, 0.5, 1, 2.5, 5, and 10 mg/mL) were used to test. The purified recombinant human (rh) MMP-9 and the Sensolyte Generic MMP assay kit (all from AnaSpec Inc.) were used to examine the effect of PAMAM-OH on soluble rhMMP-9. The thiopeptolide (substrate) in the MMP assay kit was decomposed by MMPs to gradually release a colored product, which can be quantified at a wavelength of 412 nm in a 96-well plate.
The kit inhibitor (GM6001) in the MMP kit was tested as the ‘inhibitor control’, and the “positive control” included the other concentrations of PAMAM-OH. The inhibitory effect of PAMAM-OH on rhMMP-9 was expressed as percentages of the obtained ‘positive control’ concentration, which was regarded as the maximum concentration.
In-situ zymography of resin-dentin interfaces
Following the patients’ informed consent, forty teeth from each of the control and PAMAM-OH group were cut into segments and used for in-situ zymography. The dentin segments were acid-etched with 37% phosphoric acid gel, washed with deionized water and gently air-dried. Specimens were then randomly divided into 2 groups (
n = 20). The control was pre-treated with deionized water for 60 s. The experimental sample was pre-treated with 8 mg/mL PAMAM-OH for 60 s. A 2-mm-thick layer of flowable resin composite was placed on the dentin bonded with dye-doped adhesive and light-cured. Then, specimens from each group were divided into 2 subgroups (
n = 10). The samples were stored in deionized water for 48 h, labelled as T0. The other samples were thermo-mechanically challenged, corresponding to 1 y of intraoral use [
33], labelled as T1. Each bonded specimen was cut into 1-mm-thick slices under a low-speed cutter and finally polished by wet silicon carbide to grind into approximately 50 μm-thick sections. All these sections contain the dentin-resin interface.
An important initiating factor for the degradation of collagen fibers is the activation of MMP, so we performed in-situ zymography using quenched fluorescein-conjugated gelatin (E-12055; Molecular Probes) to identify the active site of MMP in the hybrid layer [
34]. The prepared fluorescent mixture as an MMP substrate was pipetted onto each glass slide according to the previous procedure and covered with a coverslip. The glass slides were incubated in a 100% humidity chamber and kept away from light. CLSM (LSM780; Carl Zeiss) was used to evaluate endogenous gelatinolytic activity. Image channels were set at 488/530 nm. Gelatinolytic activity was quantified using ImageJ and expressed as a percentage of green fluorescence in the HLs.
Measurement of released ICTP
Cross-linked carboxyterminal telopeptide (ICTP) is released after type I collagen is degraded by MMPs. Thus, the amount of ICTP reflects MMPs activity. Fifty teeth were collected and subjected to two parallel incisions perpendicularly to the longitudinal axis of the tooth to obtain dentin samples. The excess enamel, pulp soft tissues and cementum were removed. The dentin samples were placed in a stainless-steel spiral jacking tubes and frozen in liquid nitrogen for 4 min before being transferred to a mill for repeated grinding. This process was repeated to ensure the powder with a diameter of less than 40 μm. The dentin powder was immersed in 10 wt% phosphoric acid (pH = 1.0) at 4 °C for 8 h to guarantee complete demineralization. The demineralized dentin powder was washed with deionized water three times and lyophilized. These samples were then divided into two different groups (control group and PAMAM-OH group), and each group was divided into five 10 mg aliquots subpackaged in microcentrifuge tubes. Each 10 mg aliquot of powder was immersed in deionized water or artificial saliva respectively. After 2 weeks of incubation, 50 μl of the incubation medium was gathered to quantify solubilized collagen fragments. The measurement was performed with a spectrometer (Beckman Coulter, Inc., Indianapolis, IN, USA) at the absorbance of 450 nm and the amount of ICTP was calculated according to the standard curve constructed using the standards of known concentrations provided in the ICTP ELISA kit (Catalogue no. CSB-E11224 h, Cusabio, Wuhan, China).
Analysis of the effect of PAMAM-OH on resin-dentin interface
Adhesive permeation through resin-dentin interface
Ten teeth from each of the control and PAMAM-OH groups were selected for permeability evaluation. 2.5 mm-thick midcoronal dentin surface was obtained with slow-speed saw. Fluorescent adhesive was prepared by mixing fluorescein dye with adhesive. The dentin segments to be bonded were adhered to perforated Plexiglass block. The assembly was connected to a polyethylene tubing via an 18-gauge stainless steel tube. The polyethylene tubing was immersed into a column of 0.1% blue fluorescent dye solution (Alexa Fluor™ 405, λexcitation/λemission = 401/421 nm) oriented 20 cm above the Plexiglass block to simulate the pulpal pressure. Water pressure was delivered to the bonded interface through the dentinal tubules during the acid-etching process, pretreatment, bonding and resin composite buildup. The set-up was incubated and kept away from light for 4 h to ensure water to continue permeating the resin-dentin interface.
After water pressure perfusion, the specimens were cut into 1 mm-thick slices under a low-speed cutter and finally polished by wet silicon carbide to grind into approximately 50 μm-thick sections containing the water perfused bonded interface. CLSM (LSM780; Carl Zeiss) was used to visualize adhesive permeation. Dyed adhesive permeation was quantified using ImageJ and expressed as a percentage of rose red fluorescence within the dentinal tubules.
Micro-tensile bond strength
Forty teeth from each of the control and PAMAM-OH group were cut into segments and prepared for tensile bond strength. The dentin segments were acid-etched with 37% phosphoric acid gel, washed with deionized water and gently air-dried. Specimens were then randomly divided into 2 groups (n = 20). The control was pre-treated with deionized water for 60 s. The experimental sample was pre-treated with 8 mg/mL PAMAM-OH for 60 s. A 2 mm-thick layer of resin composite build-up (Z250, 3 M ESPE, St. Paul, MN, USA) was placed on the dentin bonded with adhesive and light-cured. Then, specimens from each group were further divided into 2 subgroups (n = 10). The samples were stored in deionized water for 24 h, labelled as T0. The other samples were thermo-mechanically challenged, corresponding to 1 y of intraoral use, labelled as T1.
After storage, each bonded specimen was cut into 1 × 1 × 7 mm beams. The micro-tensile bonding test was performed by a universal testing machine with a crosshead speed of 1 mm/min. After the beam was stressed to failure, the fracture pattern of the resin-dentin interface was observed using a stereomicroscope with high magnification. Failure modes were classified as the following: adhesive failure (A), mixed failure (M), cohesive failure in resin composite (CC), and cohesive failure in dentin (CD).
Statistical analysis
For all analyses, the significance level was set at α = 0.05. Data were exhibited as means and standard deviations (SDs). For each parameter, data sets were evaluated for their normality (Shapiro-Wilk test) and homoscedasticity assumptions (modified Levene test) prior to the use of parametric statistical methods. If the assumptions were not violated, the data sets were analyzed with one-way ANOVA or one-factor repeated-measures ANOVA, which depends on the parameter tested. Post-hoc comparisons were conducted using the Holm-Sidak procedures to identify statistical significance among groups. If those assumptions were violated, the data sets were nonlinearly transformed to satisfy those assumptions prior to performing the aforementioned statistical procedures.
Discussion
PAMAM-OH has been widely used to induce biomimetic remineralization on the demineralized tooth [
30], but there has been no report whether it can induce intrafibrillar mineralization of dentin. Based on the size exclusion effect of fibrillar collagen where molecules with molecular weight (MW) between 6 and 40 kDa can partially access the intrafibrillar water compartments [
35,
36], G4-PAMAM-OH (MW = 14.279 kDa) would access the intrafibrillar spaces and enable itself to be a desirable candidate to induce intrafibrillar remineralization to protect exposed collagen fibrils within the HLs to achieve durable resin-dentin bonds. However, the remineralization process in vivo is time-consuming, during which the denuded collagen fibrils as scaffold for the growth of mineral crystallite are inevitably vulnerable to enzymatic degradation, leading to the failure of remineralization [
32]. Thereby, if PAMAM-OH (NCPs analogues) itself possesses concomitant anti-proteolytic activity during the induction of remineralization, it would be very beneficial to achieve satisfactory remineralization. In the present study, the influence of PAMAM-OH on dentin proteases was first explored, which would lay the foundation for the satisfactory intrafibrillar remineralization induced by PAMAM-OH within HLs to achieve durable resin-dentin bonds in the next work.
It is well known that the dentinal tubules are full of fluid. Intrapulpal pressure causes the intrinsic water in the pulp cavity to be continuously replenished to the dentin surface. Therefore, the adsorption capacity of PAMAM-OH on demineralized dentin is crucial for achieving its function during the remineralization process. In the study, the quantity of adsorption tests was firstly performed to measure adsorption capacity of PAMAM-OH on HA powders
(Fig.
2). With the concentration of PAMAM-OH increased from 1 to 10 mg/ml, the amount of PAMAM-OH adsorbed on HA powders quickly increased firstly and gradually reached saturation. CLSM was also performed and images
(Fig.
3A
) displayed that the strong green fluorescence was observed on both FITC-PAMAM-OH sample and FITC sample before washing. After washing, the strong green fluorescence was similarly detected on FITC-PAMAM-OH group with an intensity value of 90.4% ± 4.4%, while little fluorescence was detected on the control sample, reaching 31.2% ± 6.0% fluorescence intensity (Fig.
3A, B). The experimental samples withstood the deionized water rinse that simulated the pressure of intrinsic water from pulp cavity and exhibited strong fluorescence, indicating a strong adsorption capacity of PAMAM-OH on demineralized dentin. These results were also confirmed by a previous study [
30]. We speculated that many anionic amide groups inside PAMAM-OH may attach to the collagen fibrils by electrostatic interaction to contribute to its adsorption capacity [
30].
Dentin collagen matrix is the fundamental scaffold for the precipitation and growth of apatite mineral crystallite, which plays an important role in biomimetic remineralization [
2]. It is essential to protect the exposed collagen fibrils from degrading of MMPs during remineralization. Thereby, the purpose of this paper was to investigate whether PAMAM-OH could have inhibitory effect on endogenous MMPs. The influences of various concentrations of PAMAM-OH on soluble rhMMP-9 are represented in Fig.
4A. Quantitative analysis performed by Sensolyte assay kit showed that the degree of rhMMP-9 inhibition was proportional to the concentration of PAMAM-OH. When the concentration of PAMAM-OH was equal to or higher than 1 mg/ml, the anti-MMP-9 effect was comparable to that of the kit inhibitor control group. This experiment confirmed that the activity of exogenous rhMMP-9 is inhibited by PAMAM-OH while its effect on endogenous MMP-9 embedded within dentin collagen matrix should also be explored. In this research, gelatinolytic activities of the endogenous MMP-9 directly within HLs after incubation for 48 h (T0) and thermal cycling (T1) were detected by in situ zymography. There was a significant difference in the fluorescence intensity between the control group and PAMAM-OH group regardless of incubation for 48 h (T0) or thermal cycling (T1) (
p < 0.05) (Fig.
4B, C), indicating that PAMAM-OH has the inhibitory effect on endogenous MMPs. Additionally, the amount of type I collagen ICTP fragments was measured to determine the degradation of the demineralized collagen matrix by MMPs. The rationale for this assay was on the basis of the report that solubilized ICTP fragments are solely derived from the degradation of collagen fibrils by MMP [
37]. The quantitative assay revealed that the release of ICTP fragments from the PAMAM-OH group was significantly different from that of the control group (
p < 0.05) (Fig.
5), suggesting that PAMAM-OH has the inhibitory effect on MMPs-driven collagenolysis. Therefore, the first null hypotheses that “PAMAM-OH has no inhibitory effects on endogenous MMPs” should be rejected.
The functional mechanism of inhibitory effects of PAMAM-OH on dentin proteases remains unclear, but several factors have been speculated. The high density of nitrogen ligands in polyhydroxy-terminated dendritic polymer and the abundance of functional groups on its surface allow it to be used as a multifunctional chelating agent [
38]. MMPs are classified as Ca- and Zn- dependent endoproteases [
39]. PAMAM-OH may chelate Ca
2+ and Zn
2+ that bind to the Zn
2+- and Ca
2+- active sites of the catalytic domain of MMPs, spatially intercepting the active sites and restraining the activity of MMPs. Besides, the previous study has reported that PAMAM-COOH may electrostatically bind to proteins and inhibit the activity of MMPs [
40]. Therefore, we surmise that negatively-charged PAMAM-OH may also electrostatically bind to positively-charged domain of MMPs, which contribute to its inhibitory effect.
Sufficient infiltration of resin monomers is important for the stability of the bonded interface. The dentin samples pretreated with PAMAM-OH were observed by CLSM to evaluate whether it has adverse effect on adhesive infiltration [
41]. The CLSM images and quantitative analysis indicated that PAMAM-OH pretreatment did not decline the permeability of adhesive (Fig.
6A, B).
The microtensile bond strength was also conducted to assess if pretreatment of PAMAM-OH has adverse effect on the dentin bond strength. The result of bond strength indicated that pretreatment of PAMAM-OH on dentin did not decline the immediate tensile bond strengths. However, the bond strength of PAMAM-OH group was significantly higher than that of the control after aging (
p < 0.05) (Fig.
7A), indicating pretreatment of PAMAM-OH prolonged the resin-dentin bonds. These findings were confirmed by the results got from in situ zymography after aging. PAMAM-OH possesses anti-proteolytic activity and prevents exposed collagen fibrils within HLs from enzymatic degradation, which may conduce to the improvement of the durability of resin-dentin bonds. Hence, the second null hypotheses that “PAMAM-OH has adverse impact on the performances of resin-dentin bonds” should be rejected.
Further studies on the interaction between PAMAM-OH and dentin matrix, and studies of intrafibrillar remineralization induced by PAMAM-OH within HLs are required to support its potential clinical application.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.